`Lerner et al.
`
`IIIII
`
`I~ 1111111111111111111111111111111111111111111111111111
`US005723598A
`5,723,598
`[111 Patent Number:
`[451 Date of Patent:
`Mar. 3, 1998
`
`[54] ENCODED COMBINATORIAL CHEMICAL
`LffiRARIES
`
`[58] Field of Search .................................. 536/25.3. 24.2;
`435/6. 7.1; 5301333. 335. 336. 337
`
`[75]
`
`Inventors: Richard Lerner, La Jolla; Kim Janda.
`San Diego, both of Calif.; Sydney
`Brenner, Cambridge, England
`
`[73] Assignee: The Scripps Research Institute, La
`Jolla, Calif.
`
`[21] Appl. No.: 665,511
`Jun. 18,19%
`
`[22] Filed:
`
`Related U.S. AppHcation Data
`
`[62] Division of Ser. No. 860,445, Mar. 30, 1992, Pat. No.
`5,573,905.
`Int. CL6
`•••••••••••••••••••••••••••• C07B 21/00; C07H 21/04
`[51]
`[52] U.S. Cl •......................... 536/25.3; 530/335; 530/336;
`530/337
`
`Pri11Ulry Examiner-James Ketter
`Assistant E.xaminer-Irem Yucel
`Attome}\ Agen~ or Firm-Donald G. Lewis; Thomas
`Fitting
`
`[57]
`
`ABSTRACT
`
`The present invention describes an encoded combinatorial
`chemical library comprised of a plurality of bifunctional
`molecules having both a chemical polymer and an identifier
`oligonucleotide sequence that defines the structure of the
`chemical polymer. Also described are the bifunctional mol(cid:173)
`ecules of the library. and methods of using the library to
`identify chemical structures within the library that bind to
`biologically active molecules in preselected binding inter(cid:173)
`actions.
`
`S Claims, 2 Drawing Sheets
`
`
`
`Page 1 of 25
`
`ILMN EXHIBIT 1005
`
`
`
`U.S. Patent
`
`Mar. 3, 1998
`
`Sheet 1 of 2
`
`5,723,598
`
`Apa I
`Sty I
`JGGGCCCTATTCTTAG 3 1
`5 1 AGCTACTTCCCAAGG(coding sequence
`3 1 TCGATGAAGGGTTCC(anticoding strand]CCCGGGATAAGAATC 5'
`
`Sty I & Apa I
`
`I Step 1 Cleavage by
`t
`t Step 2 Biotinylation
`
`5'
`3 I
`
`AGCTACTTCC
`TCGATGAAGGGTTC
`
`CAAGG(coding sequence
`]GGGCC
`Q[anticoding strand)Q
`
`CTATTCTTAG
`CCGGGATAAGAATC
`
`3 I
`5'
`
`5'
`J I
`
`AGCTACTTCCC
`)GGGCC
`CAAGG[coding sequence
`TCGATGAAGGGTTC BBCQ[anticoding strand)£
`
`CTATTCTTAG
`CCGGGATAAGAATC
`
`3 I
`51
`
`FIGURE l
`
`
`
`Page 2 of 25
`
`
`
`U.S. Patent
`
`Mar. 3, 1998
`
`Sheet 2 of 2
`
`5,723,598
`
`Pl - LINK
`
`Step 1
`
`CACATG-Pl-LINK-gly
`
`ACGGTA-Pl-LINK-met
`
`t
`J,
`
`Step 2
`
`CACATGCACATG-Pl-LINK-gly.gly
`CACATGACGGTA-Pl-LINK-met.gly
`
`ACGGTACACATG-Pl-LINK-gly.met
`ACGGTAACGGTA-Pl-LINK-met.met
`
`l
`
`Step 3
`
`P2CACATGCACATGCACATGP1-LINK-gly.gly.gly
`P2CACATGCACATGACGGTAP1-LINK-met.gly.gly
`P2CACATGACGGTACACATGP1-LINK-gly.met.gly
`P2CACATGACGGTAACGGTAP1-LINK-met.met.gly
`
`P2ACGGTACACATGCACATGP1-LINK-gly.gly.met
`P2ACGGTACACATGACGGTAP1-LINK-met.gly.met
`P2ACGGTAACGGTACACATGP1-LINK-gly.met.rnet
`P2ACGGTAACGGTAACGGTAP1-LINK-met.met.rnet
`
`Pl = GGGCCCTATTCTTAG
`P2 = AGCTACTTCCCAAGG
`
`FIGURE 2
`
`
`
`Page 3 of 25
`
`
`
`5,723,598
`
`1
`ENCODED COMBINATORIAL CHEMICAL
`LmRARIES
`
`This is a divisional of application Ser. No. 07/860.445.
