CA2197242A1 - Regulatable elimination of gene expression, gene product function and engineered host cells - Google Patents

Abstract

Materials and methods are disclosed for regulated obstruction of the expression of a target gene or the biological effect of its gene product in genetically engineered cells or organisms containing them. Aspects of the invention are exemplified by recombinant modifications of host cells and their use in vitro and in vivo for the regulatable blockade of expression of a target gene, for interference with the function or effect of a target gene product or for the regulatable elimination of a target gene.

Description

21 972~12 ~ WO 96/06111 p. ~ I ~ F.limin~-~inn of Gene E~-~ulc ~ ~
~ Gene Product Function and Enb;.l~ l Host Cdls Technical Field This invention concerns materials, methods and '1'~1 .. ;.,.. c relating to the m--ltimPri7ins of chimeric proteins with a dimeric or nr~.ltim~rir preferably non-peptidic, organic rr~mpo~n~i Aspects of the invention are P~r;.. 1 by .-.. 1.;.. ~ a ~.. n.. l;r;. ~ of host oells and their use in vi~ro and in vivo for the regulatable blockade of expression of a target gene, for 10 ;uLClrC~cll~c with the function or effect of a target gene product or for theregulatable rliminvtir,n of a target gene. The materials, methods and .~.I.li. -I ;". .' of the invention provide a means for regulating expression ofgenes introduced by Ir~ .l-;..- l ...r..l;r;. ...,.c of host cells, including completely ,li.. - ;..... ~, a population of genetically engineered cells, thus 15 providing a fail-safe m~rh~nictn for controlling genetically engineered cells used in gene therapy.

Ba~Lb~uuild One approach to studying the role of a gene in a host organism is to eliminate it or replace it with a dyarull~L;olldl ~uullL~I,udl~. This can be 20 ~ hr 1~ for example, by genetic clli5;ll...;llg of cultured embryonic stem ~"ES"~ cells followed by introduction of the engineered cells into embr,vos which develop into whole organisms lacking a functional copy of the gene.
Typically such . ~1.. ;.. ...1~ are conducted in mice. However, where the geneproduct is required for normal J~lvyll~ of the host organism, simply 25 ,1"";" ~ " or vknocking out", the gene may yield severely d~arull~L;ul~dl animals, which may not be useful, or no animals at all.
A number of systems have been developed for creating "rnntlirir,nal knock-outs'~, i.e., cells or organisms in which a gene can be ablated when desired, i.e., at or after any desired stage of Ic~lvp~ L and for creating cell-21 ~724'~' -WO 96/06111 . ~ /lU:~

type specific (non-f- ~ ' l) knock-outs, i.e., organisms in which a gene is ablated in a.specific type of cell. See for example, Watson et al., ~nmhi~ 7t DNA, 2d ed. (1992), especially Chapters 14 and 24; Gu et al., Science 265 (1994) 103-106; Barinaga, "Research News: Knockout Mice: Round Two,"
S Science 265 (1994) 26-28; Lasko et al., "Targeted oncogene activation by site-specific .,.. l .:" ~l ;.. in transgenic rnice," Proc. Natl. Ac~d. Sci. USA 89 (1992) 6232-36; and Orban et al., "Tissue- and site-specific DNA .~ ' in transgenic mice," Prnc. No.tl. Ac~ Sci. USA 89 (1992) 6861-65.
T".l~ 1,....1..1 of that work, biological switches have been developed which are based on ligand-mediated .. ll :, . ,.. ;, l r n of ;. " ., .. I.. pl.:l ;.. - and other receptor-based ~ proteins. Aspects of that work are disclosed in Spencer et al., Science 262 (1993) 1019-1024 and i".,. ~ I Patent Applications PCT/US94/01617 and PCTtUS94/08008, the contents of all three of which are hl~ul,uullLed herein by reference.
ntr~r~ r ..u,,lhlLlllg of chimeric proteins by synthetic ligands has potential in basic hl~wL;~aL;u~ of a variety of cellular processw, in regulatingthe synthesis of proteins of therapeutic or ~ ...l hll,UUlLiUlUC, including the reg~llatable ob~ilu.L;u.. of the expression or function of these genw and inregulatably initiating cell death in engineered cells. rulLL~IIIlul~ ligand 20 mediated oL~,. ,.". . :, ~ ,. ,., now permits regulated gene therapy. In so doing, it providw a fresh approach to increasing the safety, expression level and overall efficacy obtained with gene therapy.
Illustrative ~ , r llC disclosing further ba~L~,luulld i"r " .. - f ll of interest are provided in PCT/US94/01617, wpecially on pages 1-4. However, 25 as will be clear from this disclosure, none of the foregoing pllhlir~tif,nc describe or suggest the present invention.

Snrnrnary Of The Invention This invention provides materials and methods for the genetic ... ~,i.. :.. ~ of host cells to render the cells, their progeny and organisms 30 containing them, susceptible to blockade of expression of a selected gene, tohlLelrelcll e with the fnnrtionine of the gene product or to ~l:...;.. ~ :. ", or - . 2 1 972~2 ~ wo 96/06~

iU~i~dL;U-I of a selected gene, in a regulated fashion. The invention further provides materials and methods for genetic P .~,:.... ;"g of host cells to render the cells and their progeny susceptible to regulated, ~,. U~l UlUlled cell death, - otherwise known as apoptosis. Such cells, as well as organisms containing 5 them, are useful as biological reagents for a variety of research purposes, as described inJ;a, including the study of diseases ,1.... ~.; ..I by the Iy~fu~ ;ull or absence of a gene or gene product of interest. Table 1 below provides a non-exclusive, illustrative list of diseases linked to the Jy~full~;uor absence of a gene or gene product.
The invention involves the adaptation of a method and materials for using homo- and hetero-m~ n of chimeric ~responder" proteins to trigger gene ~ldU~S~ ;Oll in living cells. As used herein, the terms multimer, Li,..~ and multimen2ation encompass dimers, trimers and higher order multimers and their formation. The chimeric responder proteins are 15 imra~Plll.larly expressed fusion proteins which contain a specific receptor domain and are responsive to the presence of a cullc~ulldillg ,....11 ;,.... ;,,ng agent. The mllltimPri7ing agent is a ~ulL;~ I Iigand which is capable of binding to more than one of the chimeric protein molecules to yield dimers or higher order multimers of the chimeras. The chimeric proteins are designed 20 such that the ligand-mediated mllltim~ri7~ti~m triggers tranC~ripti~n of a gene under the LldUI~UI;,t)L;UIIill control of an element which is responsive to suchn This gene may be a "blocking" gene, which encodes a blocking factor such as an anti-sense message ~ y to, and capable of interfering with I ,~..c . ,1.l .. ,.\ of, a target gene or a ribozyme. which is 2s capable of preventing the expression of a target gene. Al~tll~ ly, the blocking gene may encode an antibody moiety, such as a single-chain antibody, which is capable of binding to and preferably nPIltrali7ing or blocking a biological effect of the gene product of the target gene.

WO 96106111 ' ' A ~,~IIJ~,JI 9 Table 1 Disease Gene/Gene Product Parkinson's disease tyrosine h~Jlu~.JI~, neural d~ Lv~ diseases growth factors or ~ alhlg proteins U_~U~UUIU~ bone . -.,~ . factors anemia .,l~ll.. Ul~U;~

;"... ~A ~. i. ; ' antigen receptors, immune cell signaling proteins cancer tumor suppressor genes in animals carrying suppressor mutations cystic fibrosis CFTR
diabetes mellitus insulin, insulin receptor pituitary dwarfism growth hormone ,vh~ alphal-antitrypsin familial llyl~ ùl~ LDL receptor 15 thalassemia major and sickle beta-globin cell anemia hemophilia A Factor VIII
hemophilia B Factor IX
Gaucher's disease ~ ,ùc~,lel,lusid~
glycogen storage diseases several Ly~

severe combined ADA, purine nucleoside l~hu ~hul~l~., ;"" ". ...~1.1 ;. :.. .~.~ disease p70Z~P
Duchenne muscular dystrophin dystrophy Lesch-Nyhan syndrome H~l~w~lLll;lle ~ .l;l,r,~l transferase Lou Gehrig's Disease superoxide dismutase Glanzmann GP llb-llIa - the fibrinogen receptor LlllUllll)V~rLulJ~
Tay-Sachs disease 2~ 97.?42 ~ wo 96106~
,; S~ , The blocking gene also may encode a dominant negative gene product capable of blocking the effect of a target gene product, or it may encode a protein whose action will lead to the ,l; . . .~ nn of the selected target gene entirely.
An example of a blocking gene whose action will lead to the ~lim:-~rinn of a s target gene is the gene encoding the protein Cre, which produoes Cre 1-~ .. 1.;.. - . El~pression of the Cre ._~ .. l .;.. ~ leads to _l;.. ;.. ~ ;n.~ of a target gene A~ Iu~ dy flanked by ~lox~" sequences in the host oells.
AIL~ .,ly~ the gene may be a gene which provides a means for . l:........ ~ ;.. g a population of engineered oells, either growing in culture or in vivo, providing a m~rh~nicrn for regulating and controlling engineered cells.
Thus, this invention involves one or more chimeric responder proteins, DNA constructs ("responder" constructs) encoding them, and multi-valent ligand molecules capable of ".. Il ;.. ,i.. g the chimeric proteins. The chimeric proteins contain at least one ligand-binding (or Ureoeptor~) domain and an 15 action domain capable, upon m--ltim~ri7~tinn of the chimeric protein molecules, of initiating ~ /ll of the blocking gene or the li., .;. . ~ ;. .I, effector gene within a cell. The chimeric proteins may further contain additional domains. These chimeric responder proteins and the responder constructs which encode them are ,-...... II.;l.Al.l in the sense that their various 20 .~ .n~ l~ are derived from different sources, and as such, are not found together in nature (i.e.~ are mutually heterologous). Also provided are -.... I.;.. A.. I "blockingn constructs containing a blocking gene under the . A~ . ;l.l ;. ~al regulation of a control element responsive to the presence of..... ll,.. ;.. 1 responder proteins described above. The trAnC~riprinn~l control 25 element is responsive in the sense that trancrrirrinn of the blocking gene isactivated by the presence of the mnlrimAri7~-~ responder chimeras in cells containing these constructs. Exposure of the oells to the 1l..l1l i",. ;, ;..g ligand results in expression of the blocking gene. The constructs of this invention may contain one or more selectable markers such as a neomycin resistance 30 gene (neo~) and herpes simples virus-thymidine kinase (HSV-tk). When oells which have been genetically engineered to contain and express the responder and the blocking gene are exposed to the m1llrimrri7ing ligand, expression of 2'1 9~2~2' the blocking gene is activatet amd expression of the target gene or r l: . ;..g of the target gene product is impaired.
R~ l "Cre" constructs encoding Cre under the ~
regulation of a control element responsive to the presenoe of . ll: ;.... ; . d ~i responder proteins described above are also provided. The l .~
control element is responsive in that tranc~irti.~n of Cre is activated by the presence of the m~ im~t~7-~l responder chimeras in cells containing these constructs. That is, exposure of the oells to the ,....ll ;.... ;,:..g ligand results in expression of Cre. Cells may also contain a "target" gene, preferably, but not 10 necessarily, an ~. Iu~,.... ~ gene to be eliminated, which is flanked by loxPsequences (rfloxed~). The floxed target gene is introduoed into the oell, for example, by homnlflgollc ,r..l.l.l.;,..t;.~u using a ll-~V'''Il '''''I "target"
construct containing part or all of a copy of the rll~ target gene and/or rl~ flanking DNA sequenoe thereof, together with loxP DNA
15 elements. As in the case of other constructs of this invention, the floxed target gene construct may contain one or more selectable markers such as the neomycin resistanoe gene (neo~ or herpes simples virus-thyrnidine kinase gene ~ISV-tk). When oells, which have been genetically engineered to contain and express rhe responder, Cre and floxed target gene constructs, are exposed to the .,.. 11 .. ;,;.. g ligand, expression of Cre is activated and the target gene is eliminated.
Additionally ,~. . . l~h constructs encoding primary chimeric proteins of this invention perrnit ligand-regulated apoptosis. These chimeric constructs contain at least the uyru~ ll;c domain of the fas antigen or Apo-1 25 antigen, which when cross-linked, induces apoptosis in most cell types. (See Trauth et al. Sàence 245 (1989) 301-305; Watanaba-Fukunaga et al. Na~re 356 (1992) 314). In this way one can provide for ligand-inducible cell death for an engineered population of oells.
Modifled cells are produoed by hl.lu.lu~;llg the desired construct(s) into 30 selected host cells. This may be a~ulllluLll~d using l,Ull~ ;Ulldl vectors and techniques, many of which are commercially available. If desired, the modified cells into which one or more constructs have been sucoessfully 2 ~ 97242 ~ WO96/06111 r~_l/l)~..~/ .l introduced may then be selected, separated from other cells and cultured, agun by ~u..~ ;u~al methods.
Where the target gene is an ~ ln~ gene, hl~ ul,uO~a~;uu of a target gene construct into a host cell may be effected by h.. ln~ ;nl-, by known methods such as that of Gu et al., 1994, Science 265:103-106.
Al~ ali~ 'y, if ~ ...l~.~, .... ~ copies of the target gene are deleted or rendered n.. r.. ~ l by hnrnnlngmlc ,_.. l.i.. -~: .1, or mutation, the Ir.. ". l.. ~.A.. I
target gene construct may be introduced by any desired means. Ll~ùlluOIaL;ù
of the responder constructs and Cre constructs may also be effected using 10 UUII~ ;Ond~ r~ vectors and tPrhnir~ c Genetically engineered cells, which cAn be grown together with other cells and which can be selectively regulated in, or ablated from, the mixture ofcells by addition of an effective amount of An nll~ l nn ligand which is capable of binding to the primary chimeric protein, are an important aspect of 15 this invention. Contacting the engineered cells with the nP,~"~".. . ;~_l ;- ..1 ligand can trigger the regulatable OIJ:~Llu~ L;ùll of the expression of a specific gene through the expression of a target gene which can block or interfere with the function of a target gene. Al~cll~ .ly, the regulated gene can be one whose expression leads to specific . l: " ., . ;.~l. of a target gene, or cell death of 20 the engineered cells. For example, such cells may be permitted to produce an rlt~ or h.~clulGb~u~ product for some desired period, and may then be deleted by addition of the ligand. In such cases, the cells are engineered to produce a primary chimera in accordance with this invention.
The cclls, which may be further engineered to express a desired gene 25 under ligand-induced regulation, may be grown in Nlture by Cull~ ;ull~
means. In that case, addition to the culture medium of the ligand for the optional chimera leads to expression of the desired gene and production of the desired protein. Expression of the gene and production of the protein can then be turned off by adding to the medium an nh~,. ". .- ; ~ -l ;. ..~ antagonist 30 reagent, as is described in detail below. In other cases, production of the protein is ~ ull~L;~u~;~.. In any event, the engineered oells can then be eliminated from the cell culture after they have served their intended purpose 2 t q7242 ;
WO96/06111 r~,l/lJ.,,~ul ~1 ~

(e.g. production of a desired protein or other product) by adding to the medium an effective amount of the ~ .Jpl;~ nli~ , ligand to cause .~1;~,..,..- ;, 1:-.., of the primary chimera and induce apoptosis in the engineered cells. Engineered cells of this invention can also be used in vivo, to modify 5 whole organisms, preferably animals, including humans, such that the cells produce a desired protein or other result within the animal containing such cells. Such uses include gene therapy. Al~c~ ly, the chimeric proteins and ~.l;~,.. i, i.. ~, molecules can be used CALI~ UI uly to bring together proteins which act in concert to initiate a pl.~ ,log;~l action.
To create transgenic animals containing modified cells of this invention the desired constructs are transfected into ,~ u~.l;..s~: oell lines, for example, ES cells. Al~tlu~ ly~ the desired constructs may be directly IlUI~ tl.l into early embryos. See for example, Watson et al., Re. ..".l ~ DNA, 2d ed., 1992, especially Chapters 14 and 24. In the latter case, use of a tissue-15 specific expression control sequence, such as a promoter or enhancer sequence, in the responder construct permits tissue-specific expression of the chimeric responder protein(s), which in turn permits tissue-specific, regulatable expression of the blocking gene, and thus tissue-specific blockade of the targetgene or its gene product. It should be further noted that animals may be so 20 produced which comprise cells containing and capable of expressing the responder. Other animals may be produced which comprise cells containing a blocking gene construct. Breeding the two types of engineered animals yields offspring which contain cells containing both of the foregoing construas.
A tissue-specific expression control sequence (promoter/enhancer) in the 2~ responder construct also perrnits tissue-specific expression of the chimeric responder protein(s), which in turn permits tissue-specific, regulatable expression of Cre, and thus tissue-specific ~limin~rinn of the target gene. It should be further noted that animals may be so produced which comprise cells containing and capable of expressing the responder and Cre constructs. Other 30 animals may be produced which comprise cells containing a desired target geneconstruct. Breeding the two types of engineered animals yields offspring which contain cells containing all of the foregoing constructs. Animals and 2 1 97?42 W09610G111 '''~IlJ'''JI ' ~, their progeny may be Wll~. ~ 11J I ~ by w..~ iv,.al genetic analysis. In.addition to hl~l~ ' into ES cells or early embryos, the constructs may also be introducet by A.1.. ,.. 1.~l.. e.g. in suitable vehicles or vectors, directly into the desired tissue of the whole organisms.
The ~ , ligands useful for triggering the exprQsion of the blocking gene including in the practice of this invention are capable of bindingto two or more chimeric rQponder proteins containing such receptor domains.
The ,.. 11 :.. ;, .. ~, ligand may bind to the chimeras in either order or ... v~1y~ preferably with a Kd value below about 106, more preferably below about 10-7, even more preferably below about 10-8, and in some .... l.o.l:.. ~ below about 109 M. The ligand preferably is a non-protein and has a molecular weight of less than about 5 kDa. Even more preferably, the ,....11 ;...- . ;,;..~, ligand has a molecular weight of less than about 2 Kda, and even more preferably, less than 1500 Da. The receptor domains of the 15 chimeric proteins so ml~1rim~ti7~d may be the same or different. The chimericproteins are capable of initiating expression of the blocking gene in the host cell upon exposure to the ligand, following I~ ;lll.. ;~L;V.. of the chimeras.
Thus, L~ JL;ull of the blocking gene or the ~limin~ti~n effector gene is activated in genetically engineered cells of this invention following exposure of 20 the cells to a ligand capable of mlllrim~ti7ing the chimeras. Said differently, genetically engineered cells of this invention contain chimeric proteins as described above and are rQponsive to the presence of a ligand which is capable of ,....l. ;... ;,;..g those chimera. That rQ~v~ C55 is manifested by the initiation of ~prwsion of the blocking gene and by blockade of expression of 25 the target gene or blocking a biological effect of the target gene product.
Where regulatable ~limin~ril~rl of a gene is desired, ~ o11~ is manifQted by the initiation of Cre expression, and where a floxed target gene is also present, by the ~limin~ti~m of that target gene.
The encoded chimeric responder protein may further comprise an 30 intracellular targeting domain capable of directing the chimeric protein to adQired cellular ~v111t,..1L.,...1L The targeting domain can be a secretory leader sequence, a membrane spanning domain, a membrane binding domain or a 2 1 97~4~.
wo 96106~ , sequence directing the protein to associate with vesicles or with the nucleus, for instance.
The action domains of the chimeric proteins may be selected from any of the proteins or protein domains (preferably of the species of the desired host ceLs or organism) which upon ' are capable of activating of a gene, the blocking gene or the regulatable ,~1:.. :.~
effector gene, in our system, under the ~ u~I control of a cognate control element. For instance, the action domain of the chimeric responder protein molecules may comprise a protein domain such as a CD3~ (zeta 10 subunit) which is capable, upon exposure to the ligand and subsequent ,.. 11 :.. ;,-, :.~u, of initiating a detectable intracellular signal leading to L ....~1 activation via the IL-2 promoter. AI~ Y~ there may be a series of responder proteins, in which one responder protein contains as its action domain, a DNA-binding protein such as GAL4, while another contains 15 as its action domain a ~ ,., 1 activation domain such as VP16.
.UILI.~.;~L;OII of such responder proteins to form a GAL4-VP16 dimer activates the l.~.u~ u of the blocking genes or the l~limin~tir~n effector gene under the Ll. ~ control of elements to which the het~.ud;l.l~.;~d responder proteins can bind. Numerous other examples are 20 provided herein. In such examples" "..1l; " .. . ;, 1 :r,n activates trAn~rrirtinn of a blocking gene or an ~limin~tirln effector gene under the rrAn~rrirti.~n~l control of a rrAn~rriptirlnAl control element enhancer and/or promoter elements and the like, which is responsive to the .,.. 11:....: 1 :f.n event.
This invention further ~ DNA vectors containing the 25 various constructs described herein, whether for Lnludu~;on into host cells in tissue culture, for hlL~ullu~,L;oll into embryos or for 4.L.u...AL.A~;on to whole organisms for the ill~lUlL~,I;Ull of the constructs into cells in vivo. In either case the construct may be introduced episomally or for ~LIUIUOAUUIal im~grati~n The vector may be a viral vector, including for example, an 30 adeno-, adeno associated or retroviral vector. The constructs or vectors containing them may also contain selectable markers permitting selection of l.lAI. r~ containing the construct.

~ wo 96/06111 2 1 912 4 2 r ~"~
11 .
This invention further ~ a chimeric protein encoded by any of our DN~ constructs, as well as cells containing and/or expressing them, including ~u~vkalyuL;c and euc~ryotic cells and in particular, yeast, worm, insect, mouse or other rodent, and other ,.. ~.. -1:-.. cells, including human S cells, of various types and lineages, whether fro_en or in active growth, whether in culture or in a whole orgamsm cûntaining them.
This invention provides cells, preferably m~-nm~ n cells, which contain one or more ~ . DNA constructs encoding a responder protein, or series of responder proteins, and a blocking gene construct capable 10 of expressing a blocking gene in response to "...1~ ;on of the responder protein(s). Additionally, it provides cells which contain one or more DNA constructs encoding a responder protein, or series of responder proteins, a Cre construct capable of expressing a Cre gene in response to m~ of the responder protein(s) and optionally, a IS target gene construct UU~ UI~ g a target gene flanked by loxP sequences which is susceptible to deletion in the presence of Cre. The invention also provides cells which contain Ir~ ;..,..l DNA constructs encoding proteins whose expression is capable of leading to regulated apoptosis of the cells cont~ining the construct.
The ml~ltim~ri7ing ligands are molecules cap~ble of binding to two or more chimeric responder protein molecules of this invention to form a multimer thereof, ~nd have the formula:

linker--{rbm" rbm2, ...rbm"}

wherein n is an integer from 2 to about S, rbm(~).rbm~n) are receptor binding 2s moieties which may be the same or different and which are capable of binding to the chimeric protein(s). The rbm moieties are covalently attached to a linker moiety which is a bi or multi-functional molecule capable of being covalently linked (~_n) to two or more rbm moieties. Preferably the ligand has a molecular weight of less than about 5 Kda and is not ~ protein. Examples 30 of such ligands include those in which the rbm moieties are the same or ~1 9724?
wo 96106~ JI

different and comprise an FK506-type moiety, a ~y~ L~ul;ll-type moiety, a steroid or ~ l;,le. Cyclosporin-type moieties include l,,y~,lO~UlUl and derivatives thereof which are capable of binding to a cyclophilin, naturally occurring or modified, preferably with a Kd value below about 10~ M. In s some ,,.,1,~ it is preferred that the ligand bind to a naturally occurring reoeptor with a Kd value greater than about 10~ M and more preferably greater than about 105 M. Illustrative ligands of this invention are those in which at least one rbm comprises a molecule of FK506, FK520, rapamycin or a derivative thereof modified at C9, C10 or both, which ligands bind to a modified receptor or chimeric molecule containing a modified receptor domain with a Kd value at least one, and preferably 2, and more preferably 3 and even more preferably 4 or 5 or more orders of magnitude less than their Kd values with respect to a naturally occurring receptor protein. Linker moieties are also described in detail later, but for the sake of illustration, include such moieties as a C2-C20 alkylene, C4-C18 a7alkylene, C6-C24 N-alkylene a7alkylene, C6-C18 arylene, C8-C24 ardialkylene or C8-C36 bis~ul,u.~lldo alkylene moiety. See also U.S Patent Application Serial No. 08/481,941 entitled "Rapamycin Regulation of Biological Events", filed on June 7, 1995 and U.S. Patent Application Serial No. 08/292,598, entitled UNew Mllltimrri7ing Agents," filed 18 August 1994, the contents of which are hereby hl~ulluul ~t ;I by reference.
The ...m~..,... ;r rbm's of this invention, as well as WIU~JUU11~6 containing sole copies of an rbm, which are capable of binding to our chimeric proteins but not effecting rlim~ri7~til~n or higher order .1...ll;.,...;,- ;.
thereof (in view of the .. ,.... ;. nature of the individual rbm) are mnlrim.~ri7~tir,n ~nt~grlnictc This invention thus provides materials and methods for selectively OL~I u~hlg the expression or effects of a target gene in engineered cells in response to the presence of a .. .ll;.. ;,;,.g ligand which is added to the 30 culture medium or ù~L~uluaLt,lcd to the whole organism. The invention further provides materials and methods for selectively r~ g a target gene in engineered cells in response to the presence of a mllltim~-ri7ing ligand which ~ 1 ~7~42 wo 96/06111 1 ~I/~).,,~ .

is added to the culture medium or a.l,l.;..;~l~ r.l to the whole organism, as the case may be. The invention further provides materials and methods for selectively ablating cells in response to the addition of an ol;~,.. ;, .. g ligand.
The methods involve providing cells of this invention, or an organism 5 containing such cells, which contain and are capable of expressing (a) one or more DNA construas encoding one or more chimeric proteins capable, following,, ~ ;.au of activating ~. ~ ,. . ;l'~ ;r,n of a blocking gene; (b) a blocking gene under the ~ l regulation of an element responsive to multimers of the chimeric proteins. The method thus involves exposing the 10 cells to a m~ n Iigand capable of binding to the chimeric protein in an amount effective to result in detectable expression of the blocking gene.
The method also involves providing cells of this invention, or an organism containing such cells, which contain and are capable of expressing ('d)one or more DNA constructs encoding one or more chimeric proteins capable, 15 following mllll ;",.. ;,- ;.~.~, of activating 1, ....~ . ;l.l ;, ... of a gene encoding Cre;
(b) a gene encoding Cre which is under the tr~ncrtiption~l regulation of an element responsive to multimers of the chimeric proteins; and (c) a floxed target gene which is susceptible to lr~ and rl;~ in the presence of Cre. The method thus involves exposing the cells to a 20 muL....I;~dL;on ligand capable of binding to the chimeric protein in an amount effective to result in detectable expression of the Cre gene.
In cases in which the cells are growing in culture, exposure to the ligand is effected by adding the ligand to the culture medium. In cases in which the cells are present within a host organism, exposing them to the 25 ligand is effected by ~ tr~ ;llg the ligand to the host organism. For instance, in cases in which the host organism is an animal, in particular, a mammal the ligand is dluu~;aL~l~d to the host animal by oral, buccal, sublingual, Ll u~ad~.~lldl~ >"1.. ~ , hlLId lluauuldl~ hlLI~ uua~ intrajoint or mhalatlon dL~lU~ dLIOn m an d~lU,U~ vehlcle therefor.
Where the target gene is essential for normal r~ I; .. ,;.,g in the healthy cell or animal, ubaLIu~iulg the expression of the gene, the effects of the gene product or rl;~ ;llg the gene entirely, using the materials and methods of 2 t ~ 2 wo 96106~

this invention, provides cellular and animal models for ~ullc~l~onl;llg disease states. For ~ample, where the target gene is the gene for clyLLI~o;~ its ub~LIu~,L;On or rl;~ - provides a model for the study of anemia and potential treatments for it. Targeting genes for nerve growth factors or 5 Lu.~l.ll.~L;Ilg proteins provides models of neural LE~ L;VC diseases; for bone L~lol~hù~,_.l;c factors, of u:~Leu~uluSiS~ for LLI~ u~ , of LLIulllbo~yLu~..;a, for antigen receptors or immune cell signaling proteins, of i"",. .,o~l r;. ;. . ~: for tyrosine Lydl~ yl~c~ of Parkinson's disease; etc. See Table 1.
This invention further ~ 1 or veterinary /n~ c for obstructing the expression or effects of a gene or rl;~
a gene from genetically engineered cells of this invention, including 11:...;., -~ ;..g the engineered cells themselves from animal tissue or from a subject containing such engineered cells. Such r~ or veterinary comrn~;tinnc 15 comprise a ... ~1. ;, ....; -, ;nl~ ligand of this invention in admixture with a l~h~ ly or ~.Le~hl u;ly acceptable carrier and optionally with one or more acceptable excipients. The ~.. 11 ;........ ;,-1 ;nn ligand can be a homo-..... 11;.,.. ;,_I;nn reagent or a hetero-mllll;.. - -;, l;~-n reagent so long as it is capable of binding to a chimeric responder protein(s) of this invention, 20 triggering expression of the blocking gene or triggering Cre expression in engineered cells of this invention. Likewise, this invention further ' a l~h~ "1 ;. ~l or veterinar,v rnmpo~ nn comprising a m1llrimPri7~tinn antagonist of this invention in admixture with a r~, ., .. . " ;. ~lly acceptable carrier and optionally with one or more 25 pl ,~, ., . ,1 ;. Ily or veterinarily acceptable excipients for preventing orreducing, in whole or paït, the level of m~ n of chimeric responder proteins in engineered cells of this invention, in cell culture or in a subject,and thus for preventing or reversing the activation of transcription of the blocking gene in the relevant cells. Thus, the use of the ",..1~ ;1....;, - ;n~.30 reagents and of the 1 ,.. 11 ;l,.. . ;,~1 ;nn antagûnist reagents to prepare pl.. .,. ... 1 ;. ~1 or veterinary compositions is ~.. l ~c,. l by this invention.
This invention also offers a method for providing a host organism, 21 ~7242 WO 96/06111 ~ J~JI ~I

