CA2130100A1 - Receptor internalization signals - Google Patents

Abstract

Internalization signals and their use in modulating cell surface receptor endocytosis.

Description

~'o 93/18185 2 1 ;~ 0 1 ~ O PCr/USs3/ot66s RECEPTOR INTERNALIZATlON SLGNALS

BACKGROUN.D OF TtlE INVENTlQN ! ",, 1. Field of The Invention ~he invention relates to intemalization signals and the use of these signals -:
to modulate the transport of iigand into a cell.

. . .

2.. ~era~ A~t .

Receptor-mediated endocytosis is the mechanism by wnich a variety of nutrien~s, hormones, and ~rowth factors ars spec~ically and emciently transported into the cell. ~he pr~cess of receptor mediated endoc~tosis is complex and involves several distinct biochemical st~ps. Typically, the process proceeds by: (1) recruitmen~ of soluble coat proteins to the oell membralie and nucleation of coated pit format~on, ~2) assembly of co~t constituents and growth of the coated pit, (3) acquisition of .Specific receptors into the growing coated pit, (4) inva~ination of the cell membrane, 1~ (5) coat closure; and t6) membrane fusion wherein th~ coated pit buds in and pinches off to form a coated vesicle. The contents of the vesicles ar~
ultimately deiivered to the endosomes. Key to the entire process of receptor-mediator endocytosis is the internalization sisn~ present in the cytoplasmio ~ail of the recep~or. The ~oplasmic tail interac~s with sohlble coat proteins during forma~ion of a coated pit.

WO 93/1818~ PCl`/US93/01669 Receptors such as the transferrin receptor (TR) and the low~ensity lipoprotein (LDL) receptor are constitutively clustered in coated pits and undergo rapid internalization in the ,~resence and absence of ligand. Other receptorS such as epidermal~gFow~h factor (EGF) receptor are only concentrated in coated pRs an~ intemaleed aFter binding ligand. Pnvious studies have established ~hat there are internalization signals in the cytoplasmic domains o~ constitutively recycling receptors that are believed to interact with adaptor proteins of coated p~ts and promote high-emciency endocytosis.

tt would be desirable to use heterologous internalization signals to regulate endocytosis. In so doing, it would be possible, for example, to stimulate the internalkation of toxic substances by tumor cells or inhibit uptake of essential nutrients. However, prior to Applicants' invention, there has not been a successful transplantation of a heterologous internalization signal from one receptor to another. The present invention addresses this need and provides the means for utilizing heterologous intematkation signals and predicting the sequence and structure of heretofore unknown signals.

w~ 93/18185 pcr/us93/o1669 21301oo SUMMARY OF THE INVENTION

In accordance with the present invention, it has been discovered that transport of lisand into a cell can be modulated by introducing a het~rol~
gous intemalization signal into the cell. The signal is introduced as a nucleotide sequence or as the encoded peptide. The ability to transplant intemalkation signais from one receptor to another and retain ac~vity has important implications for control of endoc~tosis for scien~fic and medical purposes, in~uding drug delivery to cells.
, Accordingly, the present invention provides a method of modulating receptor mediated transport of ligand into a cell, which method comprises introducing a heterologous intemalkation signal into the cell.

Further, the invention provides a method ~or identlfying a sequen~e which modulates intemalization of 2 cell s~lrface receptor. This-method comprises:
(a) incubating cells having such receptors in the presence or absence of sequence and, optionally, in the presence of ligand for the cell surface receptor; and ~b) measuring internalization of the cell receptor in the presence or absence of the sequence.

The present invention also provides compositions for modulatin3 transport of ligand into a cell. Compositions embraced by the present invention comprise peptides having a tight turn conformation and a defined amino a~id sequence described in greater detail below. Provided also are nucleotide sequences encodin~ such peptides.

WO 93/1818~ - PCI`/US93/01669 In addition, the present invention proYides a method of administering gene therapy to a host subject. This method can be accomplished, for example, by introducing into a host subject, cells derived from the subject which have been modified to contain heterologous intemali~on s}gnal capable of :
modulating transport of ligand into a cell.

The present invention also pro~ides a method of gene therapy comprising ~ ;`
introducing into a host subject an expression vector comprising a nucleoffde `
sequence encoding a heterologous intemalization signal capab!e of -~
modulating transport of ligand into a cell. - :;

wo 93/18185 2 1 3 ~ Pcr/uss3/0166s BRIEF DESCRIPTION OF THF DRAWINGS

The following figur~ legends describe the first successful transplantation of heterologous internalization signals.

FIGURE 1. Uptake of 59Fe from human Tf by CEF expressing wild-type or mutant human TRs. Uptake of 59Fe from human Tf by CEF expressing human TRs was determined by in~bating cells with 59Fe-Tf for the times indicated and then washing the cells and determining their radic~ac~v ty. The ~-' results shown are for two representative experiments, A and B,- and each ~.-point represents the average values of 59Fe uptake by triplicate cul~ures of CEF expressing wild-type or mutant TRs. Panel A shows the relative 59Fe uptake for wild-type(-), 20YTFR23NPW(o)~ 1 3LSYTRF23F~NPW(-), ~RF23YKSV(o), 13LSYTRF~S CV(--), and F23M(~) m~ant TRs. Panel B shows the relative 59Fe uptake for wild-type (-), 2oyTRF23Npw(n)~
18LSYlRF23FDNPW(-), F23W(O), F231(--), and ~5g(~) mL~ human 1~ TRs.

WO 93/1818~ PCI/US93/01669 _ .
2~3010~) ' FIGURE 2. Comparisons of steady~state distributions (-) and 59Fe uptake (Cl~ of human TR mutants expressed in CEFs (mutations at the carboxy-terminal aromabc residue of the TR internalizabon motif, YTRF). PhenylaJa-nine 23 was changed to either methionine (F23M), isoleucine (F231), tryptophan (F23W), alanin~ (F23A), or glycin~ (F23G). ~he data for the F23A and F23G mutan~s is from Coilawn, et al., Cell, 63:1061-1072, 1990.
For the steady-stat~ analysis, C~Fs expressiny human TRs were incubated with 1251-labeled Tf for 60 min at 37-C then washed with buffer. The acid wash technique described in Materials and Methods was used to distinguish 10~ ' surface-bound and irlternalked Tf. The internalizabon rates represent the average of three experiments and are given as percentages + standard errors rela~ve to the wild-type (Wt) TR. .

FIGURE 3. Superimposed crystallographic turn struc~res for ir~emalization mot~f analogs with sequences matching the six-residue mannos~
t5 phosphats receptor (Man-~PF~) and the low-density lipoprotein receptor (IDLR) pattems. (A~ Man-~PR internalization moti~ analogs. (B) LDLR
intemalization m~if analogs. (C) Superimposed Man-~PR and LDLR
analogs.

w o 93/18185 . PC~r/U~93/01669 DESCRIPTION QF THE PREFERRED EMBODIMENTS

The present invention relates to a method of modulating receptor mediated transport of ligand into a cell wherein the method comprises introduang a heterologous intemaiization signai into the cell.

S ~he term Uheterologous" when used to describe the intemalkatlon signai of the invention refers to any internaiization signal that is introduced into the cell. llle term "ligand" refers to any substance capable of binding to or with a cell surface receptor. The term "tight turn" refers to ths reverse or helical turn conformation of amino acid residues within the heterolo~ous internaika-tion signal (Collawn, e~ at., Cell, 63:1061, 1990; t::ol.awn, et al., EMBO J., 10:3247, 1991). ~he term "cell-surfaee receptorU refers to any cell surface molecule whether or not it has a naturally occurring ligand. The term "sequence" refers to amino acid sequences as well as nucleic acid sequences. The tenn ~'internaiization moti~' refers to an amino acid internali~ion signal having a ti~ht turn structure. The terms intemalkation motif and internaleation signal are used interehangeably. The term "coreU
refers to the smallest sequence of amino acid residues involved in cell surface internalization.