`filed Mar. 30. 1992. now U.S. Pat. No. 5.573.905, whose
`disclosure are incorporated herein by reference.
`
`TECHNICAL FIELD
`
`The present invention relates to encoded chemical librar(cid:173)
`ies that contain repertoires of chemical structures defining a
`diversity of biological structures, and methods for using the
`libraries.
`
`BACKGROUND
`
`There is an increasing need to find new molecules which
`can effectively modulate a wide range of biological
`processes, for applications in medicine and agriculture. A
`standard way for searching for novel bioactive chemicals is
`to screen collections of natural materials, such as fermen(cid:173)
`tation broths or plant extracts, or libraries of synthesized
`molecules using assays which can range in complexity from
`simple binding reactions to elaborate physiological prepa(cid:173)
`rations. The screens often only provide leads which then
`require further improvement either by empirical methods or
`by chemical design. The process it time-consuming and
`costly but it is unlikely to be totally replaced by rational
`methods even when they are based on detailed knowledge of
`the chemical structure of the target molecules. Thus, what
`we might call "irrational drug design"-the process of
`selecting the right molecules from large ensembles or
`repertoires-requires continual improvement both in the
`generation of repertoires and in the methods of selection.
`Recently there have been several developments in using
`peptides or nucleotides to provide libraries of compounds 35
`for lead discovery. The methods were originally developed
`to speed up the determination of epitopes recognized by
`monoclonal antibodies. For example, the standard serial
`process of stepwise search of synthetic peptides now encom(cid:173)
`passes a variety of highly sophisticated methods in which 40
`large arrays of peptides are synthesized in parallel and
`screened with acceptor molecules labelled with fluorescent
`or other reporter groups. The sequence of any effective
`peptide can be decoded from its address in the array. See for
`example Geysen et al .• Proc. Natl. Acad. Sci. USA, 45
`81:3998-4002 (1984); Maeji et al .. J. lmmunol. Met.,
`146:83-90 (1992); and Fodor et al., Science, 251: 767-775
`(1991).
`In another approach. Lam et. al.. Nature, 354:82-84
`(1991) describes combinatorial libraries of peptides that are
`synthesized on resin beads such that each resin bead con(cid:173)
`tains about 20 pmoles of the same peptide. The beads are
`screened with labelled acceptor molecules and those with
`bound acceptor are searched for by visual inspection. physi(cid:173)
`cally removed, and the peptide identified by direct sequence
`analysis. In principle, this method could be used with other
`chemical entities but it requires sensitive methods for
`sequence determination.
`A different method of solving the problem of identifica(cid:173)
`tion in a combinatorial peptide library is used by Houghten
`et al .. Nature, 354:84-86 (1991). For hexapeptides of the 20
`natural amino acids, 400 separate libraries are synthesized,
`each with the first two amino acids fixed and the remaining
`four positions occupied by all possible combinations. An
`assay. based on competition for binding or other activity. is
`then used to find the library with an active peptide. Then
`twenty new libraries are synthesized and assayed to deter-
`
`2
`mine the effective amino acid in the third position. and the
`process is reiterated in this fashion until the active hexapep(cid:173)
`tide is defined. This is analogous to the method used in
`searching a dictionary; the peptide is decoded by construe-
`s tion using a series of sieves or buckets and this makes the
`search logarithmic.
