(12) INTERNATIONAL APPLICATION PUBLISHED UNDER THE PATENT COOPERATION TREATY(PCT)
`
`(19) World Intellectual Property Organization
`International Bureau
`
`(43) International Publication Date
`20 February 2003 (20.02.2003)
`
`
`
`(10) International Publication Number
`WO 03/014960 A2
`
`G)
`
`International Patent Classification’:
`
`GO6F 17/00
`
`Q))
`
`International Application Number:
`
`PCT/GB02/03512
`
`Chemistry, MRC Laboratory of Molecular Biology, ITills
`Road, Cambridge CB2 2HQ (GB). SETTANNI, Giovanni
`[IT/IT]; Strada Torino 12/A, T-10043 Orbassano, TO (IT).
`
`(22)
`
`International Filing Date:
`
`1 August 2002 (01.08.2002)
`
`(74)
`
`Agents: MASCHIO,Antonio et al.; D Young & Co, 21
`New Fetter Lane, London EC4A IDA (GB).
`
`(25)
`
`Filing Language:
`
`(26)
`
`Publication Language:
`
`English
`
`(81)
`
`English
`
`(30)
`
`()
`
`(72)
`(75)
`
`Priority Data:
`0119004.0
`0121577.1
`RM2001 A000633
`0200928.0
`0203569.9
`
`3 August 2001 (03.08.2001)
`6 September 2001 (06.09.2001)
`25 October 2001 (25.10.2001)
`16 January 2002 (16.01.2002)
`14 February 2002 (14.02.2002)
`
`GB
`GB
`IT
`GB
`GB
`
`Applicants (for all designated States except US): MEDI
`CAL RESEARCH COUNCIL [GB/GB]; 20 Park Cres-
`cent, London W1B 4AL (GB). SISSA (SCUOLA SUPE-
`RIORE INTERNAZIONALE DI STUDI AVANZATD
`[IT/IT]; Via Beirut 2-4, I-34014 Trieste (IT).
`
`Inventors; and
`Inventors/Applicants (for US only); CATTANEO, An-
`tonio [ITAT]; International School of Advanced Studies
`(SISSA), Biophysic Sector, Via Beirut, 2/4, I-34014 Tri-
`este (IT). MARITAN, Amos [IT/IT]; SISSA (Scuola Su-
`periore Internazionale di Studi Av, anzati), Via Beirut 2-4,
`1-34014 Trieste (IT). VISINTIN, Michela [IT/IT]; SISSA
`(Scuola Superiore Internazionale di Studi Av, anzati), Via
`Beirut 2-4, I-34014 Trieste (IT). RABBITTS, Terrence,
`Howard [GB/GB]; Division of Protein and Nucleic Acid
`
`Designated States (national): AE, AG, AL, AM, AT, AU,
`AZ, BA, BB, BG, BR, BY, BZ, CA, CH, CN, CO, CR, CU,
`CZ, DE, DK, DM, DZ, EC, EE, ES, FI, GB, GD, GE, GH,
`GM, HR, HU,ID,1., IN, IS, JP, KE, KG, KP, KR, KZ, LC,
`LK, LR, LS, LT, LU, LV, MA, MD, MG, MK, MN, MW,
`MX, MZ, NO, NZ, OM, PH, PL, PT, RO, RU, SD, SE, SG,
`SI, SK, SL, TJ, TM, TN, TR, TT, TZ, UA, UG, US, UZ,
`VN, YU, ZA, ZM, ZW.
`
`(84)
`
`Designated States (regional): ARIPO patent (GH, GM,
`KE, LS, MW, MZ, SD, SL, SZ, TZ, UG, ZM, ZW),
`Eurasian patent (AM, AZ, BY, KG, KZ, MD, RU, TJ, TM),
`European patent (AT, BE, BG, CH, CY, CZ, DE, DK, EE,
`ES, FI, FR, GB, GR, IE, IT, LU, MC, NL, PT, SE, SK,
`TR), OAPI patent (BE, BJ, CK CG, Cl, CM, GA, GN, GQ,
`GW, ML, MR, NE, SN, TD, TG).
