AU1042000A - Method for the exponential amplification of molecular matrices - Google Patents

Abstract

The invention relates to a method for the exponential amplification of molecular matrices, comprising the following steps: (a) binding the molecular matrices to the surface of a solid phase using a reversible linker on the matrix; (b) adding matrix fragments, one of the fragments having an optionally protected linker unit; (c) synthesising copies of the matrix; (d) removing excess matrix fragments and reaction auxiliaries; (e) detaching the copies from the matrices; and (f) populating free binding sites on the solid phase with synthesised matrix copies. The invention also relates to a method for chemical evolution through exponential amplification of molecular matrices, comprising one or more of the following evolution cycles: (1) selection of a subpopulation of molecular matrices from a combinatorial original library; and (2) amplification and mutation of the selective matrices for providing a mutant library, comprising one or more of the amplification cycles (a) to (f) defined above.

Description

SMB Process for the Exponential Amplification of Molecular Templates The present invention relates to a process for the exponential amplification of molecular templates, i.e., sequence-isomeric compounds with template properties. In this process, product inhibition is prevented by conducting the reaction at the surface of a solid phase. The present invention further relates to a method of chemical evolution using the mentioned process for the exponential amplification of molecular templates. Chemically self-replicating systems have been designed in order to identify the minimum requirements of molecular replication [D. Sievers et al., Self-Reprod. of Supramolecular Structures, Kluwer Publishers, Dordrecht (1994)], in order to transfer the principle to synthetic supramolecular systems [E.A. Wintner et al., Acc. Chem. Res. 27, 198-203 (1994)], and in order to achieve a better under standing of the scope and limitations of self-organizational processes [M. Eigen, Naturwissenschaften 58, 465-523 (1971)], which are believed to be relevant to the origin of life on earth [L. Orgel, Acc. Chem. Res. 28, 109-118 (1995)]. Current embodiments use oligonucleotide analogues [G. von Kiedrowski, Angew. Chem. Int. Ed. Engl. 25, 932-935 (1986); W.S. Zielinski et al., Nature 327, 346-347 (1987); G. von Kiedrowski et al., Angew. Chem. Int. Ed. Engl. 30, 423-426, ibid. 892 (1991); T. Achilles et al., Angew. Chem. Int. Ed. Engl. 32, 1198-1201 (1993); D. Sievers et al., Nature 369, 221-224 (1994); T. Li et al., Nature 369, 218-221 (1994); B. Martin et al., Helv. Chim Acta 80, 1901-1951 (1997); D. Sievers et al., Chem. Eur. J. 4, 629-641 (1998)], peptides [D. Lee et al., Nature 382, 525-528 (1996); K. Severin et al., Angew. Chem. Int. Ed. 37, 126-128 (1998); D.H. Lee et al., Nature 390, 591-594 (1997); S. Yao et al., Angew. Chem. Int. Ed. 37, 478 481 (1998); K.S. Severin et al., Chem. Eur. J. 3, 1017-1024 (1997)] and other molecules [T. Tjivikua et al., J. Am. Chem. Soc. 112, 1249-1250 (1990); A. Terfort -2 et al., Angew. Chem. Int. Ed. Engl. 31, 654-656 (1992); J.-I. Hong et al., Science 225, 848-850 (1992); Q. Feng et al., Science 256, 1179-1180 (1992); R.J. Pieters et al., Angew. Chem. Int. Ed. Engl. 106, 1579-1581 (1994); D.N. Reinhoudt et al., J. Am. Chem. Soc. 118, 6880-6889 (1996); B. Wang et al., Chem. Commun. 16, 1495-1496 (1997)] as templates and rely on either autocatalytic, cross-catalytic or collectively catalytic ways of template formation. A frequent problem of these systems is product inhibition, which leads to parabolic rather than exponential amplification [G. von Kiedrowski, Bioorg. Chem. Front. 3, 113-146 (1993)]. Exponential amplification is the dynamical precondition of selection in the Darwin ian sense [E. Szathmary et al., J. Theor. Biol. 138, 55-58 (1989); R.W. Wills et al., Santa Fe Institute Working Paper 97-07-065 (1997)]. Several theories of the origin of life have pointed out the importance of surfaces to chemical evolution [J.D. Bernal, The Physical Base of Life, Routledge & Kegan Paul, London, 1951; A.G. Cairns-Smitz, The Life Puzzle, Oliever & Boyd, Edinburgh, 1971; H. Kuhn et al., Angew. Chem., Int. Ed. Engl. 20, 500-520 (1981); G. W~chtershauser, Microbiol. Rev. 52, 452-484 (1988); E. Szathm ry et al., J. Theor. Biol. 187, 555-571 (1997); L. Orgel, Origins Life Evol. Biosphere 28, 227 234 (1998)]. Thus, methods involving a stepwise supply were employed before in two different chemical systems described as models of potential prebiotic proc esses [T. Li et al., Nature 369, 218-221 (1994); J.P. Ferris et al., Nature 381, 59 61 (1996); G. von Kiedrowski, Nature 381, 20-21 (1996)]. Li et al. achieves a chemical replication of duplex DNA consisting of palindromic (symmetrical) homopyrimidine and homopurine strands. The homopyrimidine strand was synthesized from its precursor fragments by triple helix linking and then served as a template for the chemical linking of the precursors of the homopurine strand. The stepwise supply of homopyrimidine and homopurine fragments allowed to prevent complexing of the fragments and thus to switch between the respective triplex and duplex linking intermediates. Ferris et al. proved the synthesis of long oligonucleotide and peptide-like materials on the surface of mineral supports. In these systems, the stepwise supply enabled the extension of activated precursors and thus allowed to overcome the length-restricting effect of precursor hydrolysis.