`A very powerful biological method has recently been
`described in which the library of peptides is presented on the
`surface of a bacteriophage such that each phage has an
`10 individual peptide and contains the DNA sequence specify(cid:173)
`ing it. The library is made by synthesizing a repertoire of
`random oligonucleotides to generate all combinations, fol(cid:173)
`lowed by their insertion into a phage vector. Each of the
`sequences is cloned in one phage and the relevant peptide
`15 can be selected by finding those that bind to the particular
`target. The phages recovered in this way can be amplified
`and the selection repeated. The sequence of the peptide is
`decoded by sequencing the DNA. See for example Cwirla et
`al .. Proc. NatL Acad. Sci. USA. 87:6378-6382 (1990); Scott
`20 et al., Science, 249:386-390 (1990); and Devlin et al ..
`Science, 249:404-406 (1990).
`Another "genetic" method has been described where the
`libraries are the synthetic oligonucleotides themselves
`wherein active oligonucleotide molecules are selected by
`binding to an acceptor and are then amplified by the poly(cid:173)
`merase chain reaction (PCR). PCR allows serial enrichment
`and the structure of the active molecules is then decoded by
`DNA sequencing on clones generated from the PCR prod(cid:173)
`ucts. The repertoire is limited to nucleotides and the natural
`pyrimidine and purine bases or those modifications that
`preserve specific Watson-Crick pairing and can be copied by
`polymerases.
`The main advantages of the genetic methods reside in the
`capacity for cloning and amplification of DNA sequences.
`which allows enrichment by serial selection and provides a
`facile method for decoding the structure of active molecules.
`However. the genetic repertoires are restricted to nucleotides
`and peptides composed of natural amino acids and a more
`extensive chemical repertoire is required to populate the
`entire universe of binding sites. In contrast. chemical meth(cid:173)
`ods can provide limitless repertoires but they lack the
`capacity for serial enrichment and there are difficulties in
`discovering the structures of selected active molecules.
`
`25
`
`30
`
`BRIEF SUMMARY OF THE INVENTION
`The present invention provides a way of combining the
`virtues of both of the chemical and genetic methods sum(cid:173)
`marized above through the construction of encoded com.bi-
`50 natorial chemical libraries, in which each chemical sequence
`is labelled by an appended "genetic" tag. itself constructed
`by chemical synthesis, to provide a "retrogenetic" way of
`specifying each chemical structure.
`In outline. two alternating parallel combinatorial synthe-
`55 ses are performed so that the genetic tag is chemically linked
`to the chemical structure being synthesized; in each case. the
`addition of one of the particular chemical units to the
`structure is followed by the addition of an oligonucleotide
`sequence. which is defined to "code" for that chemical unit.
`60 ie.. to function as an identifier for the structure of the
`chemical unit. The library is built up by the repetition of this
`process after pooling and division.
`Active molecules are selected from the library so pro(cid:173)
`duced by binding to a preselected biological molecule of
`65 interest. Thereafter. the identity of the active molecule is
`determined by reading the genetic tag. i.e .. the identifier
`oligonucleotide sequence. In one embodiment. amplified
`
`
`
`Page 4 of 25
`
`
`
`3
`copies of their retrogenetic tags can be obtained by the
`polymerase chain reaction.
`The strands of the amplified copies with the appropriate
`polarity can then be used to enrich for a subset of the library
`by hybridization with the matching tags and the process can 5
`then be repeated on this subset. Thus serial enrichment is
`achieved by a process of purification exploiting linkage to a
`nucleotide sequence which can be amplified. Finally, the
`structure of the chemical entities are decoded by cloning and
`sequencing the products of the PCR reaction.
`The present invention therefore provides a novel method
`for identifying a chemical structure having a preselected
`binding activity through the use of a library of bifunctional
`molecules that provides a rich source of chemical diversity.
`The library is used to identify chemical structures (structural 15
`motifs) that interact with preselected biological molecules.
`Thus. in one embodiment. the invention contemplates a
`bifunctional molecule according to the formula A-B-C.
`where A is a chemical moiety. B is a linker molecule
`operatively linked to A and C. and C is an identifier 20
`oligonucleotide comprising a sequence of nucleotides that
`identifies the structure of chemical moiety A.
`In another embodiment, the invention contemplates a
`library comprising a plurality of species of bifunctional
`molecules, thereby forming a repertoire of chemical diver- 25
`sity.