, i, preferably an animal, and in many cases a mammal, susoeptible to regulatable ob~LI u~L;ull of a target gene in response to a " . 11 ;., ... i 1 ;nn ligand of this invention. The method involves Ul~lUd~ll.lllE, into the organism oells which have been engineered ex vi~o in aocordance with this invention, i.e. containing 5 a DNA construct encoding a chimeric protein hereof, and so forth.
AIU.IIGLi~IY~ one can introduce the DNA constructs of this invention into a host organism, e.g. m=al or embryo thereof, under conditions permitting transfection of one or more cells of the host mammal in vi~o.
We further provide kits for producing cells susceptible to ligand-10 regulated Ob~lU~,~iUII of a target gene. One such kit contains at least one DNA construct encoding at least one of our chimeric responder proteins, ~mrricing at least one receptor domain and at least one action domain (as described elsewhere). In one ~ o~ .1 the DNA construct contains a conventional polylinker to provide the ~ ;.. a site for the ;lll.. ~ OIGL;U
15 of cell-type specific expression oontrol element(s), such as promoter and/or enhancer elements, to provide for cell-type or tissue-specific expression of oneor more of the chimeras. The kit may oontain a quantity of a ligand of this invention capable of m~lltim~ri7ing the chimeric protein molecules encoded by the DNA constructs of the kit, and may contain in addition a quantity of a mlllli.,.. i,-.;.. antagonist, e.g.. ,..... ;~ ligand reagent. Where a sole chimeric protein is encoded by the construct(s), the mnltim. ri7qti~n ligand is a homo-" .11l:".. ;,_I;n,~ ligand. Where more than one such chimeric protein is encoded, a hetero~ ; " ..., . ;ml ligand may be included. The kit may further contain a blocking gene construct linked to a rran~rrirtinn control 25 element responsive to m~lltim~ ri7qtinn of the chimeric responder protein molecules. AILC~ IY~ the kit may contain a Cre construct linked to a control element responsive to, ..1~ ;nn of the chimeric responder protein molecules and/or a target gene construct containing the target gene flanked by loxP sequence. The kit may also contain at least one DNA construct encoding a primary chimeric protein containing at least one receptor domain and one action domain where one action domain may be the ~yLuplG~Illlc domain of Fas or of a TNF receptor as described below. The 2 ~ ~724;~
WO96/06111 r~

DNA constructs will preferably be associated with one or more selectable markers for. convenient selection of l "",~t. . ~ as well as other ~,UII~ ;UII.
vector elements useful for replication in luluLIyu~t~ for expression in eukaryotes, and the like. The selection markers may be the same or different s for each different DNA construct, permitting the selection of cells which contain various i ..,..1.;., rl~ of such DNA construct(s).
For example, one kit of this invention contains vectors comprising a firstDNA construct encoding a first chimeric responder protein containing at least one receptor domain (capable of binding to a selected ligand) fused to a 0 ~I.III~,I;,~J~;UIIal activator domain; a second DNA construct encoding a second chimeric responder protein containing at least one receptor domain fused to a DNA binding domain; and optionally, a third DNA construct encoding a blocking gene under the rr~n~rriprir-~l control of an element responsive to the ",..1~ ;.,... ;, 1 ;r,n of the first and second chimeric responder proteins.15 AlitlL~ ly~ the blocking gene construct may contain a cloning site, for example, a polylinker sequence, in place of a pre-selected blocking gene to permit the lul.~L;L;O~ to insert any desired blocking gene.
The kit may also contain a DNA construct encoding Cre under the tr~n~rriptir~n~l control of an element responsive to the ",..11;,.... ;,~ n of the first and second chimeric responder proteins and optionally, a DNA construct encoding a target gene flanked by the loxP sequence, permitting deletion of the target gene in the presence of Cre. Alkl~l~L;~.ly, the fifth DNA construct may contain a cloning site in place of a target gene to permit the pll~.Li~;un~.to insert any desired target gene.
Clther kits of this invention may contain one, two, or more DNA
constructs encoding chimeric proteins in which one or more of the constructs contain a cloning site in place of an action domain, such as the Ll~ ;u~
initiation signal generator, tr~n~rriptinn~l activator, the DNA binding protein,or other domains, permitting the user to insert whichever action domain s/he wishes. Such a kit may optionally include other elements as described above, such as a DNA construct for a target gene under responsive expression control",.. ~ll;.. ;,,I;I~,n ligand, antagonist, etc.

2 1 ~4~
WO96/06111 r~

Any of the kits may also contain positive control cells which were stably ~ r .,...r.~ with coristructs of this invention such that they express a reporter gene such CAT (. l .l~ .l .... ;...l transferase), beta b,l ~ or any other ~l~fl~.lLly detectable gene product, in response to exposure of the 3 cells to ligand. Reagents for detecting and/or ~ ify;ll~ the expression of the reporter gene may also be provided.

Brief nPcrl iptinn Of ~he Figures Figure 1 is a diagram of the plasmid pSXNeo/IL2 ~iL2-SX). In NF-AT-SX, the H~ndIII-Clal DNA fragment from IL2-SX containing the IL2 enhancer/promoter, is replaced by a minimal IL-2 promoter, which confers basal and an inaucible element containing three tandem NFAT-binding sites. (These constructs are described in further detail below.) Figure 2 is a flow diagram illustrating the preparation of the intracellular signaling chimera plasmids p#MXFn and p#MFnZ7 where n indicates the number lS of binding domains.
Figs. 3A and 3B are a flow diagram of the preparation of the ~ ~tr~rrll~ r signalling chimera plasmid p#lFK3/pBJ5.
Figs. 4A, 4B and 4C are sequences of the primers used in the W~ .,L;ulls of the plasmids employed in the subject invention.
Figure 5 is a chart of the response of reporter constructs having different enhancer groups to reaction of the receptor TAC/CD3~ with a ligand.
Figure 6 is a chart of the activity of various ligands with the TAg Jurkat cells described in Example 1. For Figure 6B, see also Spencer et al., Science 262,1019, Fig 3 and caption, esp. 3B on p. 1020 therein.
Figure 7 is a chart of the activity of the ligand FK1012A ~, Figure 9B) with the extracellular receptor lFK3 ~KBPx3/CD3~.
Figure 8 is a chart of the activation of an NFAT reporter via signalling through a Illyli~L~yl.~ed CD3g/FKBP12 chimera.
Figs. 9A and 9B are the chemical structures of the allyl-linked FK506 variants and the cyclohexyl-linked FK506 variants, Ica~ Li~ly.
Figure 10 is a flow diagram of the synthesis of derivatives of FK520.
Figs. 11 A and B are a flow diagram of a synthesis of derivatives of FK520 wo 96/06111 2 l q 7 ~ 4 ~ 1 ~,lIU_,JI .1 amt chemical structures of FKS20, where the bottom structures are designet to bind to mutant FKBP12.
Figure 12 is a .I ~ depiction of mutant FKBP with 2 motified F~'i20 in the putative cleft.
Figure 13 is a flow diagram of the synthesis of L~LCIV~~of FKS20 and cy-closporin.
Figure 14 is a schematic Ic,ulc~cllLi~L;ull ûf the, I ~of chimeric proteins, illustrated by chimeric proteins containing an ;"".. ,.~ moiety as the receptor domain.
Figure lS depicts Lh~ld~ L.~kd ~ I;.. ,. of chimericproteins, showing ~ 1ly the triggering of a L~ .L;onsl initiatiorl signal.
Figure 16 depicts synthetic schemes for HED and HOD reagents based on FKS06-type moieties.
Figure 17 depicts the synthesis of (CsA)2 beginning with CsA.
IS Figure 18 is an overview of the fusion cDNA construct and protein MZF3E.
Figure 19 shows FK1012-induced cell death of the Jurkat T-cell line transfected with a Illyll~Luyl.~Lcd Fas FKBP12 fusion protein ~FF3E), as indicatet by the deaeased LI~ L;UIIaI activity of the cells.
Figure 20A is an analysis of cyclophilin Fas (and Fas cyclophilin) fusion conslructs in the transient L~ rccL;ull assay. MC3FE w~s shown to be the most effective in this series.
Figure 20B depicts T.,,. ,~ ;lir Fas antigen chimeras and results of transient eYpression ~ ...... ,L~ in Jurkat T cells stably ~I.ul~ru~ ed with large T-2s antigen. Myr: the myristylation sequence taken from pp60ffr' encoting residues 1-14 (Wilson et al., Mol ~ Cell Biol 9 4 (1989): 1536-44); FKBP: human FKBP12;
CypC: murine cyclophilin C sequence encoding residues 36-212 (Freidman et al., Cell 66 4 (1991): 799-806); Fas: intracellular domain of human Fas antigen encoting residues 179-319 (Oehm et al.,J. Biol. Chem. 267 lS (1992): 10709-1S).
30 Glls were eIe~LIU~UI~j~Cd with a plasmid encoding a secreted alkaline ~
reporter gene under the control of 3 tandem AP1 promoters along with a six fold molar el~cess of the ,r."...,....~ .. fusion construct. After 24 h (hours) the cells were stimulatet with PMA (SOng/mL), which stimulates the synthesis of the - 2~97242 ~ W096/06111 19 I~ r-.

reporter gene, and (CsA)2. At 48 hours the cells were assayed for reporter gene activity. Western blots were performed at 24 hours using anti-HA epitope antibodies.
Figure 21 depicts the synthesis of modified FKS06 type . ..~ .v S I. Discussion This invention provides chimeric proteins, organic molecules for " ... ll ;.... i~;.. p the chimeric proteins and a system for using them. The fused chimeric proteins have a binding domain for binding to the mnlrim~-ri7inp molecule and an action domain, which can effectuate a pLy~;vlo&;~.~l action or 10 cellular process. Preferably the m~ltimrri7ing molecule is a small organic molecule. The pLy~;olog;~l action or cellular process effectuated by , . " .11 ;.. ;, -l ;.~, of the chimeric proteins is generally L~ L;vll of a gene encwding a regulatable blocking gene. This gene may be a ~blocking" gene, which encodes a blocking factor such as an anti-sense message ....I,l. - ..l A~ y to, and 15 capable of interfering with L..u.~ .;on of a target gene or a ribo_yme, which is capable of preventing the e2~pression of a target gene. Alternatively, the blocking gene may encode an antibody moiety, such as a single-chain antibody, which is capable of binding to and preferably n..lLI.ll; ulg or blocking a biological effea of the gene product of the target gene. The blocking gene also may encode a 20 dominant negative gene product capable of blocking the effea of a target geneproduct, or it may encwde a protein whose action will lead to the elimination ofthe target gene entirely. An e~ample of the latter type would be the gene encoding the protein Cre, which produces Cre ICWlllblll~C and whose expression leads to elimination of a target gene d~lJI v~ tcly flanked by "lor~P" sequences in 25 the host cells. Alternatively, the constructs will encode gene products such as the .yLu~l~..,;c domain of the Fas antigen or TNF receptor, which allow the cells containing such constructs to be readily eliminated through apoptosis.
The basic concept of ligand-mediated mn~ is illustrated in figure 14. Divalent ligands which can function as heLcl~ L;vll~ or hetero-.... 1l ~ ... ;, 1 ;.. 1, (~Dn) agents and ho.. ol;---e. ;~L;vn, or homo-,.... 11;.,... ; 1;,, ("HODn) agents are depicted as dumbbell-shaped structures.
The terms ~I~.... ....l:. . ...; n... .~ and "homo-,. .. .ll ;. , ...; ~ refer to the WO96/06111 21 9 7 2 4 2 r-~lu~J~

association of like '" 'l""' ' to form dimers or multimers, which may be linked, as shown in figure 14, by the multivalent ligands of this invention. Theterms "h~LrA~ ' " and "hetero-,.. 1l;."~; -I ~1." ref~ to the association of dissimilar '' l"' l~ to form dimers or higher order multimers. ~lomo-s multimers thus comprise an association of multiple copies of a particular~nmpnnrrlt while hetero-multimers comprise an association of oopies of different AI .1l ;.. ;~ ,1 .. 11'1, ~.,,.. 1l ;.~.. ;~. ~ and Vmultimer~, as the terms are used herein, with or without prefixes, are intended to encompass ''I~ ~, "dimerize~ and ~dimer~, absent an explicit indication to the contrary.
Also depiaed in Figure 14 and in Figure 8 of Spencer et al., are fusion, or chimeric~ protein molecules containing an action domain and one or more receptor domains that can bind to the ~ - - Iigands. Lil ~llUI LA
chimeric proteins, i.e., proteins which are intended to be located within the cells in which they are produced, will in some ~ o~l:..... ~ preferably, 15 contain a cellular targeting sequenoe (e.g. including organelle targeting amino acid sequences). Binding of the ligand to the receptor domains leads to .. -11: .. ,~-~; ... of the fusion proteins. l~r~ brings the aaion domains into close proximity with one another thus triggering aaivation of of a gene which is under the 1~ control of an 20 element responsive to the mllll :"... ;,~l i....
Cellular prooesses which can be triggered by receptor ,.. 11:........... ..;,.1 nll include a change in state, such as a physical state, e.g.. ,ru. ", -~ .. ,. Al change, change in binding partner, cell death, initiation of trancrriptionl channel opening, ion release, e.g. Ca+2 etc. or a chemical state, such as an cl~zyl~ ;~
25 catalyzed chemical reaction, eg. acylation, methylation, hydrolysis, ph~ ,holyl~ n or d.l,ho~l.llolyl~iion, change in redox state, ~ AIh.lll.lli, or the like. Thus, any such process which can be triggered by ligand-mediated mllll;~ _1 nA~ is included within the scope of this ~e~Lnolv~,y~ although the primary focus here is aaivation of trancrriptir,n, direaly or indirectly, of a 30 blocking gene, including a gene whose gene produa can funaion to block expression of a seleaed target gene or the eliminate the target gene from the cell.

2 t ~2~
~WO96/06111 21 P~

:
In a central feature of this invention, cells are modified so as to be responsive to ligand molecules which are capable of binding to, and thus .,.. 11;.. ; :.. ~, the primary chimeras disclosed herein. Such engineered cells rQpond to the presence of the ,.. 11 :.. :, :.. g ligand by activating ~
5 of a blocking gene via a L~ l;r ~ I control element responsive to the ... ,1. :.. .; I chimeric protein moleculQs.
The modified cells are . hAI A~ by a genome containing (a) a genetic construct (or series thereof) encoding a primary chimeric protein (or seriQ of primary chimeric proteins) of this invention, which permits ligand-10 regulated ~IJ'''' ';~ activation of a blocking gene via a ~IlQ~ ulgl control element, and (b) a blocking gene construct l ~~ p~ r a DNA sequence containing a blocking gene under the 1 nn ~1 control of a 1. ~ ~ . ;1 l :....al control sequence which is activated by ~ of the chimeric proteins. Where the regulated blocking 15 gene encodQ the protein Cre, the oells additionally contain (c) a target geneflanked by loxP DNA sequences, permitting r~cnmhin~ti~n and .-1:",:,. :..., of the target gene in the presence of Cre protein.
Only a single construct in the first seriQ will be required where a h,.,.,,,..,.,ll;.,..., including h.,",~.l;..,..~, is involved in aaivatingthe LIA~ n of a blocking gene, while two or more constructs are required where a L~ IUI~ is involved. The chimeric proteins encoded by the first seriQs of constructs will be associated with actua~ion of gene Ll~ n and will normally be directed to the surface membrane or the nucleus.
There are two main classes of genetic construas encoding the chimeric 25 proteins of this invention. In both cases the encoded chimeric proteins contain at least one action domain and at least one ligand binding domain.
The two classQ are described as follows: (1) constructs wherein a ligand-binding domain of the encoded chimera is either extracellular or intran~ lar~
but an action domain is intracellular, such that ligand-mediated, homo- or 30 hetero-mnltim~ri7~rinn of the chimeric protein molecules inducQ a signal which results in a series of events resulting in L~ 1 activation of the selected gene; (2) constructs wherein a ligand-binding domain and an action W096106111 2 1 ~24~ I ""~

domain of the encoded chimera is ;..- A-' rl. -' ~ such that ligand-mediated, homo- or hetero ' ~ of the protein moleNles induces initiation of "". directly via ..,...~ ;n, of multimers with the initiation region of the selected blocking gene.

5 II. T , R~7nl~ tinr~
The chimeric proteins encoded by the constructs of Groups (I) and (2), above, differ somewhat in their effects. Group ~1~ chimeras activate ~I U~ ;UII indirectly and can have somewhat pleiotropic effects; that is, they may activate a number of wild-type genes in addition to the introduced 10 blocking gene. ~ulLL~ v~c, the ~ activation following m~ imPri7~tinn of group (1) chimeric proteins may be slower in onset than in the case of Group (2) chimeras. Group (2) constructs activate 1~ ;"n more directly and in a manner more narrowly limited to the selected gene.
The ~ l activation in response to ".~ I ;r~n of Group (2) 15 chimeric proteins is typically very rapid.
The chimeric protein contains a "binding~ or "receptor~ domain which is capable of binding to at least one ligand moleNle. Since the .. .11 ;.. ~;,;..... ~, ligand is ...ul~ t, in the sense that it contains more than one receptor-binding site, it can form dimers or higher order homo- or hetero-multimers 20 with the chimeric proteins of this invention. The chimeric protein can have one or a plurality of binding sites, so that, for example, h...... ll ;... ' ~ can be formed with a divalent ligamd.
The chimeric protein also contains an "aaion~ domain capable, upon ligand-mediated mllltimPti7~tinn of the chimeric protein moleNles, of 25 initiating rrancrriptinn~ whether directly or indirectly.
The chimeric proteins, whether of Group (1) or (2) will typically also contain an intraoellular targeting domain comprising a sequence or ....... II.~n~ .. 1 which directs the chimeric protein to the desired c~ alllll..lL~ for example to the surface membrane in the case of Group (1), or the nucleus in the case of 30 Group (2).
By way of illustration, a Group (1) chimeric protein may contain a 2 ~ q7~2 ~ WO96/06111 .

myristate moiety as an ~ " ' targeting domain, three FKBPl2 moieties as receptor domains; and thè action domain comprising a T cell receptor ~
subunit. See for example, Spencer et al., Scienoe (1993). Group (2) chimeric proteins generally comprise a series of at least two chimeras, where the action 5 domains of one comprise DNA-binding domains, while the action domains of another comprise tryn~rtirt~ activating domains. I~f,.ll :.......... ;, 1 ;.~n of the Group (2) chimeras brings the DNA-binding domains and L. A~
activation domains in close proximity. S ~ ,l ..1 interaction of resulting multimers with the DNA sequence to which the DNA-binding domains bind 10 results in the initiation of ~ i-,.. of the gene associated with the responsive DNA element.
In either case, a gene encoding the desired anti-sense masage, ribozyrne, antibody moiety, dominate protein or protein capable of fnnrtirning in the .1;. . :, . ~ ;. ..~ of a gene sequenoe, such as the protein Cre, must be provided in a 15 manner such that it is under the L....~ l control of a DNA element responsive, directly or indirectly, to the addition of the ,.. 1~ :.. .;,;,.g ligand, i.e., to m~ , of the Group (1) or Group (2) chimeric protein molecules. In the case of Group (l) chimeras the blocking gene is linked to a .n~1 regulatory sequence acrivated upon m~ of the 20 chimera's action domain. For example, where the action domain is a T-cell CD3~ subunit, the blocking gene may be linked to NFAT sequence. Where the desired regulatable gene encodes the protein Cre, a l~ ....~l.;..A..I construct l,Ulll~ ;llg a floxed target gene is provided to serve as the object of the Cre-mediated lrl ,..,,l...,~ " to complete the overall system.

25 A. Surface ~ -~-r Receptor Group (1) chimeric proteins of this invention are typically associated with the surface membrane of the engineered cells. Addition of the mnlrimrti7ing agent to cells containing such proteins results in the generation of a signal leading to trAn~rriptinn of one or more genes. The process involves 30 a number of auxiliary proteins in a series of ;lllr~ r1l1min~ting in the binding of IIA~ factors to promoter regions associated with the ~ Q~4~ ' WO96/06111 P~,l/L_~ .J

selected gene(s) In cases in which the ~'n factors bind to promoter regions associated with other genes"~;U is initiated there as well. A
construct encoding a chimeric protein of this f..,l~...l;,....a can encode a signal sequence which can be subject to processing and therefore may not be present 5 in the mature chimeric protein. The chimeric protein will in any event comprise (a) a binding domain capable of binding a pre~t~tmin~l ligand, (b) an optional (although in many r~ o~ , preferred) membr~ne binding element or domain which includes a ~r~a~ .llLlAu~e domain or an attached lipid for ~rA~ .g the fused protein to the cell surface/membrane and 10 retaining the protein bound to the cell surface membrane, and, (c) as the action domain, a ~y~u~ alldc signal initiation domain. The ~y~u~l Lall~;c signalinitiation domain is capable of initiating a signal which results in l ~ ;
of a gene having a ~ U~;U;~;Un sequence for the initiated signal in th I initiation region.
The molecular portion of the chimeric protein which provides for binding to a membrane is also referred to as the "retention domain". Suitable retention domains include a moiety which binds directly to the lipid layer of the membrane, such as through lipid p.~ ;,u.~.;oll in the membrane or e~tending through the membrane, or the like. In such cases the protein 20 becomes l.. =l," l~d to and bound to the membrane, PA~ UIAIIY the cellular membrane, as depicted in figure 15. Also see figure 8 of Spencer et al.

B. Nuclear T~ tiOI~ Factorâ
Group (2) chimeric proteins may contain a cellular targeting sequence which provides for the protein to be IrAnc~ t~l to the nucleus. This isignal 2s consensus" sequence has a plurality of basic amino acids, referred to as a bipartite basic repeat as reviewed in Garcia-Bustos et al., Ri~lrhimi~A et Biophysica Acta (1991) 1071, 83-101. This sequence can appear in any portion of the molecule internal or pro2~imal to the N- or C-terminus and results in the chimeric protein being inside the nucleus. One ~mh~rlim~nt of this 30 invention involves at least two such Group (2) chimeric proteins: (l) one having an action domain which binds to the DNA of the LlDIID~I;,U~;Vn 2 t 97242 WO 96/06111 r~ iG591 Y ..
initiation region associated with the blocking gene in the blocking gene construct and (2) a different chimeric protein containing as an action domain, a L.. ~ 7 activation domain capable, in association with the DNA
binding domain of the first chimeric protein, of initiating ~ n of the 5 blocking gene construct. The two action domains or ~ .L ... factors can be derived from the same or different protein molecules.
The ~ . factors can be ~-n~ nmlc or exogenous to the celular host. If the l ~ n factors are exogenous, but functional within the host and can cooperate with the ~ .. O .. ~ RNA poly.. l~.~c, rather than 10 requiring an exogenous RNA pOly~ for which a gene could be introduced, then an exogenous promoter element functional with the fused L.~ - factors can be provided within the blocking gene construct for regulating l., \~ of the blocking gene. By this means the initiation of . can be restricted to the blocking gene associated with the L~t~.ulu~ promoter region.
A large number of L.~ L;..n factors are known which require two subunits for activity. AlLe~ L;~ly~ in cases where a single I~ factor can be divided into two separate functional domains (e.g. a L~ l;r ~
activator domain and a DNA-binding domain), so that each domain is inactive 20 by itself, but when brought together in close proximity, ~ L ",~l activity is restored. T. A~'- ' ;l" ;' ~1l factors which can be used include yeast GAL4, which can be divided into two domains as described by Fields and Song, supra. The authors use a fusion of GAL4~1-147)-SNF1 and SNF4-GAL4(768-881), where the SNFI and -4 may be replaced by the subject 25 binding proteins as binding domains. Combinations of GAL4 and VP16 or ~NF-1 cari be employed. Other ~ factors are members of the Jun, Fos, and ATF/CreB families, Octl, Spl, HNF-3, the steroid receptor superfamily, and the 7ike.
As an a7ternative to using the combination of a DNA binding domain 30 and a naturally occurring activation domain or modified form thereof, the activation domain may be replaced by one of the binding proteins associated with bridging between a Iran~rrirti~m~l activation domain and an RNA

.

21 ~7~2 polyl"~ , including, but not limuted to RNA polymerase Il. These proteins include the proteins referred to as TAF's (i , ~ 1 activation factors), the TFII proteins, p~l~iuul~uly B and D, or the like. Thus, one can use any one or ~....i.;.. - ;~.~. of proteins, for example, fused proteins or binding motifs 5 thereof, which serve in the bridge between the DNA binding protein and RNA polymerase and provide for initiation of ~ hll Preferably, the protein closest to the RNA poly.ll...~vc will be employed in ..,..j - .- I ;- ~.~ with the DNA binding domain to provide for initiation of L~ If desired, the subject constructs can provide for three or more, usually not more than 10 about 4, proteins to be brought together to provide the ~ IAI.'. . ;I.L:
initiation complex.
Rather than have a ~ n~1 activation domain as an action domain, an h...~ ;oll domain, such as ssn-6/TUP-1 or K. ul,l,.l L-l~;ly suppressor domain, can be employed. In this manner, regulation results in turning off the 1, A '~ of a gene which is uull~LiluLi~.ly expressed. For example, in the case of gene therapy one can provide for .,c,llvLi~uL;~.
expression of a hormone, such as growth hormone, blood proteins, ", ,.,....h~;l.lb-llin~, etc. By employing constructs encoding one chimeric protein containing a DNA binding domain joined to a ligand binding domain ~o and another chimeric protein containing an ;.. ~ ;oll domain joined to aligand binding domain, the expression of the gene can be inhibited via ligand-mediated... 1l;,., ;,-l;. .
Constructs encoding a chimeric protein containing int~r o,lza a ligand-binding domain fused to a ~IA -~ l activating domain or subunit, or rrAncrnptihn~l h~ Lil~; domain or DNA-binding domain are designed and assembled in the same manner as described for ~he other constructs.
Frequently, the N-terminus of the transcription factor will be bound to the C-terminus of the ligand-binding domain, although in some cases the reverse will be true, for example, where two individual domains of a single transcription 30 factor are divided between two different chimeras.