-3 A selection and amplification method having an evolutionary character is further known from WO 91/19813. In this method, a subpopulation of a starting mutant library is produced by reversible binding (partitioning) to a target structure, followed by amplification of the selected mutants. However, the problem of product inhibition, which is also known from PCR, for example, is often encountered in such an amplification. It has been the object of the present invention to provide an amplification process for sequence-isomeric compounds having template properties (briefly referred to as "molecular templates" hereinbelow) which process enables an unlimited amplification of the molecular template, i.e., without product inhibition. It has been a further object of the present invention to provide a method for the evolution of molecular templates in which the amplification step is not a limiting factor, i.e., an unlimited amplification of the molecular template without product inhibition is possible. Surprisingly, an iterative stepwise process has been found which allows to expo nentially increase the available quantity of a molecular template. The process employs the surface of a solid support. Copies are synthesized from precursor fragments by chemical linking on immobilized templates, which copies are then released and immobilized to become new templates. This process can be repeated as many times as desired. Thus, the present invention relates to: (I) a process for the exponential amplification of molecular templates, comprising: (a) binding of molecular templates to the surface of a solid phase using a reversible linker on the template; (b) addition of template fragments wherein one of the fragments has an optionally protected linker unit; -4 (c) synthesis of copies of the template; (d) removing of excess template fragments and reaction auxiliary agents; (e) detaching the copies from the templates; and (f) colonization of free binding sites on the solid phase by synthesized template copies; and (II) a method of chemical evolution by the exponential amplification of molecular templates, comprising one or more of the following evolution cycles: (1) selection of a subpopulation of molecular templates from a combina torial starting library; and (2) amplification and mutation of the selected templates to provide a mutant library comprising one or more of the amplification cycles (a) to (f) as defined in (I). The role of the solid phase (also referred to as "support" hereinbelow) in the process according to the invention is to keep complementary templates, which would form stable duplexes in solution, separated from each other. Suitable support materials consist of an organic or inorganic material or of a hybrid of these materials. Organic support materials include sugar-based polymers, preferably agarose, cellulose and suitable derivatives thereof, or technical polymers, such as polystyrene, polyacrylate, polyacrylonitrile, polyalkenes or graft copolymer (e.g., PS-PEG, PAN-PEG, PAN-PAG, etc.). Inorganic support materials may be, for example, glass, functionalized hydroxyapatite and metals, special importance being attributed, in particular, to the gold surface (due to gold-thiolate interaction). As hybrids, there may be contemplated paramagnetic beads, for example. In the process according to the invention, the template copies produced in step (e) can be identical with or complementary to the starting templates. "Complemen tary" within the meaning of the invention means that a copy of the template is - 5 different from the starting template, but that a copy of this copy is again identical with the starting template. In the following, this is also briefly referred to as "(+) strand" and "(-)-strand". In the case of identity between the starting templates and the template copies, the next reaction cycle following the colonization of free binding sites (step (f)) can be performed in the same reaction vessel. On the other hand, when the produced template copies are complementary to the starting template, step (f) is preferably performed in a second reaction vessel, and only step (f) of the subsequent reaction cycle will be performed again together with the original template fragments. In a preferred embodiment of process (I), excess binding sites on the solid phase are blocked after the binding of molecular templates to the surface of the solid phase. This blocking depends on the type of linkage between the template fragments employed and the solid phase, i.e., on the nature of the linker. In the process according to the invention, the binding of the linker can be effected by both covalent and non-covalent bonding, the term "non-covalent bonding" encompassing both ionic and non-ionic binding systems and especially members of immunological binding pairs, such as avidin/streptavidin (biotin) or antigen/anti body. However, a basic requirement for the selection of appropriate linkers is its being orthogonal with respect to the other chemical bonds used in the system, especially to the chemical bonds within the templates themselves and between the template and template copy, so that the templates can be removed from the solid phase surface without the cleavage of bonds, i.e., a switchable binding to the solid phase is ensured. Regarding the blocking of the excess binding sites on the solid phase, this means that the blocking can be effected either by reaction with chemical reagents undergoing covalent bonding with free binding sites (e.g., for free thiol groups on the solid phase: reaction with suitable thiol reagents to form disulfides) or by reaction with suitable non-covalently binding compounds saturat ing the free binding sites (e.g., for free antigens in the solid phase: reaction with the corresponding antibodies). The linker can be located on either end of the template, i.e., for an oligonucleotide, it can be located at either the 5' or 3' end, and for a peptide, at either the N- or the C-terminus.