`Another embodiment contemplates a method for identi(cid:173)
`fying a chemical structure that participates in a preselected
`binding interaction with a biologically active molecule,
`where the chemical structure is present in the library of 30
`bifunctional molecules according to this invention. The
`method comprises the steps of:
`a) admixing in solution the library of bifunctional mol(cid:173)
`ecules with the biologically active molecule under
`binding conditions for a time period sufficient to form
`a binding reaction complex;
`b) isolating the complex formed in step (a); and
`c) determining the nucleotide sequence of the polymer
`identifier oligonucleotide in the isolated complex and 40
`thereby identifying the chemical structure that partici(cid:173)
`pated in the preselected binding interaction.
`The invention also contemplates a method for preparing a
`library according to this invention comprising the steps of:
`a) providing a linker molecule B having termini A' and C' 45
`according to the formula A'-B-C' that is adapted for
`reaction with a chemical precursor unit X' at termini A'
`and with a nucleotide precursor Z' at termini C';
`b) conducting syntheses by adding chemical precursor
`unit X' to termini A' of said linker and adding precursor 5o
`unit identifier oligonucleotide z· to termini C' of said
`linker. to form a composition containing bifunctional
`molecules having the structure X,-B-Z,;
`c) repeating step (b) on one or more aliquots of the
`composition to produce aliquots that contain a product 55
`containing a bifunctional molecule;
`d) combining the aliquots produced in step (c) to form an
`admixture of bifunctional molecules. thereby forming
`said library.
`
`35
`
`BRIEF DESCRIPfiON OF THE DRAWINGS
`In the drawings. forming a portion of this disclosure:
`FIG. 1 illustrates a scheme for the restriction endonu(cid:173)
`clease cleavage of a PCR amplification product derived from
`a bifunctional molecule of this invention (Step 1). and the 65
`subsequent addition of biotin to the cleaved PCR product
`(Step 2).
`
`5.723.598
`
`4
`FIG. 2 illustrates the process of producing a library of
`bifunctional molecules according to the method described in
`Example 9.
`
`DEfAllED DESCRIPTION OF THE
`INVENTION
`
`A. Encoded Combinatorial Chemical Libraries
`
`An encoded combinatorial chemical library is a compo-
`10 sition comprising a plurality of species of bifunctional
`molecules that each define a different chemical structure and
`that each contain a unique identifier oligonucleotide whose
`nucleotide sequence defines the corresponding chemical
`structure.
`1. Bifunctional Molecules
`A bifunctional molecule is the basic unit in a ltbrary of
`this invention, and combines the elements of a polymer
`comprised of a series of chemical building blocks to form a
`chemical moiety in the ltbrary. and a code for identifying the
`structure of the chemical moiety.
`Thus, a bifunctional molecule can be represented by the
`formula A-B-C. where A is a chemical moiety. B is a linker
`molecule operatively linked to A and C. and C is an identifier
`oligonucleotide comprising a sequence of nucleotides that
`identifies the structure of chemical moiety A.
`a. Chemical Polymers
`A chemical moiety in a bifunctional molecule of this
`invention is represented by A in the above formula A-B-C
`and is a polymer comprising a linear series of chemical units
`represented by the formula (X,)a• wherein X is a single
`chemical unit in polymer A and n is a position identifier for
`X in polymer A. n has the value of 1 +i where i is an integer
`from 0 to 10. such that when n is 1. X is located most
`proximal to the linker (B).
`Although the length of the polymer can vary. defined by
`a, practical library size limitations arise if there is a large
`alphabet size as discussed further herein. Typically. a is an
`integer from 4 to 50.
`A chemical moiety (polymer A) can be any of a variety of
`polymeric structures. depending on the choice of classes of
`chemical diversity to be represented in a library of this
`invention. Polymer A can be any monomeric chemical unit
`that can be coupled and extended in polymeric form. For
`example. polymer A can be a polypeptide, oligosaccharide,
`glycolipid. lipid. proteoglycan. glycopeptide. sulfonamide,
`nucleoprotein, conjugated peptide (i.e.. having prosthetic
`groups). polymer containing enzyme substrates, including
`transition state analogues. and the like biochemical poly(cid:173)
`mers. Exemplary is the polypeptide-based library described
`herein.