~ wo 96iO6111 2 ~ 9 ~ ~ 4 2 I._I/L..,.V_ ,1 III. C. , ~ Their Functions in the Chimeric Proteins There is a~.."~:~l. .AI ir amount of flexibility in the selection of ~" "I"'" ~ domains and their LU~,Ul,UUlO.~;Un into the design of chimeric proteins of this invention. For chimeric proteins intended for association with 5 the surface m~-m~r~n~-, if the ligand-binding domain is . ,....~ 1 ., the chimeric protein c;m be dQigned to contain an . I .. 11 .1~, domain selected from a variety of surface membrane proteins. Similarly, various uy I 1 or i....~ . domains of the surface membrane proteins which are able to transduce a signal can be employed, depending on which ~ "b "'"'' genes are 10 regulated by the ~yLu~ul~lu~ portion. Where the chimeric protein is to be internal to the cell, internal to the surface membrane protein or associated with an organelle, such as the nucleus or a ~yLu~ u~ vQicle, the ligand-binding domain protein will preferably be one which can bind moleculQ able to cross the surfaoe membrane or other membrane, as ~lu,u~u,u~;~Le. ThQe 15 binding domains will generally bind to ~ulLil~.IL Iigands uuIululla;llg naturally oo. urring or synthetic ligand moietiQ, which are preferably not nucleic acids or peptidic.
A. CYLUPI~ Dornains for Group (I) Chimeras A chimeric prouin of Group (1) can contain as an action domain, a 20 ~yLu,ulaaIlu~ domain from one of the various cell surface membrane receptors or variants thereof for which a .uIlQ~uonl;llg recognition sequence is known or available and which is capable of initiating LI u~ u~ion in rQponse to ,1.. 1l ... ;, 1 ;rln of the chimeric protein. Such recognition sequences include those associated with a gene responsive to rran~rriprir,n ll activation triggered 25 by such a receptor. Mutant receptors of interest will dissociate ~ ;1.L;~
activation of a selected gene from activation of genes which can be associated with harmful side effects, such as deregulated cell growth or h~ p~u~u~;~Le release of cytokines. The receptor-associated ~ym~ ;c domains of particular interQt will have the following .1~ receptor activation leads to 30 initiation of ~ "~ for relatively few (desirably fewer than lûû) and generally innocuous genQ in the cellular host; the other factors necessary for L~ L;~ initiated by receptor activation are prQent in the cellular host;

WO 96106111 2 ~ ~ ~ 2 ~ 2 genes which are activated other than the selected gene, will not affect the intended purpose for which these cells are to be used; mlll~ ;".. .; l ;. .,. of the ~yL~pl~~ domain or other available mPr~ni~m results in signal initiation;
and joining of the ~y u~pLI,~;~ domain to a desired ligand-binding domain will s not interfere with signalling. A number of different ~y .c,pl.wl..lc domains are known. Many of these domains are tyrosine kinases or are comple~ed with tyrosine kinases, and include CD3~, IL-2R, and IL-3R, among others. (See Cantley, et al., Cell (1991) 64, 281.) Tyrosine kinase reoeptors, which are activated by cross-linking, or .l;.... i,~ n include subclass I: EGF-R, 10 ATR2/neu, HER2/neu, HER3/c-erbB-3, Xmrk; subclass II: insulin-R, IGF-l-R (insulin-like growth factor receptor), IRR; subclass In: PDGF-R-A, PDGF-R-B, CSF-1-R (M-CSF/c-Fms), c-kit, STK-l/Flk-2; and subclass IV: FGF-R, flg [acidic FGF], bek [basic FGFl); n..~luLIu~ tyrosine kinases: Trk family, includes NGF-R, Rorl,2. (Based on n...~ first proposed by Yarden and Ulrich, Ann. Rev. Biochem. (1988) 57, 443.) Receptors which associate with tyrosine kinases upon cross-linking include the CD3 ~-family: CD3 and CD3 17, which are found primarily in T oells, and associate with Fyn; ~
and y chains of Fc~RI, which are found primarily in mast cells and basophils;
y chain of FcyRIII/CD16, which is found primarily in Ill.~.l.,~h..,~, 20 neutrophils and natural killer cells; CD3y, o, and ~, which are found primarily in T cells; Ig-o!/MB-1 and Ig-3/B29, which are found primarily in B cells.
Many cytokine and growth factor receptors associate with common ,~ subunits which interact with tyrosine kinases and/or other signalling molecules and which can be used as ~y~pl~lll;c domains in chimeric proteins of this 2s invention. These include (1) the common ,B subunit shared by the GM-CSF, IL-3 and IL-S receptors; (2) the ,B chain gpl30 associated with the IL-6, leukemia inhibitory factor (LIE;), ciliary ll.~ ,.pL~ factor (CNTF), oncostatin M, and IL-11 receptors; (3) the IL-2 receptor y subunit associated also with receptors for IL~, IL-7 and IL-13, and possibly IL-9; and (4) the ~
30 chain of the IL-2 receptor which is hl mnlrg~llc to the ~y~ l~lll;c domain of the G-CSF reoeptor. -=
The interferon family of receptors which include interferons o~ and y 2 1 ~
~ wo 96/06~

(whidh can activate one or m~re members of the JAK, Tyk fanuly of tyrosine kinases) as well as the reoeptors for growth hormone, ..y~Lu~u;~,hl and prolactin (whidh also c;m activate JAK2) can also be used as souroes for ~yLu~l~uuc domains.
Other souroes of ~y~u,ulQIlu~ domains include the TGF-3 family of oell surfaoe reoeptors. (~viewed by Kingsley, D., Genes and Dt /~. 1994 8 133.) This family of reoeptors contains serine/threonine kinase activity in their uy~u~l~llu~ domains, which are believed to be activated by crosslinking.
The tyrosine kinases associated with activation and hl~l~...lo.l of 10 ~ factors are of particular inurest in providing specific pathways which can be controLed and can be used to initiate or inhibit expression of an ~ogenous gene.
The following table provides a number of reoeptors and .1, .~ ;fC
associated with the receptor and their nudear response elements that activate 5 rrancrrirrir,n of genes. The list is intended to provide exemplary systems (rather than an exhaustive list) for use in the subject invention.
In many situations mutated l y~u~ uc domains can be obtained where the signal whidl is transduced may vary from the wild type, resulting in a restricted or different pathway as compared to the wild-type pathway(s). For 20 ~ample, in the case of growth factors, such as EGF and FGF, mutations have been reported where the signal is uncoupled from cell growth but is still m~im~inPrl with c-fos ~?eters, et aL, Na~r~re (1992) 358, 678).
The tyrosine kinase reoeptors can be found on a wide variety of oells LLuu~,lluuL the body. In contrast, the CD3~-fatnily, the Ig family and the 25 IyllluhoLIle ~-chain receptor family are found primarily on l-..l~ U~U;~L;C
cells, pdl~luul.llly T-oells, ~cells, mast cells, basophils, Il~ lu~h~,_, neutrophils, and natural killer oells. The signals required for NF-AT
Ll~'"'' ',1~1;... I come primarily from the ~ (zeta) chain of the antigen reoeptor and to a lesser extent CD3 y, ~

2 1 ~242 Wo 96/06111 r~ r~ .1 Table 2 Ligand DNA Binding Gene Reference Element Factors Insulin cAMP LRFI jun-B Mol. Cell Biol. (1992), 12, and others responsive many 4654 element genes PNAS, 83, 3439 (cre) PDGF, SRE SRF/SR c-fos Mol. Cell Biol. (1992), 12, FGF, TGF EBP 4769 and others EGF VL30 RVL-3 Mol. Cell. Biol. (1992), 12, RSRF virus 2793 cjun do. (1992), 12, 4472 IFN-a ISRE ISGF-3 Gene Dev. (1989) 3, 1362 IFN-~ GAS GAF GBP Mol. Cell. Biol. (1991) Il, PMA and AP-I many Cell (1987) 49, 729-739 TCR genes TNF NF~cB many Cell (1990) 62, 1019-1029 genes Antigen ARRE-I OAP/Oc many Mol. Cell. Biol. (1988) 8, t-l genes 1715 Antigen ARRE-2 NFAT IL-2 Science (1988) 241, 202 enhancer The cytoplasmic domain, as it exists naturally or as it may be truncated, modified or mutated, will be at least about 10, usually at least about 30 amino acids, more usually at least about 50 arnino acids, and generally not more than about 400 amino acids, usually not more than about 200 amino acids. (See Romeo, er ~1., Cell (1992) 68, 889-893.) While any species can be employed, the species .... 1~",.. ,~ to the host cell is usually preferred. However, in many cases, the cytoplasmic domain from a different species can be used effectively. Any of the above indicated cytoplasmic domains may be used, as well as others - ~ 2 1 ~724~;
~ WO 96/06111 . . ~1ll.,,~., ~ ~ .1 which are presently known or may ~..h~. ~1 ly be discovered.
For the most part, the other chimeric proteins associated with ..... factors, will differ primarily in having a cellular targeting - sequence which directs the chimeric protein to the internal side of the nuclear s membrane and having ~ n factors or portions thereof as the action domains. Usually, the rr~An~rtirrirn factor action domains can be divided into DNA binding domains and activation domains. One can provide for a DNA
binding domain with one or more ligand binding domains and an activation domain with one or more ligand binding domains. In this way the DNA
10 binding domain can be coupled to a plurality of binding domains and/or activation domains. Otherwise, the discussion for the chimeric proteins associated with the surface membrane for signal tran~r!~-rrir~n is applicable tothe chimeric proteins for direct binding to genomic DNA.

B. Cellular Targeting Dornains IS A signal peptide or sequence providc-s for transport of a chimeric protein to the cell surface ml-nnhmAnP where the same or other sequences can result in binding of the chimeric protein to the cell surface membrane. While there is a general motif of signal sequences, two or three N-terminal polar amino acids followed by about 15-20 primarily hydrophobic amino acids, the 20 individual amino acids can be widely varied. Therefore, s~ t~nti~lly any signal peptide can be employed which is functional in the host and may or may not be associated with one of the other domains of the chimeric protein.
Normally, the signal peptide is processed and will not be retained in the mature chimeric protein. The sequence encoding the signal peptide is at the 5'-2s end of the coding sequence and will include the initiation mrtllirninr codon.The choice of membrane retention domain is not critical to thisinvention, since it is found that such membrane retention domains are sl~hct~nri~lly fungible and there is no critical amino acid required for bindingor bonding to another membrane region for activation. Thus, the membrane 30 retention domain can be isolated from any convenient surface membrane or ~yLu~ dc protein, whether ~ , .,...,c to the host cell or not.

~1 q72~
wo 96/06~

There are at least two different membrane retention domains: a ..1-, ... retention domain, which is an amino acid sequence which extends across the m~mhranf and a lipid membrane retention domain, which lipid associates with the lipids of the cell surface membrane.
For the most part, for ease of ~ vlla~l u~L;oll~ the ~ ." . .f domain of a ~y ~pLvllll~ domain or a receptor domain can be employed, which may tend to simplify the ~ll~ru~ L;Oll of the fused protein. However, for the lipid membrane retention domain, the processing signal will usually be added at the 5' end of the coding sequence for N-terminal binding to the membrane and, proximal to the 3' end for C-terminal binding. The lipid membrane retention domain will have a lipid of from about 12 to 24 carbon atoms, particularly 14 carbon atoms, more p~lL;~uL~ly myristoyl, joined to glycine. The signal sequence for the lipid binding domain is an N-terminal sequence and can be varied widely, usually having glycine at residue 2 and Iysine or arginine at residue 7 (Kaplan, et aL, MoL CelL BioL (1988) 8, 2435). Peptide sequenoes involving post-Ll~lsl.lL;cllldl processing to provide for lipid membrane bindingare described by Carr, et aL, PNAS USA (1988) 79, 6128; Aitken, et aL, FEBS
Lett. (1982) 150, 314; Henderson, et aL, PNAS USA (1983) 80, 319; Schu12~
et aL, Virology (1984), 123, 2131; Dellman, et aL, Nature (1985) 314, 3~4; and reviewed in Ann. Re~. of Biochem. (1988) 57, 69. An amino acid sequence of interest includes the sequence M-GS-S-K-S-K-P-K-D-P-S-Q-R. Various DNA
sequences can be used to encode such sequence in the fused receptor protein.
Generally, the Ll~ ....,l.,,..f domain will have from about 18-30 amino acids, more usually about 20-30 amino acids, where the central portion 25 will be primarily neutral., non-polar amino acids, and the termini of the domain will be polar amino acids, frequently charged amino acids, generally having about 1-2 charged, primarily basic amino acids at the termini of the 1 .~"~ ,.. .,~1.,.."~ domain followed by a helical break residue, e.g. pro- or gly-.

C. Tissue Specific Expression of the Chimeric Proteins It will be preferred in certain ~mho~limfntc that the target gene be regulatably eliminated in a cell-specific or tissue-specific manner. To achieve 2~ q7242 ~ WO 96/nV61 1 1 P~. I 1-).,,~ Iv~ -such specificity, one m~y render the expression of tbe cbimeric proteins cell-type specific Such specificity of expression m~y be ~ieved by linking one ore more of the DNA sequences encoding the c~imeric protein(s) to c cell-type specific i , ~ ' regulc~ory sequence (e.g. promoter/enhmcer).
5 Numerous cell-type specific i ~r ' I reguhtory sequences ~re known.
Od~ers m~y be obuined from genes which ire expressed in ~ cell-specific mcnner.
For e~mple, constructs for expressing the cbimeric proteins m~y conuin regul~tory sequences derived from known genes for specific expression in selected tissues. R.~ , excmples cre ubuhted below:
Tissue Gene Reference lens y2-cryst~llin Breiem~n, Ml., Chpoff, S., Ross~nt, J., Tsui, L.C., Golde, L M~ M~vell, I~I., Bernstein, A., Genetic Abhtion: ~rgeted e~pression of i ~o~in gene c~uses , . ' ' ' in ~r~nsgenic mice, Sciencc238 (1987) 1563-1565.

c~A-cryst~llin L~ndel, C.P., Zh~o, J., Bok, D., Ev,ns, G.A., Lens-specif~c expression of ~ ICW~ . ricin induces ..~ defeGs in ~he eyes of ~nnsgenic mice, Ge#c-sD~. 2 (1988), 1168-1178.

K~ur, S., Key, B., Ssock, J., McNeish, J.D., Akeson, R., Potter, S.S., T~rg~ed ~bl~ion of ~Iphl cryst~llin-cells producev lens~eficien~ eyes in tnnsgenic mice, ~ ' r los (1989) 613-619.

~BSTITUTE SHEET ff~LE 26) 2 1 ~
WO96/Vn6111 l~l/v~,~

pitvitlry Growtb Be~ringer, RR., M~tbews, L5., Pdrniter, RD., Brinster, . ' hormone R.L., Dw~rf mice produced by genetic ~bhtion of ceLIs growth ~ v ceils, Ga~cs Dcv. 2 (1988) ~53~61.
.

pm~ers Insulin- Ornitz, D M., P~ltniter, RD., H~mtner, RE., Brinst~, Elrs~se - ~ r Rl., Swift, G H., r ~ r~n~ , R.,J., Specific expression orll specific of rn l~v_ ~ growth fnsion in pm~e~tic ~n~r cells of tr~nsgenic rnice, h~lm 131 (1985) 600-603.

P~lmiter, R.D., Bebringer, RR., Qu~ife, CJ., M~well, F., M~xwell, I.H., Brinster, RL., Cekl line~ge ~ tion in trmsgenic mice by cell-specific espression of ~ toxin gene, CcLI 50 (1987) 435~3.

T cells lck promoter Cb~ffin, l~E., Be~ls, CR., Wivlcie, T.M., Forbush, K A.,Simon, Ml., Perlmutter, RM., EMBO Journ~ 9 (1990) 3821-3829.

B cells T o~ ' Borelli, E., Heym~n, R., Hsi, M., E~nns, R.M., T~rgetinglin hpp~ lignt of ~n inducible toxic pnenotype in ~ l cells, Proc.
cv~in Ahz~L Ac~d. Sa. USA 85 (1988) 7572-7576.

Heym~n, R~, Borrelli, E., Lesley, J., Anderson, D., Ricllmond, DD., B~ird, S.M., Hym~n, R., Ev~ns, R.M.
Thymidine kin~se ' ' ~e~tion of tr~nsgenic mice witn controlled ~ r ~ ~ . Proc NatL
Ac~L Sa. USA 86 (1989) 2698-2702.

~STITUTE SHEET (RU' E 26) ~ . ' .
~ WO 96106111 .

Schw~nn cells P0 promoter Messing, A., Behringer, RR., H~mm~ng, JP. P~lmiter, RD, Brinster, RL, Lemke, G., P0 promoter tirects e~pression of reporter ~nd to~in genes to Scbw~nn cells of tr~nsgenic mice, Nel~ron 8 (1992) 507-520.
Myeiin b~sic Mislcimins, R. Rn~pp, L., Dewq, MJ., Zbmg, X. Cell protein ~nt i ".~iL. e~pression of ~ i ~ ' O O~ene unter conrrol o1f tne myelin b~sic protein gene promot~
in tr~nsgenic mice, Br,~zn Ra. Deu. 65 (1992) 217-21.

sperm~tits prot~mine Brei=, M L., Rombol~, H., M~welL LH., R1intwortn, G.R., Bernstein, A., Genetic ~bl~tion in tr~nsgenic mice with ~ttenu~tet tipbtneri~ to~in A gene, MoL CclL BioL 10 (1990) 474~79.

Iung Lung surf~nt Ornitz, D M., P~lmiter, RD., H~mmer, R.E., Brinster, gene R L., Swif~, G H., ~ RJ., Specific e~pression of ~n l~__ ! growth fusion in p~ncre~tic ~cin~r cells of tr~nsgenic mice, N~ 131 (1985) 600-603.
~dipocyte P2 Ross, S.R, Br~ves, RA, Spiegelm~n, B.M., Targeted e~pression of ~ to~in gene to ~oipose tissue: tr~nsgenic mice resis~nt to obesity, Gn~ ~md Dcv. 7 (1993) 1318-2~.
muscle myosin light Lee, RJ., Ross, R.S., Rockm~n, HA, H~rr~s, AN, ch~in O'Brien, TX, v~n-Bilsen, M., Shubeit~, HE, R~ntolf, R., Brem, G., Prices ~ ~1., J. BioL C~ 267 (1992) 15875-85.

SUBSTITUTE SHEET (RUEr 26) wo 96106111 2 T 9 7 2 4 2 Alphs sctin Muscst, GE., Perry, S., Prentice, H. Redes, L. The hum~n skelet~l slphs~ctin gene is regulsted by 5 muscle-rpecific enh~ncer thst binds three nucle~r fsctors, Gene ~s~on 2 (1992)111-26.
nwrOnS i~ r- Reeben, M. Hsknekyto, M. Alhonen, L. Sinernrcs, R.
proteins Sssrms, M. Jsnne, J., T -r~ e~pression of rlt ~t , rl promoter driven reporter gene in trsnsgeric mice, BB~C 192 (1993) 4~70.
liver tyrosine sse, slbumin, r li, . . ' T~ -r of tissue specific promoters To identify the sequcnces th~t control the tissue- or cell-type specific 5 expression of ~ gene, one isolstes ~I genomic copy of the selected gene including sequences "upstresm" from the eYons th~t code for the protein.

5' flsnking sequences coding sequences These upstreim sequences sre then usu711y fused to ~n essily detect~ble 10 reporter gene like bet~ ..ci.i in order to be ~ble to follow the expression of the gene under the control of upstre~m regulstory sequences.

5' ll~nking sequences reporter gene I

To est~blish which upstresm sequences sre necesssry ~nd sufficient to 15 control gene expression in 5 cell-type specific msnner, the complete upstreunsequences sre introduced into the ceUs of interest to determine whether the initial clone cont~ins the control sequences. Reporter gene eYpressor is SUI~STITUTE S~tEET (~LE 26) 2 1 q7242 96/06111 1 ~, I/~J~ 71 monitored as evidence of expression.

s I I ----- I

I- ~ I -! - ~- ,, , -I

10 If these sequences contain the necessary sequences for cell-type specific expression, deletions may be made in the 5' flanking sequences to determine which sequences are minimally required for cell-type specific expression. This can be done by making transgenic mice with each construct and ".."~ ., ;,.g beta-gal expression, or by first examining the expression in specific culture cells, 15 with, v.,.~ .,. to ~pression in non-specific cultured cells.
Several successive rounds of deletion analysis normally pinpoint the minimal sequences required for tissue specific expression. Ultimately, these sequences are then introduced into transgenic mice to confirm that the expression is only detectable in the cells of interest.

20 D. Ligand Binding Domain The ligand binding (~limf~ri7~til~n~ or "receptor") domain of any of the chimeric proteins of this invention can be any convenient domain which binds to a natural or preferably, to a synthetic ligand. The location of the binding domain in the chimeric protein, as expressed and located within a cell, can be 2s internal or external to the oellular membrane, depending upon the nature of the chimeric protein and the choice of ligand. A wide variety of binding proteins are known, including receptors and binding proteins associated with the cytoplasmic regions indicated above Binding proteins for which ligands are known or may be readily produced are of particular interest. These ligands are 30 preferably small organic molecules. The receptors or ligand binding domains include the FKBPs and cyclophilin receptors, the steroid receptors, the 2 ~ 2 WO 96106111 ~ J..,àl L~LI~y~lulC receptor, the other receptors indicated above, and the like, as wellas ~unnaturalr receptors, which can be obtained from antibodies, pd~ uLly the heavy or light chain subunit, mutated sequences thereof, random amino acid sequences obtained by stochastic procedures, ....,.1,;..-1~.,; I syntheses, and the like. For the most part, the receptor domains will be at least about 50 amino acids, and fewer than about 350 amino acids, usually fewer than 200 amino acids, either as the natural domain or truncated active portion thereof.
Preferably the binding domain will be small, usually less than 25 kDa, to allow efficient """~F ~ ~.. ,., in viral vectors, and will be .. ;- n.~.. ;.... h~,.. ;~
10 Additionally, it should bind synthetically accessible, cell permeable, nontoxic ligands that can be configured for .1;.. ;,-l ;
The receptor domain of the chimeric proteins expressed by a host cell can be h~ clluL or ~ ra~ r, depending upon the design of the construct enco&g the chimeric protein and the availability of an ~lu~ ligand.
For l-yll~pho'o;~ ligands, the binding domain can be on either side of the membrane, but for l~ydl~pll;l;c ligands, p.u~;~uldlly protein ligands, the bin&gdomain will usually be external to the cell membrane, unless there is a transport system for ;",,", 1;,;"~ the ligand in a form in which it is availablefor bin&g. For an intracellular receptor, the construct can encode a signal 20 peptide and ~1 ulalll..lL~ domain 5' or 3' of the receptor domain sequence orby having a lipid attachment signal sequence 5' or 3' of the receptor domain sequence. Where the receptor domain is between the signal peptide and the e domain, the receptor domain will be extracellular.
The portion of the construct enco&g the receptor can be subjected to 2~ for a variety of reasons. The ~ ' b. '';'f~l protein can provide for higher binding affinity, allow for .1;~. .;",;,~.I;hn by the ligand of the naturally occurring receptor and the 1.. l ~,. .;~. ~I receptor, provide ~ ,o..u..;~;w to design a receptor-ligand pair, or the like. The change in the receptor can involve changes in amino acids known to be at the binding site, random 30 ... ~ ..- - ' using ~UIIIb;ll~ ;.ll techniques, where the codons for the amino acids associated with the binding site or other amino acids associated with . ,.. r.. , .. ;.~"~1 changes can be subject to ~.. , ~,.. -.~ by changing the codon(s) 2 ~ q72~
~ W0 96/O~
, 39 for the particular amino acid, either with known changes or randomly, expressing the resulting proteins in an .~Lu~uluuli~le pluL~yuLic host and then screening the resulting proteins for binding. Illustrative of this situation is to modify FKBP12's Phe36 to Ala and/or Asp37 to Gly or Ala to ~ u~ a 511hcti~n.on~ at positions 9 or 10 of FK506 or FK520 or related ligands. In particular, mutant FKBP12 moieties which contain Val., Ala, Gly, Met or other small amino acids in place of one or morejof Tyr26, Phe36, Asp37, Tyr82 and Pheg9 are of particular interest as receptor domains for FK506-type and FK520-type ligands containing .. I:r;., n~.C at C9 and/or C10.
Antibody subunits, e.g. heavy or light chain, particularly fragments, more p~uL;~ul~llly all or part of the variable region, or fusions of heavy and light chain to create high-affinity binding, can be used as the binding domain.
Antibodies can be prepared ag.unst haptenic molecules which are ,ul.y,;ologi.dly acceptable and the individual antibody subunits screened for binding affinity. The cDNA encoding the subunits can be isolated and modified by deletion of the constant region, portions of the variable region, of the variable region, or the like, to obtain a binding protein domain that has the .~ ulul;.~Le affinity for the ligand. In this way, almost any physiologically acceptable haptenic compound can be employed as the ligand or to provide an epitope for the ligand. Instead of antibody units, natural receptors can be employed, where the binding domain is known and there is a useful ligand for binding.
The ability to employ in vitro .1..ll ~,. .1. :~ or ...1l.1,;ll~
m<~lifir~rir~nc of sequences encoding proteins allows for the production of 25 libraries of proteins which can be screened for binding affinity for different ligands For example, one can totally randomize a sequence of 1 to 5, 10 or more codons, at one or more sites in a DNA sequence enco&g a binding protein, make an expression construct and introduce the expression construct into a unicellular ..u..uu.~,~..~..., and develop a library. One can then screen30 the library for binding affinity to one or desirably a plurality of ligands. The best affinity sequences which are compatible with the cells into which they would be introduced can then be used as the binding domain. The ligand WO 96/06111 PCT/US9~/10591 would be screened with the host cells to be used to determine the level of binding of the ligand to _,..1..".... ~ proteins. A binding profile could be defined by weighting the ratio of binding affinity to the ... l ,,. ..: I binding domain with the binding affinity to _.. 1.. ~,.. ~ proteins. Those ligands which 5 have the best binding profile could then be used as the ligand. Phage display terhniq~Pc~ as a n~ m;ring example, can be used in carrying out the foregoing.

E. ~ n The transduced signal will normally result from ligand-mediated ",..1~ n of the chimeric protein molecules, i.e. as a result of 1ll..l~ ;I.... ;~1 ;l~n following ligand binding, although other binding events, for e~ample allosteric activation, can be employed to initiate a signal. The construct of the chimeric protein will vary as to the order of the various domains and the number of repeats of an individu?l domain. For the 15 extracellular receptor domain in the 5'-3' direction of ~ , the construct will encode a protein comprising the signal peptide, the reoeptor domain, the 1..,,,~ ,...,.1.,~,.. domain and the signal initiation domain, whichlast domain will be intracellular, either nuclear or ~y~ l;c. However, where the receptor domain is ;I.~ different orders may be employed, 20 where the signal peptide can be followed by either the receptor or signal initiation domain, followed by the remaining domain, or with a plurality of receptor domains, the signal initiation domain can be ~dllJw;l~. I between receptor domains. Usually, the active site of the signal initiation domain will be internal to the sequence and not require a free carboxyl terminus. Either of 2~ the domains can be present in multiple copies, particularly the receptor domain, usually having not more than about S repeats, more usually not more than about 3 repeats.
For mlllrimPri7ing the receptor, the ligand for the receptor domains of the chimeric surface membrane proteins will usually be multimeric in the sense 30 that it will have at least two binding sites, each of which being capable of binding to a receptor domain. Desirably, the ligand will be a dimer or higher 2 ~ 4 2 W096106111 P~

order multimer, usually not greater than about tetrameric, of small synthetic organic molecules, the imdividual molecules typically being at least about 150 Dand fewer than about 5 kD, usually fewer than about 3 kD. A variety of pairs of synthetic ligands and receptors can be employed. For example, in ~,.. 1.~.1;.. 1~ involving natural reoeptors, dimeric FK506 can be used with an FKBP receptor, dimerized ~.y~lu~lJolhl A can be used with the cyclophilin receptor, dimerized estrogen with an estrogen receptor, dimerized ~luuu~ulL;~u;L with a l;lu~u~u~L;~u;d receptor, dimerized ~cL~ ;uc with the ~c~ld~y~l;nc receptor, dimerized vitamin D with the vitamin D receptor, and iO the like. Al~c~ ly~ higher orders of the ligands, e.g. trimeric can be used.
For ~.,.l~o~ involving unnatural receptors, such as antibody subunits, modified antibody subunits or modified receptors and the like, any of a large variety of . .." l,u~ can be used. A significant ~Ld~d~c~;~L;~ of these ligand units is that they bind the receptor with high affinity (preferably with a Kd S
15 10-8M) and they are able to be dimerized chemically.
The ligand can have different receptor binding molecules with different epitopes which are referred to as "HED~ reagents, since they can mediate hetero~ l:, .... ;,-l ;, .,~ or hetero-,.. l. ;., .. ;, -l ;..., of chimeric proteins having the same or different binding domains. For example, the ligand may comprise FK506 or an FK506-type moiety and a CsA or a ~y~lu~l~ulhl type moiety. Both moieties are covalently attached to a common linker moiety. Such a ligand would be useful for mediating the ...ulL.,.~ ion of a first and second chimeric protein where the first chimeric protein contains a receptor domain such as an FKBP12 which is capable of binding to the FK506-type moiety and 2s the second chimeric protein contains a receptor domain such as cyclophilin which is capable of binding to the ~y~lo~olhl A-type moiety.

IV. Cells For d~J~liu~L;ulls in which one wishes to engineer all or C111~Ct~nti~lly all cells of an organism, standard IlI;~,lU;llj~ ;UII of embryos or use of ES cells is 30 preferred. For other ~l,~.l i. l ;.., ,~ the cells may be procaryotic, but are preferably eucaryotic, including plant, yeast, worm, insea and ",..,"" l;.., 2 ~ q 72 4 2 wo96106111 P~ , v ~

Those cells may be ., ~ ; . cells, from any mammal of interest, particularly nin~lC~ such as horses? cows, pigs, dogs, cats, rats, mice and so forth. Among these species, various types of cells can be involved, such as osteoclasts, nct~nblq~tc~ neuronal, h~ -n'- I r, neural, ~ Lyu1al, s cutaneous, mucosal, stromal, muscle, spleen, 1c~ l. epithelial, rnrlrthrl;~l~ hepatic, kidney, pancreatic, b~L1.,L1~c,Lu1al, pulmonary, e~c.
cells, which include any of the nucleated cells which may be involved with the Iymphoid or ~ O1IO~V L;~ lineages and members of the T- and B-cell lineages, 11.~..,pl. b-~ and monocytes, myoblasts and fibroblasts 10 are of particular interest. Also of interest are stem and progenitor cells, such as L~.1~LUPU;~.;C, neural, stromal, muscle, hepatic, pulmonary and br~LI.~L.Lc~L;..dl progenitor cells.
The cells can be autologous cells, syngeneic cells, allogenic cells and even in some cases, xenogeneic cells. The cells may be modified by changing 15 the major l: u ""1 ~ hility complex ("MHC") profile, by L1~L;v~.~;llg ~?2-..u~..Jglol,ul;.. to prevent the formation of functional Class I MHC molecules, by L~L;v.~L;11b Class II molecules, by providing for expression of one or more MHC molecules, by enhancing or L1..~L;v~L;llb cytotoxic r~p~hiliti.oc by . . .
enhancmg or mhlbltmg the expressmn of genes assoaated wlth the cytotPxlc 20 activity, or the like.
In some instances specific clones or oligoclonal cells may be of interest, where the cells have a particular specificity, such as T cells and B cells having a specific antigen specificity or homing target site specificity.