-6 Further, the colonization of free binding sites on the solid phase by synthesized template copies requires that free binding sites are available for the next reaction cycle. This can be effected, on the one hand, by exposing new binding sites on the solid phase which is already present in the synthesis, or on the other hand, by the addition of new solid phase material, or by transferring the template copy formed to a new solid material. Of these three possibilities, the former two are particularly suitable for a situation where the starting template and the template copy are identical. The third possibility is more appropriate for systems in which the produced template is complementary to the starting template. In another embodiment of process (I), protected linker units of the template copies can be deprotected at this point of the reaction cycle (i.e., prior to, during or after step (e)). Then, after the deprotection, these linker units will react with free binding sites on the support material. In this embodiment, free binding sites may also be present on the solid phase during steps (b) to (e). The blocking of excess binding sites after step (a) is not necessary in this embodiment. With respect to the nature of these linker protecting groups, the same require ments apply as above for the linker chemistry. In this case too, orthogonality to the remaining chemical compounds must be ensured. Protecting groups suitable for the respective amplification system can be seen, for example, from T.W. Green, Protective Groups in Organic Synthesis, Wiley & Sons, N.J. Finally, in a special embodiment, two or more orthogonal immobilization methods are employed rather than one method. This implies that two or more surfaces are used rather than one surface. Preferably, the (+)-strand is immobilized by a different method from that used for the (-)-strand, so that two orthogonal immobilization methods on two appropriately functionalized surfaces have to be performed. Orthogonal immobilization methods can be based on different specific non-covalent interactions, e.g., on the biotin/avidin pair or the digoxygenin/anti digoxygenin pair, but also on orthogonal covalent linking methods, such as Diels Alder reactions and condensation reactions. When A-A' and B-B' are used to refer to two orthogonal covalent or non-covalent interactions, then the replication process comprises the following steps: - 7 (la) the A'-functionalized (+)-strand is immobilized on surface A; (1b) the B'-functionalized (-)-strand is immobilized on surface B; (2a) fragments are added to the (+)-strand, one of them being functionalized with B'; (2b) fragments are added to the (-)-strand, one of them being functionalized with A'; (3a) the fragments on the (+)-strand are subjected to ligation; (3b) the fragments on the (-)-strand are subjected to ligation; (4a) the B'-functionalized copy of the (+)-strand is detached; (4b) the A'-functionalized copy of the (-)-strand is detached. In the preferred technical embodiment, the A surface always remains separated from the B surface. The substeps (a) and (b) in the reactions (1a,b), (3a,b) and (4a,b) can then be respectively performed together for both surfaces, i.e., in the same liquid phase. Only the substeps (2a) and (2b) must be conducted separately because it must be prevented that the A' fragment contacts the A surface and that the B' fragment contacts the B surface. In the processes according to the invention, templates having a length of from 2 to 2000, especially from 4 to 50, monomer units can be used. As "template frag ments" in step (b) of the process according to the invention, there can be em ployed individual monomer units as well as oligomers of such monomer units. However, this requires that one of these monomer units or oligomer units com prises a protected linker unit, wherein the nature of the linker protective group is defined as above. The concentration of the templates employed in the process according to the invention depends on the solid phase material used, especially on the size of the particles and the number of binding sites on its surface. Another limiting factor is the length of the templates. The concentration of the template molecules in the reaction mixture should generally be from 10-15 to 101, especially from 10-12 to 10-3 mol/l. The template fragments used should be employed in a concentration of from 10-12 to 1 mol/l, especially from 10~ 9 to 10-' mol/l, some excess of template fragments being employed as compared to the amount of monomer units neces- -8 sary for the corresponding template. An excess of from 1 : 1.1 to 1 : 100 of the monomer units is particularly preferred, the reaction being economically inefficient at an excess of above 10. "Molecular templates" within the meaning of the present invention means all sequence-isomeric compounds having template properties. These include, in particular, oligonucleotides and oligonucleotide derivatives (including PNA, pRNA, 2'-5' nucleotides and RNA/DNA mirromers ("Spiegelmere")), oligopeptides and oligopeptide derivatives (especially in coiled-coil (e.g., leucine zipper) and pleated sheet arrangements). Among the oligonucleotide and oligopeptide derivatives, there may be mentioned, in particular, non-natural structures, such as oligonucleo tides having 2'-5' structures and peptides comprising D-amino acids, which have an increased stability towards cleaving enzymes. As the backbone of the template, there may be preferably used phosphoric acid diesters, amidates, thioates, pyrophosphates and other phosphoric acid derivatives, as well as amides, but also esters, ureas, urethanes, disulfides, imines, acetals and C-C-linked structures. The molecular recognition taking place in the templates can be effected through heterocycles (especially nucleobases), salt bridges (especially amidinium carboxy late interaction), ion-ion interaction, ion-dipole interaction, dipole-dipole interac tion, inclusion complexing (especially via cyclodextrines and so-called molecular tweezers), stacking and charge-transfer interaction (especially via the stacking of electron-excess and electron-deficient aromatics) and ligand-metal-ligand interac tion. The templates can have linear, cyclic, branched (dendrimer-like) and two dimensional structures. Suitable solvents for the process according to the invention include water and organic reaction media or mixtures thereof. When the templates and template fragments contain ionic groups (such as phosphate mono- or diesters), aqueous solvent systems are particularly preferred. The solvents may further contain suitable buffer systems. The bond formation in