WO 93/1818~, PCr/US93/01669 2~3~1 Amino acids referred to herein may be identified accordin~3 to 'the followin~
.three-letter or on~letter abbreviations:
.
Thre~L~tter On~Letter Amino Acid Abbreviation Abb~on L-Aianine Ala A
L-Arginine Arg R
L-Aspara~ine Asn N
L-Aspartic Acid Asp D
L Cysteine Cys C
10 ~ L~llnamine Gln Q
L-Glutamic Acid Glu E
L-Glycine Gly G
L-Histidine His H
L-lsoleucine lle L-Leucine Leu L
L-Lysine Lys K
L-Methionine Met M
L-Phenylalanine Phe F
L-Proline Pro P
L~erine Ser S
L-Thre~nine Thr T
L-Tryptophan Trp W
L-Tyrosine Tyr Y
L-Valine Val V

2~ The heterologous intemaiization signals utilized herein mzy be the same as or different from intemali~a~ion signal already present in the cell. When DNA
encoding heterologous sign~l is introduced into a target cell, the introdueed nucleotide sequence may further comprise addrtional nucleotid~ sequencP
encoding a cell surface receptor. This surface receptor may be the same as or different ~rorn receptor already Qresent in the target cell.

WO 93/18185 PCI`/US93/01669 Thus, thc present invention envisions embodiments wherein heterologous intemalization signal is introd~ced irlto cells having receptors contaWng intemalization signals as well as those having receptors without internaliza-tion signals. Further, when the internalization sequenc~ is introduced as s part of a cell surface receptor, the receptor may be the same as or dfflerent from a receptor already present in the target cell.

Introduction of heterologous internalization signal into target cells according to the invention can be accomplished by several means. For example, a signal can be introducsd as a peptide or as a nucleotide~se~uence encoding such a peptide. Moreover, the introduced sequence can furth~r comprisa addftion~l nanking sequence. Such additional flanking sequence can encode or can be a cell surface receptor.
.
Internalization signal peptide or nucleic acid sequence encoding the peptide may be introduced into a cell surface receptor by several methods including:
1~ subs~uting residues in the receptor by residues in the introduced internalization sequence (one for one replacement of residueslbases); 2) inserting an internalization ~ignal between ~wo residues/bases in the resident receptor so that the receptor sequence becomes !onger; 3) replacing some residues in the resident reeeptor with a greater number of intemalization si~3nat residues such that the receptor becomes longen The method selected for introduction of interna~ization sign~ peptide will be dictated by strucitural and expenmental considerations on a case by case basis.

Preferred locations for introduction of internalization signal into a receptor sequence can be ascertained by locating endogenous internalization signal and by locating positions in the receptor sequence ha~rin~ t}ght turn W0 93/1818~ PCl'lUSg3/01669 -1~

conformation and paclcing interactions that can accommodate introducad signal. Structure and pac~ing infor~mation can be obtained from crystallog-raphy, NMR, or high electron resolution microscopy. In situations where such information is not available, commonly used and availablo secondary st~ucture predic~on algorithms may be used to locate sequencs regions that likely fold as tight turns. In cases involving substitution of internalization sequence for receptor sequence, tight tum regions of the receptor which ar~
most sequence similar to the internalization sequence are preferred sites for subs~tution since substitutions in such sites minirnkes structural destablia-bon. Sequence similari~y can be determined by mutation data matrices and other measures of sequence similarity, such as residue size and polari~y.
:
The introducbon of heterologolJs intemalization signal serves to modulate the transport of ligand into a cell having a surface receptor reactive with that ligand. This modulation can inducs either an increase or a decrease in endocytosis, depending upon the choice of heterologous intemalization signal. As disc~ssed earlier, the key to receptor-mediated endocytosis is the intemalization signal present in the cytoplasmic tail of the surface receptor. It is the internalization signal that regulates the uptake of cell surface receptor.

Identification of a heterologous internalization sisn21 which can modulate intemalization of cell surface receptor is accomplished by incubating oells having such receptors in the presence or absence of the sequence suspected of being an internalization signal. Cells are incubated according to the method of Jing, et al., J. Cell Biol., 110:283, 1990. Optionally, inc~bation is carried out in the presence of ligand and internalization of cell surface receptor, in the presence or absence of secuence, is measured.

~ 93/1818~ 21 3 01 0 0 PCl`/US93/01669 The invention also provides a method for inhibiting intern~ ion of cell surface receptor. This is accomplished by introducing into a cell having a r~ceptor that clusters in coated pits, an effec~ve amount of heterolosous intemalkation signal peptide, wherein the hetero!ogous peptide competes with intemal~ation signal peptide of the receptor for binding with adaptor protein. The term "effec~ve amounr' refers to the amount of peptide which results in inhibition of endocytotic vesicle forrnation. Delivery of an effecffve amount of the intemalization signal peptide can be accomplished by one of the m~chanisms described herein, such as by encapsulation in liposomes, or other methods well known in the art.

Intemalkation signal peptides that can be utilized in the present invention are characterized by having a tight turn and the amino acid sequence:

X~ (3x4x5x6x7x8x9xl 0X
where X5 - X8 constitutes the core sequence, X1 - X4 residues and Xg - Xl 1 residues are op~ional and wherein:

Xl, when present, is leucine or glutarnic acid;

X2, when present, is isoieueine, methionine or proline;

X3, when present, is any amino acid residue;

wo 93/1818~ - Pcr/uss3/ol66s 2~30~oo X4, when present, is selected from alanine, polar amino acids, or aromatic amino acids, and when X3 is al50 present, at least one of the X3 or X4 residues is polar;

X5 is an aromatic amino acid when r~sidues X1 - X4 are not present, or is selected from aromatic amino acids or polar amino acids when - at least residu~ X4 or additional upstream residue~s) is present;

X6 is a po'ar arnino acid or alanin~; .

X7 is selected from polar amino acids or alanine when residues X1 - X4 are not preser~, or is any amino acid residue when at ieast ~ -~
X4 or addi~onal upstrearn residue(s) is present;

X8 iS selected from aromatic amino acids or hydrophoblc amino `-acids;

Xg, when present, is sarine or alanine;

X10, wherl present, is alanine or leucine; and X11, when present, is alanine or phenylalanine;

wherein at least one of residues X3, X5, and X8 is an aromatic amino acid and further wherein residues X1, X2, X3, X1 0, and X~ 1 can only be presen~
when the naxt adjacent residue(s) relative to the core is presant.

w~ 93/1818~ 2 1 3 ~ I U O PCl/US93~01669 - A first preferred group of internalization signal peptide~ is chara~terized as having the amino acid sequence:

X5X~8 -wherein X5 - X8 are as previously defined with the proviso that X6 and X7 oannot both ba alanine.

A more preferred group of intemalization signal peptides having the ;
sequence X5X6X7X8 are those wherein:

X5 is phenylalanine or tyrosine;
X6 is alanine, arginine, glutamine, serin~, or threonine;
X7 is alanine, arginine, aspartic acid, glycine, glutarnic acid, histidine, Iysine, or threonine; and X8 is isoleucine, leucine, rnethionine, phenylalanine, valine, or tryptophan.

Most preferred internalization signal peptides havin~ the X~X6X7X8 sequence are selected from the group consisting of YTRM, YARF, YTRI, YQDL, YTKF, YSKV, YTRW, YRHV, YSAF, YQTI, YTAF, YTGF, YTEF, t I PIF, and YTRF.

WO 93/1818~ PCI'/US93/01669 ~,~ 30~

A s2cond preferred group of internalization signal peptid~s is characterized as having the amino acid ~equence:

wherein X3 - X8 are as defined pr~viously.

A more preferred group of internaleation signal peptides having the ,, q nce X~X4X5X6X7X8 are those wherein: .

X3 is asparagine, leucine, methionine, phenylaJanine, proline, or ~ :~
tyrosine;
X4 is alanine, aspartic acid, gluta mine,lysine, phenylalanine, or serine;
X5` is aspara~ine, gl~rtamine, or tyrosine;
X6 is arginine, glycine, proline, serine, or thr~onine;
X7 is alanine, ~rginine, Iysine, phenylalanine, or valine; and X8 is isoleucine, leucine, m~thionine, phenylalanine, tryptophan, tyrosine, or valine.

1~ Most preferred intemalization signal peptides having ~e sequenc~
X5X~X7X8 are selected from the grou~ consisting sf Y ~S K~/, NF~tFUU
I AYTP~F, PQQGFF, FDNPW, MSYTRF, or LSYlRF.