`Where the library is comprised of peptide polymers. the
`chemical unit X can be selected to form a region of a natural
`protein or can be a non-natural polypeptide. can be com(cid:173)
`prised of natural D-amino acids. or can be comprised of
`non-natural amino acids or mixtures of natural and non(cid:173)
`natural amino acids. The non-natural combinations provide
`for the identification of useful and unique structural motifs
`60 involved in biological interactions.
`Non-natural amino acids include modified amino acids
`and L-amino acids. stereoisomer of D-amino acids.
`The amino acid residues described herein are preferred to
`be in the "L" isomeric form. NH 2 refers to the free amino
`group present at the amino terminus of a polypeptide. COOH
`refers to the free carboxy group present at the carboxy
`terminus of a polypeptide. In keeping with standard
`
`
`
`Page 5 of 25
`
`
`
`5,723,598
`
`5
`polypeptide nomenclature, J. BioL Chem, 243:3552-59
`(1969) and adopted at 37 C.P.R. § 1.822(b)(2)), abbrevia(cid:173)
`tions for amino acid residues are shown in the following
`Table of Correspondence:
`
`TABLE OF CORRESPONDENCE
`
`SYMBOL
`
`!-Letter
`
`3-Letter
`
`AMINO ACID
`
`y
`G
`F
`M
`A
`s
`I
`L
`T
`v
`p
`K
`H
`Q
`E
`w
`R
`D
`N
`c
`
`Tyr
`Gly
`Phe
`Met
`Ala
`Ser
`lie
`Leu
`Tbr
`Val
`Pro
`Lys
`His
`Gin
`Glu
`'frp
`Arg
`Asp
`Asn
`Cys
`
`tyrosine
`glycine
`phenylalanine
`methionine
`alanine
`serine
`isoleucine
`leucine
`threonine
`valine
`proline
`lysine
`histidine
`glutamine
`glutamic acid
`ttyptophan
`arginine
`aspartic acid
`asparagine
`cysteine
`
`The phrase "amino acid residue" is broadly defined to
`include the amino acids listed in the Table of Correspon(cid:173)
`dence and modified and unusual amino acids, such as those
`listed in 37 C.P.R. §1.822(b)(4). and incorporated herein by
`reference.
`The polymer defined by chemical moiety A can therefor
`contain any polymer backbone modifications that provide
`increased chemical diversity. In building of a polypeptide
`system as exemplary, a variety of modifications are
`contemplated, including the following backbone structures:
`-NHN(R) CO-, -NHB(R) CO-, -NHC(RR')CO-,
`-NHC 6 H 4 CO-, 40
`-NHC(=CHR)CO-,
`-NHCH2CHRCO-, -NHCHRCH 2CO-, and lactam
`structures.
`In addition. amide bond modifications are contemplated
`including -COCH2- . -COS-. -CONR. -COO-.
`-CSNH-, -CH2 NH-. -CH 2 CH2 - . -CH2 S-,
`-CH 2 SO-. -CH 2 S0 2 - . -CH(CH 3 )S-,
`-CH=CH-. -NHCO-.
`-NHCONH-,
`-CONHO-. and -C(=CH~CH2-.
`b. Polymer Identifier Oligonucleotide
`An identifier oligonucleotide in a bifunctional molecule of
`this invention is represented by C in the above formula
`A-B-C and is an oligonucleotide having a sequence repre(cid:173)
`sented by the formula (Z,.)a, wherein Z is a unit identifier
`nucleotide sequence within oligonucleotide C that identifies
`the chemical unit X at position n. n has the value of 1 +i
`where i is an integer from 0 to 10. such that when n is 1. Z
`is located most proximal to the linker (B). a is an integer as
`described previously to connote the number of chemical unit
`identifiers in the oligonucleotide.
`For example. a bifunctional molecule can be represented
`by the formula:
`x..x,x,x,-B-z,z,z,z •.