step (c) can be effected either chemically or enzymatically. Chemical linking involves the reaction of the reactants with activation reagents and condensation agents suitable for the respective bond formation. Thus, when the -9 reaction is conducted in aqueous media, the condensation reaction of phosphoric acid derivatives is preferably effected with the use of water-soluble carbodiimides. For enzymatic linking, all enzymes may be used which have the corresponding catalytic properties (ligases, polymerases, esterases, etc.). The reaction temperature for the above mentioned reaction steps, especially the bond formation step (c), can be between -20 and 100 0 C, depending on the reaction system. Particularly preferred are temperatures of between 0 and 50 0 C. Especially for the linking reaction using polymerases, the amplification system according to the invention enables the reaction to be conducted at below 50 oC. After the synthesis of the template copies, in step (d), the excess of the template fragments and reaction auxiliary agents (i.e., condensation agents, activation reagents, buffers, enzymes etc.) needed for the linking are removed. The detaching of the template copies from the templates is dependent on the respective template system and must be performed in such a way that both the formed chemical bonds and the linkage between the linker and the surfaces or appropriate linker protective groups on the template copies are not adversely affected. Suitable measures include, for example, treatment of aqueous or organic aqueous solutions with denaturing reagents, increase of temperature, electric and magnetic fields, mechanical methods and combinations thereof. Thus, the present invention offers a wide variety of possible denaturing methods, i.e., heating to above 90 OC as with Taq polymerase in PCR is not absolutely required. In the method (II), the selection of the subpopulation of molecular templates is effected by reversible binding to one or more target structures (molecules). This is preferably done by adsorption chromatography. In the evolution cycle according to the invention, if the starting mutant library does not comprise a suitable linker, the selected molecular templates are provided with a reversible linker in an additional step (1'). The mutant library obtained in step (2) of the method is the starting library of the next evolution cycle. The mutation in the amplification cycle is controlled by the reaction parameters in steps (b) and (c). The number of evolu tion cycles is determined by the complexity (i.e., the number of possible variables - 10 in the structure to be selected), the decrease ratio in the selection step, which is between 0.01 and 50% in the present invention, and the countereffect of the mutation in the amplification step. In addition to the selection step in (1), an additional selection may be effected by the detaching step (2)(e). The method (II) may further comprise the additional steps of isolation of the individual template molecules obtained, followed by analysis, either after each evolution cycle or after the last evolution cycle. The present invention is illustrated in more detail by the following Figures. Description of the Figures Figure 1: General scheme of the process according to the invention: (1) Immobilization step: A template is immobilized to the surface of a solid support by an irreversible reaction. (2) Hybridization step: The template binds complementary fragments from the solution. (3) Ligation step: The fragments are linked together by chemical bonding. (4) Detaching step: The copy is released and again immobilized to a different portion of the solid support to become a template for the next cycle of steps. Figure 2: Oligonucleotide analogues and reactions employed in the experiment: The individual steps of the process, (1) to (4), were performed separately in different tubes for the complementary templates X and Y. PySS represents a 2-pyridyl disulfide structural unit cleaved in the immobilization (la) and capping (1b) to form 2-thiopyridone. The skeletal modification X on the central internucleo tide linkage of X and Y is X = N-H, except for the first immobilization where X = 0. Ax, Bx, AY and By represent the corresponding template fragments. The hybridiza tion step (2) results in a termolecular complex of the immobilized templates and the respective fragments Ax, BX, AY and By. In the presence of the water-soluble carbodiimide EDC, the chemical bonding (3) results in a 3'-5' phosphoramidate - 11 linkage between a 5'-amino and a neighboring 3'-phosphate group. Each resulting double-stranded complex is denatured (4), resulting in the template-bearing support and a PySS-modified copy immobilized on a fresh SH support. Figure 3: HPLC analysis of products under denaturing conditions obtained in the successive steps of an amplification cycle. All HPLC samples with no reduction buffer (RB) were brought to a concentration of 100 mM DTT prior to analysis (except for (a)) in order to ensure the reproducibility of the HPLC quantification. tR = retention time in minutes; (a) = reaction mixture containing X and GAATCCATGGTAAG (Xref) as an internal standard, before immobilization; (b) = supernatant after immobilization; (c) = reaction mixture after hybridization of immobilized X with AY and By and treatment of the support with RB. (d) Reaction mixture after chemical bonding, and reductive cleavage from the support using RB; (e, f) = HPLC analysis of the set of sixteen samples obtained by the reductive cleavage of immobilized templates produced by three amplification cycles. The indications on the right-hand side of the HPLC profiles are explained in more detail in the following under Figure 4. Note that small contaminant peaks visible in the front profiles gradually disappear in the course of the amplification. Figure 4: Pathway of template transfers in the course of three amplification cycles: Boxes and letters symbolize reaction tubes or template-bearing supports. X-axis: number of amplification cycles; Y-axis: number of samples. The arrows indicate which copy is produced from which template. The shown amount of the template material was quantified by HPLC analysis after cleavage of the disulfide bridges by reduction with DTT. From the above data, 14 individual yields were calculated. The average yields and standard errors are: px = 0.763 +/- 0.071 and py = 0.855 +/ 0.079. The molar quantity of the respective template in cycle n in nanomoles is compared with its theoretical value (in brackets) calculated from (3a,b): xO = 38.87 (38.87), yo = 45.44 (45.44), x 1 = 72.81 (73.54), yi = 80.71 (78.68), x 2