V. Ligands ~5 A wide variety of ligands, including both naturally occurring and synthetic substances, can be used in this invention to effect "...l~ ;.,... ;, 1:..n of the chimeric protein molecules. Applicable and readily observable or measurable criteria for selecting a ligand are: (A) the ligand is physiologically acceptable (i.e., lacks undue toxicity towards the cell or animal for which it is 30 to be used), (B) it has a reasonable Ll~ Lic dosage range, (C) desirably (for~IJIi~L;ul1~ in whole animals, including gene therapy ~pFlir~t;r~nc), it can be 21 ~7242 ~ WO 96/06111 . ~ ~iC~I

~ .
uken orally (is suble in the ~;~OuulLe~L;IIll system and absorbed into the vascular system), (D) it c~n cross the cellular and other .. I~.,.U., as necessary, and (E) binds to the receptor domain with reasûnable affinity for thedesired ~prlir~tirn A ftrst desirable criterion is that the compound is relatively 5 pLy~;olu~lly inert, but for its activating capability with the receptors. The less the ligand binds to native receptors and the lower the proportion of total ligand which binds to native receptors, the be,tter the response will normally be. P uL;~ uL~Ily~ the ligand should not have a strong biological effect on native proteins. For the most part, the ligands will be non-peptide and nu"
10 acid.
The subject ~U~I~JUUIIL will fûr the most part have two or more units, where the units can be the same or different, joined together through a central linking group. The ~units~ will be individual moieties (e.g., FK506, FK520, cyclosporin A, a steroid, etc.) capable of binding the receptor domain. Each of 15 the units will usually be joined to the linking group through the same reactive moieties, at least in h.. ~.. l;.. ~ or higher order homo-multimers.
As indicated above, there are a variety of naturally-occurring receptors for small non-ulu~f~ u .~ organic molecules, which small organic molealles fulfill the above criteria, and can be dimerized at various sites to provide a ligand accor&g to the subject invention. Substantial mrr1ifir~tirnc of these UUIII~JUUlldS are permitted, so long as the binding capability is retained and with the desired specificity. Many of the ~UIIIpUUnL will be ll~ lu~d;~s~ e.g.
macrolides. Suitable binding affinities will be reflected in Kd values well below 10 ~, preferably below 10-6, more preferably below about 10-7, although bin&g 2s affinities below 109 or 10-1~ are possible, and in some cases will be most desirable.
Currently preferred ligands comprise multimers, usually dimers, of .ulul.uullL capable of binding to an FKBP protein and/or to a cyclophilin protein. Such ligands indudes homo- and h.L~ul~lulLhll~.~ (usually 2~, more usually 2-3 units) of cyclosporin A, FK506, FK520, and rapamycin, and derivatives thereof, which retain their binding capability to the natural or J. ..i~ binding domain. Many derivatives of such rr,mpollnrlc are already WO 96106111 21~ 2 r known, including synthetic high affinity FKBP ligands, which can be used in the practice of this invention. See Holt et al., J. Am. Che7n. Soc. 1993, 115, 992s,993s. Sites of interest for linking of FK506 and analogs thereof include positions involving annular carbon atoms from about 17 to 24 and substituent positions bound to those annular atoms, e.g. 21 (allyl), 22, 37, 38, 39 and 40, or 32 (cyclohexyl), while the sarne positions except for 21 are of interest for FK520. For ~ O~JUl;n, sites of interest include MeBmt, position 3 and position 8.
Of particular interest are rn~..lifi~tin"c to the ligand which change its 0 binding ~ ;rC, p ~L;l-u~ with respect to the ligand's naturally occurring receptor. f~~ n~ y~ one would change the binding protein to u~ .n.l ~ the change in the ligand. For example, one can modify the groups at position 9 or 10 of FK506 ~see Van Duyne et ~1. (1991) Science 252, 839), so as to increase their steric ICl.lUh~lll.ll~, by replacing the hydroxyl with 1~ a group having greater steric IC~IUhC~ , or by modi~ying the carbonyl at position 10, replacing the carbonyl with a group having greater steric ICU,U;I~ or flln~ .,. l;,;..g the carbonyl, e.g. forming an N-substituted Schiff's base or imine, to enhance the bulk at that position. Various filnntinrl~1iti~c which can be ~ull~..d~.lLly introduced at those sitcs are alkyl 20 groups to form ethers, acylamido groups, N alkylated amines, where a 2-LrLu~..hyl;nullc can also form a 1,3-oxazoline, or the like. Generally, the sllbstit... ntc will be from about I to 6, usually 1 to 4, and more usually 1 to 3 carbon atoms, with from I to 3, usually 1 to 2 1L~CIU_LUUI_~ which will usually be oxygen, sulfur, nitrogen, or the like. By using different derivatives of the 2~ basic structure, one can create different ligands with different .. r~.. -~ ;.~.. ~l rcu~uhc~ for binding. By n...l ~,~.8,;ng receptors, one can have different receptors of c~ksr mti~lly the same sequence having different affinities for modified ligands not differing S;~,~l;r;~ u-Lly in structure.
Other ligands which can be used are steroids. The steroids can be 30 mnlrinn~ti7~ ~I so that their natural biological activity is cllh~t~mi~lly diminished without loss of their binding capability with respect to a chimeric protein containing one or more steroid receptor domains. By way of non-limiting ~wo 96/06111 2 1 q 7 2 4 2 ~ ~ r~
4s example, ijlu~.u~ulLi~uilD and estrogens can be so used. Various drugs can also be used, where the drug is known to bind to a particular receptor with high affinity. This is particularly so where the binding domain of the receptor is known, thus permitting the use in chimeric proteins of this invention of only 5 the binding domain, rathe~ than the entire native receptor protein. For this purpose, enzymes and enzyme inhibitors can be used.

A. Linkers Various filn~ri~n~liti~-c can be involved in the linking, such as amide groups, including carbonic acid derivatives, ethers, esters, including organic and ~0 inorganic esters, amino, or the like. To provide for linking, the particular monomer can be modified by oxidation, Lydlu~yl.lL;ul~ n. reduction, etc, to provide a site for coupling. Depending on the monomer, various sites can be selected as the site of coupling.
The multimeric ligands can be DyllLL~;~I by any convenient means, 15 where the linking group will be at a site which does not interfere with the binding of the binding site of a ligand to the receptor. Where the active site for PIIYD;OIO~D;~I activity and binding site of a ligand to the receptor domain are different, it will usually be desirable to link at the active site to inactivate the ligand. Various linking groups can be employed, usually of from 1-3û, 20 more usually from about 1-20 atoms in the chain between the two molecules (other than hydrogen), where the linking groups will be primarily composed of carbon, hydrogen, nitrogen, oxygen, sulphur and phosphorous. The linking groups can involve a wide variety of fnn~if)n~liri~c, such as amides and esters,both organic and inorganic, arnines, ethers, thioethers, disulfides, quaternary 2s .,.Il~,~.n;l.ll. salts, hydrazines, etc. The chain can include aliphatic, alicyclic, aromatic or heterocyclic groups. The chain will be selected based on ease of synthesis and the stabilit,v of the multimeric ligand. Thus, if one wishes to maintain long-term activity, a relatively inert chain will be used, so that the mnlrimrrir ligand link will not be cleaved. Alttlll~ .ly, if one wishes only a 30 short half-life in the blood stream, then various groups can be employed which are readily cleaved, such as esters and amides, particularly peptides, where 2 4i ~
WO 96106111 ~ J

circulating and/or int~rr~ r proteases can cleave the linking group.
Various groups can be employed as the linking group between ligands, such as alkylene, usually of from 2 to 20 carbon atoms, azalkylene (where the nitrogen will usually be between two carbon atoms), usually of from 4 to 18 s carbon atoms), N alkylene azalkylene (see above), usually of from 6 to 24 carbon atoms, arylene, usually of from 6 to 18 carbon atoms, ardialkylene, usually of from 8 to 24 carbon atoms, bis-..ubu..,.u;do alkylene of from about 8 to 36 carbon atoms, ete. Illustrative groups include decylene, o~L~ yl~ll., 3-~ LyLlle~ 5-~l..yl~lle, N-butylene 5-~ulul~yl~lle~ phenylene, xylylene, p-10 d;l~lu~yll l~l .; .r bis-benzoyl 1~8-~p~ ro~ r and the like. ~ UILiV~ IIL
or other ligand molecules containing linker moieties as described above can be evaluated with chimeric proteins of this invention bearing ~ull~,und;llg receptor domains using materials and methods described in the examples which follow.

15 B. Ligand Ch~l...L~ Li.~
For intracellular binding domains, the ligand will be selected to be able to be transferred across the membrane in a bioactive form, that is, it will be membrane permeable. Various ligands are hydrophobic or can be made so by appropriate mr.~lifi~ ~ti~7n with lipophilic groups. P~IL;I UI.111Y~ the linking20 bridge can serve to enhance the lipophilicity of the ligand by providing aliphatic side chains of from about 12 to 24 carbon atoms. Altc~ y~ one or more groups can be provided which will enhanoe transport across the mPmhr~nP desirably without endosome formation.
In some instances, multimeric ligands need not be employed. For 2s example, molecules can be employed where two different binding sites provide for .l;",. .;,,li,." of the receptor. In other instances, binding of the ligand can result in a ~ullfu~lll.~dunal change of the receptor domain, resulting in activation, e,g. m~ n~ of the receptor. Other "~P. l.~ may also be operative for inducing the signal., such as binding a single receptor with a 30 change in ~,.,.r~..",.~8~. resulting in activation of the ~yLupl~lluc domain.

--WO 96/06111 I~_",J~ ." ., C. Ligand A
M..~ ... . Iigands can be used for reversing the effect of the m.-ltinn~r~r ligand, i.e., for preventing, inhibiting or disrupting multimer formation or ~ Thus, if one wishes to rapidly terminate the effect of cellular activation, a ~.. ,n~... Iigand can be used. Cullv~m~.. ly~ the parent ligand moiety can be modified at the same site as the multimer, using the same prooedure, exoept ~ ;"g a ,.,.. ,~.~.. I .nal compound for the polrruu~io.ml ~ _ ' Instead of the POIY~",L~ ,"...~nr. ..:. c p u~;~ uLly of from 2 to 20 (although they can be longer), and usually 2 to 12, 0 carbon atoms can be used, such as ethylamine, hw,yLIuue, b~ yl~l..lle~ etc.
AIL~ V~IY~ the Luouvvdl~.lL parent compound can be used, in cases, or at dosage levels, in which the parent compound does not have undue ,..,.1. I lr pllr~;olo~i~l activity such as ;l l ll ~ c~;OII, ~D toxicity, and so forth.

15 D. Ilhlstrative hetero . .11. - :,;..~, (HED) and horno ' ~ ~ -(HOD) reagents with "bumps" that can bind to mutant receptors . ~ g r mUtatiOnS
AS disalssed above, one can prepare modified HED/HOD reagents that will fail to bind appreciably to their wild type receptors (e.g., FKBP12) due to20 the presenoe of ~ ("bumps") on the reagents that sterically clash with sidechain residues in the receptor's binding pocket. One may also make ~u~c~luv~Lug receptors that contain mutations at the interfering residues (".. 1.. ~ ~ ., y mutations") and therefore gain the ability to bind ligands with bumps. Using "bumped" ligand moieties and reoeptor domains bearing 25 ~ c ry mutations should enhance the specificity and thus the potency ofthese reagents. Bumped reagents should not bind to the ~nfl~grnrllc wild type receptors, which can otherwise act as a "buffer" toward dimerizers based on natural ligand moieties. In addition, the generation of novel receptor ligand pairs should C:..,..ll~.,..,~.~lr yield the HEI) reagents that will be used when30 hLelu~1 .. . :, 1 ;on is required. example, regulated vesicle fusion may be achieved by inducing the hete..J~1;. I .. : . ;r~rJ of syntaxin (a plasma membrane ~, q724~
wo 96106~

fusion protein) and sy~ ub.~hl (a vesicle membrane fusion protein) using a HED reagent. This would not only provide a research tool, but could also serve as the basis of a gene therapy treatment for diabetes, using .l~lv~JI;.l~ly modified secretory cells.
As an illustration of "Bumped FK1012s" C10 acetamide and formamide derivatives of FK506 were prepared. See Figure 16 and Spencer et al., "Controlling Signal T. . ~ l l 8~n with Synthetic Ligands,~ Scienoe 262 (1993) 1019-1024 for additional details concerning the syntheses of FK1012s A-C and FK506M. Two classes of bumped FK1012s were also created: one with a bump at C10 and one at C9. The R- and S-isomers of the C10 acetamide and formamide of FK506 have been ~yl~L~ d aocording to the reaction sequence in Figure 16B. These bumped derivatives have lost at least three orders of magnitude in their binding affinity towards FKB1~12 ~igure 16B). The affinities were ~. rPrmin~l by measuring the ability of the derivatives to inhibit FKBP12's rotamase activity.
An illustrative member of a second class of C9-bumped derivatives is the spiro-epoxide (depicted in Figure 16C), which has been prepared by adaptation of known procedures. See e.g. Fisher et al., J Org Chem 56 8(1991): 2900-7 and Edmunds et al., Tet. Lett. 32 48 (1991):819-820. A particularly interesting series 20 of C9 derivatives are ~ L~II~d by their sp3 hyb~id;~ ;OII and reduced oxidation state at C9. Several such l ulu~uullla have been ~yllLll~ d according to the reactions shown in Figure 16C.
It should be appreciated that h.~cl Ud;ll.~..S (and other hetero-mnltim~ri7. rc) must be ~..ull~LIu~,~a.l differently than the homorlim~ rc, at least 2s for Appli~ ~ ;....c where hom~ulim~r II.I~ could adversely affect their successful use. One illustrative synthetic strategy developed to overcome this problem is outlined in Figure 16D. Coupling of mono alloc-protected 1,6-1,. .u~. .1: ~ ..;..~ (Stahl et al., J Or~ Chem 43 11 (1978): 2285-6) with a derivati_ed form of FK506 in methylene chloride with an excess of L~;..Lyl ulullc gave an alloc~mine-substituted FK506 in 44% yield. This ;.. l~ can now be used in the coupling with any activated FK506 (or bumped-FKSû6) molecule.
D~lvLe~;vn with catalytic tetrakis-Ll;,uh..lyl~l.o~ e palladium in the 2 1 972~2 1~ WO96106111 r~l,.J~ 71 presence of dimedone at room Lcl~ Lulc in THF removes the anune protecting group. Immediate treatment with an activated FKS06 derivative, followed by desilylation leads to a dimeric product. This technique has been used to synthesize the illustrated HOD and HED reagents.

s E. Illustrative Cy~lu~ bo~r~ reagerlts Cyclosporin A (CsA) is a cyclic ...,,I,~. p. l~l ;,lf that binds with high affinity (6nM) to its j~tr~r~ r receptor cyclophilin, an 18 kDa .... n.~
protein. The resulting complex, like the FKBP12-FK506 complex, binds to and inactivates the protein pl~ f calcineurin resulting in the 0 ;I.. n~ yylc~;vc properties of the drug. As a further illustration of this invention, CsA has been dimerized vi~ its MeBmtl sidechain in 6 steps and 35% overall yield to give (CsA)2 (E~igure 17, steps 14 were conducted as reported in Eberle et al., J Org Chem 57 9 (1992): 2689-91). As with FK1012s, the site for .1;.... ;, 1 ;~... was chosen such that the resulting dimer can bind to IS two molecules of cyclophilin yet cannot bind to calcineurin following ~yluyll;l;u-binding. (CsA)2 was found to bind to cyclophilin A with 1:2 ~I.. ;. I.;~.. : ~y. Hence, (CsA)2, like FK1012s, does not inhibit signaling pathways and is thus neither ;......... u~ .ylc~;vc nor toxic at ~ull cuLl.~L;ullS
useful for practicing this invention.

20 VI. Genes for R,g~ Blocking of Gene Expression or Function and for Gene Flimin Iti~n A. Tranc. riptioll Initiation Region for the Blocking Gene Constructs Blocking gene constructs encode gene products which are capable of blocking expression of target genes, the interfering with the function of the 25 gene products encoded by target gene or fl;,l,il, l ;.,g the gene from the host cell entirely. Blocking genes of this latter type may genes which encode gene products such as the Cre 1'''1.'.1.;.1-'~, whose expression leads to ~-limin~til~n of a target gene ~yyluyl;~lkly flanked by loxP sequences in host cells. Constructs encoding the blocking gene products of the present invention have a responsive 30 element in the 5' region, which responds to ligand-mediated m~ ;.n.; ~l;..n of ~ ~ q7242 wo 96106~
~0 the chimeric receptor protein, j,.c~u.. ably via the generation and ~ ~n of a ~ L;~Jn initiation signal as discussed infia. Therefore, it will be necessary to select at least one 1~ ;n~ itiation system, e.g. 1,, .~ -L rn factor, which is activated either directly or indirectly, by the l yLu~
domain or can be activated by association of two domains. It will also be necessary to select at least one promoter region which is responsive to tbe resulting l~ initiation system. Either the promoter region or the gene under its ~ 1 control need be selected. In other words, an action domain can be seiected for the chimeric proteins, which is encoded by a 10 "first~ series construct, based on the role of that action domain in initiat;ng rr~ncrrirtir,n via a given promoter or responsive element. See e.g. Section m(A) "Cytoplasmic domains~, above.
Where the responsive element is known, it can be included in the blocking gene construct to provide an e~ression cassette for integration into 15 the genome either as an episome or by chromosomal integration. It is not necessary to have isolated the particular sequence of the responsive element, solong as a gene is known, which is trancrriptirln~lly activated by the ~yLu~l~llllc domain upon natural ligand binding to the protein comprising the ~y~u~ lic domain. Homologous rcc~llb;l~.~L;on is then used for insertion of the gene of 20 interest d~,wll~LI~II from the promoter region such that the inserted gene is under the LIA~ I regulation of the endogenous promoter region.
Where the sequence of a specific responsive element is known, it can be used in ~t~njnnrtjr~n with a different L~ ;oll initiation region, which has other 'ddV.Lll~_b~ . or desired properties, such as a high or low rate of 1, All~
25 binding by particular tt3ncrrirticln factors, including tissue specific binding factors, and the like.
The expression construct will therefore have at its 5' end in the direction of transcription, the responsive element and the promoter sequence which allows induced transcription initiation of a target gene of interest, 30 usually a therapeutic gene. The trancrr;rtinn~l t~-rTnin~ir,n region is not as important, and can be used to enhance ~he lifetime of or make short half-lived mRNA by inserting AU sequences which serve to reduce the stability of the - 2 ~ 97242 ' ~ wo 96/06111 r~
51 ~
mR~A and, as a ~ , limit the period of activity of the protein. Any region can be employed which provides for the necessar,v rranc,~tirtin- ~1 ~; - - ., and as ..~ L~
The responsive element can be a single sequence or it can be a repeated s sequence, but it would usually not be repeated more than about S times, often not more than about 3 times.
T~m.. l~ge1~< ".. 1.;.. - ;-.. , can be used to remove or inactivate endogenous 1.~ .~ ;1. ...-1 control sequences, including promoter, enhancer and other responsive elements, which are responsive to the .... 11 i.,.. ;,-1 ~1l of 10 Group (1) and Group (2) chimeric proteins as discussed above. Al" 11~"
h ~ . can be used to insert responsive IIA-~ L;~
control sequences upstream of a desired ~,.rl~.~,.. ~ gene.

B. Target Gene A wide variety of genes can be employed as the target gene. The target 15 gene can be any sequence of interest, the absence of which provides a desiredphenotype. The target gene can encode a surface membrane protein, a secreted protein, a ~yLu~ ldc protein, or there can be a plurality of target genes which can express different types of products. The encoded proteins can be involved in homing, ~y~uL~ ;Ly, proliferation, immune response, ;... .n~..... .~ ,.y 20 response, clotting or dissolving of clots, hormonal regulation, or the like (See Table 1). The target gene may alLtlll..Li~ly encode a product of unknown function, in which case the modified cells or organisms in which the target geneis regulatably obstructed can be used to identify or study that function. In addition, the target gene need not be r ~ b~ ""'' to the engineered cell--in 2s some r~ it is a gene of an infective agent e.g. bacteria or viruses. In such ~mho~!irn~nt~ it is a viral or bacterial gene or gene product which is regulatably obstructed.

C. Blocking Gene In the practice of this invention there are several options and ~ 1.1. fle2~ibility with regard to choice of blocking agent. In all cases the -2 ~ ~7~
wo 96106~ .J/

gene encoding the blocking agent is expressed under the ~
regulation of an element responsive to the ... ~ ;.. of chimeric proteins, as described in detail elsewhere. The blocking gene may encode an anti-sense message or a ribozyme, an amtibody or related form thereof, or a 5 dominant negative form of the target gene product. Additionally, the blocking gene can be a gene that is capable of rl;~ the target gene from the host genome completely, such as the gene encoding the protein Cre, which produces Cre I .. !~ Expression of this l~ .. 1.;.. Ieads to rl;~ 8 .. of a target gene ~ u,u~ tly flanked by "loxP~ sequences in the host cells.

l O (i) Antisense messages and ribozymes for blocking target gene expression When the target gene sequenoe is known, its expression can be blocked by ligand-regulated expression of an antisense message or ribozyme. An amtisense message or a ribozyme contains sufficient sequence ...".I.i .... ~.y to the target gene such that it specifically recognizes the target message and blocks its expression. For a recent review containing useful b.. L~;Iuulld ;,,r,.. ~;
and guidance, see Altman, "RNA enzyme-directed gene therapy," Proc. Natl.
AGZ~ Sci USA 90 (1993) 10898-10900 and papers cited therein, including Yu et al., "A hairpin ribozyme inhibits expression of diverse strains of human ;,...... ~.~. r;,;.. y virus type 1," Proc. Natl. Acad Sci. USA 90 (1993) 6340-20 6344. See also Efrat et al., Ribozyme-mediated ~ l of pancreatic ,B-cellgln~okin~c~- expression in transgenic mice results in impaired glucose-induced insulin secretion., Proc. Natl. Acad. Sci. USA 91, 2051-2055.

(ii) Tn~r~ r E~yll ~ of Antibodies to Block Gene Function The function of a target gene product cm be blocked by ligand-2s regulated expression of an antibody, antibody fragment or other amibody likemoiety, that specifically recognizes the encoded target protein and blocks its cellular function. This is done by expressing a gene encoding a neutralizing or otherwise blocking antibody against the target gene product. By regulatably expressing such an antibody gene, one regulatably obstructs the r,.... ~ ..;,.g of 30 the target protein. As in other r~ O~ , this may be effected in a cell-type - 2t 97242 ~WO96/06111 I._l~U~.,~/~,I

spccific manner by expressing the chimeric proteins using a cell-type specific promoter.
In this way, only oells that express the ",..ll ;.... ;, I,lr chimeric proteins are capable of expressing the antibody, and do so only in the presence of or 5 following exposure to the m~ agent.
Examples of; ~ ; IA d Y ~pressed antibody moieties blocking a gene product function include an antibody to HIV I gpl20 protein expressed in a , ...,.,. 1.~.. cell (Marasco et al. PNAS 90:7889 1993) and an anti-p21ras antibody expressed in Xenopus oocytes (Biocca et al. BBRC 197: 422 1993).
10 Intr~Pllnl~r expression of the anti-gpl20 antibody blocked processing of the envelop precursor and reduced the infectivity of HIV-1 particles. The anti-ras antibody blocked insulin-mediated meiotic m~ r~tinn of Xenopus oocytes.
The antibody should be selected such that it binds to the target gene or gene product and blocks its cellular function. A preferred form of the antibody 15 is the so called single chain form, since this ensures that the heavy and light chains will associate within the cell. Other uu~r;6~ L;olls~ such as two chain antibodies, particularly two chain F,bs, or F~ fragments, or antibody chains fused to other proteins are also possible. The antibody may need to be targeted to the same .ull.lu...LIIl~ll. as the desired gene product to be disrupted. This can 20 be a~ul~ kcd by fusing a l~ nll sequence to the antibody coding sequence. This could be at the N-terminus, C-terminus, or embedded witkin the antibody moiety amino acid sequence.
The Lc~ -oloi;y of generating 1~ 18 ~ antibodies, including single-chain antibodies is well known in the art. For a recent review, see Huston et al., Int Rev. Imm~snol. 10 (1993) 195.

(iii) I..L~lr~.cll.c by A Dominant Negative Gene Product Protein-protein hlLcl~cL;ull~ that are critical for a cellular process can be selectively blocked by expression of a non-functional variant of one of the protein partners. For example, raf-1 is a serine/threonine protein kinase that 30 functions in gro vth ~actor-stimulated proliferation pathway (Schaap et al. J.
Biol. Chem. 268: 2û232 1993). It is composed to two domains, an N-terminal 21 ~72~

s4 regulatory domain and C-terminal kinase domain. Constitutive ~ A~lca~ n of the N-terrninal domain of p74raf-1 in cultured oells blocked induced by growth factors. This domain also interfered with an oncogenic - variant of p21ras. Such a system could be useful for models of cancer or the s role of growth faaors on cellular ~ l;L.~;ul~.
Other examples of dominant negative gene produas include certain variants of steroid receptors, growth faaor receptors having an inaaive protein kinase or lacking the protein kinase domain altogether, cell surface receptors having a non-funaional . Yrrarrllnl~r ligand binding domain or intrarrlllll~r 10 ~,ytul~ luc domain, ~ ;r,n faaor variants that lack a DNA binding domain and/or a ~l u~ai~ iOn domain.
Dominant negative proteins typically disrupt the normal funaion of a target protein by ~ g it away from its normal partner. Dominant negative proteins can be Wna~luaCI by random ",..~ by seleaive 15 deletion of gene segments, or by a rational protein r~ l.g where the domain structure and funaion of the protein is ~....l..~.}o~ Often the dominant negative protein is an inaaive version of a protein with enzymatic aaivity.
One important lc~luh~ .lL is that the dominant negative protein be 20 ~..cA~c~cd relative to its normal c~JUllLclt~ L~ The increased expression afforded by the ligand-regulated rran<rriprir,nal aaivation of our invention makes this a particularly useful application of the technology.

iv. Intracellular Expression of Cre to Elirninate a Target Gene The Cre construas and floxed target gene construas for use in this ~5 c uboL~ lL of the invention will generally be as described by Barinaga (1994)and by Gu et al. (1994), cited above, and references cited therein, with the proviso that the Cre-encoding DNA sequence will be linked to a promoter/enhancer sequence responsive to mlllrimrri7arirn of Group (1) chimeras or to a DNA sequence to which m--lrim~ri7r~1 Group (2) chimeras are 30 capable of binding and initiating Cre l l ~ll~ ' ;I'' l'''' The modified cells or organisms in which the target gene is regulatably wo 96~06~
5s eliminated can be used to identify or study the function of the gene that is eliminated either at the cellular level or at the level of the organism.
Cre constructs will have a responsive element in the 5' region, which responds to ligand-mediated mllll :.,...; -: ,., of the chimeric receptor protein, ~cau~ bly via the generation and ~IAI~.~dl,~ ;O~l of a lldll'- ';1'1 .~ initiation signal as discussed in~Sa. Therefore, it will be necessary to select at least one ;n~ initiation system, which utilizes at least one ~ "' factor, which is activated either directly or indirèctly, by the uy~uplAa1lu~ domain or can be activated by association of two domains. It will also be necessary to select at least one promoter region which is responsive to the resulting n A~ initiation system. Either the promoter region or the gene under its ~ I r~ control need be selected. In other words, an action domain can be selected for the chimeric proteins (encoded by a ~first" series construct) based on the role of that action domain in initiating ~ via a given promoter or responsive element. See e g. Section III(A) "Cytoplasmic domains", above.
Where the responsive element is known, it can be included in the Cre gene construct to provide an expression cassette for integration into the genomewhether as an episome or by incorporation into the CIIIUIIIOAUIUC~ It is not necessary to have isolated the particular sequence of the responsive element, solong as a gene is known which is tran~rriptinnAlly activated by the uyLuplAaullcdomain upon natural ligand binding to the protein ~UIII~ ;IIg the cyLu~l~111;c domain. Homologous 1~', .,,,1/,,. . :,~1~ could then be used for insertion of the gene of interest duw11~L1c~11 from the promoter region to be under the 1, A~ n~ regulation of the c11d~D~.1uus promoter region. Where the specif~c responsive element sequence is known, that can be used in ~mjnn~tinn with a different trancrriptifm initiation region, which can have other aspects, such as a high or low activity as to the rate of trancrrirtinn) binding of particular trancrriptinn factors and the like.
The expression construct will therefore have at its 5' end in the direction of 1 . All~ ' ;1'l ;"~, the responsive element and the promoter sequence which allows for induced LIAII~ JL;UII initiation of a target gene of interest, -WO96106~ 9 7 2 4 2 "~

usually a Ih. ~ gene. The l,~ ~..;1.1:...,~l 1~....:.. ~;.... region is not as important, and can be used to enhance the lifetime of or make short half-lived mRl~A by inserting AU sequences which serve to reduce the stability of the mRNA and, therefore, limit the period of action of the protein. Any region 5 can be employed which provides for the necessary LIA~ n~1 n .~ In~
and as .,~ c, 1~
Tke responsive element can be a single sequence or can be ",.. 11 :.. :,. ~1 usually having not more than about ~ repeats, usually having about 3 repeats.
I T.. nl.. ~,.. , ,~ .. 1- .. -l ;~ .. can also be used to remove or inactivate Ll~ ;Onal control sequences, including promoter and/or responsive elements, which are responsive to the u~ .;~Lion event, and/or to insert such responsive ~ I control sequences upstream of a desired rl.,ll.L,..1....c gene.
A wide variety of genes can be employed as the target gene to be eliminated from selected host cells. The target gene can be any sequence of interest, the absenoe of which provides a desired phenotype. The target gene can encode a surface membrane protein, a secreted protein, a uy~
protein, or there can be a plurality of target genes which can express different20 types of products. The encoded proteins can be involved in homing, cytotoxicity, proliferation, immune response, intl~mm~tnry response, clotting or dissolving of clots, hormonal regulation, or the like (See Table 1). The target gene may l~tlll.~Li~,ly encode a produa of unknown function, in which case the modified cells or organisms in which the target gene is regulatably 2s eliminated can be used to identify or study that function.
v. Regulated Apoptosis In many situations it may be desirable to kill the genetically modified Ir....,.l.:..~..l cells, such as where one wishes to terminate the treatment provided by the modified cells, where the cells have become neoplastic, 30 p~lL;cula~ly in a patient, where a genetic therapy has been deleterious rather than beneficial or where the removal of the cells from a subject, particularly a~cuu~l~kh~lL animal, after the engineered cells have expressed a desired protein, 2~ ~72~2 ~WO 96/06111 1 ~

, may be of interest for research. For this purpose one may provide for the expression of the Fas antigen or TNF receptor fused to a binding domiun. (See Watanable-Fukunaga et al., Nature 356 (1992) 31~317.) Cells containing such constructs are readily eliminated through apoptosis following exposure of the cells to a ligand capable of ol;~,. ,.. ;,;.. g the primary chimeras. Constructs encoding the primary chimera may be designed for ~un~LiLuL;Ye expression using uu~ L;uu~l materials and methods, so that the modified cells have such proteins on their surface or present in their cytoplasm. AIL~ Y~ one can provide for controlled expression, where the same or different ~1;~,.,.. ;,;.. ~;
10 ligand can initiate expression of the primary chimera and initiate apoptûsis. By providing for the .rLu~l,D.luc portions of the Fas antigen or TNF receptor in the cytoplasm joined to binding regions different frûm the binding regions associated with expression of a target gene of interest, one can kill the modified cells under controlled conditions.