`
`Declaration under Rule 4.17:
`ofimventorship (Rule 4.17(iv)) for US only
`
`Published:
`
`without international search report and to be republished
`upon receipt ofthat report
`
`For two-letter codes and other abbreviations, refer to the "Guid-
`ance Notes on Codes andAbbreviations" appearing at the begin-
`ning ofeach regular issue ofthe PCT Gazette.
`
`O03/014960A2
`
`Title: INTRACELLULAR ANTIBODIES
`
`G4)
`
`(57) Abstract: A method of identifying at least one consensus sequence for an intracellular antibody (LCS) comprising the steps
`of: creating a database comprising sequences of validated intracellular antibodies (VIDA database) and aligning the sequences of
`validated intracellular antibodies according to Kabat; determining the frequency with which a particular amino acid occurs in each
`of the positions of the aligned antibodies; selecting a frequency threshold value (LP or consensus threshold) in the range from 70 %
`to 100 %; identifying the positions of the alignment at which the frequency of a particular amino acid is greater than or equal to the
`LP value; and identifying the most frequent aminoacid, in the positions of said alignment.
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`Intracellular antibodies
`
`The present invention relates to molecules which can function in an intracellular
`environment. In particular, the invention relates to the characteristics of immunoglobulin
`molecules which can bind selectively to a ligand within an intracellular environment.
`
`Uses of these molecules are also described.
`
`Backgroundto the Invention
`
`Intracellular antibodies or intrabodies have been demonstrated to functionin antigen
`recognition in the cells of higher organisms (reviewed in Cattaneo, A. & Biocca, 8.
`(1997) Intracellular Antibodies: Development and Applications. Landes and Springer-
`Verlag). This interaction can influence the function of cellular proteins which have been
`successfully inhibited in the cytoplasm, the nucleusor in the secretory pathway. This
`efficacy has been demonstrated for viral resistance in plant biotechnology (Taviadoraki,
`P., et al. (1993) Nature 366: 469-472) and several applications have been reported of
`intracellular antibodies binding to HIV viral proteins (Mhashilkar, A.M., et al. (1995)
`EMBO J 14: 1542-51; Duan, L. & Pomerantz, R.J. (1994) Nucleic Acids Res 22: 5433-8;
`Maciejewski, J.P., et al. (1995) Nat Med 1: 667-73; Levy-Mintz,P., et al. (1996) J. Virol.
`70: 8821-8832) and to oncogene products (Biocca, S., Pierandrei-Amaldi, P. & Cattaneo,
`A. (1993) Biochem Biophys Res Commun 197: 422-7; Biocca, S., Pierandrei-Amaldi, P.,
`Campioni, N. & Cattaneo, A. (1994) Biotechnology (N Y) 12: 396-9; Cochet, O., ef al.
`(1998) Cancer Res 58: 1170-6). The latter is an important area because enforced
`expression of oncogenes often occurs in tumourcells after chromosomaltranslocations
`(Rabbitts, T.H. (1994) Nature 372: 143-149). These proteins are therefore important
`intracellular therapeutic targets (Rabbitts, TH. (1998) New Eng. J. Med 338: 192-194)
`which could be inactivated by binding with intracellular antibodies. Finally, the
`international efforts at whole genome sequencing will produce massive numbers of
`potential gene sequences which encodeproteins about which nothing is known.
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`Functional genomicsis an approach to ascertain the function ofthis plethora ofproteins
`andthe useofintracellular antibodies promises to be an importanttool in this endeavour
`as a conceptually simple approach to knocking-out protein function directly by binding an
`antibody inside the cell.