=

- 12 126.29 (133.58), Y2 = 138.97 (141.57), x 3 = 238.90 (241.61), y3 = 249.77 (255.80). Figure 5: General scheme of the method according to the invention: (1) Immobilization step: A template is immobilized to the surface of a solid support by an irreversible reaction. (2) Hybridization step: The template binds complementary fragments from the solution. (3) Ligation step: The fragments are linked together by chemical bonding. (4) Detaching step: The copy is released and again immobilized to a different portion of the solid support to become a template for the next cycle of steps. (5) Selection step: Binding of the template to the target molecule. (6) Discarding of the non-binding templates. In detail, Figure 5 describes two particular embodiments of the process (I) with the following steps: (1) a template immobilization step; (2) a hybridization or capping step; (3) a ligation or copying step; and (4) a detaching or denaturing step. The method comprises two process variants (I and II) of which (I) involves a non enzymatic copying step and (II) involves an enzymatic copying step. (I): In the first process variant, oligodeoxynucleotide or oligoribonucleotide derivatives are used as a template; they contain from 10 to 100 nucleotide units and either a natural backbone of 3'-5' phosphoramidate linkages, 3-5' pyrophos phate linkages or other linkages including 2'-5' internucleotide linkages. Natural and synthetic linkages can be mixed in the backbone. (Step 1) The immobilization of the template is effected at the 5'-terminus through a disulfide bridge or through the biotin/avidin interaction. In the first case, the oligonucleotide template to be immobilized contains a 2-pyridyl disulfide modifica tion at the 5'-terminus, and the support material contains free thiol groups. Immobilization takes place through disulfide exchange with the cleaving off of 2 thiopyridone. In the second case, a 5'-biotinylated oligonucleotide template is employed wherein the biotin modifier may also contain a disulfide bridge. Immobi- - 13 lization takes place through the binding of biotin to avidin, which is itself bound to solid supports. As support materials, there may be used tentagels, agarose or other suitable polymers including glass or paper. (Step 2) After the immobilization, the unoccupied surface sites are deactivated by treatment with a suitable reagent. Thiol groups are preferably deactivated by treatment with 2-hydroxyethyl 2-pyridyl disulfide, but also with 2,2'-dipyridyl disulfide, a mild alkylant or by oxidation. Free avidin is deactivated with biotin. (Step 3) The copying step is performed by chemical ligation: dimeric to 50-meric complementary oligodeoxynucleotides, oligoribonucleotides or derivatives thereof containing either 5'-phosphate and 3'-hydroxy groups, or 5'-hydroxy and 3' phosphate groups, or 5'-phosphate and 3'-amino groups, or 5'-amino and 3' phosphate groups, or 5'- and 3'-phosphate groups, or other chemically linkable groups are bound to the immobilized template and linked together in the presence of a water-soluble carbodiimide, preferably EDC, or another condensation agent, such as bromocyanogen or N-cyanoimidazole. The oligodeoxynucleotide comple mentary to the 3'-end of the immobilized template serves the function of a 5' primer. At the 5'-end, it contains one of the modifications mentioned under (1), and at the 3'-end, it contains one of the groups mentioned under (3). (Step 4) The denaturing with detachment of the copy is effected either by setting a suitable high temperature or by elution with a urea solution or another denaturing agent. The detached copies are immobilized on a fresh support, as described under (1), and serve as a template in the next step. The newly coated support material is combined with that used in former cycles. Immobilized (+)- and (-)-strands can be combined separately or together. When they are combined separately, the steps (3) are performed alternately with (+)- and (-)-primers; when they are combined together, each step (3) is performed with both primers. (II): In the second process variant, steps (1), (3) and (4) are the same as those mentioned under (I). Step (2) involves the enzymatic extension of a primer or the - 14 enzymatic ligation of two or more oligonucleotide fragments. The target sequence to be amplified is first transcribed, using a 5'-modified primer, into a sequence which can be immobilized at the 5'-terminus. Immobilization of the template is followed by hybridization with the 5'-modified primer and extension thereof by dNPT in the presence of a polymerase. Alternatively, the ligase-catalyzed linking of 5'-phosphorylated oligonucleotides can take place. Figure 4 summarizes the pathway of each individual support and the quantity of material obtained after each copying cycle for the example described. In general, the yield p of a replication cycle need not necessarily reach p = 1 in order to enable an exponential amplification mode. In the case of palindromic templates in which X = Y, Ax = AY, Bx = By, the quantity of material xn obtained from a starting quantity xo after n replication cycles is given by: xn = xo-(1 + p)" (1), which allows exponential growth for p > 0. For non-palindromic templates, which underlie our experiments, each replication cycle consists of two copying cycles with the individual yields px and py for the synthesis of X and Y, respectively. Here, the quantities of the immobilized templates X and Y follow the iteration progression: xn+1 = xn + px-yn and yno1 = yn + py.xn (2a,b), which requires both px > 0 and py > 0 for an exponential growth. Equations (2a,b) yield: xn= x.,+y-. y- -(+jp-p1, + 2 x.- y". .12p,. py (3a,b), 1n ( .+X P . 1 (Y __ (. 