Novel compositions of the present invention comprise peptides having a tight turn and the amino acid sequence:

X1 X2X3X4XsX6X7X8XgX1 oXl 1 ~1~ 93/18185 2 1 3 ~ 1 0 0 PCI/US93/01669 wh~re X5 - X8 constitutes the core sequence, X1 ~ X4 residues and Xg - X1 1 residues are optional and wherein:

X1, when present, is leuane or glutamic acid;

X~, when pres~nt, is isoleucine, methionine or proline; ; -~ ,,.
X3, Arhen present, is any amino acid residue;

X4, when present, is s~iected from alanine~ polar amino acids, or aromatic amino acids, and when X3 is also present, at least one of the X3 or X4 residues is polar; -X5 is an aromatic amino acid when residues X1 - X4 ar~ not pr~sen~
or is selected from aromatic amino acids or polar amino acids when at least residue X4 or additional ups~ream residue(s) is present;

X6 is a polar amino acid or alanin~;

X7 is sele~ed from polar amino acids or alanine when residues X1 - X4 are not presen~, or is any amino acid residue when at least ;
X4 or additional upstrearn residue(s) is present;

X8 is selected from aromatic amino acids or hydrophobic amino acids;

Xg, when present, is serine or alanine;

WO 93/1818~ PCI/US93/01669 ~, ~3 o ~

X10, when present, is alanine or leucine; and X11. when present, is alanine or phenylalanine;

wherein at least one of residue X3, X5, and X8 is an aromatic amino add and further wherein residues Xl, X2, X3, X1 0, and X1 1 can only be pres~nt 5~ wh~n the next adjacent residue(s) relative to the core is present, provided th~t sequences selec~ed from the group ccnsisting of FXN~X;Y, GPLY.
PPGY, and YXYXKV, where X stands for any arnino acid, are ~xciuded.

As employed herein, the term "polar amino acid" includes glycine, serine, threonine, his~dine, tyrosine, proline, aspartic acid, asparagine, glutamic acid, giutamine, arginine, and Iysine. (Rose, et al., Scien~e, 9:834, 1985~.

~he term Uaromatic amino acidU refers to any amino acid containing an unsaturat~d carbocycli_ ring including phenylalanine, tyrosine, and tryptophan.

The term "hydrophobic amino a~idU inciudes cysteirle, valine, isoleucine, leucine, methionine, phenylalanine, and tryptophan. (Rose, et al., Scienc~, æg:834. 198~). llle term "large hydrophobic amino acid' refers to amino acids seiected from the grsup consisting of valine, isoleucine, leucine, methionine, phenylalanine, and tryptophan.

In addition to increasing endocytosis by addins3 ac~ive in~ernalization signat (where "active" is defined as having at least twice the internaliz~ion emciency of a receptor lac~ing a signal), the present invention also embraces inhibition of endocytosis. Inhibition is accomplished by 1) introducing signal shown ~O 93~18185 2 1 3 ~) 1 0 0 P~T~US93/01669 -17- i ~ '`

to be inac~ive when subsffluted to an active resident signal or 2) using activa internalization peptides as competitive inhibitors of receptor internalization (i.e., the peptide competes with intact receptor for bindins3 to adaptor complex). Inactive sisnals which can be used according to the first method include sequences cor~sponding to the core X5X6X7X8 and selected from the group consisting of ATP~F, GTRF, ATRA, YTRG, FQDI, IGSY, NTLY, HLAF, PPGY and NPW.

Generation of inactive internalization signals can be accomplished by mutating aromatic residues occurring in positions X3, X4, X5 and X8-to non~
aromatic residues.

The present invention can utilize any nucleotide sequence encoding peptides -~
described above. These peptide encoding nucleotide sequences can further comprise flanking nucleotide sequence encoding a cell surface receptor.

In the present invention, nucleotide sequences encoding internaliza~ion 1~ signal may be introduced into a host C211 by means o~ a recombiriant expression vector. The tsrm "recombinant expression vec~or" refers to a plasmid, virus or other vehicle known in the art ~hat has been manipulated by insertion or incorporation of intemaiization signal sequences. Such expression vectors typically contain a prsmotor sequence which facilitates emcient transcription of ~he inserted sequence in the host. The expression vecsor also typically contains specific genes which allow phenotypic selec~}on of the transformed cells.

Alternatively, nucteotide sequences encoding intemalization signal can be introduced directly as a plasmid, for example, by microinjection, or WO ~3/1818~ PCI`/US93/01669 ?.,~30~G

'.

transfection.

The present invention also provides methods for the treatment of disease employing gene therapy that modulates receptor mediated transport of ligand into a cell or intemalization of cell surface receptor. Such therapy can be effected by introduction of internalization signal into cells of a subject having the disease. Delivery of internalkation signal can be achieved usinç~
techniques well known in the art. For example, when an int~malization signal is introduced by means of a nucleotide sequence, a recombinant expression vector, such as a chimeric virus, or a colloidal dispersion system can be employed.

Various viral vectors which can be utilized for introduction of intemalization signal according to the present invention, include adenovirus, herpes virus, vaccinia, or, preferably, an RNA virus such as a retrovirus. Preferably, the retroviral vector is a derivative of a murine or avian retrovirus. Example~ of retroviral vectors in which a single foreign ~ene can be inserted include, but are not limited to: Moloney murine leuksmia virus (MoMuLV), Harvey murine sarcoma virus (HaMuSV), murine mammary tumor virus (MuMlV), and Rous Sarcoma Virus (RSV). A number of additional retroviral vectors can incorporate multiple genes. All of these vectors can incorporate a gene for a selectable marker so that transduced celis can be identified and cultured.

By inserting a polynucleotide encoding the internalization signal of interest into a target specific viral vector, the internalization signal becomes targetedto specific cells. Retroviral vectors can be made tar~et specific by inciuding in the retroviral vector a polynucleotide encoding a su~stance which binds to cell surface tar~et. Preferred targeting is accomplished by using an ~vo 93~18185 213 01 0 0 PCT/US93/01669 -1 9~

antibody to target the retroviral vector. Those of skill in the art will know of, or can readily ascertain without undue experimentation, specific polynucleo-tido sequences which can be inserted into the retroviral genome to allow target specific delivery of the retroviral vector containing the intemalization S signal polynucleotide.

Sincs recombinant retroviruses are defective, they require assistance in order to produce infectious viral particles. Tliis assistance can be provided, for example, by using helper cell lines that contain plasmids encoding all of the structural genes of the retrovirus under the control of regulatory sequences within the LTR (long terminal repeat). These plasmids are missing a nucleoffde sequence which enables the packaging mechanism to recognize an RNA ~ranscript for encapsidation. Helper cell lines which have delebions o~ ~e packaging signal include, but are not limited to, ~2, PA317 and PA12, for example. lllese cell rmes produce empty virions, since no senome is paclcaged. If a retroviral vector is introduced into such cells in which the packaging signal is intact, b~t the structural genes are replaced by other genes of interest, the vector can be pack~ed and vector virion produced. The vector virions produced by this method ran then be used to infect a tissue cell line, such as NIH 3T3 cells, to produce large quantitiesof chimeric retroviral virions.

Alternatively, NIH 3T3 or other tissue culture cells can be directly transfectedwith plasmids encoding the retroviral s~ructural ~enes gag, pol and env, by conventional calcium phosphate transfe~ion. These cells are then transfected with the vector plasmid containing the genes of interest. The ~c resuning cells release the retroviral vector into the cunure medium.

. ~,.

,~

WO 93/1818~ - PCI`/US93/01669 ~,~3~
-2~

An altemative use for recombinant retroviral vectors compris~s the introduction of polynucleotide sequences into the host by means of skin transplants of cells containing the virus. Long term expression of foreign genes in implants, using cells of fibroblast origin, may be achieved if a strong housekeeping gene promoter is used to drive transcription. For example, the dihydrofolate reductase (DHFR) gene promoter may be used.
Cells such as fibroblasts, can be ir~ec~ed with virions containing a retroviral construct containing the intemalization signal gene of interest together wi~h a gene which allows for specific targeffng, such as tumor-associated anffgen (TM), and a strong promoter. The infected cells can be embedded in a colla~en matrix which can be grafted into the connecbve tissue of the dermis in the recipient subject. As the retrovirus proliferates and escapes the matrnc it will speci~ically infect the target cell population. In this way the transplanta-tion results in increased amounts of internalization signal peptide being produced in cells manHeshng the disease.

Another targeted delivery system for introduction of intemaleation signal peptides, or nucleotides encoding same, is a colloidal dispersion sys~em.
Colloidal dispersion systems include nacromolecular complexes, nanocapsules, mfcrospheres, beads, and lipid-based systems includin~ oil-2~ in-water emulsions, micelles, mixed micelles, and liposomes The preferred colloidal system of this invention is a liposome.