`In this example. the sequence of oligonucleotides Z 1, Z2 , Z 3 65
`and z4 identifies the structure of chemical units xl. x2. x3
`and X4 , respectively. Thus, there is a correspondence in the
`
`50
`
`10
`
`20
`
`6
`identifier sequence between a chemical unit X at position n
`and the unit identifier oligonucleotide Z at position n.
`The length of a unit identifier oligonucleotide can vary
`depending on the complexity of the library, the number of
`5 different chemical units to be uniquely identified, and other
`considerations relating to requirements for uniqueness of
`oligonucleotides such as hybridization and polymerase
`chain reaction fidelity. A typical length can be from about 2
`to about 10 nucleotides. although nothing is to preclude a
`unit identifier from being longer.
`Insofar as adenosine (A). guanosine (G), thymidine (T)
`and cytosine (C) represent the typical choices of nucleotides
`for inclusion in a unit identifier oligonucleotide. A. G. T and
`C form a representative "alphabet" used to "spell" out a unit
`identifier oligonucleotide's sequence. Other nucleotides or
`!5 nucleotide analogs can be utilized in addition to or in place
`of the above four nucleotides, so long as they have the ability
`to form Watson-Crick pairs and be replicated by DNA
`polymerases in a PCR reaction. However. the nucleotides A.
`G. T and Care preferred.
`For the design of the code in the identifier
`oligonucleotide. it is essential to chose a coding represen(cid:173)
`tation such that no significant part of the oligonucleotide
`sequence can occur in another unrelated combination by
`chance or otherwise during the manipulations of a bifunc-
`25 tional molecule in the library.
`For example. consider a library where Z is a trinucleotide
`whose sequence defines a unique chemical unit X. Because
`the methods of this invention provide for all combinations
`and permutations of an alphabet of chemical units. it is
`30 possible for two different unit identifier oligonucleotide
`sequences to have closely related sequences that differ by
`only a frame shift and therefore are not easily distinguish(cid:173)
`able by hybridization or sequencing unless the frame is clear.
`Other sources of misreading of a unit identifier oligo-
`35 nucleotide can arise. For example, mismatch in DNA
`hybridization, transcription errors during a primer extension
`reaction to amplify or sequence the identifier
`oligonucleotide. and the like errors can occur during a
`manipulation of a bifunctional molecule.
`The invention contemplates a variety of means to reduce
`the possibility of error in reading the identifier
`oligonucleotide, such as to use longer nucleotide lengths for
`a unit identifier nucleotide sequence as to reduce the simi(cid:173)
`larity between unit identifier nucleotide sequences. Typical
`45 lengths depend on the size of the alphabet of chemical units.
`A representative system useful for eliminating read errors
`due to frame shift or mutation is a code developed as a
`theoretical alternative to the genetic code and is known as
`the commaless genetic code.
`Where the chemical units are amino acids. a convenient
`unit identifier nucleotide sequence is the well known genetic
`code using triplet codons. The invention need not be limited
`by the translation afforded between the triplet codon of the
`genetic code and the natural amino acids; other systems of
`55 correspondence can be assigned.
`A typical and exemplary unit identifier nucleotide
`sequence is based on the commaless code having a length of
`six nucleotides (hexanucleotide) per chemical unit.
`Preferably. an identifier oligonucleotide has at least 15
`60 nucleotides in the tag (coding) region for effective hybrid(cid:173)
`ization. In addition, considerations of the complexity of the
`library. the size of the alphabet of chemical units. and the
`length of the polymer length of the chemical moiety all
`contribute to length of the identifier oligonucleotide as
`discussed in more detail herein.
`In a preferred embodiment. an identifier oligonucleotide
`C has a nucleotide sequence according to the formula
`
`
`
`Page 6 of 25
`
`
`
`5.723.598
`
`7
`P1-(Zn)0 -P2. where P1 and P2 are nucleotide sequences that
`provide polymerase chain reaction (PCR) primer binding
`sites adapted to amplify the polymer identifier oligonucle(cid:173)
`otide. The requirements for PCR primer binding sites are
`generally well known in the art. but are designed to allow a 5
`PCR amplification product (a PeR-amplified duplex DNA
`fragment) to be formed that contains the polymer identifier
`oligonucleotide sequences.