0 - --. ( ', 22 - 15 wherein x 0 and yo represent the initial quantities of X and Y, respectively. A comparison of the experimental and theoretical values for xn and yn as given in the legend of Figure 4 confirms that the data shown in Figure 4 are consistent with an exponential mode of amplification of the template materials. Process (I) combines the advantages of solid phase chemistry with chemical replication and can be developed further for the non-enzymatic and enzymatic amplification of RNA, peptides and other templates, and for Darwinian evolution in artificial chemical ecosystems. Similar processes may have played a role in the origin of life on earth, since the primordial replication systems could have propa gated by spreading on mineral surfaces. Further, the process according to the invention is a first example of exponential amplification and solid phase chemistry in chemical replication. Solid phase chemistry is a precondition of a practicable automation of the process because, in essential, only pipetting and filtration steps are involved. The exponential amplifi cation offers the possibility of a well-aimed molecular evolution. In the latter context, the process has the potential of a controllable mutation rate, since the releasing of the copies in step (4) should depend on the thermodynamic stability of the immobilized duplexes, and thus, mutants should be eluated more quickly as compared to perfect copies. Protocols relying on temperature cycles (or other environmental parameters) and solution chemistry can be less readily applied to chemical systems based on short templates. Quick negative temperature jumps and a fast condensation chemistry would be necessary to prevent re-equilibration of the participant oligonucleotide complexes. Such conditions would then necessar ily limit the accuracy of template copying since the thermodynamic control of molecular recognition determines the threshold for the copying accuracy in the absence of an energy-dissipating proof-reading mechanism. From the above, it follows immediately that the process according to the invention is of substantial value for the designing of evolutionary chemical systems. Other templates, surfaces and template-surface combinations can be employed, enzy matic variants can be explored, and different steps can be combined to reach a higher level of process integration and autonomy. Protocols are possible in which - 16 related templates are immobilized in close proximity so that the conception of quasispecies [M. Eigen, Naturwissenschaften 58, 465-523 (1971)] obtains a spatial dimension. Further, protocols in which two or more different classes of template molecules, such as oligonucleotides and peptides, are involved can contribute to the selection and discovery of more complex cooperative systems. Over the short term, the process can already be applied to the chemistry of peptide and pRNA replication which has been described by the groups of D.H. Lee et al, Nature 382, 525-528 (1996), and B. Martin et al, Helv. Chim. Acta 80, 1901-1951 (1997), and to the synthesis of so-called mirromer RNA (Jens FOrste et al., Nature Biotechnol ogy 14, 1112, 1116; 15, 229). The invention is further illustrated by the following Example. Example Materials: The synthesis of the two complementary 14-meric templates X and Y and of the four template fragments Ax, Bx, AY and By was effected using standard phosphoramidite chemistry, for example, L. Beaucage and R.P. Iyer, Tetrahedron 48, 2223-2311, or L. Beaucage et al., Tetrahedron Lett. 22, 1859-1862 (1981). The thiol-modified support was obtained from commercially obtainable 2-pyridyl disulfide activated Agarose* (Sepharose® 6B, Pharmacia) by reduction with dithiothreitol (DTT). Reduction buffer (RB): 100 mM DTT, 40 mM Tris-HCI, 0.5 M NaCI, 1 mM EDTA, pH 7.9; immobilization buffer (IB): 370 mM sodium acetate/acetic acid, 0.5 M NaCI, 1 mM EDTA, pH 4.4, degassed with argon to expulse oxygen. Capping buffer (CB): 60 mM S-(2-thiopyridyl)-2-mercaptoethanol, 40 mM MES, 0.5 M NaCl, 1 mM EDTA, pH 6.5. Hybridization buffer (HB): 100 mM MES, 40 mM MgCI 2 , 1 M NaCI, 1 mM 2,2'-di pyridyl disulfide, pH 6.1. Linking buffer (LB): HB with 0.2 M EDC, freshly prepared before the chemical linking. General procedures: If not stated otherwise, all supported reactions were per formed at 25 1C in micro-spin centrifuge filters (0.5 ml, cellulose acetate with - 17 0.45 pm pores, Roth) stacked in Eppendorf tubes (1.5 ml). Each filter of a series of 16 filters was charged with 50 mg of moist Sepharose* 6B (washed according to the manufacturer's directions) and then centrifuged (3 min at 4000 x g) to remove the supernatant. Prior to an immobilization step, the beads were vortexed for 1 h in 400 pl of RB to produce the thiol form of Sepharose 6B. Then, RB was replaced by IB, and the addition of buffer, ultrasonication (10 s) and centrifugation were repeated six times. Immobilization: A suspension of SH-Sepharose* in 400 pl of IB containing from 20 to 40 nmol of the respective template was stirred by vortexing at 1000 rpm for 1 h. Capping: 100 pl of CB was added to the above suspension. After 1 h of stirring, the beads were separated by centrifugation, resuspended in 400 pl of CB and vortexed at 25 *C for 1 h. Then, CB was replaced by a 200 mM solution of S-(2-thiopyridyl) 2-mercaptoethanol in ethanol. After 2 h of vortexing at 70 *C, the beads were washed with ethanol (5 x 200 pl) and 1 M NaCi (5 x 200 p1). Hybridization: The beads were resuspended in 250 pl HB containing the comple mentary heptamers in a concentration of 400 pM. The temperature was increased to 85 0 C, decreased to 4 OC within 1 h and maintained at this value for 2 h. Excess heptamers were removed by washing the beads with three 200 pl portions of 1 M NaCl at 4 0 C. Chemical linking: The beads were resuspended in 375 pl of LB, vortexed at 4 4C for 40 h and washed three times with 200 pl of 1 M NaCl at 4 0 C. Denaturing: A micro-spin filter with the corresponding beads was inserted in an Eppendorf tube containing 150 pl of 1 M acetate buffer (pH 4.4). The beads were resuspended in 50 pl of 0.1 M NaOH and slightly stirred, and the supernatant was transferred to an Eppendorf tube by centrifugation (7000 x g, 30 s). After three repeats of this process step, the resulting solution (400 pl) was ready for the next cycle. The beads were recycled by washing with four portions of HB.