15 VII. I~iudu_L;ûll of Constructs into Cells The constructs described herein can be introduced as one or more DNA
molecules or constructs, where there will usually be at least one marker and there may be two or more markers, which will allow for selection of host cells which contain the constructs. The constructs can be prepared in ~ull~..L;u..al 20 ways, where the genes and regulatory regions may be isolated, as d~lu~ Le, ligated, cloned in an ~lu~l;.LLe cloning host, analyzed by restriction or c~ neing~ or other convenient means. Particularly, using PCR, individual fragments including all or portions of a functional unit may be isolated, where one or more mutations may be introduced using "primer repair", ligation, in 25 v~tro ... ~ , etc as ~ u~ Le. Once the constructs are completed and have been .1. ~Il~ to contain the ~ lu~ e, desired sequences, they may be introduced into the host cell by any convenient means. The constructs may be integrated and packaged into non-replicating, defective viral genomes like s adenovirus, adeno-associated virus (AAV), or herpes simplex virus ~SV) or others, including retroviral vectors, for infection or rr~ncrlllrtieln into cells.
The constructs may include viral sequences for transfection, if desired.

WO96/06111 2 1 q 7 2 4 2 AIL~I.I...;~IY, the construct may be introtuced by fusion, elc~ ,p.,l..L;on, biolistics, i ' lipflf~f~ n or the like. The host cells will usuaLy be grown and expanded in culture before introduction of the construct(s), followed by the ~ L~ treatment for introduction of the construct(s) and 5 integration of the construct(s). The cells will then be expanded and screened by virtue of a marker present in the construct. Various markers which may be used a~l~afully include hprt, neomycin resistance, thymidine kinase, L~ bl~ y~LI resistance, etc.
In some instances, one may have a target site for h,,..,..l..b..,.~
,,. ",,,1,;" ~ where it is desired that a construct be integrated at a particular locus. For example, an ,.. fl.. " .. ~ gene, at the same locus or elsewhere, can be deleted and/or replaced with a ~ ,8.;~ construct of this invention, using materials and methods known in the art for hOIlloh~b~.la .,....,.1.;., -; ...
The ,,. f..,.l.;..~..l constructs of this invention also can be used to introduce the 15 floxed target gene into particular cells, using materials and methods known in this art. For h-.",.-l. ,,,.~"~ .,,, ...,1,;. -~ ;fln~ one may generally use either n or O-vectors. See, for example, Thomas and Capecchi, Cell (1987) Sl, 503-512;
Mansour, et aL, Nature (1988) 336, 348-352; and Joyner, et al., Nature (1989) 338, 153-156. Gu et al. (1994) provide additional methods which can be used 20 for introduction of the floxed target gene.
The constructs may be introduced as a single DNA molecule encoding all of the genes, or as different DNA molecules having one or more genes. The constructs may be introduced ~h~ ly or ~ UL;~IY~ each with the same or different markers. In an illustrative example, one construct would 25 contain a LL~ UL;~ gene under the control of a specific responsive element (e.g. NFAT), another encoding the receptor fusion protein comprising the signaling region fused to the ligand receptor domain (e.g. as in MZF3E). A
third DNA molecule encoding a homing receptor or other product that increases the efficiency of delivery of the therapeutic product may also be 30 introduced.
Vectors containing useful elements such as bacterial or yeast origins of replication, selectable and/or amplifiable markers, promoter/enhancer elements 72~

~wo 96/06~

for expression in prokaryotes or eukaryotes, etc. which may be used to prepare stocks of construct DNAs and for carrying out Ll u.~L~L;v~, are well known in the art, and many are commercially available.

VIII~ A ' ~ ~ of Cells and Ligands s The cells which have been modified with the DNA constructs may be grown in culture under selective conditions and cells ~which are selected as having the construct may then be expanded and further analyzed. For example, the polyll...d,e chain reaction may be used for verifying the presence of the construct in the host oells. Once the modified host cells have been 10 identified, they may then be grown in culture or introduced into a host organism, as d~ v~ te for the purpose for which they were developed.
Depending upon the nature of the particular modified cells, they may be introduced into a host organism, such as a mammal, in a wide variety of ways IT.... i .~ ,k ~ cells may be a~Luul~clcd by injection into the vascular 15 system, there being usually at least about 10~ cells and generally not more than about 101~, more usually not more than about 108 cells. The number of cells which are dLIIhli,Lclcd will depend upon a number of ~ , including the purpose of hlLIvlu~llg the modified cells, the life~ime of the cells and the;rn protocol used. For example, the number of d.LIIIII;~LIdL;ons, the 20 ability of the cells to multiply, the stability of the ~II..d~ L;. agent, the~Ly~;ologic need for the ~L..d~ lc agent, and the like will all be considered in...... i.. g the number of modified cells to be aLI.ill;,~clcd. With skin cells which may be used as a graft, the number of modified cells used would depend upon the size of the layer to be applied to the burn or other lesion. Generally,2s for myoblasts or fibroblasts, the number of cells will be at least about 10~ and not more than about 103 and may be applied as a dispersion, generally being injected at or near the site of interest. The cells will usually be in a pky~;olvg;~dlly-acceptable medium.
Instead of ex vivo mrflifir~ n of the cells, in many situations one may 30 wish to modify oells in vivo. For this purpose, various techniques have been developed for mr.~lifir~ir,n of target tissue and cells in vivo. A number of virus ~ 1 ~t242 ~vo 96106~

vectors have been developed, such as adenovirus and 1cLluvhua~a~ which allow for i r - and random integration of the virus into the host. See, for ~ample, Dubensky et al. (1984) Proc. Natl. Acad. Sci. USA 81~ 7529-7533;
E~aneda et al., (1989) Science 243,375-378; Hiebert et al. (1989) Proc. Natl.
Acad. Sci. USA 86~ 3594-3598; Harvoglu et al. (1990) J. Biol. Chem. 265~
17285-17~93 and Ferry, et al. (1991) Proc. Natl. Acad. Sci. USA 88,8377-8381.
The vector may be d~Lu11;aLt1~d by injeaion, e.g. h1~ v~ulduly or ;U~1 u~ua~uLly~ inhalation, or other parenteral mode.
In accordance with in vivo genetic .,.r,1l;r.. 1;n11~ the manner of the 10 mr.~lifr~tinn will depend on the nature of the tissue, the efficiency of cellular "~r..l;r;. 1;,.,~ required, the number of O~u1Lu11;L;~ ta modify the particular cells, the a~ucaa;lJ;l;Ly of the tissue to the DNA rrmpnc;tir,n to be introduced, and the like. By employing an attenuated or modified retrovirus carrying a target ~ L~ initiation region, one can activate the virus using one of the subject ~IA..~. ';I.i;'.l. factor constructs, so that the virus will be produced and transfect adjacent cells.
The DNA h1~1ulu L;v11 need not result in integration in every case. In some situations, transient l-lAIII~ - r of the DNA introduced may be sufficient. In this way, one could have a short term effect, where cells could be introduced into the host and then turned on after a lu1~ 1 1.. ,.~.~ time, forexample, after the cells have been able to home to a particular site.
The ligand providing for activation of the ~y~u,ul lalluc domain may then be d~L~t~ d as desired. Depending upon the binding affinity of the ligand, the response desired, the manner of dd11u11;aL1.lL;o1l~ the half-life, the 2s number of cells present, various protocols may be employed. The ligand may be d~luuu;aL~I parenterally or orally. The number of a~L1ull;aLl~;ull~ will depend upon the factors described above. The ligand may be taken orally as a pill, powder, or dispersion; bucally; sublingually; injected h1L1~auul.~ly, intr~prri~nn~lly, ~ ly~ by inhalation, or the like. The ligand (and 30 monomeric rr mpo1ln~T) may be formulated using ~u11~.1L;u..al methods and materials well known in the art for the various routes of dd~1u~1;aL~.~L;on. Theprecise dose and particular method of A~1..1;~8~l ~1 ;r~n will depend upon the 21 97~42 ~WO 96/V6111 . ~ 1V~Y

above factors and be rlPrprminpd by the attending physician or human or animal healthcare provider. 'For the most part, the manner of A. 1. ~;~' '~1 Al ;....
will be ~ l~ ".;. d, In the event that the activation by the ligand is to be reversed, the ......... ;. compound may be - ' ~ cd or other single binding site compound which can compete with the ligand. Thus, in the case of an adverse reaction or the desire to terminate the 11,.. ,1.. l ;f effect, the .. ~... ;r binding compound can be ' ~ cd in ary convenient way, pdlL;~Il~ly intravascularly, if a rapid reversal is desired. AIL~ IY~ one may provide for 10 the presence of an hI~L;~L;UII domain (or L. A I .~. ~ ;1~1 ;~ Il ~al silencer) with a DNA
binding domain. In another approach, cells may be eliminated through apoptosis via signalling through Fas or TNF receptor as discussed Example 4~B) below.
The particular dosage of the ligamd for any application may be 15 ~ ".;.,. d in accordance with the procedures used for therapeutic dosage ....~l,;l..r;,.y~, where ,- -:,.lr A ,-, of a particular level of expression is desired over an extended period of times, for example, greater than about two weeks, or where there is repetitive therapy, with individual or repeated doses of ligand over short periods of time, with extended intervals, for example, two weeks or more. A dose of the ligand within a plrlL Ir~l6l~ range would be given and monitored for response, so as to obtain a LhII~IG~IOII level .r1~ as well as observing therapeutic response. Depending on the levels observed during the time period and the therapeutic response, one could provide a larger or smaller dose the next time, following the response. This process would be 2s iteratively repeated until one obtained a dosage within the therapeutic range.
Where the ligand is chronically ddll~ LtlGd, once the I~ .lrlla.~lr dosage of the ligand is drrr~rminf rl, one could then do assays at extended intervals to be assured that the cellular system is providing the d~ ll;d~G response and level of the expression product.
It should be appreciated that the system is subject to many variables, such as the cellular response to the ligand, the efficiency of expression and, as d~ Jl;dLC~ the level of secretion, the activity of the expression product, the 2 ~ ~7242 -WO 96/06111 I ~ O:~YI

particular need of the patient, which may vary with time and .h~
the rate of loss of the cellulàr aaivity as a result of loss of cells or expression activity of individual cells, and the like. Therefore, it is expeaed that individual patient would be monitored for the proper dosage for the individual, s even if there were universal cells which could be ~ 1 to the population at large.
The subject m~rllr,rlr~lrgy and ~...~ iOI~ may be used for the study and/or treatment of a wide variety of conditions and in~irqtinn~ For example, ~ and T-cells may be used in the hl~..L;g~lun and/or treatmene of cancer, 10 infeaious diseases, metabolic ~ firirnriPc, ~ud;uv~_uLu disease, hereditary cr~qg~llqtir,n tl~-firiPnrirc A.. n, .,.. r diseases, joint d~,.. c.. ~;~. diseases, such as arthritis, pulmonary disease, kidney disease, endocrine ~ , etc.
Various cells involved with struaure, such as fibroblasts and myoblasts, may be used in the treatment and/or investigation of genetic ~Irfirirnri~c such as 15 connective tissue .1. r;r .... ;., arthritis, hepatic disease, etc.
The following examples contain important additional ;"r...,., ~ m~
~-mplifirqtinn and guidance which can be adapted to the practice of this invention in its various embodiments and the equivalents thereof. The examples are offered by way illustration and not by way limitation.

20 Examples Cellular Tl..-l~f ,l Ill.lLi.Jll, and Evaluation Example 1: Induction of Isolated IL-2 Enhancer-Binding T.~ ~ 'I''''''' Factors by Cross-Linking the CD3 Chain of the T-Cell Receptor.
The plasmid pSXNeo/ll ~ (IL2-SX) ~igure 1), which contains the placental 2~ secreted alkaline ph~.cphqtq~ gene under the control of human IL-2 promoter (-325 to +47; MCB(86) 6, 3042), and related plasmid variants (ie. NFAT-SX, NFB-SX, OAP/Oal-SX, and AP-1-SX) in which the reporter gene is under the transcriptional control of the minimal IL-2 promoter (-325 to -294 and -72 to +47) combined with synthetic oligomers containing various promoter elements 30 (ie. NFAT, NKB, OAP/Oa-1, and AP1, ~c~ ly)~ were made by three piece ligations of 1) pPL/SEAP (Berger, et aL, Gene (1988) 66,1) cut with SspI

21 972~2 md Hindm; 2) pSV2/Neo (Southern and Berg, J. h~oL AppL Gene~ (1982) 1, 332) cut with NdeI, blunted with Klenow, then cut with PvuI; and 3) various promoter-containing plasmids (ie. NFAT-CD8, B-CD8, cxl21acZ-Oct-1, API-L~CIF3H, or cx15IL2) (described below) cut with PvuI and Hindm. NFAT-S CD8 contains 3 copies of the NFAT-binding site (-286 to -257; Genes and Dev.
(1990) 4, 1823) and cxl21acZ-Oct contains 4 copies of the OAP/Oct-1/(ARRE-1) binding site (~CB, (1988) 8, 1715) from the human IL,2 enhancer, B-CD8 contains 3 copies of the NFB binding site from the murine light chain (I~MBO
(1990) 9, 4425) and AP1-LUCIF3H contains 5 copies of the AP-1 site (5'-TGA-CTCAGCGC-3') from the mPt~llnrl~ n~n promoter.
In each ~ r-l ~ nl.~ 5 ~g of expression vector, pCDL-SR ~CB 8, 466-72) (Tac-IL2 receptor-chain), encoding the chimeric receptor TAC/TAC/Z
(TTZ) (PI~S 88, 8905-8909), was co-transfected along with various secreted aL~aline pl.r~l.h~ ~based reporter plasmids (see map of pSXNeo/IL2 in Figure 1) in TAg Jurkat cells (a derivative of the human T-cell leukemia line Jurkat stably transfected with the SV40 large T antigen (Northrup, et aL,J. BioL Chem.
[19931D. Each reporter plasmid contains a .,-~ ol,l,...,..rl. V~ of the binding site for a distinct IL-2 enhancer-binding ~ L;OII faaor within the context of the minimal IL-2 promoter or, ILC.I1..Li~.lY, the intact IL-2 20 enhancer/promoter upstream of the reporter gene. After 24 hours, aliquots of cells (~ lu~hlldkly 105) were plaoed in microtiter wells containing log dilutions of bound anti-TAC (CD25) mAb (33B3.1; AMAC, Westbrook, ME).
As a positive control and to control for I ran~fi-~ti~m efficiency, ionomycin (1 ~Lm) and PMA (25 ng/ml) were added to aliquots from each ,... ,~r~
25 After an additional 14 hour in~lh~ri~m~ the cllprrn~t~nt~ were assayed for the alkaline pho~ t~c activity and these activities were expressed relative to that of the positive control samples. The addition of 1 ng/ml FK506 dropped all activity due to NFAT to b~h~,.u.,..d levels, .1 .,..",~ ..g that deactivations are in the same pathway as that blocked by FK506. Each data point obtained 30 was the average of two samples and the e"~,l;lll..lL was performed several times with similar results. See Figure 5. The data show that with a known .~ rtrarl~ lar reoeptor, one obtains an ..~,. u~ c response with a reporter gene 2 ~ 97~4~

and different enhancers. Similar results were obtained when a MAb against the TcR complex. (i.e. OKT3) was employed.
Example 2: Inhibitory Activity of the T.. ~.. p~.. ~.,IlL Drugs FK506 arld Cyclosporin A (CsA) or the Dimeric Derivative ('rmE ' FK1012A
(8), FK1012B (5), and CsA dim~ (PB-1-218).
Ionomycin (1 ~m) and PMA (25 ng/ml) were added to 105 TAg-Jurkat cells. In addition, titrations of the various drugs were added. After 5 hours the cells were Iysed in mild detergent (i.e. Triton X-100) and the extracts were incubated with the ,~ gol~ ll.ci l ~ substrate, MUG (methyl galactosidyl umbelliferone) for 1 hour. A glycine/EDTA stop buffer was added and the extracts assayed for nuolca.~ll.c. Each data point obtained was the average of two samples and the ~..;IU~,IIL was performed several times with sirnilar results. Curiously, FK1012B appears to augment mitogen activity slightly at the highest ~n~cllLl~L;ull (i.e. 5 ,~g/ml); however, a control ~ shows that FK1012B is not stimulatory by itself. See Figure 6.

Example 3. Activity of the Dimeric FK506 Derivative, FK1012A, on the Chimeric F~BP12/CD3 (IFK3) Receptor.
S ~g of the eukaryotic expression vector, pBJ5, (based on pCDL-SR
with a polylinker inserted between the 16S splice site and the poly A site), containing the chimeric receptor (lFK3), was co-transfected with 4 ~g of the NFAT-inducible secreted alkaline ~ reporter plasmid, NFAT-SX. As a control, 5 ~Lg of pBJ5 was used, instead of lFK3/pBJ5, in a parallel tranc~rtir~n After 24 hours, aliquots of each L~ c~L;~ containing d~ h~l."ely 105 cells were incubated with log dilutions of the drug, FK1012A, as indicated. As a positive control and to control for l.~ t ,1;~
efficiency, ionomycin (1 ,um) and PMA (25 ng/ml) were added to aliquots from each tran~rrirln After an additional 14 hour inr~ .tirn, the ~u~
were assayed for alkaline phosphatase activity and these activities were expressed relative to that of the positive control samples. The addition of 30 2 ng/ml FK506 dropped all ctim~ll.tir~nc to background levels, dllon~Ll~,L;ngthat the activations are in the same pathway as that blocked by FK506. Hence, ~WO 96/06111 2 ~ 9 7 2 4 2 . ~

FK506 or ~yluD~uliul will serve as effective antidotes to the use of these Each data point obtained was the average of two samples and the -l' ;.. ,- was performed several timQ with similar results. See Figure 7.

Example 4A. Acti~ity of the Dimeric FR506 Derivative, FKî012B, on the 5 l~Iy~;D~Jyl~tcd Chimeric CD3/Fl~BP12 (MZF3E) Receptor.
A number of approachQ to ligand dQign and synthQQ have been succQsfully J. ~,~...,~1.,.1~.1, including positive rQults with FK506-based HOD
reagents named aFK1012~s. FK1012s were found to achieve high affinity, 2:1 binding ,~..i I.i.., . ~y (Kd(l) 0.1 nM; Kd(2) - 0.8 nM) and do not inhibit 10 calcineurin-mediated TCR signaling. The ligands are neither Q~ivr" nor toxic (up to 0.1 rnM in cell culture). Similarly, we have prepared a ~y~lo,~Jolh~ A-based h.. f.. l;.l.~ . ;,;.. ~, agent, a(CsA)2" which binds to the CsA receptor, cyclophilin, with 1:2 ~1..;.1,;,. . . ~y~ but which doQ
not bind to r~lr;n~l1rin Thus, like FK1012s, (CsA)2 doQ not inhibit signalling 15 pathways and is thus neither ;~ Q~ nor toxic.
ThQe and other of the el~amplQ of ligand-mediated protein association resulted in the control of a signal ~ n pathway. In an illustrative case, this was ~ ~....l.l ~h~ l by creating an intracellular receptor comprised of a small fragment of Src sufficient for pu~u~ l.,L;ol.al myristoylation ~vl), the 20 cytoplasmic tail of zeta (~; a ~ uun-.lL of the B cell receptor was also used), three consc~uLiv~ FKBP12s (F3) and a flu epitope tag O Upon exprQsing the construct MZF3E (Figure 18) in human Uurkat) T cells, it was confirmed that the encoded chimeric protein underwent FE1012-mediated oli~,..,..- ,, l-rn The attendant ~ , LiWl of the zeta chains led to signaling via the endog,~nrllq 25 TCR-signaling pathway (Figure lS), as evidenced by secretion of alkaline hrcph~ ~cP (SEAP) in response to an FE1012 ~Cso - 50 nM). The promoter of the SEAP reporter gene wac constructed to be transcriptionally activated by nuclear factor of activated T cells (NFAT), which is assembled in the nucleus following TCR-signaling. FKlol2-induced signaling can be terminated by a 30 deaggregation procQs induced by a nontoxic, monomeric version of the ligand called FKS06-M.

WO96106111 2 ~ q 7 2 4 2 66 Specifically, 5 ,ug of the eukaryotic expression vector, pBJ5, containing a Illyl;~Luyl~e~l chimeric reOEptor was co-transfected with 4 ~Lg NFAT-SX. MZE, MZFlE, MZF2E and MZF3E contain 0, 1, 2, or 3 copies of FKBP12, lc~ ly~ dUwl~L.c_... of a lllyl;~LuyLI~eJ CD3 ~y~ulJl~lluc domain (see S Figure 2). As a control, 5 ~g of pBJ5 was used in a parallel ~ . All. r 1 ;l ... After 24 hours, aliquots of each L. ~ ~rr . I ;.... containing d~JlU duu~Lcly 105 OEl15 were incubated with log dilutions of the drug, FK1012B, as indicated. As a positive control and to control for l ' efficiency, ionomycin (1 ~m) and PMA
(25 ng/ml) were added to aliquots from each i ' rin After an additional 10 12 hour inr~h~tirm, the '"l ~ were assayed for alkaline pl.,.~
activity and these activities were expressed relative to that of the positive control samples. The addition of 1 ng/ml FK506 dropped all ~;.. l ~;r~ i~ to near b.~ ;lUUIIII levels, ~~ l ;..g that the activations are in the same pathway as that blocked by FK506. This result is further evidence of the 15 reversibility of the subject cell activation. Each data point obtained was the average of two samples and the - l ; . . ., was performed several times with similar results. See Figure 8. The lllyfl~iuyl~Lcd derivatives respond to lower uullccllLl.lL;ullS of the ligand by about an order of magnitude and activate NF-AT dependent Ll-..' ' ;1'1 .~n to ~ulll~dl.~blc levels, but it should be noted that 20 the ligands are different. Compare Figs. 7 and 8.
In vivo F1~1012-induccd protein dimerization (~ ril~l~ was then obtained that intr~rr~ r aggregation of the MZF3E reOEptor is indeed induced by the FK1012. The influenza l ~r ,.~.~ l epitope-tag (flu) of the MZF3E-construct was therefore exchanged with a different epitope-tag (flag-2~ M2). The closely related chimeras, MZF3Efl~, and MZF3Efl,~, were co-expressedin Jurkat T cells. Immunoprecipitation c~t...;lll..lL~ using anti-Flag-antibodies coupled to agarose beads were performed after the cells were treated with FK1012A. In the presenOE of FK1012A (l~M) the protein chimera MZF3Efl~6 interacts with MZF3EflU and is co-i..ul.u..op.~ ;LdL~d with MZF3E~e. In absence of FK1012A, no co-;lLulluuu~ dL;ull of MZF3EflU is observed.
Related ~ fll.l.llL~ with FKBP monomer constructs MZFlEflU and MZFlEflqp which do not signal., revealed that they are also dimerized by FK1012A. This 2 t 9724~
ilO 96/06111 ~J I
~ 67., reflects the lcLIIIh~ L for ~bblL~,..L;vn observed with both the cl~dob ~.us T
cell receptor and our artificial receptor MZF3E.
FK1012-ir duced protein-t~rosine pllv..~l~v~ iv~l The i,u ....11 ,1~.
domains of the TCR, CD3 and zeta-chains interact with ~ylu~ld ~lltl~ protein tyrosine kinases following antigen ~ .. l .~ ;~,u Specific members of the Src family (Ick and/or fyn) phu~llulyl~Lc one or more tyrosine residues of activation motifs within these intracellular domains (tyrosine activation motif,TAM). The tyrosine kinase ZAP-70 is recruited (via its two SH2 domains) to the tyrosine phu t~hL lyl.-~cl T-cell-receptor, activated, and is likely to be 10 involved in the further d~ LlL~u activation of phospholipase C. Addition of either anti-CD3 mAb or FK1012A to Jurkat cells stably transfected with MZF3E resulted in the lc~lu;L~ lL of kinase activity to the zeta-chain as measured by an in ~,itro kinase assay following hlllllL'llOplCI ', ~' ~nn of thet nfln~nml~ T cell receptor zeta chain and the M~F3E-construct, ~c~ .,ly.
15 Tyrosine phu~lJhulyl~L;on after treatment of cells with either anti-CD3 mAb or FK1012 was detected using l...,"n~lnn~l alpha-phosphotyrosine antibodies.
Whole cell Iysates were analyzed at varying times after ctim~ tinn A similar pattern of tyrosine-phu~Lolyl.~Lcl proteins was observed after ctim~ tion with either anti-CD3 MAb or FK1012. The pattern consisted of a major band of 70 kDa, probably ZAP-70, and minor bands of 120 kDa, 62 kDa, 55 kDa and 42 kDa.