`The presence of the two PCR primer binding sites. Pl and
`P2. flanking the identifier oligonucleotide sequence (Zn)a 10
`provides a means to produce a PeR-amplified duplex DNA
`fragment derived from the bifunctional molecule using PCR.
`This design is useful to allow the amplification of the tag
`sequence present on a particular bifunctional molecule for
`cloning and sequencing purposes in the process of reading 15
`the identifier code to determine the structure of the chemical
`moiety in the bifunctional molecule.
`More preferred is a bifunctional molecule where one or
`both of the nucleotide sequences Pl and P2 are designed to
`contain a means for removing the PCR primer binding sites 20
`from the identifier oligonucleotide sequences. Removal of
`the flanking P1 and P2 sequences is desirable so that their
`sequences do not contribute to a subsequent hybridization
`reaction. Preferred means for removing the PCR primer
`binding sites from a PCR amplification product is in the
`form of a restriction endonuclease site within the PeR(cid:173)
`amplified duplex DNA fragment
`Restriction endonucleases are well known in the art and
`are enzymes that recognize specific lengths of duplex DNA
`and cleave the DNA in a sequence-specific manner.
`Preferably. the restriction endonuclease sites should be
`positioned proximal to (Zn)a relative to the PCR primer
`binding sites to maximize the amount of P1 and P2 that is
`removed upon treating a bifunctional molecule to the spe(cid:173)
`cific restriction endonuclease. More preferably. P1 and P2
`each are adapted to form a restriction endonuclease site in
`the resulting PeR-amplified duplex DNA. and the two
`restriction sites. when cleaved by the restriction
`endonuclease. form non-overlapping cohesive termini to
`facilitate subsequent manipulations.
`Particularly preferred are restriction sites that when
`cleaved provide overhanging termini adapted for termini(cid:173)
`specific modifications such as incorporation of a biotinylated
`nucleotide (e.g .. biotinyl deoxy-lffP) to facilitate subse(cid:173)
`quent manipulations.
`The above described preferred embodiments in an iden(cid:173)
`tifier oligonucleotide are summarized in a specific embodi(cid:173)
`ment shown in FIG. 1.
`In FIG. 1. a PCR-amplified duplex DNA is shown that is
`derived from an identifier oligonucleotide described in the
`Examples. The (Zn) sequence is illustrated in the brackets as
`the coding sequence and its complementary strand of the
`duplex is indicated in the brackets as the anticoding strand
`The Pl and P2 sequences are shown in detail with a Sty I
`restriction endonuclease site defined by the Pl sequence 55
`located 5' to the bracket and an Apy I restriction endonu(cid:173)
`clease site defined by the P2 sequence located 3' to the
`bracket.
`Step 1 illustrates the cleavage of the PCR-amplified
`duplex DNA by the enzymes Sty I and Apa I to form a 60
`modified identifier sequence with cohesive termini. Step 2
`illustrates the specific biotinylation of the anticoding strand
`at the Sty I site. whereby the incotporation of biotinylated
`lffP is indicated by a B.
`The presence of non-overlapping cohesive termini after 65
`Step 1 in FIG. 1 allows the specific and directional cloning
`of the restriction-digested PCR-amplified fragment into an
`
`8
`appropriate vector. such as a sequencing vector. In addition.
`the Sty I was designed into PI because the resulting over(cid:173)
`hang is a substrate for a filling-in reaction with dCfP and
`biotinyl-dUTP (BTP) using DNA polymerase Klenow frag(cid:173)
`ment. The other restriction site. Apa L was selected to not
`provide substrate for the above biotinylation. so that only the
`anticoding strand can be biotinylated.
`Once biotinylated the duplex fragment can be bound to
`immobilized avidin and the duplex can be denatured to
`release the coding sequence containing the identifier nucle(cid:173)
`otide sequence. thereby providing purified anticoding strand
`that is useful as a hybridization reagent for selection of
`related coding strands as described further herein.
`c. Linker Molecules
`A linker molecule in a bifunctional molecule of this
`invention is represented by B in the above formula A-B-C
`and can be any molecule that performs the function of
`operatively linking the chemical moiety to the identifier
`oligonucleotide.