- 18 HPLC: All HPLC samples without reduction buffer (RB) were brought to a concen tration of 100 mM DTT prior to analysis in order to ensure reproducibility. All separations were performed on an RP-C-18 column (250/4, Nucleosil* 120-5 AB, Macherey & Nagel). Eluates: 0.1 M Triethylammonium acetate (pH = 7)/MeCN 1% (A) and MeCN (B). Analytical measurements were made at 50 OC by using a gradient of from 2% to 6% A in 2 min, from 6% to 25% A in 30 min and a flow rate of 1 ml/min. The eluate was monitored simultaneously at 254 nm and 273 nm. Equipment (Kontron): Two pumps 422, autosampler 465, diode-array detector 440. Kroma 2000 was used as a data acquisition system. Experiment: To perform the process outlined in Figure 2, the 14-meric templates X and Y were first immobilized using disulfide exchange reactions [J.R. Lorsch, Nature 371, 31-36 (1994)] on one charge each of the SH support to obtain the template-bearing supports X0 and YO. The efficiency of the immobilization step was determined by an HPLC analysis of the supernatant (Figure 3a,b). The remaining thiol groups of the supports were capped by reaction with S-(2-thio pyridyl)-2-mercaptoethanol. Then, fragments Ax, Bx and AY, By were hybridized with the immobilized templates YO and XO, respectively. To determine hybridization efficiency, part of the support was reduced with DTT and analyzed by HPLC (Figure 3c). The chemical linking [N.G. Dolinnaya et al., Nucleic Acids Res. 19, 3073-3080 (1991); K.D. James et al., Chem. Biol. 4, 595 605 (1997)] was achieved by exchanging the hybridization buffer by a linking buffer containing N-ethyl-N'-(dimethylaminopropyl)carbodiimide hydrochloride (EDC) as a condensation agent. To determine the efficiency of ligation, product formation was monitored by the HPLC analysis of the mixture of products obtained upon cleavage of the disulfide linkages of an aliquot of the template-bearing supports with DTT as a reductant (Figure 3d). Then, the copies were released by washing the supports with 0.1 M NaOH, analyzed by HPLC and immobilized again on two new charges of the SH support to provide the template-bearing supports X1 (copy of template-bearing support YO) and Y1 (copy of template-bearing support XO).