`
`Simple approaches to derivation of antibodies which function in cells are therefore
`necessaryif their use is to have any impacton the large number of protein targets. In
`normal circumstances, the biosynthesis of immunoglobulin occurs into the endoplasmic
`reticulum for secretion as antibody. However, when antibodies are expressed in the cell
`cytoplasm (where the redox conditions are unlike those found in the ER) folding and
`stability problems occurresulting in low expression levels and the limited half-life of
`antibody domains. These problems are most likely dueto the reducing environment of
`the cell cytoplasm (Hwang,C., Sinskey, A.J. & Lodish, H.F. (1992) Science 257: 1496-
`502), which hinders the formation ofthe intrachain disulphide bond of the VH and VL
`domains (Biocca, S., Ruberti, F., Tafani, M., Pierandrei-Amaldi, P. & Cattaneo, A. (1995)
`Biotechnology (N Y) 13: 1110-5; Martineau, P., Jones, P. & Winter, G. (1998) J MolBiol
`280: 117-127) important for the stability of the folded protein. However, some scFv have
`been shown to tolerate the absence of this bond (Proba, K., Honegger, A. & Pluckthun, A.
`(1997) JMol Biol 265: 161-72; Proba,K., Worn, A., Honegger, A. & Pluckthun, A.
`(1998) JMol Biol 275: 245-53) which presumably depends on the particular primary
`sequence ofthe antibody variable regions. No rules or consistent predictions until the
`present invention, been made about those antibodies which will tolerate the cell
`cytoplasm conditions. A further problem is the design of expression formats for
`intracellular antibodies and mucheffort has be expended on using scFv in which the VH
`and VL segments(i.e. the antibody combiningsite) are linked by a polypeptide linkerat
`the C-terminus of VH and the N-terminus of V,, (Bird, R.E., et al. (1988) Science 242:
`423-6). While this is the most successful form for intracellular expression, it has a
`drawback in the lowering ofaffinity when converting from complete antibody (e.g. from
`a monoclonal antibody) to a scFv. Thus not all monoclonal antibodies can be made as
`scFv and maintain function in cells. Finally, different scFv fragments have distinct
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`properties of solubility or propensity to aggregate when expressedin this cellular
`
`environment.
`
`The antigen binding domain of an antibody comprises two separate regions: a heavy
`chain variable domain (V}) and a light chain variable domain (VL: which can beeither
`
`Vkappa or Vlambda). The antigen bindingsite itself is formed by six polypeptide loops:
`
`three from VY domain (H1, H2 and H3) and three from Vi, domain (L1, L2 and L3). A
`diverse primary repertoire of V genes that encode the Vy and V{, domains is produced by
`the combinatorial rearrangement of gene segments. The V} geneis produced by the
`
`recombination of three gene segments, Vij, D and Jy7. In humans, there are
`
`approximately 51 functional VY segments (Cook and Tomlinson (1995) Immunol Today,
`
`16: 237), 25 functional D segments (Corbett et al. (1997) J. Mol. Biol., 268: 69) and 6
`functional Jy segments (Ravetch et al. (1981) Cell, 27: 583), depending on the haplotype.
`The VH segment encodes the region of the polypeptide chain which formsthefirst and
`
`second antigen binding loops of the Vj] domain (H1 and H2), whilst the Viz, D and Jy
`
`segments combine to form the third antigen binding loop of the VH domain (H3). The VJ,
`gene is produced by the recombination of only two gene segments, VT, and Jz.. In
`
`humans, there are approximately 40 functional Vy segments (Schible and Zachau (1993)
`Biol. Chem. Hoppe-Seyler, 374: 1001), 31 functional Vi, segments (Williamset al.
`
`(1996) J. Mol. Biol., 264: 220; Kawasaki et al. (1997) Genome Res., 7: 250), 5 functional
`Jkappa segments (Hieter et al. (1982) J. Biol. Chem., 257: 1516) and 4 functional
`
`Jlambda segments (Vasicek and Leder (1990) J. Exp. Med., 172: 609), depending on the
`
`haplotype. The VL, segment encodesthe region of the polypeptide chain which forms the
`
`first and second antigen binding loops of the Vl, domain (L1 and L2), whilst the V1, and
`
`JL segments combine to form the third antigen binding loop of the Vy, domain (L3).