Since internalization signal peptide may be indiscriminant in its action with respect to cell type, a targeted delivery system offers a si~nificant improve-ment in therapeutic applications over randomly injected non-spec~ic liposomes. A number of procedures can be used to covalent)y altach either polyclonal or monocJonal antibodies to a liposome containin~ internalization ~
.'.:

~ 93/18185 PCI'/US93/01669 ~ 1 3 ~ 0 signals to render the liposome ~rget specific. Antibody-targeted liposomes can include monoclonal or poly~,lonal antibodies or fragments thereof such as Fab, or F(ab')2, as long as they bind efficiently to an epitope on the target cells. Uposomes may also be targeted to cells expressing receptors 5- for hormones or other serum factors.

Uposomes are artificial membrane vesicles which are useful as in vitro and in ~nvo delively vehicles. It has been shown that lars~e unilamellar vesicles (LUV), whic~h range in size from 0.2~.0 um can encapsulate a substantial percentage of an aqueous buffer containing large macromolecules. RNA, DNA, intact virions and peptides can be encapsulated within the aqueous interior and be delivered to cells in a biologically active form (Fraley, e~ al.Trends Biochem. Sci., 6:77, 1981). In addition to mammalian cells, liposomes have been used for delivery of polynucleotides into plant, yeast and bacterial cells. In order for a liposome to be an emcient transfer vehicle, it should be capable of: (1) encapsulation of pep.ides or nucleotides of interest at high efficiency without compromisin~ biological ac~ity; (2) preferential and substan~al bindin~ to target cells relative to non-target cells;

(3) delivery of aqueous contents of vesicle to the target cell cytoplasm at high emciency; and (4) accurate and effective expression of genetic information (Mannino, et al., ~iotechniques, ~:682, 1988).

~he cornposition of the liposome is usually a mixture of phospholipids, particularly high-phase-transition-temperature phospholipids, usually in combination with steroids, especially cholesterol. Other phospholipids or other lipids may also be used. The physical characteristics of liposomes depend on pH, ionic strength, and the presence of divalent cations.
,.

WO 93/1818~ 0 1 ~) PCl`/US93/01664 ` ~ .

z Examples of lipids useful in liposome production include phosphatidyl compounds, such as phosphatidy!glycerol, phosptlatidylcholine, phos-phatidylserine. phosphatidylethanolamine, sphingolipids, cerebrosides, and gangliosides. Particularly useful are diacylphosphatidylglycerols, where tho s lipid moiety contains from 1~18 carbon atoms, particularly from 1~18 carbon atoms, and is saturated. Illustrative phospholipids includs egg phosphatidylcholine, dipalmitoylphosphatidylcholine and distearoylphos-phatidylcholine.

The surface of the targeted delivery system may be modified in a varie~y of ways. In the case of a liposomal tar~eted d~ ery system, specific lipid groups can be incorporated into the liposome in order to maintain the target directed binding substance in stable association with the liposome. Various linking groups can be used for joining the lipid chains to the target directed binding substance.
.,. -In general, the targeted delivery system will be directed to cell surface receptors thereby allowing the delivery system to find and-"home jnu on the desired cells. Alternatively, the delivery sys~ern can be directed to any cell surface molecule preferentially found in the cell populabon for which treatment is desired provided that the cell surface molecule is capable of association with the delivery system. Antibodies can be used to target liposomes to specific cell-surface molecules. For example, certain antigens expressed specifically on tumor cells, referred to as ~mor-associated antigens (TAAs), may be exploited for the purpose of targeting antibody-internalization signal-containing liposomes directly to a malignant tumor.

93/1818~ - 2~301Q;~ Pcr/Us93/ol669 . ' Because tho present invention identifies sequences involved in the intemalization of cell surface receptors, it is possible to design therapeutic or diagnostic protocols utilizing these sequences. Thus, where maintenance of a disease stats is related to intemalization of a cell su~ace receptor, ths S na~e intemalization sequ~nce can be utilked to design sequences which interfere with the function of the native intemalkation signal. This approach utilkes, for example, antisense nucleic acid and ribozymes to block transtation of specific receptor mRNA, either by masking the mRNA with antisense nucleic acid or by cleaving it with ribozyme.

Antisense nucleic acids are DNA or RNA molecules that are complementary to at least a portion of a specific mRNA molecule (VVeintraub, Scient~fic Arneric~n, ~:40, 1990). In the cell, the ar,tisense nucleic acids hybridize to the corresponding mRNA, forming a double-s~anded ~olecule. The an~bsense nucleic acids in~erfere with the translation of the mRNA since the cell will not translate a mRNA that is double-stranded. Antisense oligomers of about 15 nucleotides are preferr~d, since they are easily synthesized and are less likely to cause problems than iarger molecules when introduced into the target receptor-producing cell. The use of antisense methods to inhibit the in vi~ro translation of genes is well known in the art (Marcus-Sakura, An~l.Biochem., 172:289, 1988).

Ribozymes are RNA molec~les possessing the ability to specifically cleave other single-stranded RNA in a manner analogous to DNA res~riction endonuc~eases. Through the modification of nucleotide sequences which encode these RNAs, it is possible to en~ineer molecules that recosnize specific nuc~eotide sequences in an RNA molecule and cteave it (Cech, JAmer.Med~ssn., 260:3030, 1988). A major advantage cf this approach is WO 93/18185 PCJ/US93/01669 _~

2,~3 ~ -24-that, because ribozymes are sequence-specific, only mR~As with particular sequences are inactivated.

There are nNo basic types of ribozymes namely, tetrahymena-typ~
(Ha~elhoff, Natur~, 334:585, 1988) and Uhammerhead~-type. Tetrahymena-type ribozymes recognke sequences which are four bases in length, while Uhammerhead"-type ribozymes recognke base sequences 11-18 bases in Iength. The longer the recognition sequence, the greater the likelihood that that sequencs will occur exclusively in the target mRNA species. Cons~
quently, hammerhead-type ribozymes are preferable to tetrahymena-type ribozymes for inactNating a specific mP~NA species and 18-based recogni-tion sequences are preferable to shorter recognition sequences. -Antisense sequences can be therapeutically administered by techniques as described above for the administration of heterologous intemalization signal polynucleotides. Targeted liposomes are especially preferred for therapeutic delivery of antisense sequences.

!~'0 93/18185 PCI`/US93J01669 -25- ~:

MATERIALS AND METHODS

The below described materials and methods were utilized in transplanting internalization signals and analyzing their structure.

A. Oligonucleotide Site-Directed Muta~enesis :

A Cla I fragment containing the entire coding region of the human TR was cloned into ~he phagemid pBluescript SK (Stra~agene, La Jolla, CA) (Jing, ,- et al., .1. Cell 8iol., 110:283294, 1990). Oligonucleotide site~irected mutagenesis was performed with pBluescript SK phagemid templates ôf the wild-type TR by the method of Kunkel, (Proc. Natl. Acad. Sci. USA, ~:48&~
492, 1985) usin~ the Muta-~ene mutagenesis kit (Bio-F'~ad, Richmond, CA) as described previously (~::ollawn, etal., Cell, 6~:1061-1072, 1990). Mutants were selected by restriction mapping and Cta I fragments encoding the mutant receptors were then excised and cloned into the retroviral expression vector RCAS (Hughes, et al., J. Viro/ogy, 61:300~3012, 1g87~ or BH-RCAS. -:
The mutations were verified by dideoxynucleotide sequencing ~Sanger, et al., Prw. Natl. Acad. Sci. IJSA, 74:5463-5467, 1977; Tabor and Richardson, Proc. Natl. Acad. Sci. U~A, 84:4767~771, 1987) of the retroviral construc~s using the Sequenase kit (USB Corp., C:leveland, Ohio) according to the manL facturer's direc~ions. ~;

wo 93/1818~ PCr/uss3/0166s ~

2~3~ -26-B. Cell Culture and Expression of Human TRs in CEF

Primary chicken embryo fibroblasts (CEF) were obtained from nins day embryos (SPAFAS, Norwich, Conn.) and grown in Dutbecco's modifie:l Eagle's medium (DMEM) supplemented wit~ 1% (vh) chicken s~rum, 1%
(v/v) defined calf bovine serum (Hyclone, Logan, Utah), 2% (v/v) tryptose phosphate broth (D-~co, Detroit, Ml). CEF were transfected with retroviral construct prepared in (A) using 30 micrograms DNA per 1 Ocm tissue culture plate of ~40% confluent cells using the polybrene-dimethyl sulfoxide method ~ (Kawai and Nishizawa, Mol. Cell. Biol., 4:1172-1174, 1984).