- 19 For the next generation, the whole cycle of steps was repeated with each of the four charges XO, YO, X1 and Y1 to provide Y2, X2, Y3 and X3, respectively. A further cycle yielded the template-bearing supports X4..X7, Y4..Y7 (Figure 3e,f).

- 20 Sequence Listing <110> von Kiedrowski, Gi~nther NOXXON Pharma AG <120> Process for the Exponential Amplification of Molecular Tem plates <130> 992074wo/JH/ml <140> <141> <160> 1 <170> PatentIn Rel. 2.1 <210> 1 <211> 14 <212> DNA <213> Artificial sequence <220> <223> Description of the artificial sequence: reference template <400> 1 gaatccatgg taag 14

Claims (20)

1. A process for the exponential amplification of molecular templates, compris ing: (a) binding of molecular templates to the surface of a solid phase using a reversible linker on the template; (b) addition of template fragments wherein one of the fragments has an optionally protected linker unit; (c) synthesis of copies of the template; (d) removing of excess template fragments and reaction auxiliary agents; (e) detaching the copies from the templates; and (f) colonization of free binding sites on the solid phase by synthesized template copies.

2. The process according to claim 1, wherein the template copies produced in step (e) are identical with or complementary to the starting templates.

3. The process according to claim 1, wherein two or more orthogonal immobili zation methods are employed on two or more surfaces.

4. The process according to one or more of claims 1 to 3, wherein excess binding sites on the solid phase are blocked after step (a).

5. The process according to one or more of claims 1 to 4, wherein the free binding sites on the solid phase in step (f) are produced by exposing new binding sites on the solid phase, by the addition of new solid phase material, - 22 or by transferring the template copy formed in step (e) to a new solid phase material.

6. The process according to claim 1, wherein the templates produced in step (e) are complementary to the starting templates and are transferred to a new support material.

7. The process according to one or more of claims 1 to 3, wherein the linker unit of the template fragment in step (b) is protected and the template cop ies formed in step (e) are deprotected.

8. The process according to one or more of claims 1 to 7, wherein the tem plates have a length of from 2 to 2000, especially from 4 to 50, monomer units, and monomer units or oligomeric monomer units are used as said template fragments.

9. The process according to claim 1 or 8, wherein the amount of templates employed in step (a) is from 10-15 to 10-' mol/l, and the concentration of the template fragments in step (b) is from 10-12 to 1 mol/l.

10. The process according to one or more of claims 1 to 9, wherein said molecular templates are selected from nucleotides and nucleotide deriva tives, peptides and peptide derivatives.

11. The process according to one or more of claims 1 to 10, wherein the linkages formed in step (c) are selected from phosphoric acid diester, phos phoric acid amide ester, ester, amide, disulfide, acetal and C-C linkages.

12. The process according to claim 1 or 11, wherein the bond formation in step (c) is effected chemically or by enzymatic linking.

13. The process according to one or more of claims 1 to 12, wherein the linker binds to the solid phase through covalent bonds or non-covalent binding. - 23

14. The process according to one or more of claims 1 to 13, wherein the solid phase material is selected from organic or inorganic materials or from a hy brid of such materials.

15. A method of chemical evolution by the exponential amplification of molecu lar templates, comprising one or more of the following evolution cycles: (1) selection of a subpopulation of molecular templates from a combina torial starting library; and (2) amplification and mutation of the selected templates to provide a mutant library comprising one or more of the amplification cycles (a) to (f) as defined in claims 1 to 14.

16. The method according to claim 15, wherein the selection of said subpopula tion of molecular template molecules in step (1) is effected by reversible binding to one or more target structures.

17. The method according to claim 15 or 16, wherein the mutant library obtained in step (2) is the starting library for the next evolution cycle.

18. The method according to one or more of claims 15 to 17, wherein the mutation is controlled by the reaction parameters in steps (b) and (c) of the amplification cycle.

19. The method according to one or more of claims 15 to 18, wherein the first evolution cycle additionally comprises the following step (1'): (1') providing the selected molecular templates with a reversible linker.

20. The method according to one or more of claims 15 to 19, wherein an additional selection is achieved in detaching step (e) by selective detaching.