Examplc 4(B): Rt~g~ inn of Plv~,...llllll.~ Cdl Dcath with T~ hilin-Fas Antigen Chimeras The Fas antigen is a member of the nerve growth factor (NGI')/tumor 2s necrosis factor ~lNF) receptor ~U~[;~lll;ly of cell surface receptors.
Crosslinking of the Fas antigen with antibodies to its cxtracellular domain activates a poorly und~ uod signaling pathway that results in plU~;ldllllll~,d cell death or apoptosis. The Fas antigen and its associated apoptotic signaling pathway are present in most cells including possibly all tumor cells. The 30 pathway leads to a rapid and unique cell death ~2 h) that is ~LdldLLcl;~l by condensed cy-toplasm, the absence of an ,,,n.~,,""-n,,y response and WO96/06111 2 ~ q 72 ~ 2 1~ OaYl ~

r",t;" .,~ ,r~l~ Of ~ rl v~,...al DNA, none of which are seen in necrotic cell death.
We have also developed a second, inducible signaling system that leads to apoptotic cell death. Like the MZF3E pathway, this one is initiatet by 5 activating an artificial receptor that is the product of a LUU~Li~uLi~ly expressed "responder" gene. However, the new pathway differs from the first in that our HOD reagents induce the synthesis of products of an fn~ pathway rather than of the product of a tranc~rt~l, inducible (e.g., reporter) gene.
Gaining control over the Fas pathway presents significant ul~ul-ull;L;..
10 for biological research and medicine. Transgenic animals can be designed with"death" responder genes under the control of cell-specific promoters. Target cells may then be chemically ablated in the adult animal by a~ . ing a HOD reagent to the animal. In this way, the role of specific brain cells in memory or cognition or immune cells in the induaion and ...-...~-.,- ..- of ~.,1,.;.. disorders could be assessed. Death responder genes may also be introduced into tumors using the human gene therapy technique developed by M. Blaese and co-workers (Culver et al., Science 256 5063 (1992) 15~0-2) and then ~ 1y activated by treating the patient with a HOD reagent in a manner similar to the "gancyclovir" gene therapy clinical trials recently 20 reported for the treatment of brain tumors. Finally, we ....a, ~ the co-ddlll;ll;~Lld~;Ull of a death-responder gene together with the therapeutic gene in the practice of gene therapy. This would provide a "failsafe" ~ to gene therapy. If something were to go awry, for example, if an integration -induced loss of a tumor suppressor gene which could lead to cancer were to 25 oo ur, the gene therapy patient could take a "failsafe" pill that would kill all transfected cells. We have therefore designed a system of orthogonal ,.",. ;,;,.g reagents for such purposes. Thus we provide for the use of one set of ligands and chimeric responder proteins for regulating apoptosis in the host cells, and another set for regulating the rran~rrip-i~n of therapeutic genes.
30 The ligands used for regulating transcription of a therapeutic Ol desired gene are designed or selected not cross-react and initiate apoptosis.
An exemplary chimeric cDNA has been ~unsLlu~Ll consisting of three 2~ 97242 ~ WO 96106111 ';~ . If Fl~BP12 domains fused to the ~yLupl~~ signaling domain of the Fas antigen (Figure 19). This construct, v hen expressed in human Jurkat and murine D10 T cells, can be induced to dimeri_e by an FK1012 reagent and initiate a signaling cascade resu]ting in FK1012-dependent apoptosis. The LDX for FK1012A-media~ed death of cells ~ransiently transfected with MFF3E is 15 nM
as fl~-rf-rminPfl by a loss of reporter gene activity (Figure 19; for a discussion of the assay, see legend to Figure 20). These data coincide with Ill~ulcul..lL~ of cell death in stably transfeaed cell lines. Since the stable 1l~" r ,~ representa I - - ~g population of cells, they have been used to ascertain that death 10 is due ro apoptosis rather ehan necrosis as evidenoed by membrane blebbing and ..-,-1- .,~ ....~ DNA ~r~ ) Because the transient Ll f protocol is more convenient, it has been used as an initial assay system, as is describedbelow.

Example 4(C): Rfg~ tinn of Plogl.~ul~ed Cell Death with Cyclulull;l;n ras 15 Antigel~ Chimeras We have also prepared a series of cyclophilin C-Fas antigen constructs and assayed their ability to induce (CsA)2-dependent apoptosis in transient expression assays (Figure 20A). In addition, (CsA)2-dependent apoptosis has been d~ulull~LIdLcd with humanJurkat T cells stably transfected wirh the most 20 active construct in the series, MC3FE (M ll,yl;~Luyl.~L;on domain of Src, C ecyclophilin domain, F - cytoplasmic tail of Fas, E - flu epitope tag). The cytoplasmic tail of Fas was fused either before of after 1, 2, 3, or 4 CU~ 1,UL;VC
cyclophilin domains. Two control constructs were also prepared that lack the Fas domain. In this case we observed that the signaling domain functions only 25 wben placed after the ~limf ri7qtinn domains. (The zeta chain constructs signal when placed either before or after the dhll..;,d.;ùn domains.) Both the e~cpression levels of the eight signaling constructs, as ascertained by Western blotting, and their activities differed u,u~llLiLdL;~.ly (Figure 20B). The optimal system has thus far proved to be MC3F~ The LDs~ for (CsA)2-mediated cell 30 death with IvIC3FE is ~200 nM. These data d. llull~LIdLl the utiliry of the cyclophilin-~yclu~ulhl im~frafrinnc for regulating intracellular protein wo 96106111 ~ ~ 9 7 2 4 ~ r~,,vv~l ~

association and illustrate an orthogonal reagent system that will not cross-react with the FKBP12-FK1012 system. Further, in this case, the data show that only djm~'ti7'~ti~n and not ~ b ~;u~l is required for initiation of signal l,,,.~.l... Ii.", by the Fas ~.y~U,U~ tail.
5 Mutation of the N-terminal glycine of the l~ly~;~Luyl.~ion signal to an alanine prevents ~y~ uyl.~;on and hence membrane lo~ 1 ;m~ We have also observed that the mutated construct (~MFF3E) was equally potent as an inducer of FK1012-dependent apoptosis, indicating that membrane ~ i7~1til'~n is not necessary for Fas-mediated cell death.

10 Example 5. Co..~L._ ~ of Murine Signalling Chimeric Protein.
The various fragments were obtained by using primers described in Figure 4. For identifying specific primer, reference should be made to Figure 4. A cDNA fragment of .~ u~l.~dy 1.2 kb ~u~ h~v the I-E chain of the murine class II MEIC receptor (Cell, 32, 745) was used as a source of the signal peptide, employing P#6048 and P#6049 to give a 70 bp SacII-XhoI
fragment using PCR as described by the supplier (Promega). A second fragment was obtained using a plasmid comprising Tac t[L2 receptor chain) joined to the L~ .IIbIAIIC and ~yLupLI~llc domains of CD3 (Pl\lAS, 88, 8905). Using P#6050 and P#6051, a 320 bp XhoI-EcoRI fragment was obtained 20 by PCR ~u~ hlg the ~IAl.~....,,l.~....r and cytoplasmic domains of CD3.
These two fragments were ligated and inserted into a SacII-EcoRI digested pBluescript (Stratagene) to provide plasmid SPZ/KS.
To obtain the binding domain for FK506, plasmid rhFKBP (provided by S. Schreiber, Nature (1990) 346, 674) was used with P#6052 and P#6053 to 2s obtain a 340 bp XhoI-Self fragment containing human FKBP12. This fragment was inserted into pR~ crrirt digested with XhoI and SalI to provide plasmid FK12/KS, which was the source for the FKBP12 binding domain. SPZ/KS was digested with XhoI, pho,~L.~L4,~v (cell intestinal alkaline ~ tA~, CIP) to prevent self-annealing, and combined with a lo-fold molar excess of the Xhol-30 SelI FK'oP12-containing fragment from FK12/KS. Clones were isolated that contained mrn~mArc, dimers, and trimers of FKBP12 in the correct r,ri.-nrArir,n 2 ~ 97242 . . .
WO 96106111 r~ n 71 The clones lFK1/KS, lFK2/KS, and lFK3~KS are comprised of in the direction of rr~nc~rirti~ n the signal peptide from the murine MHC class II
gene I-E, a monomer, dimer or trimer, Ic~ ly~ of human FKBP12, and the Ll....'~ and ~y LUIJI~IIIIC portions of CD3. Lastly, the SacII-EcoRI
5 fragments were excised from p~l , using restriction enzymes and ligated imto the polylinker of pBJ5 digested with SacII and EcoRI to cr~te plasmids lFKl/pBJ5, lFK2/pBJ5, and lFK3/pBJ5, lc~ ly. See Figs. 3 and 4.

E~mple 6 A. CUII~LIU~L;U.I of lntrr~ r Signaling Chimera.
A ulyl;.k~yldL;ull sequence from c-src was obtained from Pellman, et al., l~lature 314, 374, and joined to a . ~ AIY sequence of CD3 to provide a primer which was .. l l.. I ,y to a sequence 3' of the Ll~,.~ .. ,.1 domain,-narnely P#8908. This primer has a SacII site adjacent to the 5' terminus and a XhoI sequence adjaoent to the 3' terminus of the Illyl;~LuyL~;Dn sequence. The other primer P#8462 has a Sall IC.UI;II;L;UII site 3~ of the sequence ~:u~ ..y to the 3' terminus of CD3, a stop codon and an EcoRI
Iccu~,ll;L;ull site. Using PCR, a 450 bp SaclI-EcoRI fragment was obtained, which was comprised of the Illyl;~Luyl~L;ull sequence and the CD3 sequence fused in the 5' to 3' direction.- This fragment was ligated into S~cII/EcoRI-20 digested pBJ5(Xhol)(Sall) and cloned, resulting in plasmid MZ/pBJ5. Lastly, MZ~pBJ5 was digested with SalI, rh..~l.h.. ~-.1 and combined with a lo-fold molar excess of the XhoI-Sall FKBP12-containing fragment from FK12/KS and ligated. After cloning, the plasmids comprising the desired constructs having the lllyl;~Luyl.~Lhm sequence, CD3 and FBP12 multimers in the 5'-3' direction 25 were isolated and verified as having the correct struaure. See Figures 2 and 4.

B. Con~Llu.L~,ll of expression cassettes for hlLI...ell.lldl signaling chimeras The construct MZ/pBJ5 (MZE/pBJ5) is digested with restriction enzymes XhoI and SalI, the TCR ~ fragment is removed and the resulting vector is ligated with a 10 fold excess of a monomer, dimer, trimer or higher order multimer of FKBP12 to make MFlE, MF2E, MF3E or MFnE/pBJ5.

2 1 972~2 wo 96/06~

Active domains designed to contain compatible flanking restriction sites (i.e.
XhoI and Sall~ can then be cioned into the unique XhoI or SalI restriction sitesof MF=E/pBJ5.

~ Ex~mple 7. COII~L ~ of Nuclear Chimer~
5 A. GAL4 DNA binding domain - FKBP dornain(s) - epitope tag. The GAL4 DNA binding domain (amino acids 1-147) was amplified by PCR using a 5' primer (#37) that contains a SacII site upstream of a Ko_ak sequence and a sl~L;ull31 start site, and a 3' primer (#38) that contains a SalI site. The PCR
product was isolated, digested with SacII and SaII, and ligated into Fr'l lPsrrirt 10 Il RS (+) at the Sacll and Sall Sites, generating the construct pBS-GAL4. Theconstruct was verified by ~PqllPnring The SacII/SalI fragment from pBS-GAL4 was isolated and ligated into the IFK1/pBJ5 and IFK3/pBJ5 constructs (containing the ..lyfl~Lc~ ;on sequence, see Example 6) at the SacII and Xhol sites, generating constructs GFlE, GF2E and GF3E.

15 5' end of PCR amplifled product:

SacII ~ GAh4(1-147)---~M K L L S S
5' CGA Q CCGCGGCCACCATGA~GCTACTGTCTTCTATCG.

Kozak 3' end of PCR amplified product~
~<----GAh4~1-147)--~
R Q h T Y S
5' GACAGTTGACTGTATCGGTCGACTGTCG - .
25 3~ CTGTC~CTGACATAGCCAGCTGACAGC : :
. . = . =, SalI

B. HNF1 Dimrri7~ti~n/DNA- Binding Domain - FKBP Domain(s) - Tag.
The ~Fla ~im~ri7~rir,n/DNA binding domain (amino acids 1-282) 30 was amplified by PCR using a 5' primer (#39) thal contains a SacII site upstream of a Ko_ak sequence and a translational start site, and a 3' primer 21 q7242 ~WO 96/06111 r~

(#40) that contains a Sall site. The PCR produa was isolated, digested with Sac~ arld Sall, and ligated into rr~ II KS (+) at the SacU and Sall sites, generating the construct pBS-HNF. The construct was verified by , ~
The Sac~/Sall fragment from pBS-HNF was isolated and ligated into the 5 IFKl/pBJ5 and IFK3/pBJ5 constructs at the SacJI and XhoI sites, generating constructs HFlE, HF2E and HF3E.

5' end of PCR arnplified product:

SacII ~ NFl (1-281) -->~
M V S K L S
5 ' rr.ArArrrcGGCCACCA~ AAGcTGAGc Kozak 3' end of PCR amplified product:
<<----~ ----EIl~l (1--282~ ------------A F R H K L
5' C~llC~:~.~C~AA~ll~i~LC~ACTGTCG
3 ' GGaA~ L,l~;ll~A~CCaGCTGACAGC _l 20 C. FKBP domain(s)-VP16 ~ n activation domain(s)-epitope tag.
These constructs were made in three steps: (i) a construct was created from IFK3/pBJ5 in which the Illyl;~Luyl.L~;ull sequence was replaced by a start site h~ led;..k ly upstream of an XhoI site, generating construct SF3E; ~li) a nuclear 1~1i7'1ti--n sequence was inserted into the XhoI site, generating 2s construct NF3E; ~iii) the VP16 activation domain was cloned into the SalI site of NF3E, generating construct NF3VIE.
(i). (~,,"~1~1 ~ ~ y uliL,.~ .1. (#45 and #46) encoding a Kozak sequence and start site flanked by SacII and Xhol sites were annealed, phosphorylated and ligated into the Sadl and XhoI site of MF3E, generating 30 construct SF3E.

.. . .. :

W096~6111 2 ~ ~7242 PCT~S951105~1 ~
, 74 Insertion of generic start site Ko_ak _ M L E ,, .
5' GGCCACCATGC
5 3' ~C~lG~lACGAGCT
SacII XhoI
uv~ g overhang (ii). C~ y ~,1 ~,.",..~lr..~ r: (#47 and #48) encoding the SV40 T
10 antigen nuclear lrr~li7~tir,n sequence flanked by a 5' SalI site and a 3' Xhol site were annealed, pl.o~ olylatcd and ligated into the XhoI site of SFlE, generating the construct NFlE. The construct was verified by DNA
cPrillPnring A construct containing the mutant or defective form of the nuclear lrr~li7~ n sequence, in which a threonine is substituted for the Iysine at 15 position 128, was also isolated. This is designated NFlE-M. Multimers of the FKBP12 domain were obtained by isolating the FKBP12 sequence as an XhoI/S~lI fragment from pBS-FKBP12 and ligating this fragment into NFlE
lineari_ed with XhoI. This resulted in the generation of the constructs NF2E
and NF3E.

20 Insertion of NLS into generic start site T (ACN) L D P K K K R K V L E .=
5' TCGACC~T~r.~ Z~ TAC
3' GGGA~ L~ ATGAGCT
SalI XhoI

Threonine at position 128 results in a defective NLS.
~lii). The VP16 rrancrriptir,n~l activation domain (amino acids 413-490) was 30 amplified by PCR using a 5' primer (#43) that contains Sall site and a 3' primer (#44) that contains an XhoI site. The PCR product was isolated, digested with SalI and XhoI, and ligated into MF3E at the XhoI and Sall sites, generating the construct MVlE. The construct was verified by SPqllPnring MnltinnPri7P~I
VP16 domains were created by isolating the single VP16 sequence as a 096/06111 .

Xhol/Sall fragment from MVlE and ligating this fragment into MVlE
linearized with Xhol. Constructs MV2E, MV3E and MV4E were generated in this manner. DNA fragments encoding one or more multiple VP16 domains were isolated as Xhol/Sall fragments from MVlE or MV2E and ligated into S NFlE linearized with San, generating the constructs NFlVlE and NFlV3E.
Multimers of the FKBPl2 domain were obtained by isolating the FKBP12 sequence as an Xhol/Sall fragment from pBS-FKBP12 and ligating this fragment into NFlVlE linearized with Xhol. This resuited in the generation of the constructs NF2VlE and NF3VlE.

10 5' end of PCR amplified product:
SalI ¦--VP16(413-490)___>~
A P P T D V
5' cGAcAGTrr~rç~lc(~(cc~ Arrr7~TGTc 3' end of PCR amplified product:
~<---VP16~413-490)~
D F ~ G G
5' GACGAGTA~G~l~GG~l~AGTGTCG
3' cTGc~cATGrr~rrrr~r~rTcAcAGc Xho ol ;,;.., .. " 1. v~
#37 38mer/0.2um/OFF 5~rr~r~rrr7cGr7rr~rr~TGA~GcTA~
ATCG
=38 28mer/0.2um/OFF 5~cGAcAGTcGAcrr~T~r~r7TcAAcTGTc 2s =39 34mer/0.2um~OFF 5~cGAcAcrrrrçrr~rr~ L~AGcTGAGc =40 28mer/0.2um/OFF 5~cGAcAGTrr7~rr~ l~l~r~ r7G
=43 29mer/0 2um/OFF 5~cGAcAGTcGAcGrrrrrrrr~Arrr~TGTc =44 26mer/0.2um/OFF 5~cGAcAcTcGAGrrr~rrr7T~rTcGTc =45 26mer/0.2um/OFF 5~Gc~rr~rr~Tr7c =46 18mer/0.2um~0FF 5'TCGAGCATQGTGGCCGC
=47 27mer/0.2um/OFF 5~Trr7~rrrT~r7A-~c/A)-r7~r7~r~r7r7T~r =48 27mer/0.2um/OFF 5~Trr~r~T~ r~ Lc-~G/T)-TcTTAGGG
Exarnple 8. D~ . of T.. ~ Induction.
Jurkat TAg cells were transfected with the indicated constructs (5 ILg of each construct) by I~ UPVI~L;UII (960 ,uF, 250 v). After 24 hours, the cells were l'~ .~p ..1. .1 in fresh media and aliquoted. Half of each ~ rt:~-ion was WO96/06111 2 l q 7 ~ ~ 2 1 ."J~ s .1 ~

incubated with the dimeric FK506 derivative, (Exarnple 14) at a final ~r nr~-ntrarirm of 1 ILM. After 12 hours, the cells were washed and cellular extracts were prepared by repeated freeze thaw. ('.hll ~p '- ' -~e~y ' ' ~e (CAT) activity was measured by standard protocols. Molecular Cloning: A Laboratory Manual., Sambrook et al. eds. (1989) CSHLaboratory, pp. 16-59 ff. The data l~.IIUII~LI~ILCI CAT activity present as expected (in sarnple 2, with or without ligand; and in samples 5 and 6 in the presence of ligand) in 70 ILL of extract (total extract volume was 120 ~L) after incubation at 37~C for18 hours. The samples employed in the assays are as follows:
10 1. G5E4TCAT (GAL4-CAT reporter plasmid) 2. G5E4TCAT, GAL4-VP16 3. G5E4TCAT, NF3VlE

4. G5E4TCAT, GF2E ~ :

5. G5E4TCAT, GF2E, NF3VlE

6. G5E4TCAT, GF3E, NF3VlE

Synthetic Chemistry Exatnples As indicated elsewhere, .. 1.. ,.. ,.l~ of particular interest at present asoLt,~,.l...;~L;un agents have the following structure:

linker--{rbml,rbm2, . . .rbm,~ .

20 wherein "linkerr is a linker mûiety such as described herein which is covalently linked tû "n" (an integer from 2 to about 5, usually 2 or 3) reoeptor binding moieties ("rbm"'s) which may be the same or different. As discussed elsewhere herein, the receptor binding moiety is a ligand (or analog thereof) for a known receptor, such as are cl~ Led in Section V(C), and including FK506, FK520, 2~ rapamycin and analogs thereof which are capable of binding to an FKBP; as well as cyclosporins, LeLI.lcy~ l.s, other antibiotics and macrolides and steroids which are capable of binding to respective receptors.

2 ~ 97242 ~W096/U6111 1~ 1~)_,.., .I

The linker is a bi- or multi-functional molecule capable of being covalently linked ("--~) to two or more receptor binding moieties. Typically thelinker would comprise up to about 40 atoms and may include nitrogen, oxygen and rulfur in addition to carbon and hydrogen. Illustrative linker moieties are s disclosed in Section VI(A) and in the various Examples and include among others C1-C30 alkyl, alkylene, or arylalkyl groups which may be ~h~in~ 1 or l and may be straight-chain, branched or cyclic. For example, alkyl ~..h~ritu~ nr~ are satunted straight-chain, cyclic or branched Lyllu~ ubon moieties, preferably of one to about twelve carbon atoms, including methyl, 10 ethyl, n-propyl, i-propyl, I y~ Iv~lv~yl~ n-butyl, i-butyl, t-butyl, cyclobutyl, ~y~lv~lv~yLIl~llyl~lc~ pentyl, hexyl, heptyl, octyl and so forth, and may be optionally substituted with one or more ~ such as lower alkoxy, carboxy, amino (~ .1 or ....~ I), phenyl, aryl, mercapto, halo (fluoro, chloro, bromo or iodo), azido or cyano.
These .. I.v .. I~ may be prepared using uvl~ull~w,dly available materials and/or procedures known in the art. Engineered receptors for these I,VIII~JVUllla may be obtained as described infra. G~nnrollnrlc of particular interest are those which bind to a receptor with a Kd of less than 10-6, preferably less than about 10-7 and even more preferably, less than 107M.
One subclass of nl ~,~".... ;~;.~g agents of interest are those in which one or more of the receptor binding moieties is FK506, an FK506-type compound or a derivative thereof, wherein the receptor binding moieties are covalently attached to the linker moiety through the allyl group at C21 (using FK506 LIUI.lJ~..;III;) as per compound 5 or 13 in Fig 9A, or through the cyclohexyl ring 2s (C29-C34), e.g. through the C32 hydroxyl as per ~v.ll~vu.lL 8, 16, 17 in Fig 9B.
Compounds of this class may be prepared by adaptation of methods disclosed herein, including in the examples which follow.
Another subclass of ~ ,.... ;,;.. g agents of interest are those in which atleast one of the reoeptor binding moieties is FK520 or a derivative thereof, 30 wherein the molecules of FK520 or derivatives thereof are covalently attachedto the linker moiety as in FK1040A or FK 1040B in Figure 10. (~omrol.n.lc of this class may be prepared by adaptation of Scheme 1 in Figure 10, Scheme 2 in ~t ~ 2 WO96/06111 r~

Figs. llA and llB or Scheme-3 in Fig 12 and Fig 13.
A further subclass of ~ ,.. ;,;.. g agents of interest are those in which at least one of the receptor binding moieties is ~y~ lo~t)o~ A or a derivative.
It should be ~JIc~id~cl that these and other ~ ..,... ;,;..g agents of this s invention may be homo~ ; reagents (where the rbm's are the same) or hetero~,l;~,.. ;,:i.g agents (where the rbm's are different). Hetero-oL~,.,IIl..~llg agents may be prepared by analogy to the procedures presented herein, including Scheme 3 in Figure 13 and as discussed elsewhere herein.
The following synthetic examples are intended to be illustrative.

10 A. General Pl~cduA~. All reactions were performed in oven-dried glassware under a positive pressure of nitrogen or argon. Air and moisture sensitive ..,,.I,u ,,.l~ were introduced via syringe or cannula through a rubber septum.
B. Physical Data. Proton magnetic resonance speara ~H NMR) were recorded on Bruker AM-500 (500 MHz), and AM400 (400 MHz) ~ IUIIA~
Chemical shifts are reported in ppm from ~e~l.ull.,Lyl~;lane using the solvent resonance as an internal standard (chloroform, 7.27 ppm). Data are reported as follows: chemical shift, Illuhi~ ;Ly (s - singlet, d - doublet, t triplet, q quartet, br - broadened, m multiplet), coupling constants (Hz), integration.
Low and high-resolution mass spectra were obtained.
C. CL~ y. Reactions were monitored by thin layer ~LI.JllldtogldlJlly (TLC) using E. Merck silica gel 60F glass plates (0.2~ mm).
Components were visualized by illnmin~ri~n with long wave ultraviolet light, exposed to iodine vapor, and/or by dipping in an aqueous ceric ~.,.. ;.. ~
molybdate solution followed by heating. Solvents for ~LI~Jllld~ y were 2s HPLC grade. Liquid III~ O~ hY was performed using forced flow (flash ~LI~ U~ Y) of the indicated solvent system on E. Merck silica gel 60 (230-400 mesh).
D. Solvents and Reagents. All reagents and solvents were analytical grade and were used as received with the following exceptions. TcLIdllydl~)fuldll (THE~
30 benzene, toluene, and diethyl ether were distilled from sodium metal benzophenone ketyl. Tri..llyldllulle and aoetonitrile were distilled from calcium - 21~72~
~WO 96/06111 r~ lil!i91 ~ ~ ~ =; 79 ~
hydride. Di~Lh~uul~LL~ was distilled from phv~Lvlv~la pentoxide.
Dhl.. ~I-yl~ [)MF~ was distilled from calcium hydride at reduced pressure and stored over 4A molecular sieves.

Plcr ~ of FR506 Derivatives Example 9. IIy~Luho~.~L;o.. /Oxidation of Fl~506-TBS2 (I to 2).
The Ly~LubullL;ull was performed according to the procedure of Evans (Evans, et al., JACS (1992) 114, 6679; ibid. (1992) 6679-6685). (See Harding, et al., NatKre (1989) 341, 758 for mlmh~-ring ) A 10-mL flask was charged with 24,32-bis[(tert-1uLyl.:lh..~ yl,;lyl)oxy]-FK506 (33.8 mg, 0.033 mmol) and [Rh(nbd)(diphos-4)]BF~(3.1 mg, 0.ûû4 mmol, 13 mol ~h). The orange mixture was dissolved in toluene (2.0 rnL) and the solvent was removed under reduced pressure over four hours. The flask was carefully purged with nitrogen and the orangish oil was dissolved in THF (3.0 mL, 10 mM finai .~ c....~L;un) and cooled to 0~C with an ice water bath. C;~c.Lolbu.~ie (98 ~L, 0.098 mrnol, 15 1.0 M solution in THF, 3.0 c.~u;~ ' ' was added via syringe and the resultingsolution was stirred at 0~C for 45 min. The reaction was quenched at 0~C with 0.2 mL of THF/EtOH (1:1) followed by 0.2 mL of pH 7.0 buffer (Fisher;
0.05 M phosphate) then 0.2 mL of 30% H2O2. The solution was stirred at room Lcll~ dLulc for at least 12 h. The solvent was removed under reduced pressure 20 and the remaining oil was dissolved in benzene (10 mL) and washed with saturated aqueous sodium bicdlLoll~Le solution (10 mL). The phases were separated and the aqueous phase was back-extracted with benzene (2 x 10 mL).
The organic phases were combined and washed once with saturated aqueous sodium L dfl)olldLc solution (10 mL). The benzene phase was dried with 25 MgSO~, uull cllLIdLcd~ and subjected to flash cll.~ y (2:1 L~A~le.~ ylacetate) providing the desired primary alcohol as a clear, colorless oil (12.8 mg, 0.012 mrnol, 37%).
Preparation of Mixed Carbonate (2 to 3). The plc~dldL;Oll of the mixed carbonate was a~ulll~ llel by the method of Ghosh (Ghosh, et al., Tetrahedron Lett. (1992) 33, 2781-2784). A 10-mL flask was charged with the primary alcohol 2 1 9~242 WO 96106111 r~ u.. ,~