Stable expression of wild-type and mutant human TR5 was achieved using a helper-independent retroviral vector, BH-RCAS, derived from Rous sarcoma virus (Hughes, et ~I., J. Viro/ogy, 61:300~3012. 1987). 1-2 wk after transfection, the CEF stably expressed wild-type and mutant receptors on their cell surface as a result of infection by recombinant virus. Cell surface expression was confirmed by 1251-transferrin binding at 37~ C under steady~
state conditions.

To select for high expression of mutant human T~s, eells were grown in selecti~n media (DMEM supplemented with 3% (~Jiv) horse serum, 2%
tryptose phosphate broth and 50 micrograms/ml diferric human transferrin ~Tfl (Miles Scientific, Naperville, IL)) starting 10-14 days after transfec~ion (Jins, et al., ~1. CeJI Biol., 110:283-294, 1990).

~''" 93/18185 PCI`/US93/01669 213~0~

C. 125l-T~ Binding Diferric human transferrin (Tf) was labeled wi~ 1251 to a specific activ~ty of 24 ,uCi/mg using Enzymobeads (Bio-Rad, Richmond, CA) according to the manufacturer's directions. Cells were plated at a densfty of 7.5 x 104 cells/cm' in 2~well Costar tissue culture plates 24 hr prior to the binding assay. Cells were incubated in serum-free DMEM for 1 hr at 37^ C and then washed once with ic~cold 0.1 5M NaCI-0.01 M Na phosphate buffer (pH 7.4) c~ntainin~ 0.1% bovine serum albumin (BSA-PBS~. 1251-Tf (4 micro-~ grams/ml) in 0.15 rnl BSA-PBS was added to triplicate wells and incubated at 4- C for 60 min. Cells were then washed three times with ice-cold 0.5 ml BSA-PBS, removed from the wells with 0.15 ml 1M NaOH, and the radioactivity counted in a gamma counter.

D. Analysis of Transferrin Internalization at Steady-State For studies of the steady-state distribution of mu~ant and wild-type human TRs, CEFs expressing TR mutants were plated in triplicate wells as des~ribed for the binding studies. The cells were first inoubated for 1 hr at 37 C in serum-free DMEM, and then incubated with 4 micrograms/ml 1 251-Tf in 0.1% BSA in DMEM for 1 hr a~ 37 C. llle media was removed and the cells were washed three times with 1 ml of ice-cold BSA-PBS. The cells were then incubated twice for 3 min with 0.5 ml of 0.2 M acetic acid-0.5M
NaCI (pH 2.4) to remove surface~bound 1251-Tf (Hopkins and Trowbridge, J. Cell. Biol., 97:508-521, 1983). Cells were then removed from the wells with 1 M NaOH and radioac~ivity in the acid wash and the cell Iysate was WO93/1818~ . PCr/US93/01669~
~3~oo determined. More prolonged incubation with the acid wash did n~t affect the radioactivity released (Jing, et al., J. Cell Biol., 110:283 294, 1990).

int [TR]Sur = keXt ~TR]jnt,~the rate of intemalization k of cell surface Tf-TR complexes, [TR]sUr, equals the rate of e)*ernalkation of the intemal pool of apoTf-TR complexes, [TRljnt, assuming an insignificant rate of i~.tracellular degradation of receptors (Collawn, et al., Cell, ~:1061-1072, 1990). The term "apoTf" refers to transferrin having no bound iron. ;
:.
,- E 59Fe Uptake Measurements ; . .
Human apoTf was labelled with 59Fe (FeCI3; Amersham Corp., Arlington Heights, IL) to a specific activity of ~10,uCi/mg using nitrilotriacetate (Batesand Schlabach, J. Biol. Chem., 248:322~3232, 1973). Cells were plated at a density of 7.5 x 104 cellslcm2 in 2~well Costar tissue culture plates 24 hr before the assay. The following day, cells were washed twice in prewarmed (37 C) serum-free DMEM and then were incubated in DMEM containing 0.1% BSA and 4 ~lg/ml 59Fe-Tf at 37 C for 0, 1, 2, 3, and 4 hrs. After the indicated times, ~he media was removed, and the cells were washed three times w~h ice cold 0.1% BSA in PBS. Cells from triplicate wells for each time point were removed in 0.5 ml 1M NaOH, and the radioactivity was counted in a gamma counter. The relative levels of human TRs expressed on the various C~F populations were deterrnined in each experiment. After pre-incubation for 1 hr at 37- C in serum-free DMEM, triplicate weJls of cells were incl~bated with 4 yglml 1251-labelled Tf on ice for 1 hr and then washed three times with 1 ml of ice-cold BSA-PBS, and the radioactivity bound to the cells was determined.

~- 93/181~5 PCl`/US93/01669 213~1~0 F. Comparison of Protein Structure Analogs of Six-Residue Internal~
ization Motifs Analogs of the Man-~PR and LDLR sK-residue signals were identihed by searching the ~550 protein structures in the October 1990 Protein Data Bank (PDB) (Bemstein, et al., J. Mol. Biol., 193:77~791, 1977) for sequences matching the intemalization pattems and having the final fouF residues in a s1urfac~exposed tight turn ~a helical tum or a reverse turn). The computer program Sequery (developed by Leslie Kuhn and Michael Pique as the ~ successor to Searchwild, described in Collawn, et al., Cel/, 63:106t-1072, 1990) searched PDB sequences for the six-residue Man-6-PR pattern (Y,F) p()(Y,F)(X) (polar:K,R,H,N,Q,D,E) (large hydrophobic:v~l~LM~F~y~vv) and the six-residue LDLR pattern (F,Y)(X)(N)p()(X)~`f,F,W) with X bein~ otherthan Gly for both patterns. Homologous matches and matches lacking side chain ~dimensional coordinates were deleted, and those structures containing su~face-exposed tight turns in the final four positions were examined using molecular ~raphics to ascertain preferred posit~ons for the side chains. The backbone atoms of the turn positions in each analog were superimposed on an ideal type I turn using th~ molecular graphics program Insight ll (Biosym Technologies, San Diego, CA) on a Silicon Graphics Iris 4D310VGX
workst~tion.

The following examples are intended to illustrate but not limit the invention.
While they are typical of those that mi~ht be used, other procedures known to those sl<illed in t~e art may alternatr~Jely be used.

WO 93~1X185 PCl /US93/016S9 . -~,~3~0 . 3~ ~

:

TRANSPLANTATION OF HETEROLOGOUS INTERNALIZATION SiGNALS

. . .
Studies wer~ performed to determine wh~hér intemalkation signals of cell surface receptors are interchangeable. These studies involved the transplantation of the internalkation signal~ from low densi~y lipoprotein receptor (LDLF~) and mannose-~phospate receptor (Man-6-PR) into TR.
Oligonucleotid~directed mutagenesis was used to replace the four residue YTR~ internalization signal from the cytoplasmic tail of the wild-type TR with the four-residue sequences from LDLR and Man-~PR signals and the six-residue TF~ sequence LSYrRF by the six-residue LDLR and Man-6-PR
signals (TABLE 1).

Wf' 93/18185 PCI`/US93/01669 21 30 L ~ o TABLE I

- INTERNALIZATION~I~QUENCE~
IN TRANSFERRIN REGEPTOF~ Ç:QNST~

Length ot `
Sourçe Sequence Sequencea -~ -1 1 2 3 4 '- W~TR 4 20~ T R F23 ` Man-~PF~ 6 8y K Y S K V23 Man~PF~ 4 2~ S K V23 LDL~ 6 F D N P V y23 LDL~ 4 20N p V y23 a For the four-residue substitutions, the TR signal YTRF, was chan~ed ~o YSI~V for the bovine Man-~-PR si~nat and NPW for the human LDLR signal. For the six-residue subs~ ions, the TR se~uence, LSYTRF, was changed ~o YKYSKV or to FDNPW. Numbers ab~v~
the sequences (-2 through 4~ refer to positions rela~ve to the tum reglon implicated in TR and LDLR ~Collawn, etal., GeJI, 63:1061-1072.
1990) with position 1 sta~ing the turn. TR residue numbers appear as superscripts. Critical residues of internalization sequences identified by mutation to alanine are underlined (TR, Collawn, et al., Cell, 63:1061-1072, 1990; Man-6-PR, Canfield, et al., ~1. BJol. Chem., ~:5682-5688, 1991; LDLR, Chen, et ~I., J. Bfol. Ghem.~ ~:311 3123, 19~0), and aromatic and large hydrophobic residues are shown in bold type.