`
`Antibodies selected from this primary repertoire are believed to be sufficiently diverse to
`bind almost all antigens with at least moderate affinity. High affinity antibodies are
`produced by "affinity maturation" of the rearranged genes, in which point mutations are
`generated and selected by the immune system on the basis of improved binding.
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`Analysis of the structures and sequences of antibodies has shown that five of the six
`
`antigen binding loops (H1, H2, L1, L2, L3) possess a limited number of main-chain
`
`conformations or canonical structures (Chothia and Lesk (1987) J. Mol. Biol., 196: 901;
`Chothia er al. (1989) Nature, 342: 877). The main-chain conformations are determined by
`Gi)the length of the antigen binding loop, and (ii) particular residues, or types ofresidue,
`at certain key position in the antigen binding loop and the antibody framework. Analysis
`of the loop lengths and key residues has enabledus to the predict the main-chain
`conformations of H1, H2, L1, L2 and L3 encoded by the majority of human antibody
`sequences (Chothia et a/. (1992) J. Mol. Biol., 227: 799; Tomlinson ef al. (1995) EMBO
`J., 14: 4628; Williamsef al. (1996) J. Mol. Biol., 264: 220). Although the H3 region is
`much more diverse in terms of sequence, length and structure (due to the use of D
`segments), it also forms a limited number of main-chain conformations for short loop
`lengths which depend on the length and the presenceofparticular residues, or types of
`residue, at key positions in the loop and the antibody framework (Martin et al. (1996) /
`Moi. Biol., 263: 800; Shirai et al. (1996) FEBS Letters, 399: 1.
`
`Recently, the present inventors have devised a technique for the selection of
`immunoglobulins which are stable in an intracellular environment, are correctly folded
`and are functional with respect to the selective binding of their ligand within that
`environment. This is described in WO00/54057.In this approach, the antibody-antigen
`interaction method uses antigen linked to a DNA-binding domain as a bait and the scFv
`
`linked to a transcriptional activation domain as a prey. Specific interaction ofthe two
`facilitates transcriptional activation of a selectable reporter gene. An initial in-vitro
`binding step is performed in which an antigen is assayed for bindingto a repertoire of
`immunoglobulin molecules. Those immunoglobulins which are found to bindto their
`
`ligand in vitro assays are then assayedfortheir ability to bind to a selected antigen in an
`intracellular environment, generally in a cytoplasmic environment.
`
`The present inventors found that often, a significant number of those immunoglobulins
`which bind in vitro fail to bind specifically to their ligand in vivo. Therefore, there
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`remains a need in the art for methods and procedures for predicting whether a given
`
`antibody will function within an intracellular environment.
`
`Summary of the invention
`
`The invention relates to a method for a priori
`
`identification of stable antibodies, capable
`
`of functioning as intracellular antibodies in reducing intracellular environments and for
`
`the design of libraries enriched with intracellular antibodies. In particular, the invention
`
`describes consensus sequencesfor intracellular antibodies (intrabody consensus
`sequences, ICS) that characterise the intracellular antibodies and the use of these ICS
`consensus sequences for the design and construction oflibraries that are enriched with
`
`intracellular antibodies.
`
`The presence of ICS sequences in an antibodyis diagnostic of its property of being a
`
`functional intracellular antibody, without having to undertake experimental selection with
`
`IACTorintracellular expression in other systems, with a considerable saving of
`
`experimental work. The ICS sequence can be used for the optimisation of antibodies of
`
`interest, as well as for the design and construction oflibraries that are enriched with
`
`intracellular antibodies.