`Preferably. a linker molecule has a means for attaching to
`a solid support. thereby facilitating synthesis of the bifunc(cid:173)
`tional molecule in the solid phase. In addition. attachment to
`a solid support provides certain features in practicing the
`screening methods with a library of bifunctional molecules
`25 as described herein. Particularly preferred are linker mol(cid:173)
`ecules in which the means for attaching to a solid support is
`reversible. namely. that the linker can be separated from the
`solid support.
`A linker molecule can vary in structure and length. and
`30 provide at least two features: (1) operative linkage to chemi(cid:173)
`cal moiety A. and (2) operative linkage to identifier oligo(cid:173)
`nucleotide C. As the nature of chemical linkages is diverse.
`any of a variety of chemistries may be utilized to effect the
`indicated operative linkages to A and to C. as the nature of
`35 the linkage is not considered an essential feature of this
`invention. The size of the linker in terms of the length
`between A and C can vary widely. but for the pUtposes of the
`invention. need not exceed a length sufficient to provide the
`linkage functions indicated Thus. a chain length of from at
`40 least one to about 20 atoms is preferred
`A preferred linker molecule is described in Example 3
`herein that contains the added, preferred. element of a
`reversible means for attachment to a solid support. That is.
`the bifunctional molecule is removable from the solid sup-
`45 port after synthesis.
`Solid supports for chemical synthesis are generally well
`known. Particularly preferred are the synthetic resins used in
`oligonucleotide and in polypeptide synthesis that are avail(cid:173)
`able from a variety of commercial sources including Glen
`so Research (Herndon, Va.). Bachem Biosciences.
`(Philadelphia. Pa.). and Applied Biosystems (Foster City.
`Calif.). Most preferred are teflon supports such as that
`described in Example 2.
`2. Libraries
`A library of this invention is a repertoire of chemical
`diversity comprising a plurality of species of bifunctional
`molecules according to the present invention. The plurality
`of species in a library defines a family of chemical diversity
`whose species each have a different chemical moiety. Thus
`the library can define a family of peptides. lipids. oligosac(cid:173)
`carides or any of the other classes of chemical polymers
`recited previously.
`The number of different species in a library represents the
`complexity of a library and is defined by the polymer length
`of the chemical moiety. and by the size of the chemical unit
`alphabet that can be used to build the chemical unit polymer.
`The number of different species referred to by the phrase
`
`
`
`Page 7 of 25
`
`
`
`5,723,598
`
`9
`"plurality of species" in a library can be defined by the
`formula va. i.e .. v to power of a (exponent a). v represents
`the alphabet size. i.e .. the number of different chemical units
`X available for use in the chemical moiety. "a" is an
`exponent to V and represents the number of chemical units
`of X forming the polymer A, i.e., the length of polymer A.
`For example. for a bifunctional molecule where polymer
`A is a peptide having a length of 6 amino acids. and where
`the amino adds utilized can be any of the 20 natural amino
`acids. the alphabet (V) is 20 and the polymer length (a) is 6.
`and the library size is 206 or 64 million. This exemplary
`library provides a repertoire of chemical diversity compris(cid:173)
`ing 64 million different hexameric polypeptides operatively
`linked to corresponding unique identifier oligonucleotides.
`Because the complexity of the library will determine the
`amount of a particular species of bifunctional molecule
`relative the other species in the library. there are theoretical
`limits to the maximum useful complexity in a library.
`Therefore it is useful to consider how large (complex) a
`library should be. This size limit is dictated by the level of
`sensitivity for detecting the presence of a polymer identifier
`oligonucleotide after a screening procedure according to this
`invention. Detection sensitivity is dictated by the threshold
`of binding between an acceptor molecule to be assayed and
`a bifunctional molecule.
`If. for example, the binding threshold is 10-6M
`(micromolar), then there must be at least one nanomole of
`each species in a library of 1 milliliter (ml) volume. At this
`threshold, a library having a complexity of 104 could contain
`10 micromoles of each species. Because of the reciprocal
`relationship betwe