WO 93/1818~ PCI/US93/01669 .
2~3~

-32- . :

.::
The four-residue sequence, NPW, from the LDLR was tested becauso the NPXY pattern is also found in the cytoplasmic tails of other transmembrane proteins including the EGF and insulin receptors and therefore may play a role in the intemalization of these proteins (Chen, et al., J. Biol~ Chem., ~:311~3123, 1990; Backer, et at., J. Biol. Chem., ~:1645~16454, 1990)..
The YS~Y sequence from Man-~PR was tested because mutational analysis of Man~PRs (Canfield, et al., J. Biol. Chem., 266:5682-5688, 1991) indicated that this tetrapeptide was almost as efficient as the complete intemalization sequence in promoting high emciency endocytosis.

Efficient expression of wild-type and mutant human TRs in chicken embryo fibroblasts (CEF) was achieved using a helper-independent retrovir~ vector, BH-RCAS, derived from Rous sarcoma virus (see Materials and Methods).
~s~
The relative intemalization efficiencies of wild-~ype and mutant receptors were .

determined ~rom Fe uptake m~asurements and their intracellular accumula-ffon under steady-state condWons.

Two representative Fe uptake experiments whioh include all-the mutant TRs ana~zed are shown in FIGUR~ 1, and the results of steady-stat~ internaliza-tion studies and Fe uptake experiments are summarized in TABLE ll. 80th assays gave similar values for the internalization emciencies of mutant TRs relative to wild-type receptors, and the capacity of mutant receptors to mediate Fe uptake was highly reproducible.

~"- 93/1818~ PCr/VS93/01669 2130100 :~

-33~

COMPARISONS OF STEADY~S~LD~=~DS~
AND S9F~ UPTAKE EFFIC ENCIES OF HUMAN TR MlJTANTS

H~nTR % lntemaleed HTR Internalkation Fe Uptake Inten~ization ~u~ts at Steady~tat~Efficiency (%) (atornslreceptor/h) Emdency (%) ~:

~a 6~2+ l.gb (7)C 100 34 + 2 (6) 100 :~

~59 'railless^ 24.7 + 1.6 (2) 17 6 + 1.2 (5)d 18 ~;
LDLR 4mer 25.7 + 0.9 (7) 18 8 + 0.6 ~;) 24 IDIR 6mer 40.9 + 1.6 (7) 37 19 + 1.9 (5) ~6 Man~PR 4mer 61.7:t 3.~ (4) 86 30* 3.4 (3~
Man~PR 6mer 68.5 + 1.8 (4) 1~6 33 + 4.~ (3) 97 F23M 7Q.9 + ~8 (3) 130 37 ~ 2.7 (33 109 F231 60.5 ~ ~5 ~3) 82 33 + 3.8 (3) 97 F23W 58.7 + 5.7 (3) 76 27 + 4.1 (3) 79 a ~Id-~pe Mean _ standard enor d Number of independent experiments From Collawn et al., Cell, 63:1 Q61, 1990 WO 93/~8185 PC1'11~S93/01669 : . ~

~,~30~

Strikingly, mutant TRs containing the Man-~PR six-residue signal, YKYS CV, were intem~lized as emciently as wild-type receptors. Mutant receptors containing the Man-~PR four-residue signal, YSICV, were also rapidly intemalized and had an internalization efficiency of approximately 87Yo that of wild-type receptors (TABLE ll, FIGURE 1A). Mutant TRs containing the LDLR six-residue signal, FDNPW, were intemalized at ~50% the rate of wild-type receptors, indicating that the LDLR six-residuc signal had significant ac~ty when transplanted into the TR cytoplasmic tail. In contrast, mutant , lRs containing the four-residue LDL~ signal, NPW, were intemalized at essentially the same rate as tailless receptors (TABLE ll, FIGURE 1A and B).
The results of these experiments indicate that L~LR and Man-~PR
intemalization signals can subs~tute for the TR signal and mediate high offiaency intemaleation of the TR. The lack of activity of the transplanted LDLR four-residue sequence, NPW, implies that the complete LDLR-i~temalizatfon signal is the six-residue sequence, FDNPW.

~he analysis of mutant TRs with transplanted four- or six-residue Man-~PR
and IDLR intemali7ation signals shows that an amino-terminal aromatic residue is required for activi~y, but that a large non-aromatic hydrophobic residue may substitute for an aromatic residue at the carboxy-terminal position. To investigate whether a carboxy-terminal aromatic residue is required for actNity of the TR intemalization signal, Phe-23 was altered to Met, then lle, then Trp. Functional analysis of mutant TRs with Phe-23 altered to Met or lle indicated that their internalization rate and ability to mediate Fe uptake were identical tn wild-type TRs (TABLE 11). Mu~ant TRs with Phe-23 altered to Trp were also rapidly internalized (FIGURES 1 and 2) ..'~,'~

~'') 93/18185 PCT~US93/01669 213010~

-35- :

STRUCTyRAL ANALYS!S O~ INTERNALIZATION SIGNALS ~:

The ac~ o~ lDLR and Man-~PF( six-residu~ intemai za~on sign~s transplanted into the TR suggests that their actrve confonnations ara related to th~t of ~e na~ve TR intemalization motif. ( iiven the tight tum structural preference of the four-residue TR and LDLR motifs ~Collawn, et al., Cell, 63:1061, 1990) and th~ TR internalization activity of the six-residue L.DLR and Man-~PFl sequences, the three-dimensional arrangement of critica~ side chains was investiga~ed in c~stallographic protein struc~res containing six-residue sEquences matching the internalization pattems, wi~h the four :-carboxy-terrninal residues in ti~ht turns.

Wlthin their native protein sequences, the TR, U)LR, and Man~PR
intemalization signais were all predicted by a seconda~ ru~re algor~hm (Chou and Fasman, Adv. ~nzymo~., 47:4~, 1978~ to be turns before or at the start of an alpha helix. Four stru~ural anatogs were iden~fied for the Man-~PP/ signal and five ~r the LDLR signal (Figure 3).

In Figure 3A, the Man-~PR internalization mo~if analogs matching the sequence pa~tern YXYX~V (detailed in M~terial and Methods) are shown.
Side chains are shown for critical residues at positions -2, 1, 3, and 4. Man ~PR intemalka~on analogs are: blue, FTFSDY, immunoglobulin 4~20 Fab fragment (PDB cade 4FAB), residues H27-H32; red, FEFEKF, phospho~lyc-era$e kinase (3PGK), residues 340 345; purple, FMFNQF, concanavalin A
(1CN1), residues A128-A133; green, FAFIRL D-xylose isomerase (4XIA), residues A377-A382. The sirnilar, roughly parallel orientations of the ` -WO 93/1818~ - PCr/US93/01669 t .

2~ C3~

aromatic rings paired in positions -2 and 1 (see pairs of same-colored rinys, lower left of Figure 3A) suggest how one of the aromatic rings oould compensate for the loss of the other, as indicated by mLJtagenesis. All four functionally important side chains at positions -2. 1, 3, and 4 are simulta-neously accessible from one side of the tum, requiring at most a single bond rotation.

In the four Man-~PR analogs, the position -2 and 1 aromatic side chains (with position 1 being the first position in the tum) weré in pairs with similarring orientations (see like~olored rin~s, lower leR). Of the four analogs, three were reverse tums and one was an initial turn of alpha helix.
Distar.ces between C~3's of the aromatic side chains in positions -2 and 1 averaged 6.6A and were comparable to those found between C~3's in positions 1 and 4 of the turn (averaging 5.8A), showing a similar spacing ~or the pairs of critical residues in positions -2 and 1 and in positions 1 and 4.
As was also found for the sequence-similar TP~ motif (Collawn, et al., Cell, 63:1061, 1990), the polar side chains in position 3 extended in front of the plane of the turn (projecting out of the page), and th~ position 4 hydropho-bic side chains formed a fan of positions perpendicular to this plane.

It is noteworthy that a truncated mutant Man-~PR with ~ alanine residues amino-terminal to the internalization signal is signifioantly more ac~ve than the wild-type receptor (Canfield, et al., J. Bjol. Chem., 266:5682, 1991), as the alanine residues would be expected t~ stabilize a reverse or helical turn.

In Fi~ure 3B, $he LDLR internalization motif analo~s matching the LDLR
pattern FXNPXY are shown. Side chains at positions -2, 3, and 4 are shown. LDLR analoss identified by Se~uery are: yellow! YANLVF, trypto-) 93/1818~i 2 1 3 ~ 1 0 1) PCI'/US93/01669 phan synthetase (1WSY), resldues A102-A107; red, FYNMW, lectin (2LTN), residues A123-A128; orange, FrNEFY, c~tochrome c peroxidase (2CYP), residues 19~203; blue, FKNSKY, chymotrypsinogen a (2CGA), residues A89A94; and green, YDNI~YW, calcium-free phospholipase A2 (1PP2), residues L11~L118. The aromatic groups at positions -2 and 4 and the position 3 side chain were exposed on the same side of the struc~ure.