`
`Thusin a first aspect the present invention provides a method ofidentifying at least one
`
`consensus sequence for an intracellular antibody (ICS) comprising the steps of:
`
`a) creating a database comprising sequencesofat least a proportion of a variable heavy
`
`chain domain and/or variable light chain domain of validated intracellular antibodies
`
`(VIDA database) and aligning the sequences of the variable heavy chain domains or
`
`variable light chain domains ofvalidated intracellular antibodies;
`
`b) determining the frequency with which a particular amino acid occurs in each of the
`
`positions of the aligned antibodies;
`
`c) selecting a frequency threshold value (LP or consensus threshold) in the range from
`
`70% to 100%;
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`d) identifying the positions of the alignment at which the frequency of a particular amino
`acid is greater than or equal to the LP value;
`
`e) identifying the most frequent aminoacid, in the positions of the alignment defined in
`
`d).
`
`According to the above aspect of the invention, advantageously the sequences ofthe
`variable heavy chain domainsor variable light chain domains of validated intracellular
`
`antibodies present in the VIDA database are aligned according to Kabat.
`
`As used herein, the term ‘database’ means any collection of data. Those skilled in the art
`
`will appreciate that there are many ways in which such data may bestored. Suitable
`
`methods include but are not limited to storage in electronic form and in paper form. Those
`skilled in the art will be aware of other suitable methodsof data storage.
`
`According to the present invention, the term ‘creating’ means generating or producing.
`
`Asherein defined, The VIDA database containsall the sequences of antibodies selected
`
`with IACT,in particular the sequences of the anti-TAU antibodies described herein. In
`
`addition, it comprises those antibodies reportedin the literature to bind specifically to one
`or more antigen/ligand/s within an intracellular environment.
`
`By ‘aligning the (amino acid) sequences’, it is meant that the (amino acid) sequences are
`
`arranged or lined up such that those amino acid residues which are the sameor similar
`
`between the sequences are apparent. Thus ‘aligning the sequences’ as herein defined
`
`permits the simple and efficient comparison of the residue similarities and differences
`
`between two or more amino acid sequences. Sequences are advantageously aligned as set
`forth in Kabat, “Sequences of Proteins of Immunological Interest”, US Department of
`Health and Human Services, using the Kabat numbering system which is known to those
`
`skilled in the art.
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`It should be appreciated that although reference is made throughout to the Kabat database,
`
`other databases of antibody sequences could be used as an alternative, or in addition to
`
`this database.
`
`Asused herein the term ‘frequency’ denotes the frequency (that is the numberof times)
`
`with which a specific amino acid occurs in each of the residue positions of the aligned
`
`sequences. The % frequency meansthe percentage of identical amino acid residues at
`
`any given residue position in the sequence andis calculated as a percentageofthe total
`
`number of amino acids at that position to be compared (that is the total number of
`
`sequences to be compared). Thus, for example, if 10 amino acid sequencesare to be
`
`compared and at amino acid residue number1, 7 out of 10 of the residues are arginines,
`
`then the percentage frequency of arginine at that position is 70%.
`
`As used herein the term ‘frequency threshold value’ or ‘LP’ value refers to a selected
`
`minimum % frequency as herein defined for each aminoacid position within the aligned
`
`sequences. Advantageously, the frequency threshold value or (LP) value selected is the
`
`same for each and every residue within the aligned sequences. The selection of a
`
`‘frequency threshold value’ creates a cut-off point at each residue position for the
`
`allocation of a consensusresidue at that position. That is, the % frequency of one or more
`
`identical amino acids at any given position is compared with the ‘frequency threshold
`
`value’ or (LP) value at that position, and if the % frequency of one or more identical or
`
`similar amino acids at any given residueposition is at least the sameas the selected
`
`frequency threshold value, then that residue will be assigned the ‘consensusresidue’ for
`
`that residue position.
`
`Those skilled in the art will appreciate that the higher the ‘frequency threshold value’ or
`
`“LP value’ selected, then the greater the ‘% frequency’ (as herein defined) is required to
`
`be for a given residue at any given residue position, for it to be assigned a ‘consensus
`
`residue’ and hence part of the consensus sequence.