(29.2 mg, 0.0278 mmol) and benzene (4 mL). The solvent was removed under reduced pressure over 60 min; The oil was dissolved in ~. r~ (2.0 mL, 14 mM final ~u~ L;u~l) and stirred at 20~C as tl;.~Lyl (77 ~LL, 0.56 mmol) was added. N,N'~ . 8~;...; lyl carbonate (36 mg, 0.14 mmol) was 5 added in one portion and the solution was stirred at 20~C for 46 h. The reaction mixture was diluted with l;.LIv.u~ .l.~,e and washed with saturated aqueous sodium l,;~L solution (10 mL). The phases were separated and the aqueous layer was back-extracted with l;lflolu~ fl~ e (2 x 10 mL). The organic phases were combined and dried (MgSO~ n ~ and subjected to flash ~LI""' ~r~ y (3:1 to 2:1 to 1:1 h~ hyl acetate). The desired mixed carbonate was isolated as a clear, colorless oil (16.8 mg, 0.014 mmol, 51%).
D;.. ;, -l .. , of FK506 (3 to 4). A dry, 1-mL conical glass vial (Kontes Scientific Glassware) was charged with the mixed carbonate (7.3 mg, 0.0061 mmol) and acetonitrile (250 ~L, 25 mM final ~ ;r~
Tri.. l.yL.. I.. e (10 ~LL, 0.075 mmol) was added followed by ~ylyl.. I~ .. r (8.3 ~L, 0.0027 mmol, 0.32 M solution in DMF). The reaction stirred 22 h at 20~C and was quenched by dilution with l;.Llo.ul...~ e (10 mL). The solution was washed with saturated aqueous sodium b;~ voli~se solution (10 mL). The phases were separated and the aqueous layer was back-extracted 20 with l;~hlu~ul~l~.hane (2 x 10 mL). The organic phases were combined and dried (MgSO,~ u~ e~ and subjected to flash ch,u....,~u~ ul.y (3:1 to 2:1 to 1:1 h~ Lyl acetate) providing the desired protected dimer as a clear, colorless oil (4.3 mg, 1.9 I~mol, 70~h).
D~lute~t;ull of the FK506 Dimer (4 to 5). The protected dimer (3.3 mg, 25 1.4 ~Lmol) was placed in a 1.5-mL polypropylene tube fitted with a spin vane.Acetonitrile (0.5 mL, 3 mM final concentration) was added and the solution stirred at 20~C as HF (55 ~L, 48% aqueous solution; Fisher) was added. The solution was stirred 18 h at room LCIll~ Ul~. The d~lJIute~ed FK506 derivative was then partitioned between dichlvlvl.l.,L~.Ie and saturated aqueous30 sodium bicarbonate in a 15-mL test tube. The tube was vortexed extensively tomix the phases and, after separation, the organic phase was removed with a pipet. The aqueous phase was back-extracted with di~lllulu~ ll~le (4 x 2 mL), 0 96106~ 7 and the combined organic phases were dried (MgSOd, .lln.~ rll and subjected to flash .L... - ~r~ y (1:1:1 hexane:THF:ether to 1:1 THF:ether) providing the desired dimer as a clear, colorless oil (1.7 mg, 0.93 ,umol, 65%).Following the above procedure, other ....~nn --... - and diamines may be used, such as L.~yl~ (14) o.~ul.. Lyl.. 1 - .. ~-, l~uu~.~llyl~ .. I. ~.. ~r, etc Exarnple 10. Reduction of Fl~S06 with L-Sdectride (FK506 to 6).
D~llL7L~r7LY and coworkers have shown that the treatment of FK506 with L-Selectride provides 22-dihydro-FK506 with a boronate cster engaging the C24 and C22 hydroxyl groups (Coleman and Danishefsky, Heterocycles (1989) 28, 157-161; Fisher, et al., J. Org. Chem. (1991) 56, 2900-2907).
P~c~d dL;u.. of the Mixed Carbonate (6 to 7). A 10-mL flask was charged with 22-dihydro-FK506-sec-.,u~yl.,o.~,..dlc (125.3 mg, 0.144 mmol) and ~7frrnnitril.
(3.0 mL, 50 mM final ~ n) and stimd at room ~.Ill~J..dLu.c as L~;~LLyldllullc (200 ~L, 1.44 mmol, 10 eu,u;~ I ' was added to the dear solution. N,N' .1;. ....:...,.lyl carbonate (184.0 mg, 0.719 mmol) was added in one portion, and the clear solution was stirred at room Lell~ ,dLulc for 44 h.
The solution WdS diluted with ethyl acetate (20 mL) and washed with saturated aqueous sodium L~ ...,LC: (10 mL) and the phases were separated. The 20 aqueous phase was then back-extracted with ethyl acetate (2 x 10 mL), and the organic phases were combined, dried ~MgSO,,), and the resulting oil was subjected to flash L~- .. n.O,-.~ y (1:1 to 1:2 h~d~ d~yl acetate) providing thedesired mixed carbonate as a clear, colorless oil (89.0 mg, 0.088 mmol, 61%).
D;.. ,.. :, - .nll of FK506 Mixed Carbonate (7 to 8). A dry, 1-mL conical 25 glass vial (Eiontes Scientific Glassware) was charged with the mixed carbonate (15.0 mg, 0.0148 mmol) and dichlo.~ .l.~.c (500 ILL, 30 mM final ~n~cuLIdL;uu). The solution was stirred at room Ltlll~ dLulc as Ll;~.Lyldllullc (9 ~L, 0.067 mmol, 10 equiv.) was added followed by p-~ylyl. .,~ (0.8 mg, 0.0059 mmol). The reaction stirred 16 h at 20~C and was quenched by dilution 30 with dichlc,~ .l.du.e (5 mL). The solution was washed with saturated aqueous sodium b; dlbondLe solution (5 mL). The phases were separated and the aqueous layer was back-extracted with l;.Llolull...ll~le (2 x 5 mL). The organic phases were combined and dried IMgSO,,), ~I.n. .~ and subjected to flash luulllao~ lly (1:1 to 1:2 hw,ule.~.Lyl acetate) providing the desired dimer as a clear, colorless oil (7.4 mg, 3.8 ~mol, 65%).
s Following the above procedure, other, ~ n. --.. , diamines or triamines may be used in place of the ~Lylyl ...I:- .~i.., such as 1,..1~14luule (15), o"yh ....I:-...;,.r d~ull~Lyl..l..I -.l8~r- (16), bis-p d;b~ ylululle~ N-methyl d;~dlyl~ uluue~ tris-uldl.u.llyL.u..c (17), tris-~.lu..ùp.u~uylul..l.ec, 1,3,5-Ll;~ld..o..l.Lllyl~y.lohexane, etc.

Evample 11. Oxidative Cleavage and Reduction of FK506 (I to 9). The osmylation was performed according to the procedure of Kelly (VanRheenen, et al., Tetrahe~ron Lett. (1976) 17, 1973-1976). The cleavage was performed according to the procedure of Danishefsky (~ell, et al., J. Org. Chem. (1986) 51, 5032-5036). The aldehyde reduaion was performed according to the procedure Of ~ h.~ LLy (~. Org. Chem., (1981) 46, 46Z8-4691). A 10 mL flask was charged with 24,32-bis[tert-Lu.yl.L..~ yl~;lyl)oxy]-FK506 (84.4 mg, 0.082 mmol),4-ul.Lllylll.ul~holine N-oxide (48 mg, 0.41 mmol, 5 equiv), and THF (2.0 mL, 41 mM final ~ull~ellLldL;un)~ Osmium tetroxide (45 ~LL, 0.008 mmol, 0.1 equiv) was added via syringe. The clear, colorless solution was stirred at room 20 tc~ Lulc for 5 hr. The reaction was then diluted with 50Yo aqueous methanol (1.0 mL) and sodium periodate (175 mg, 0.82 mmol, 10 equiv) was added in one portion. The cloudy mixture was stirred 4û min at room temperature, diluted with ether (10 mL), and washed with saturated aqueous sodium b;wllJoll.lLe solution (5 mL). The phases were separated and the aqueous 2s layer was back-eYtracted with ether (2 x 5 mL). The combined organic layers were dried (MgSO~ and treated with solid sodium sulfite (50 mg) The organic phase was then filtered and ..~n. ,~ f.l and the oil was subjected to flash chn-m~r~lgrarhy (3:1 to 2 1 h.,.ule..Lyl acetate) providing the ;~
unstable aldehyde (53 6 mg) as a clear, colorless oil. The aldehyde was 30 immediately dissolved in THF (4.0 mL) and cooled to -78~C under an dn..o~ elc of nitrogen, and treated with lithium tris[(3-ethyl-3-2 ~ ~2~2 ~wo 96106111 P~ J

pentyl)oxy]aluminum hydride (0.60 mL, 0.082 mmol, 0.14 M solution in TEIF, 1.0 equiv). The clear solution was allowed to stir for 10 min at -78~C then quenched by dilution with ether (4 mL) and addition of saturated aqueous - ..... .. chloride (0.3 mL). The mixture was allowed to warm to room 5 LeLulu.l L~ule and solid sodium sulfate was added to dry the solution. The mixture was then filtered and .. ~ and the resulting oil was subjected to flash ULI.~ y (2~ IIrI acetate) giving the desired alcohol as a clear, colorless oil (39.5 mg, 0.038 mmol, 4i%).
P~c,u u~-;uLu of Mixed Carbonate (9 to 10). The lul~ ~,~...;ull of the mixed 10 carbonate was a~wll~ hed by the method of Ghosh, et aL, ~etrahedron Lett.
(1992) 33, 2781-2784). A 10 mL flask was charged with the primary alcohol (38.2 mg, 0.0369 mmol) and ~rrtrnittil~ (2.0 mL, 10 mM final l,ul~cllLI~L;un) and stirred at room L~u~ Lulc as 2,6-lutidine (43 ~L, 0.37 mmol, 10 equiv) was added. N~N~ u~ .. lyl carbonate (48 mg, 0.18 mmol) was added in one 15 portion and the solution was stirred at room LeLllu~ul~: for 24 h. The reaction mixture was diluted with ether (10 mL) and washed with saturated aqueous sodium l;~lull~l~e solution (10 mL). The phases were separated and the aqueous layer was back-extracted with ether (2 xlO mL). The organic phases were combined and dried (MgSOd, wll~clnl~Led~ and subjected to flash ~hlu~u,~Lu~ Ly (2:1 to 1:1 IL~A~U1~ ~IIYI acetate). The desired mixed carbonate was isolated as a clear, colorless oil (32.6 mg, 0.028 mmol, 75%).
Plelu~u,L~;un of Benzyl Carbamate (10 to 11). A dry, 1 mL conical glass vial (Kontes Scientific Glassware) was charged with the mixed carbonate 10 (8.7 mg, 0.0074 mmol) and acetonitrile (500 ,~L, 15 mM final eun~c~lLl~L;
The solution was stirred at room Leul~u.l.l-ulc as Lfl~LllyLII;ne (10 ILL, 0.074 mmol, 10 equiv) was added followed by l~ yLlune (1.6 ~LL, 0.015 mmol, 2 equiv). The reaction stirred 4 h at room Leullu..~ul~:. The solvent was removed with a stream of dry nitrogen and the oil was directly subjected to flash eLulll~lLu~ ully (3:1 to 2:1 L~ LIIYI acetate) providing the desired protected monomer as a clear, colorless oil (6.2 mg, 5.3 ~Lmol, 72%).
The protected monomer (6.2 mg, 5.3 ~mol) was placed in a 1.5 mL
poly~ululuyl.lle tube fitted with a spin vane. Acetonitrile (o.s mL, 11 mM final WO 96tO6~11 2 1 ~ 7 2 ~ 2 CU1~CIILI~L;Un) was added and the solution stirred at room ICII~ LUIC as HF
(55 t~L, 48% aqueous solution; Fisher, 3.û N final ~ ;rl,~) was added. The solution was stirred 18 h at room ~tl~lu~ Lulc. The d~lu~c~Lcl FK506 derivative was then partitioned between l;~LIul u..l~ ule and saturated aqueous 5 sodium b;~ulvoll~Le in a 15 mL test tube. The tube was vortexed extensively tomix the phases and, after separation, the organic phase was removed with a pipet. The aqueous phase was back-extracted with L~ l.lc,lu.l,.~l~.e (4 x 2 mL),amd the combined organic phases were dried (MgSOd, cun~cl~LI.~Lcd and subjected to flash clllu~u.~u~ uLy (1:1 to 0:1 hc~ c.~ yl acetate) providing the10 desired ~luLtclcd l..l~yL.Ilb~ll.~Le as a clear, colorless oil (3.9 mg, 4.1 ~mol, 78%).
By replacing the b~ y~llllle with a diamine such as ~ylyl~ ....l ..;..
(12), hw~ull~ yl.... ~.1 - .. ~ o~ llyl... ~ .r decamethyl.. 1~ (13) or other diamines, dimeric CUIII~UU~IU~ of the subject invention are prepared.

Example 12. r~c~..... Lùn of the Mixed Carbonate of FK506 (12). A 10-mL
flask was charged with 24, 32-bis [(tert-bu.ykl;lu.dlybllyl)oxy~-FK506 (339.5 mg., 0.329 mmol), 4~ ..EIyLllul~holine N-oxide (193 mg, 1.64 mmol, 5 equiv), water (0.20 mL) and THF (8.0 mL, 41 mN final concentration). Osmium tetroxide (0.183 mL, 0.033 mmol, 0.1 equiv, 0.18 M solution in water) was added via 20 syringe. The clear, colorless solution was stirred at room LellllJ..rLulc for 4.5 h.
The reaction was diluted with 50~h aqueous methanol (4 0 mL) and sodium periodate (700 mg, 3.29 mmol, 10 equiv) was added in one portion. The cloudy mixture was stirred 25 min at room tC.~IY~ LUIC~ diluted with ether (20 mL), andwashed with saturated aqueous sodium b;~..llvon.-Le solution ~10 mL). The 25 phases were separated and the aqueous layer was back-extracted with ether (2x10 mL). The combined organic layers were dried over MgSO~ and solid sodium sulfite (50 mg). The organic phase was then filtered and cullccllll~Led and the resulting aldehyde was hlull~ tely dissolved in THF (8.0 mL) and cooled to -78 ~C under an ~LIuu~luk..= of nitrogen, and treated with lithium tris [(3-ethyl-3-pentyl)oxv] aluminum hydride (2.35 mL, 0.329 mmol, 0.14 M solution of THF, 1.0 equiv). The clear solution was allowed to stir for 60 min at -78 ~C

2 ~ 97242 ~ WO 96/06111 r~,l/lJ ,.vl .l (monitored closely by TLC) then quenched at -78 ~C by dilution with ether (5 mL) and addition of saturated aqueous -...... ... chloride (0.3 mL). The mixture was allowed to warm to room ~e.l~/d~UlC and solid sodium sulfate was added to dry the solution. The mixture was stirred 20 min, filtered, r.l and the resulting oil was h~ el;lLely dissolved in ~rrtrnitrilt. (10 mL). To the solution of the resulting primary alcohol in CH3CN was added 2,6-luti&e (0.3R0 mL, 3.3. mmol, 10 equiv) and N,N'~ ..;...:.lyl carbonate (420 mg, 1.65 mmol, 5 equiv). The h.t~.uS .~ mixture was stirred at room Lcu~ Lulc for 19 h, at which time the solution was diluted with ether (30 mL) 10 and washed with saturated aqueous sodium b;~dllJo~ e (20 rnL). The aqueous phase was back~tracted with ether (2xlO mL). The organic phases were combined and dried (MgSO,,), ...,.. -. ~ and subjected to flash ~L"" ~5~l8~y (3:1 to 2:1 to 1:1 hexane/ethyl acetate). The desired mixed carbonate 12 was isolated as a clear, colorless oil (217 mg, 0.184 mmol, 56%
15 overall for 4 steps).

Example 13. P~c~J~dth3n of 24, 24', 32, 32'-tetrakis [(tertbutyl-d;~ yb;lyl)Oxy]-FKl012-A (p-~lyl- -- l ~ bridge).
A dry, 1-rnL conical glass vial was charged with the mixed carbonate (23.9 mg, 0.0203 mmol) and :~rttr~nitrilr (500 ILL, 41 mM final ...n._~ .n rn) Tli.,llyl.u~ulle (28 ~L, 0.20 mmol, 10 equiv) was added followed by p-xylyl..,t.li~....n, (46 f~L, 0.0101 mmol, 0.22 M solution in DMF). The reaction stirred 18 h at room ~C~ UIC, the solvent was removed with a stream of dry nitrogen, and the oil was directly subjected to flash lu~ Lu5ld~1ly (3:1 to 2:1 to 1:1 hexane/ethyl acetate) affording the desired protected dimer as a clear, colorless oil (11.9 mg, 5.3 &mol, 52%) Example 14. Preparation of FK1012-A (p-~ylyl. .. I ...:n. bridge) (13). The protected dimer (11.0 mg, 4.9 ~mol) was placed in a 1.5-mL POIY~JIU~YI~ ..c tubefitted with a spin vane. Acetonitrile (0.50 mL, 10 mM final concentration) was added, and the solution stirred at 20 ~C as HF (55 IbL, 48% aqueous solution;
30 Fisher, 3.0 N final ~l~n~C.md~hJll) was added. The solution was stirred 16h at 21 9~24~
WO96106111 .~.II.J..,~'i.-.

room k.~ Lulc. The d~ u~c~ed FK506 derivative was then partitioned between dichlu.uu.~Lll~lc and saturated aqueous sodium bi~ubull~e in a 15-mL
test tube. The tube was vortexed extensively to mix the phases and, after separation, the organic phase was removed with a pipet. The aqueous phase was 8 back-extracted with d;clllulull~ll ult (4x2 mL), and the combined organic phases were dried ~MgSO4), ~ rd and subjected to flash ~LI.~ Y (1:1:1 hexane/THF/ether to 1:1 TE~F/ether) providing FK1012-A as a clear, colorless oil (5.5 mg, 3.0 ~mol, 63%).

Example 15. Plc~ L~JII of 24, 24', 32, 32'-tetrakis[(ter-bulyld;lll~lLyl~;lyl)ûxy]-FK1012-B (~ . bridge). A dry, 1-mL
conical glass vial was charged with the mixed carbonate (53.3 mg, 0.0453 mmol) and a. ~'1....;11 ,1~ (2.0 mL, 11 mM final ....~. ,.,U~I ..I~) Triethylamine (16 ~L, 0.11 mmol, 5 equiv) was added followed by .l ..i.~ (61 ~L, 0.0226 mmol, 0.37 M solution in DMF). The reaction stirred 12 h at room ~LIll,u...l-ule~ the solvent was removed with a stream of dry nitrogen, and the oil was directly subjected to flash ~LIulll~to~ ully (3:1 to 2:1 to 1:1 hexane/ethyl acetate) affording the desired protected dimer as a clear, colorless oil (18.0 mg, 7.8 ~mol, 35%).

Example 16. Plc,u.~ ;on of FE1012-B (~ -1,10 bridge) (14).
The protected dimer (18.0 mg, 7.8 ~mol) was placed in a 1.5-mL poly,ulu,u.y1~.letube fitted with a stirring flea. Acetonitrile (0.45 mL, 16 mM final UUI~ ;On) was added, and the solution stirred at room ~tlll,U...~UII~ as HF
(55 ~bL, 48~h aqueous solution; Fisher, 3.6 N final ~un~ll~ ;oll) was added. Thesolution was stirred 17 h at 23 ~C. The product FK1012-B was then partitioned 2s between dichloromethane and saturated aqueous sodium l,i~fl,oll~te in a 15-mLtest tube. The tube was vortexed extensively to mix the phases and, after separation, the organic phase was removed with a pipet. The aqueous phase was back-extracted with dichh~lulll..ll...l. (4x2 mL), and the combined organic phases were dried (MgSOd"l,.,.~ .l and subjected to flash ~ILulll.~ ,ully (100% ethyl aoetate to 20:1 ethyl acetate/methanol) affording FK1012-B as a ~wo 96/06111 I'_I/U~

dear, colorless oil (5.3 mg, 2.9 ~mol, 37%).
, Exarnple 17. P~ tioll of 24, 24', 32, 32'-tetrakis[(tert-buLy~ ' ' yl~ilyl)Oxy-]-FKl012-C lbis-p .~ ylb~ UJI ~
bridge). A dry 25-mL tear-shaped flask was charged with the diamine linker s (151 mg, 0.0344 mmol) and 1.0 mL of DMF. In a separate flask, the mixed carbonate and ~;. .Lyl~lliut (0.100 mL, 0.700 mmol, 20 equiv) were dissolved in 2.0 mL of l' ~ ~.l..~I.~.e then added slowly (4xO.S0 mL) to the stirring solution of bis-p- ."i,...,.,. l~ylb...~vyl~ nninf~ ~n~ -1,10. The flask containing the mixed carbonate 12 was washed with dichlv.u~.L~.t (2xO.50 10 mL) to ensure complete transfer of the mixed carbonate 12. The reaaion stirred 16 h at 23 ~C, the solvent was removed with a stream of dry nitrogen, and the oil was directly subjected to flash d~ Ly (1:1 to 12 hexane/ethyl acetate) to afford the desired protected dimer as a dear, colorless oil (29.6 mg, 11.5 ~mol, 34%).

Example 18. P~.. L;.~.. of FK1012-C (15). The protected dimer (29.6 mg, 1~5 ~mol) (17~ was placed in a 1.5-mL poly~l~",rlclle tube fitted with a stirring flea. Acetonitrile (0.45 mL, 23 mM final ron~l~ntrati~n) was added, and the solution stirred at room Lelllu...l~ulc as HF (55 ILL, 48% aqueous solution;
Fisher, 3.6 N final uull~ m~L;ull) was added. The solution was stirred 17 h at 20 room tClu~.l..~Ul~. The desired sy~ i~l dimer was then partitioned between dichlol~,u...L~ule and saturated aqueous sodium b;~dlbun~k in a lS-mL test tube.The tube was vortexed extensively to mix the phases and, after separation, the organic phase was removed with a pipet. The aqueous phase was back-extraaed with dichh~.ulll.LllOl~e (4x2 mL), and the combined organic phases were dried 2s (MgSOd, . ~,n. a ~ and subjeaed to flash ~L~ hy (100% ethyl acetate to 15:1 ethyl acetate/methanol) affording FK1012-C as a clear, colorlessoil (11.5 mg, 5.5 ~Lmol, 47%).

.. :

21 q~42 WO96106111 F~

P~ ;ùn of CsA Derivatives Example 19. MeBmt(OAc)--OHICsA (2). MeBmt(OAc)-OAcl-CsA (I) (161 mg, lZ4 mmol) (see Eberle and Nuninger, J. Org. C~em. (1992) 57~ 2689) was dissolved in Methanol (10 mL). KOH (196 mg) was dissolved in water S (8 mL). 297 ~L of the KOH solution (.130 mmol, 1.05 eq.) was added to the solution of (1) in MeOH. This new solution was stirred at room ~tlll~d~
under an inert YLIllua~ for 4 hours at which time the rea~ion was quenched with acetic acid (2 mL). The reaction mixture was purified by reversed phase HPLC using a 5 cm x 25 cm, 12 ~L, 100 A, C18 column at 70~C eluting with 70% ~rprnnitri~p/H2o containing O.l~Yo (v/v) T~ uu~u~eL;c acid to give 112 mg ~2%) of the desired ,~n,~n~ r (2).
MeBmt(OAc)-OCOImlCsA (3). MeBmt(OAc)-OHI-CsA (2) (57 mg, 45.5 ~mol) and ~,dfl~ollyl~ P (15 mg, 2 eq., 91 Ibmol.) were transferred into a 50 mL round bottom flask and dissolved in dry THF (6 mL).
Du~u~lu~yl~.llyldll~ (32 ~L, 4 eq., 182 ~mol) was added and then the solvent was removed on a rotary evaporator at room Le~ ..dLu~:. The residue was purified by flash ~LIulll~u~,ld~lly on silica gel using ethyl aoetate as eluent to give 45 mg (73%) of the desired carbamate (3).
Tris-(2-du.u.lo..hyl)amine CsA Trimer Triacetate (6). MeBmt(OAc)-OCOIml-CsA (3) ~.5 mg, 5.54 ~mol, 3.1 eq.) was dissolved in THF (100 ~L).
D;;au~lu~yl~llyLul~lle (62 ~L, 5 eq., 8.93 ~mol of a solution containing 100 ~L
of amine in 4 mL THF) was added followed by tris(2-aminoethyl)amine (26,~L, 1.79 ,umol, 1 eq. of a solution containing 101 mg of tris-amine in 10 mL THF).
This solution was allowed to stir under Nl atmosphere for 5 days. The rea~ion Z~ mix was evaporated and then purified by flash chrnm1rngr~rhy on silica gel using 0-5% methanol in chloroform to give 4.1 mg of desired product (6).
Example 20. D.- ,.;no~ CsA Dimer (8). Solid Na metal (200 mg, excess) was reacted with dry methanol (10 mL) at 0~C. Di~minn~lPr~nP CsA Dimer Diacetate (5) (4.0 mg) was dissolved in MeOII (5 mL). 2.5 mL of the NaOMe solution was added to the solution of (5). After Z.5 hours of stirring at room Lclll~.ldLulG under an inert YLIlloa~h~lc~ the solution was quenched with aoeticacid (2 mL) and the product was purified by reversed phase HPLC using a 2 ~ 97242 ~WO 96/06111 5 mni x 25 mm, 12 ~, 100 A, C18 column at 70~C eluting with 70-95%
~-rnnitril. /H1O over 20 minutes containing 0.1% (v/v) T~ UOIU~L;C acid to give 2.5 mg (60%) of the desired diol.
The d: - ~; r rl~ f CsA Dimer Diacetate (5) was prepared by replacing the tris(2-~uu.lo~.Lyl)amine with 0.45 eq. of l,10-1 -, i"r~ l . .r E~ample 21. p-Xyl~' ' ~ ~ CsA Dimer (4).
The p-xylene diamine CsA Dimer (4) was prepared by replacing the tris(2-aminoethyl)amino with 0.45 eq. of p-xylylene diamine.
Following procedures described in the literature other derivatives of 10 cyclophilin are prepared by linking at a site other than the l(MeBmt 1) site.Position 8 D-isomer analogues are produced by feeding the producing organism with the D-amino analogue to obtain incorporation specifically at that site. See Patchett, et ~1., J. Antibiolics (1992) 45, 943 G~-MeSO)D-Ala~-CsA);
Traber, et aL, ibid. (1989) 42, 591). The position 3 analogues are prepared by 15 poly-lithiation/alkylation of CsA, specifically at the -carbon of Sac3. See Wenger, Trf~nsplant Prooeeding (1986) 18, 213, supp. 5 (for cyclophilin binding and activity profiles, p.~lL;cul.uly D-MePhe3-CsA); Seebach, U.S. Patent No. 4,703,033, issued October 27, 1987 (for ~lclua~ ;oll of derivatives).
Instead of ~lo~uul;ll A, following the above-described procedures, other 20 naturally-occurring variants of CsA may be ml~lrimrri7~-~ for use in the subject invention.
E~ample 21A. Alternative synthesis for CsA dimer.
MeBmt(OH)-~OCOIm1 -CsA
MeBmt(OH)-rl-OH'-CsA (38 mg, 31 mmol, 1218.6 glmol) and carbonyl-II;i.,- -l .,.,lr (20 mg, 4eq., 124 mmol, 162.15 g/mol) were transferred into a 10 mL round bottom flask and dissolved in dry THF (2 mL). Diiau~lu~ ;Lylolllillc (22 mL, 4 eq., 125 mmol, 129.25 g/mol) was added and then the solvent was removed on a rotary evaporator at room t~ Lu-c. The residue was purified by flash c l... O ,' ~ on silica gel usmg 0-20% acetone in ethyl acetate as 30 eluent to give 32mg (78% yield) of a white solid.

2 t q~
WO96106111 ~ ~3 Me O
HO~ ~'o ~ N~N
Me Me Me ~=i Me ~ ~ Me ~ ~ Me y ~ N ~ ~ NH ~ N~
~ ~ ~ Me Me ~O
Me INMeMeO Me ~ o MeN

~' X~ ~MN~
Me Me ~ Me Me Me Me (CsA)2 .yl~ ~ CsA dimer MeBmt(OH)-~-OCOlm'-CsA (12.5 mg, 9.52 mmol, 1312.7 glmol) was dissolved in DCM (200mL). To this solution was added 22ml (0.Seq., 4.75 mmol) of a solution of xylylene diamine (14.7 mg, 136.2g/mol) in DMSO (0.5 5 mL) and the reaction mixture was stirred for 72 hours at room t~ UlC under a nitrogen atmosphere . .. ~ g slowly. The reaction was diluted with acetonitrile (2 mL) filtered through glass wool and purified by reverse phase HPLC (Beckman C18, 10m, 100A, Icm x 25cm, 5mL/min, 50 to 90%
ACNlH10(+0.1%TFA) over 30 minutes, 70~C) to give 6.1 mg (49% yield) of a 10 white solid.