~he LDL analoçls included three surface exposed reverse turns and two inibal tums of alpha helix. For each of the analogs, the important aromatic groups in positions -2 and 4 were accessible from the same face of the structure. The Asn side chains in position 1 ~shown in C) fell in ~vo families, positioned in front or in back of the tum according to whether the incoming secondary structure was to the back or in front. The Asn side chain (position 1) ean stabilize the turn by forming hydrogen bonds to adjacent rnain chain atoms, suggesting that this side chain has a structural role. Asn-Pro, the first two residues in the LDLR tetrapeptide, are the most favored residues for positions 1 and 2 in the most prevalent type of reverse turn ~filmot, et al., J. Mol. Biol., 203:221, 19883 and Asn-Pro is also a potent -helix initiator (Richardson, et al., Science, 240:1648, t9383.

Figure 3C shows superimposed Man-~PR and LDLR analogs. The Man-~PR analogs from A are blue and the LDLR analogs from B are yellow with side chains shown for positions -2, 1 and 4. The -2 position aromatic groups extend to the left of the turn, and the position 1 Asn (IDl R) and aromatic (Man-6-PR) side chains form perpendicular fans like those found for the aromatic (LDLR) and hydrophobic (Man-6-PR) side chains in position WO 93/18185 PCI/US93/01669 ~

30~-~ -38-4. Position 2, Ser in Man-~PR and Pro in LDLR, appears primarily structural, with the S~er side chains in Man-~PR in YS~CV analo~s (L.A.K and J.A.T., data not shown) forming backbone hydrogen bonds (Richardson, et al., Pr~diction of Protein Structure ~nd t~7e Princ~ples of Protein Conforrna-bon, G.D. Fasman, ed., (New York:Plenurn Publishing Corp.), pp., 1-98, 1989) mimicking Pro conformations (Witmot, et al., Protein Eng~neenng, ~:479, 1990), and the Pro side chains covalently stabilizing the turn. The structural role of this posi~ion is supported by the tendency of th~se side chains in the six-residue analogs to be somewhat behind the turn, away from the critical side chains. In both the Man-~PR and I DLP~ analogs, the side chains in positions -2, 1, and 4 formed a line (.. ) across the base of the turn, with the line of critical side chains in LDLR analogs at an angle of ^'35 to that for Man-~PR.

An overall structural pattern emerged from superposition of the Man-~PR
and LDLR analogs ~Figure 3C): the critical amino-terminal aromahc groups were localized to the l~ft and somewhat to the front of the tum, the important carboxy-terrTlinal aromatic and hydrophobic side chains formed a sin~le cluster, and all the critical side chains in each structure could b~
recognized simuttaneously from one side of the tum. The analogs predicted a structural role for the position 2 side chain in stabilizing the turn, with direc~ rccognition of this side chain unlikely because it was oriented away from the critical side chains. Whereas a tight turn accounts for the pattern of critical residues by being a self-contained structure that places critical resldues so that they are simultaneously accessible, it is difficult to see how the motifs could be reconciled with positions within a ~B sheet. Extended ,B
conformation would place critical residue side chains on opposite sides of the seconda~y structure, making their simultaneous recognition difficult, and ~:~) 93~1818~ . PCI`/US93/01669 ~ ..

would require more extensive structure for stability; this has been n~led out :
by previous deletion mutants. :~

TRANSPlANTATlON OF MULTIP~ E INTERNALlZATlON SIGNALS

Sh~dies were performed to examine the effect of multiple internalization si~nals on TR endocytosis. 5tructural analysis of th~ TR cytoplasmic domain usin~ the Chou-Fasman algorithm (Chou and Fasmar~, Adv.
Enzymol., 47:45, 1978) predicts that this domain is composed of a series of helices interrupted by tums with the wild-type internalization signal in one of : :~
these turns (tum 2, see Table lll).

wo 93/18185 pcr/us93/o1669 ~ 3 ~ 4~

TAE~
TR CYTOPLASMIC TAIL SEaUENÇE SHOWING PREDICTEI:) TURN POSmONS
Tum 1 Twn 2 Turn 3 Turn 4 MMDt~ARSAFSNUGGEPLSYTRFSt ARQVDGDNSHVEMKLAVDEEENADNNTKANVr~KR~

MUTANTS

Tum 1 YTRF YrRF
Tum 3 ~'TRF Y'rRF

Tum 4 YTRF `rrRF .. , : .
Transmembrane region '`~' .

As shown in TABLE lll, mutant TP~s cor~aining an additional copy of the TR
ir~temalization signal YTRF at putative turn positions 1, 3, or 4 were prepared using oligonucleotide-directed mutagenesis. tach mutant retained the wild-type signal at turn 2, and therefore ~ontained two internalization signals.
Mutant TRs were tested in steady~state distribution assays in chic~en em~ryo fibroblasts as describecl. The resu~ts of th~ analysis are shown in Table IV and indicate that additional signals have variable effects and appear to be positionally dependent.

~' ~ 93/1818~ 2 1 3 0 1 PCI/US93/0166 TABLE IV
COMPARISONS O~ STEADY-STATE D!STRIBUTIONS OF HUMAN
l~:~G TWO !NTERNALIZATION~SIGNALS
Rel~ re % HTR InternalkedIntemallzation Human TR Constructs at steady-stateEfffciency (YO) \~t TR [l8LSYTRF23~ 66.0+1.3'(11)b 100 Turn 1 [9YTRF 3 6~ 3.6 (3) 96 Tum 3 l3tYT~ 79.9+0.8 (3) 205 ~ ~:
Tum 4 [~7~frRF~0l 73.8 ~ 1.4 (3) 145 ;:

a Mean + S.E.
b Number of independent experimen~s.

~he sequence shown in brac~ets represents the addXional internaJiza~ion signat that was transplanted into the TR using oligonucleotide~irected muta~enesis. Each of the mutants contains the wild-type signal at Turn 2 [20YTRF23~. The superscripts represent the sequence position in the TR
where the changes were made. The distribution of mutant TRs was determined from the surface bound and internai pools o~ radiolabeled Tf under steady-state conditions.

Additicn of an internalization si~nal at turn 1 creates a TR that is indis~in-guishable from a wild-type receptor (96% internalization rate compared to wild-type). Addition at turn 3, however, cr~ates a mutant with twice the WO 93/1818~i - PCr/US93/0166? ` `.

?,~33~ 42-.

internalization rate of the wild-type (206% ~ 30%; 3 experiments, mean ~ S.E).
Adpition of a signal at turn 4 has~a less dramatic eff~, bL t still creates a TRmutant with 145% activity. The results demonstrate that Internalization rates can be modula~ed by ~ansplantation of additional internalization signals.

EXAMPLE 4 ~:

IDENTlFIÇATlON OF HETEROLOGOUS INTERNALIZATION S!GNALS `

Potential signals from heterologous receptors were tested by transplantat;ion into the TR. In these studies putative internalization signals from other receptors were transplan~ed into TR using oligonucleotid~directed `:
mutagenesis and tested as deseribed above. The signals examined by this method are shown in Table V.

~' ~ 93/1818~ 21 3 ~1 ~ O PCl/US93/01669 TABLE V
COMPARISONS OF STEADY STAT~ DISTR18UT10NS OF HUMAN
TR MUTANTS CONTAINING PUTAllVE INTERNAUZATIC~N
SEQUENCES FROM QTHER RECEPTORS

RdaUv~
% HTR Internalked Internalization Human TR Constructsat steadv state Efficlencv (%~

Wt T~ E"LSYTRF~66.0i1.3'(11)~ 100 ASGPR H1 l~YQDLZ'l54.6:t29 (4) `62 ~, ASGPR H2 12"FaDI~23.6~æl (5) 16 Poly lg R l~YSAFZ~~3.5 (5) 57 EGFR ["NFYRALZ'~1.6 (5) 119 EGFR 1'LlPaQGFFS2~1542+1.5 (4) 59 Mannose R l"FENll.Y2~ 31.5~2 (4) 24 W [2"YRHN~60.3~t.3 (5) 78 LAP l23PPGYZ']27.8~.9 (3) 20 LAMP-1 laYQTlZ~48.0~1.5 (6~ 48 Tailless lR [~59124.7lt.6 (2) 17 M~an + S.E.
b Number of independent expenmen~s W0 93/18185 ~ PCI'/USg3/01669 `
-ln TABLE V, the sequences shown in brackets represent the putative signals that were transplanted Into the TR using oligonucleotide-directed mutagen~
sis. rhe superscripts represent the sequence position in the Tl~ where the changes were made. The distribution of mutant TRs was determined from the surface bound and internal pools of radiolabeled Tf under steady-statR ~`
conditions. ,.. -.

Wlthin this system, some sequences function very well as internalization signals, e.g., sequences from the asialoglycoprotein receptor (ASGPR H1 subunit, sequence YQDL), polymeric immunoglobulin (poly l~, sequence YSAF), epidermal growth factor receptor (EGFR, sequence NFYRAL), Iysosomal acid phosphatase (lAP, sequence YRH\I), Iysosome-associated membrane glycoprotein (LAMP-1, sequence YQTI), while others, e.g., ASGPR H2 subunH (sequence FQDI), mannose receptor (sequence FENTLY), and LAP (sequence PPGY) ~unction poorly, with activity similar to that of tailless receptors (TABLE V). Once a signal has been localked to a region within the c~noplasmic tail, this method is usef~JI for mapping the exact site. For example, analysis of the region near the tyrosine of LAP, sequence PPGYRHV, shows that one sequence, YRHV, h~nctions well and another, PPGY, does not.

This strategy is also useful for mapping internalization signals from ligand-induced internalized receptors such as the E~FR. This is significant since none of these internalization signals have been ctearly identified simply because the large cytoplasmic domains of these receptors makes Ule interpretation of deletional analysis more difficult. Identifying an internaliza-tion signal in a ligand-induced receptor may require additional transplanta-tion experiments with the T~, and parallel mutagenesis experiments within ') 93/18185 2 ~ 3 0 1 0 ~ PCI`/US93/01669 45- :

the receptor in question may have to be performed. The putative EGFR
signal, NFYRAL which functions very well in the context of the TR, appears to have a sequence related to the cation-independent mannose-~phosphate receptor. The implication from these studies is that potential sign~s can be tested in the TR and.that ligand-induced receptors may have irltemalization ~ ~-signals similar to those of the constitutively recycling receptors. ` -The foregoing is meant to illustrate, but not to limit, the scope of the invention. Indeed, those of ordinary skill in the art can readily envision and produce further embodiments, based on the teachings herein, without undue experimentation.

Claims (24)

1. A method of modulating receptor mediated transport c ?nd into a cell, said method comprising introducing heterologous ?nalization signal into the cell.

2. A method according to claim 1 wherein the internaliz? signal is introduced as a nucleotide sequence.

3. A method according to claim 2 wherein the nucleoti? sequence further comprises flanking sequence encoding a cell surf? receptor.

4. A method according to claim 3 wherein the receptor encoded by introduced nucleotide sequence is different from cell surface receptor already present in the cell.

5. A method according to claim 1 wherein at least two internalization signals are introduced.

6. A method according to claim 3 wherein the receptor encoded by introduced nucleotide has substantially the same ligand specificity as cell surface receptor already present in the cell.

7. A method according to claim 6 wherein the internalization signal of receptor encoded by introduced nucleotide differs from the internaliza-tion signal of cell surface receptor already present in the cell.

8. A method according to claim 1 wherein modulation results in increased transport of ligand into the cell.

9. A method according to claim 2 wherein the nucleotide sequence encoding the internalization signal is incorporated in an expression vector.

10. A method according to claim 9 wherein the expression vector is incorporated in a carrier system.

11. A method according to claim 1 wherein the internalization signal is introduced as a peptide.

12. A method according to claim 11 wherein the peptide is incorporated in a carrier system.

13. A method according to claim 1 wherein the internalization signal is characterized by having a tight turn and the amino acid sequence:

where X5 - X8 constitutes the core sequence, X1 - X4 residues and X9 - X11 residues are optional and wherein:
X1, when present, is leucine or glutamic acid;
X2, when present, is isoleucine, methionine or proline;

X3, when present, is any amino acid residue;

X4, when present, is selected from alanine, polar amino acids, or aromatic amino acids, and when X3 is also present, at least one of the X3 or X4 residues is polar;

X5 is an aromatic amino acid when residues X1 - X4 are not present, or is selected from aromatic amino acids or polar amino acids when at least residue X4 or additional upstream residue(s) is present;

X6 is a polar amino acid or alanine;

X7 is selected from polar amino acids or alanine when residues X1 - X4 are not present, or is any amino acid residue when at least X4 or additional upstream residue(s) is present;

X8 is selected from aromatic amino acids or hydrophobic amino acids;

X9, when present, is serine or alanine;

X10, when present, is alanine or leucine; and X11, when present, is alanine or phenylalanine;

wherein at least one of residues X3, X5, and X8 is an aromatic amino acid and further wherein residues X1, X2, X3, X10, and X11 can only be present when the next adjacent residue(s) relative to the core is present.

14. A method according to claim 13 wherein the internalization signal is characterized by having the amino acid sequence:

wherein X6 and X7 cannot both be alanine.

15. A method according to claim 14 wherein:

X5 is phenylalanine or tyrosine;

X6 is alanine, arginine, glutamine, serine, or threonine;

X7 is alanine, arginine, aspartic acid, glycine, glutamic acid, histidine, lysine, or threonine; and X8 is isoleucine, leucine, methionine, phenylalanine, valine, or trypto-phan.

16. A method according to claim 15 wherein the amino acid sequence is YTRM, YARF, YTRI, YQDL YTKF, YSKV, YTRW, YRHV, YSAF, YQTI, YTAF, YTGF, YTEF, FTRF, or YTRF.

17. A method according to claim 13 wherein the internalization signal is characterized by having the amino acid sequence:

18. A method according to claim 17 wherein:

X3 is asparagine, leucine, methionine, phenylalanine, proline, or tyrosine;

X4 is alanine, aspartic acid, glutamine, lysine, phenylalanine, or serine;

X5 is asparagine, glutamine, or tyrosine;

X6 is arginine, glycine, proline, serine, or threonine;

X7 is alanine, arginine, lysine, phenylalanine, or valine; and X8 is isoleucine, leucine, methionine, phenylalanine, tryptophan, tyro-sine or valine.

19. A method according to claim 18 wherein the amino acid sequence is:

YKYSKV, NFYRAL, LAYTRF, PQQGFF, FDNPVY, MSYTRF, or LSYTRF.

20. A method for identifying a sequence which modulates internalization of a cell surface receptor, the method comprising:

(a) incubating cells having such receptors in the presence or absence of the sequence and, optionally, in the presence of ligand for the cell surface receptor; and (b) measuring internalization of cell surface receptor in the presence or absence of the sequence.

21. A composition comprising a peptide having a tight turn and the amino and sequence:

where X5 - X8 constitutes the core sequence, X1 - X4 residues and X9 - X11 residues are optional and wherein:
X1, when present, is leucine or glutamic acid;
X2, when present, is isoleucine, methionine or proline;
X3, when present, is any amino acid residue;
X4, when present, is selected from alanine, polar amino acids, or aromatic amino acids, and when X3 is also present, at least one of the X3 or X4 residues is polar;
X5 is an aromatic amino acid when residues X1 - X4 are not present, or is selected from aromatic amino acids or polar amino acids when at least residue X4 or additional upstream residue(s) is present;
X6 is a polar amino acid or alanine;

X7 is selected from polar amino acids or alanine when residues X1 - X4 are not present, or is any amino acid residue when at least X4 or additional upstream residue(s) is present;

X8 is selected from aromatic amino acid s or hydrophobic amino acids;

X9, when present, is serine or alanine;

X10, when present, is alanine or leucine; and X11, when present, is alanine or phenylalanine;

wherein at least one of residues X3, X5, and X8 is an aromatic amino acid and further wherein residues X1, X2, X3, X10, and X11 can only be present when the next adjacent residue(s) relative to the core is present. provided that sequences selected from the group consisting of FXNPXY, GPLY, PPGY, and YXYXKV, where X stands for any amino acid, are excluded.

22. A composition comprising a nucleotide sequence encoding a peptide according to claim 21.

23. A method of gene therapy comprising introducing into a host subject, cells derived from the subject and modified to contain heterologous internalization signal capable of modulating transport of ligand into the cells.

24. A method of gene therapy comprising introducing into the cells of a host subject an expression vector comprising a nucleotide sequence encoding a heterologous internalization signal capable of modulating transport of ligand into the cells.