Me O O Me HO~--C~N_~_ N o J~ ,OH

Y~' ~ ~ H~'f NH~ ~ M~

U~ M~o M~ O M~N MeN O j~M~ o M~ NUe\<~M
NH U~
Me M- ~ Me Me Me Me Me Me Mé Me ~ Me Me 21 97242:

~wo s6/06lll r~

Exnmple 21B. Synthesis of a FK506-CsA dimer - MeBmt(OAc)-~
CH2COOEt-CsA
MeBmt(OAc)-rl-Br'-CsA (26 mg, ~80% pure, 15.7 mmol, 1323.57 g/mol) was dissolved in THF (500 mL). This solution was added by syringe pump over 5 15 hours to a THF solution of the magnesium enolate of ethyl hydrogen malonate(excess) prepared by the addition of iPrMgCI (2.15 mL, 2.34 M in ether) to a 0~Csolution of ethyl hydrogen rnalonate (Lancaster, 2.5 mmol, 332 mg, 132.12 g/mol) in THF (4.7 mL) followed by warming to room i , The reaction mixture was quenched with IN HCL (50 mL) and extracted with ethyl acetate (2 10 x 50 mL). The organic layers were dried over Na2SO4, filtered and evaporated.The crude product was dissolved in DMF (I mL). Et~NOAc.4H20 (150 mg. excess) was added and the mixture was heated at 90~C for 2 hours. The reaction mixture was cooled to room h.~ Lu~c, diluted with H20 (50 mL) and extracted with ether (2 x 50 mL). The combined organics were dried over 15Na2SO4, filtered and evaporated. The residue was purified by flash , .' y on silica gel, eluting with 75-100% ethyl _ n to give 11.4 mg (55~/O) of a white solid.
Me O
AcQ,~ ~ O'~' Me~ Me Me Me ~ ~ Me ~ ~ Me ~ ~ ~ N ~ ~ NH ~ ~
Me ~ ~ ~ Me Me ~0 Me NMeMeO Me ~ O MeN

N ~ 0~ Me ~ ~
Me Me Me Me Me Me MeBmt(OH)-~CH2COOH' -CsA
MeBmt(OAc)~ CH2COOEt'-CsA (11.0 mg, 8.27 mmol, 1330.76 glmol) - 20 was dissolved in MeOH (2 mL) and added to a solution of NaOMe (I .30 M in MeOH, 10 mL). The reaction mixture was stirred at room t~ lC under a nitrogen atmosphere for 5 hours at which time H2O (2 mL) was added and the mixture was stirred for another 2 hours. The reaction was quenched with glacial 2 t ~4~
WO96/06111 1'~

acetic acid (I mL), filtered through glass wool and purified by reverse phase I~LC (Raimn C18 dynamax5 5m, 300A, 21.4 mm x 250 mm, 20 mL/min, 50 to 90% ACN/H2O(+0.1%TFA) over 30 minutes, 70~C) to give 5.5 mg (53% yield) of a white solid. M e o H O.,~~~'O H
C M e Me Me Me ,: ~¦ Me ~l ~ Me O~N~ ~NH~

Me I M~C Me~< O MeN
=~--N/~ XN H~
Me Me ~ Me Me Me Me ~ bis-TBS-N-(6-(B ' -)hexyl) FK506 cnrbnmnte bis-TBS-FK506 Iyl carborlate (also a precursor to (tbs)4-FK1012) (5.8 mg, 1177.62 g/mol, 4.93 mmol) was dissolved in DCM. To this was added N-Boc-1,6-d ' - (7.25 mg, excess). After stirring for 10 min at room the reaction mixture was evaporated and the product purified by flash ~ S~y eluting with 10 to 40% ethyl ~ to provide 5.9 mg (94 % yield).
Me~~MeO
MeO~ .~O~=~

H ~OMe BOC-NH' ~ ~ ~ Me Me ~OTBS

N-(6 ~ 1) FK506 cnrbnmnte bis-TBS-N-(6-(Boc-arnino)hexyl) FK506 carbamate (5.9 mg, 1278.88 g/mol, 4.61 mmol) was transferred to a POI~Y~ .~ tube in ACN (700 ml) 15 followed by aqueous HF (49%, lOOmL). The reaction was complete after six hours at room i , ~ amd was quenched by the slow addition of a saturated solution of NaHCO3. The mixture was diluted with saturated NaHCO3 (4mL), -2 1 972 42 ~
~ wo s6/06~

H20 (4mL) and extracted with DCM (3 x 10 mL). The combined organic phases were dried witb MgSO4, filtered and evaporated to give 3.6 mg (82% yield) of crude product.
Me~~MeO
Me ~O~O I ~

H ~OMe H2N ~ Nb,O M e M e "OH

FKCsA
MeBmt(OH)-~-CH2COOH'-CsA (2.86 mg, 2.27 mmol, 1260.66 g/mol) and N-(6 ' ~1) FK506 carbamate (crude, 2.16 mg, 2.28 mmol, 949.21 g/mol) were dissolved in DCM (900mL). To this solution was added 127ml (3.0eq., 6.8 mmol) of a solution of BOP (11.9 mg, 442.5 g/mol) in DCM (500 mL), followed by 45 mL (2.25 eq., 5.1 mmol) of a solution of di;su~ hLh.yl amine (20 mL, d=0.74 2129.25 g/mol) in DCM (1.0 mL). Finally DMF (40 mL) was added and tbe reaction mixture was evaporated slowly at room h~ Lu~c under a stream of nitrogen over 12 hours. The reaction mixture was diluted with acetonitrile (ImL) filtered through glass wool and purified by reverse phase HPLC (Beckman C18, Icm x 25cm, 5mL/min, 50 to 90~/O ACN/H2O over 25 minutes, 50~C) to give 2.4 mg (48% yield) of a white solid.

M40----Mto M~

N~lX~N'~--NH~; O H ~OM~
4r ~ M o= ~H N~O~,~' M-- M-- OH

~M--~ ~1 M- M- ~A~M~ M~M-WO 96/06111 2 ~ ~ 1 2 ~ J l ., ~

Exnmple 22. (A) Structure-Based Design and Synthesis of F1~1012-nBump"
Compounds and FKBP12s with C~ .-- y Mutations .S..I.~ at C9 and C10 of FK506, which can be and have been accessed by synthesis, clash with a distinct set of FKBP12 sidechain residues.
5 Thus, one class of mutant receptors for such ligands should contain distinct , one creating a ..,..,~ ..y hole for the C10 5llhstit~nt and one for the C9 5~hsrit.-~nt Carbon 10 was selectively modified to kave either an N-acetyl or N-formyl group projecting from the carbon (vs. a hydroxyl group in FK506). The binding properties of these deri-atives clearly reveal thatthese C10 bumps effectively abrogate binding to the native FKBP12. Figure 21 depicts syntheses of FK506-type moieties containing additional C9 bumps. By assembling such ligands with linker moieties of this invention one can construct HED and HOD (and antagonist) reagents for chimeric proteins containing u~ olld;llg binding domains bearing '~ .y rnutations.
5 An illustrative HED reagent is depicted in Figure 21 that contains mn~lifir ~rinn~ at C9 and C10'.
This invention thus r~ ' a class of FK506-type . ~
comprising an FK506-type moiety which contains, at one or both of C9 and C10, a func~ional group comprising -OR, -R, -(CO)OR, -NH(CO)H or -20 NH(CO)R, where R is substituted or ~.~h~ , alkyl or arylalkyl which may be ~ L.l.dill~ branched or cyclic, including substituted or nncnh~ritmf-r pero~ides, and carbonates. "FK506-type moietiesr include FK506~ FR520 and synthetic or naturally occurring variants~ analogs and derivatives thereof ~mcluding rapamycin) which retain at least the ~-h~rir--rP~I or nnc~h~titnrr~1) C2 through C15 ponion of the ring structure of FK506 and are capable of binding with a natural or modified FKBP~ preferably with a Kd value belov.
about 10-6M.
This invention fmther ~....".~ s homo- and hetero-dimers and higher order oligomers containing one or more of such FK506-type ~ IJUII~b 30 covalently linked to a linker moiety of this invention. Monomers of these FK506-type . ,,..~p,.l,..~l~ are also of interest, whether or not covalently attached to a linker moiety or otherwise modified without abolishing their binding 2 t ~7~2 ~WO 96/06111 ,~ ~
affinity for the w.lGq~,on.L.Ig FKBP. Such ~,.r nn.. ;r . r. l.u , l~ may be used as .,li".. ;,.I;rn antagonist reagents, i.e., as .ulL;~bu~ for ,....;,.,.g reagents based on a like FK506-type ~rlmpo~lnrl Preferably the ~nTnrmln~lc and oligomers ~OIl~ ;llg~ them in accordance with this invention s bind to natural., or prefenbly mutant, FKBPs with an affinity at least 0.1%
and preferably at least about 1% and even more preferably at least about 10~b as great as the affinity of FK506 for FKBPl2,. See e.g. Holt et al., in~a.
Receptor domains for these and other ligands of this invention may be obtained by structure-based, site-directed or random ... ,~ methods. We 10 . ~o~ , a family of FKBP12 moieties which contain ~al., Ala, Gly, Met or other small amino acids in place of one or more of Tyr26, Phe36, Asp37, Tyr82 and Phe99 as receptor domains for FK506-type and FK52Q-type ligands containing mr~lifirqrinn.c at C9 and/or C10. ln particular, we . l),0r.,.l.l lr using FI~BP's with small ~ ,.. o, such as Gly or Ala for Asp37 in rrJnj~mrtir~n with FK506-type and FK520-type ligands containing ~.. l .~. ;l .. l ~ at C10 (e.g., -NHCOR, where R is alkyl, preferably lower alkyl such as methyl for example; or -NHCHO), and FKBP's with small lrl8~ such as Gly or Ala for Phe36, Phe99 and Tyr26 in rrnj~lnrr;r,n with FK506-type and FK520-type ligands containing le~ at C9 (e.g., oxazalines or imines).
Site-directed ~ ~ b ~ may bc conducted using the .,.. ~
mlltqgPnPcic protocol (see e.g., Sakar and Sommer, BioTechn.i~uci 8 4 (1990):
404-407). cDNA sequencing is performed with the Sequenase kit. Expression of mutant Fl~P12s may be carried out in the plasmid pHN1~ in the ~ coli strain XA90 since many FKBP12 mutants have been expressed in this system 2s efficiently. Mutant proteins may be uull~..li~.lily purified by frqrrirnqri~n over DE52 anion exchange resin followed by size exclusion on Sepharose as described elsewhere. See e.g. Aldape et al., J. Biol. Chem. 267 23 (1992): 16029-32 and Park et al., J. Biol. Chem. 267 5 (1992): 3316-3324. Binding constants may be readily ~iPtPrminPrl by one of two methods. If the mutam FKBPs ~ 30 maintain sufficient rotamase activity, the standard rotamase assay may be utilized. See e.g., Galat et al., BiocheJnis~ 31 (1992): 2427-2434. Otherwise, the mutant FKBP12s may be subjected to a binding assay using LH20 resin and wo 96/06111 2 ~ ~ 7 2 4 ;~ r rarl;~ H~-dihydroFK5v6 and 3HrdihyroCsA that we have used previously with FKBPs and cyclophilins. Bierer et al., Proc N~zl. Acad. Sci.
U.S.A. 87 4 (1993): 555-69.

(B) Selection of ~C'-~ n--I'J~ Muta~ions in FKBP12 for Bump-Fl~i06s s Using thc Yeast Two-Hybrid System One approach to obtaining variants of receptor proteins or domains, including of FKBP12, is the yeast "two-hybrid" or "interaction trap" system.
The two-hybrid system has been used to detea proteins that interact with each other. A "bait" fusion protein c.onsisting of a target protein fused to a 10 t....~ activation domain is co-expressed with a cDNA library of potential "hooks" fused to a DNA-binding domain. A protein-protein (bait-hook) interaction is detected by the appearanoe of a reporter gene product whose synthesis requires the joining of the DNA-binding and activation domains. The yeast two-hybrid system mentioned here was originally 15 developed by Elledge and co-workers. Durfee et al., Genes & D~IV~ IL 7 4 (199}): 555-69 and Harper et al., Cell 75 4 (1993): 805-816.
Since the two-hybrid system per se cannot provide insights into receptor-ligand interactions involving small molecule, organic ligands, we have developed a new, FKlG12-inducible ~. .n~ ,I;..n~l activation system (discussed 20 below). Using that system one may extend the two hy brid system so that smallmolecules (e.g., FK506s or FK1012s or FK506-type molecules of this invention) can be ;~ ted One first generates a cDNA l;brary of mutant FKBPs (the hooks) with mutations that are regionally localized to sites that surround C9 and C10 of FK506 For the bait, two different strategies may be pursued.
2s The first uses the ability of FK506 to bind to FKBP12 and create a composite surface that binds to c~lrin~nrin The sequence-specific tr~n~rripti~n~l activator is thus comprised of: DNA-binding domain-mutant FKBP12--bump-FK506--calcineurin A-activation domain (where--refers to a noncovalent binding interaction). The second strategy uses the ability o{ FK1012s to bind two 30 FKBPs Cimlllt~n~o~lcly~ AA HED version of an FK1012 may be used to screen for the follow-ing ensemble: DNA-binding domain-mutant FKBP12--bump-i o 96/06111 Y~l/~)v~ .

FK506-normal FK506--wildtype FKBP12-activation domain.

1. Calcineurin-GAL4 activati(Jn domain fufion As A bait: A derivative of pSE1107 that contains the GAL4 actiVatiOn domain and calcineurin A subunit - fusion construct has been ~u~ . Lcl Its ability to act as a bait in the s proposed manner has becn verified by studies using the two-hybrid system to map out calcineurin-s FKBP-FK506 binding site.
2. hFKBP12-GA14 ACtiVAtion domainfs~sion As A bait: hFKBP12 cDNA may be excised as an EcoRI-Hindm fragment that covers the entire open rcading frame, blunt-ended and ligated to the blunt-ended Xho I site of pSE1107 to 10 generate the full-length hFKBP-GAL4 aaivation domain protein fusion.
3. Afutant h~BP12 cDNA libranes hFKBP12 may be digested with EcoRI and HindlII, blunted and cloned into pAS1 ~Durfee et al., s~pra) that has been cut with Ncol and blunted. This plasmid is further digested with Ndel to elirninate the Ndel fragment betwcen tbe NdeI site in the polylinker sequence 15 of pAS1 and the 5' end of hFKBP12 and religated. This generated the hFKBP12-GAL4 DNA binding domain protein fusion. hFKRvP was reamplified with primers #11206 and #11210, Primer Table:

ElZe6 NdeI
SNdFK: s~-GGAATrc CAT AT6 GGC GTG CAG G-3' H M G V Q
11207 SmaI
35mFK37: S'-CTGTC CCG GGA NNN NNN NNN m CTT TCC ATC TTC AAG C-:
R S X X X K K G D E L
11208 Sm~I
35mFK~7: S'-CTG~ GA GGA ATC AAA m crT TCC ATC TTC AAG CA
R S S D F K K G D E L M
NNN NNN NNN GTG CAC CAC GCA GG-3' X X X H V V C
117~9 8cn~I
38mFK98: S'-C OE GGA rcc T U TTC CAG m TAG AAG CTC CAC ATC NNN
END E L K L L E V D X
NNN NNN AGT GGC ATG TGG-3' X X T A H P
llZ10 8~mHI
3EmFK: S'-C OE GGA TCC TCA TTC CAG m TAG AAG C-3' END E L K L L
Primer T ' ' ?lhll..~ used in the ~.UI~IU~.-;UII of a regionally localized hFKBP12 cDNA library for use in screening for ''''''I' '' ''~'Y mutations.

~ . , WO96/06111 2 t, 2 4 ~ 98 r~

Mutant hFBP12 cDNA fragments were then prepared using the primers listed below that contain IA;~.L~ d mutant sequences of hFKBP at defined positions by the polymerase chain reaction, and were inserted into the GAL4 DNA binding domain-hFKBP~deI/BamHI) construct.

5 4. Yeast strain S. cerevisiae Y153 carries two selectable marker genes ~his3/~B-"~ ) that are integrated into the genome and are driven by GAL4 promoters. ~Durfee, supra.) Using ~~ n~llrin~GAL4 Activation Dornain as Bait The FKBP12-FK506 complex binds with high affiniry to calcineurin, a type 2B pro~ein phosrh~r~P
10 Since we use C9- or C10-bumped ligands to senre as a bridge in the two-hybridsystem, onlyr those FKBPs from the cDNA library that contain a ~ Y
mutation generate a ~ lls~ L;ull.11 a~ivator. For l,Ul~V~.~l;ClI~_C, one may prepare at least three distin~ libraries (using primers 112û7-11209, Primer Table) that v ill each contain 8,000 mutant FKBP12s. RAnrir)mi7~1 sites were 1~ chosen by inspecting the FKBP12-FK506 structure, which suggested clusters of residues whose mutation might allow binding of the offending C9 or C10 ,..l.,l;l". ."~ on bumped FKSû6s. The libraries are then individually screened using both C9- and ClO-bumped FK;06s. The interaction between a bumped-FKS06 and a '''"'l" ~ ry hFKBPl2 mutant can be detected by the ability of 20 hos~ yeast to grow on h~s drop-out medium and by the expression of ,B-~".l.. ro~irl ~ gene. Since this selection is dependent on the presence of the bumped-FK506, false positives can be eliminated by subtra~ive screening v,ith replica plates that are supplemented with or without the bumped-FK506 ligands.

25 Using hFKBP12-GAL4 Activation Domain as 3~ait Using rhe calcineurin A-GAL4 activasion domain to screen hFKBP12 mutant cDNA libraries is a simple way to identify ~ lly mutations on FKBP12. However, mutations th;3t allow bumped-FK506s to bind hFKBP12 may disrupt the interaction between the mutant FKBP12--bumped-FK506 complex and calcineurin. If the ~wo 96/06111 r~l,o~

initial screening with calcineurin as a bait fails, the wild type hFKBP12-GAL4 activation domain will instead be used. An FK1012 E~ED reagent consisting of:
native-FK506-bumped-FK506 (Figure 16) may be synthesized and used as a hook. The FK506 moiety of the FK1012 can bind the FKBP12-GAL4 s activation domain. An interaction between the bumped-FK506 moiety of the FK1012 and a ""'l" ~ ~~lr mutant of FKBP12 will allow host yeast to grow on his drop-out medium and to e press ~ In this way, the selection is based solely on the ability of hFKBPL2 mutant to interact with the bumped-FK506. The same subtractive screening strategy can be used to 10 eliminate false positives.
In addition to the in vitro binding assays discussed earlier, an in vivo assay may be used to determine the binding affinity of the bumped-FK506s to the ~ r hFKBP12 mutants. In the yeast two-hybrid system, ~-gal activity is ~ r~rrnin~d by the degree of interaction between the ~bait" and the 15 aprey~. Thus, the affinity between the bumped-FK506 and the """l" ~ r FKBP12 mutants can be estimated by the UUllC~ ULlll;ng ~ O'~ f activities produced by host yeasts al different HED (native-FK506-bumped-FK506) concentrations.
Using the same strategy, additional randomized mutant FKBP12 cDNA
20 libraries may be created in other bump-contact residues with low-affinity ~r~mp.-nc~tr~ry FKBP12 mutants as templates and may bc screened similarly.

Phage Display Scrcening for High-Affinity C.~ n .. r FKBP Mutations Some high-affinity hFKBP12 mutants for bump-FK506 may contain several combined point mutations at discrete regions of the protein. The size of the 2s library that contains d~ ul~lidm: combined mu~ations can be too large for the yeast two-hybrid sys~em s capacity (e.g., > lo8 mutations). The use of b ~ r as a vehicle for exposing whole functional proteins should greatly enhance the capability for screening a large numbers of mutations. See e.g. Basset al., Proteins: Structure, Function & Genetics 8 4 (1990): 309-14i h,IcCafferty et al., Nature 348 6301 ~1990): 552~; and Hoogenboom, Nucl. Acids Res. 19 15 (1991): 4133-7. If the desired high-affinity '""'l ~ n~lr mutan~s is not be ~ i WO 96/06111 2 1 ~ 7 2 4 ~ PCTIUS95/10591 ~

identifed with the yeast two-hybrid system, a large number of combined mutations can be created on hFKBP12 with a phage vector as a carrier. The mutant hFBP12 fusion phages can be screened with bumped-FK50~5epharose as an affinity matrix, which can be ~yll~L~;~l in analogy to our originai .F~i506-based affinity matrices. Fretz et al., J Am Chem Soc 113 4 (1991): 1409-1411. Repeated rounds of binding and phage amplification should lead to the ;.1..,I;t;..l;.~ of high-affinity ...."~ y mutants.

(C) Synthesis of "Bumped (CsA)2s: Motlifir~t;nn of Meval(ll)csl~
As detailed above, we have d~lnallr,LI.lLell the feasibility of using 10 cyclophilin as a ,I;".. .;~;nl. domain and (CsA)2 as a HOD reagent in the context of the cell death signaling pathway. However, to further optimize the cdlular activity of the (CsA)2 reagent one may rely upon similar strategies as described with FK1012s Thus, modified (bumped) CsA-based ol;~... ;,;..~, reagents should be preferred in ~pFlir~Tinnc where it is particularly desirable for 15 the reagent to be able eo differentiate its target, the artificial protein constructs, from fnringl~nmlc cyclophilins.
One class of modified CsA derivatives of this invention are CsA analogs in which (a) NMeValll is replaced with NMePhe (which may be substituted or ....~..i.~l ;1.~1 f-l) or NMeThr (which may be ..,.~ I ;O ~lrcl or substituted on the 20 threonine beL~llydlo~yl group) or (b) the pro S methyl group of NMeValll is replaced with a bulky group of at least 2 carbon aloms, preferably three or more, which may be straight, branched and/or contain a cyclic moiety, and may be alkyl (ethyl, Ol preferably propyl, butyl, including t-butyl, and so forth), aryl, or arylalkyl. These rnmpolln~lc include those CsA analogs which ~5 contain NMeLeu, NMene, NMePhe or specifically ~he unnatural NMelbeiaMePhe], in place of Mevalll The r(b)" CsA .cJIll~ou.lL are of formula 2 where R represents a functionai group as discussed above.

.
~WO96/06111 r. ~ ~,J~

.

Me \ R."", ~ Me Me N ~ ~ N~'o Me ~
~ o ~ O Me ~< )--\r Me N ~ Me H ~ \ Me I Me I l ~ O

''~ ~ 11 Me o ~ Me O Me Me Me I le Me 1 (R = Me): CsA
2 (R = Me): Modified Lr MeVall1]CsA

R"~ 1e "
R'O
NH
O Me This inven~ion further e n~nmr~ceS homo- and hetero-dirners and higher order oligomers containing one or more such CsA analogs. Preferably the ~uLu~vuLIda and oligomers ~nmrricing them hl accordance with this invention bind to natural,, or preferably mutant, cvclophilin proteins with an affinity at5 least 0.1% and preferabiy at least about 1% and even m, ore preferably at least about 10% as great as the affinity of CsA for c,vclophilin.
A two step strategy may be used to prepare ~he modified [MeVal~]CsA
derivatives starting from CsA. In the first step the residue MeValll is removed from the LLL~Iu~ele~ In the second step a selected amino acid is introduced at 10 the (for ner) MeValll site and the linear peptide is cyclized. The advantage of this strategy is the ready access to several modified [MeVaP~]CsA derivatives ineulll,u u;aulL with a total synthesis. The synthetic scheme is as follows:

~ .

2l q7~4?~
WO96/06111 I~,lIU~

CH3S03H, THF, =0-C ~r NH~N~e~Mec M Th - ~ACO)2, Fyddine,THF, r; I G ~ ~ ~,,f 0 M e ,~
Me Me C Me Me ,MeG M~ ~ Me O Me r~NH~, NAC ~NJ~,N~ l- THF3FSrOe3nH s'NH~ HO~
M6_ OHOH Me~J Z NaOH aq 1.1 ~OH ~J
Me M ~ Me Me Me Me O Me . (BOCi O.DMAP. r~ ,~NBOC R.~Me_ ~e 3~R'. au), PyBrop, M~J OH ~,~ ~ 1 TFA CH2C12 2 To d,rrc.cllL;~tc the amide bonds, an N,O shift has been achieved between the amino and the hydroxyl groups from MeBmtl to give IsoCsA
(Ruegger et al., Helv. Chim. Acta 59 4 (1976): 1075-92) (see scheme above).
The reaction was carried out in THF in the presence of .,...ll ....,lr.,.li. acid.
(Oliyai et al., Pharm Res 9 5 (1992): 617-22). The free amine was protected with an acetyl group with pyridine and acetic anhydride in a one-pot procedure. The overall yield of the .~-acetyl protected IsoCsA is 90YO. The ester MeBmtl-MeValll bond is then reduced selectivel-y- in the presence of the N-methyl amide bonds, e.g. using DIBAL-H. The resulting diol is then (.. -r .. ,.~.. I to the ~.~.. c~onL.. g di-ester with another acid-induced hl,O shift.
This will prepare both the N-aoetyl group and MeValll residues for removal through hydrolysis of the newly formed esters with aqueous base.
After protection of the free amino group the new amino acid residue is introduced e.g. with ~he PyBrop coupling agent. Deproteclion and cycliza~ion 15 of the linear peptide with BOP in presence of DMAP ~AIberg and Schreiber, Science 262 5131 ~1993): 24~-250) completes the synthesis of 2. The binding of bumped-CsAs to cyclophilins can be evaluated by the same methods described for FK506s and FK1012s. Once cyclophilins are identified with ~ . ., y mutations, bumped (CsA)2 HED and HOD reagents may be ~yll~L~;~I

-~ q7242 ~:
~WO 96/06111 PCI/US9~/10591 according to the methods discussed previously. Of particular interest are bumped CsA ~ v .-i~ which can form dimers which themselves can bind to a cyclophilin protein with 1:2 ,lu: l.:.,~.,. ;~y Homo dimers and higher order homo-oligomers, LLcludulA~.~ and hetero-higher order oligomers containing at 5 least one such CsA or modified CsA moiely may be designed and evaluated by the methods developed for FKlû12A and (CsA)2, and optimize the linker element in andogy to the FK1012 studiesj.
Mutant uyclopl~;l;l" that bind our position 11 CsA variants (2) by .. ~ . .. , .. ~ ~~ :.. g the extra bulk on the ligand may be now be prepared.
10 Cy~lu~Lilills with these ~ .u y mutations may be identified through the structure-based site-directed and random ~ /screening protocols described in the FK1012 studies.
It is evident from the above results, that the subject method and ~Ulll~o~;L;u~ provide for great versatility in the production of cells for a wide 15 variety of purposes. By employing the subject constructs, one can use cells for I l....1....: ,. or t .~ "...l .l purposes, where the cells may remain inactive until needed, and then be activated by a. 1. . .~ .. . of a safe drug. Because oells can have a wide variety of lifetimes in a host, there is the U~uolLuu;Ly to treat both chronic and acute indications so as to provide short- or long-term 20 protection. In addition, one can provide for cells which will be directed to a particular site, such as an anatomic site or a functional site, where therapeutic effect may be provided.
Cells can be provided which will produce a wide variety of proteins or other gene products which may serve to correct a deficit or inhibit an 25 undesired result, such as activation of cytolytic cells, to inactivate a destructive agent, to kill a restricted cell population, or as is the focus here, to provideregulatable ol;,L~ u~,L;on of the expression of a target gene or r,.... I ;.... ~ ; of the target gene product. By having the cells present in a host over a defined periodof time, the cells may be readily activated by A.l~ , the mnlrimtri7ing 30 drug at a dose which can result in a rapid response of the engineered cells.
Cells can be provided where the expressed chimeric receptor is ;.,n~. _11..1 "
avoiding immune response due to a foreign protein on the cell surface.

-;e 2t 972~ ' WO 96/06111 1 ~,I/~)~,..;l .
~04 r~lLh .I.lOIC~ the intracellular chimeric receptor protein provides for efficientsignal ~ f~n upon ligand binding, apparently more efficiently than the receptor binding at an ~ receptor domain.
By using relatively simple molecules which bind to chimeric membrane 5 bound receptors, resulting in the expression of products of interest or inhibiting the expression of products, one cam provide for models for the study of disease and for cellular therapeutic treatment. The .,...11 ;" .. ;,;..g and related agents which may be aLullu;~Lelcd are safe, can be dLIIIIL~Lel~d in a variety of ways, and can ensure a very specific response, so as not to upset homPostocic All ~ and patent ~ Liu~ cited in this spffifir~tif~n are herein hll,UI~JUI.I~Cd by reference as if each individual ~ubl;~aL;ull or patentapplication were specifically and individually indifated to be hl~ul~ I~Lcl by referenoe.
Although the foregoing invention has been described in some detail by way of illustration and ~ample for purposes of clarity of .. ~ ... l;.. g, it will be readily apparent to those of ordinary skill in the art in light of the teachings of this invention that certain changes and mf rlififotifl~c may be made thereto without departing from the spirit or scope of the appended claims.

Claims (15)

Claims

1. Chimeric responder proteins comprising one or more ligand-binding domains and one or more action domains, which are capable of multimerizing in the presence of the ligand, and upon multimerization, are capable of activating transcription of a blocking gene under the transcriptional control of an expression control element which is activated in response to the presence of the ligand, wherein the blocking gene encodes an anti-sense message or ribozyme directed to a selected gene, a neutralizing antibody against a selected protein, a dominant negative form of a selected protein or the protein Cre.

2. A DNA construct encoding a chimeric responder protein of claim 1.

3. A cell containing and capable of expressing a DNA construct of claim 2.

4. A cell of claim 3 which further contains a DNA construct comprising a blocking gene under the transcriptional control of an expression control element which is activated in response to the presence of the ligand.

5. A cell of claim 4 which further contains a DNA construct comprising a blocking gene, wherein the blocking gene encodes a protein whose function leads to the elimination of a selected gene.

6. A cell of claim 5 which further contains a DNA construct comprising a blocking gene, wherein the blocking gene encodes the protein Cre.

7. A cell of claim 6 which further contains a target gene flanked by loxP
sequence permitting elimination of the target gene in the presence of Cre.

8. A cell of claim 4 which further contains a DNA construct comprising a blocking gene, wherein the blocking gene encodes an anti-sense message.

9. A cell of claim 4 which further contains a DNA construct comprising a blocking gene, wherein the blocking gene encodes a ribozyme directed to a selected gene.

10. A cell of claim 4 which further contains a DNA construct comprising a blocking gene, wherein the blocking gene encodes a neutralizing antibody moiety directed against a selected protein.

11. A cell of claim 4 which further contains a DNA construct comprising a blocking gene, wherein the blocking gene encodes a dominant negative form of a selected protein.

12. An organism containing at least one cell of claim 4.

13. An organism of claim 12 wherein the organism is a mouse.

14. An organism of claim 12 which further contains a target gene flanked by loxP sequence permitting elimination of the target gene in the presence of Cre.

15. An organism of claim 14 wherein the organism is a mouse.