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`Analysis of the antibody sequences contained in VIDA makesit possible to identify a
`subset of the amino acid residues that are conserved in human and murine intracellular
`
`antibodies. This subset of residues is designated ICS (intrabody consensus sequence), and
`
`enables us to define an ICS for the VL chain and one for the VH chain for each species
`
`(human ICS-VH, human ICS-VL, mouse ICS-VH, mouse ICS-VL). Comparative analysis
`of the ICSs of different species for the same chain madeit possible to identify the amino
`
`acids in commonand therefore an ICS-VH hxm (man mouse) and an ICS-VL hxm,Le.
`
`(minimum) general ICSs. Obviously the ICSs will be different depending on the threshold
`of homology betweenall the antibodies present in the VIDA database (absolute
`
`consensus, 90% consensus,etc.).
`
`The present invention also provides a procedure for finding the optimum ICSfor each
`reference group. The optimum ICSis obtained with an algorithm, described below, which
`changesthe threshold of homology between the antibodies of the VIDA datasetiteratively
`and defines an optimum homology threshold.
`
`Thus, in a further aspect the present invention provides a methodofidentifying at least
`
`one optimum consensus sequence for an intracellular antibody (optimum ICS) comprising
`
`the steps of:
`a) identifying different ICSs fordifferent LP values;
`b) for each of said ICSs: constructing a frequency distribution of the numberof identical
`amino acids between that particular ICS and each of the antibodies making up the VIDA
`
`database (VIDA distribution);
`c) for each of the ICSs, constructing a frequency distribution of the numberofidentical
`amino acids between that particular ICS and each of the antibodies that make up the
`
`Kabat database (Kabat distribution);
`d) defining a "distance" D between the VIDA distributions and the Kabat distribution
`corresponding to a value of LP;
`e) for each LP value, determining the value of the "distance" D between the VIDA
`distributions and the Kabat distribution correspondingto that value of LP;
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`f) identifying the optimum ICSas the ICS corresponding to the value of LP for which the
`
`calculated value of the distance D defined in d) is maximum.
`
`According to the above aspect of the invention, advantageously, the ICSs are generated
`
`according to one or more methods described herein.
`
`Analysis of the antibody sequences contained in VIDA makesit possible to identify a
`
`subset of the amino acid residues that are conserved in human and murine intracellular
`
`antibodies. This subset of residues is designated ICS (intrabody consensus sequence), and
`
`enables us to define an ICS for the VL chain and one for the VH chain for each species
`
`(human ICS-VH, human ICS-VL, mouse ICS-VH, mouse ICS-VL). Comparative analysis
`
`of the ICSsof different species for the same chain madeit possible to identify the amino
`
`acids in common and therefore an ICS-VH hxm (man mouse) and an ICS-VL hxm,i.e.
`
`(minimum) general ICSs. Obviously the ICSs will be different depending on the threshold
`
`of homology between all the antibodies present in the VIDA database (absolute
`
`consensus, 90% consensus, etc.).
`
`A procedureis described for finding the optimum ICS for each reference group. The
`
`optimum ICS is obtained with an algorithm, described below, which changes the
`
`threshold of homology between the antibodies of the VIDA datasetiteratively and defines
`
`an optimum homology threshold.
`
`Accordingly, an intracellular antibody is identified as an antibody that has an optimum
`
`ICS, defined as above, on the positions of the chain where the said ICS is defined,
`
`whereas hypotheses are not made regarding other positions, nor are constraints placed.
`
`Comparison between the ICS and the Kabat consensus sequence (for the same group)
`showsthat the ICS is highly homologous(but not completely identical) to the Kabat
`consensus, in those positions of the chain where the ICSitself is defined.
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`10
`
`The ICSis used for predicting the property of any given antibody of being a functional
`
`intracellular antibody. In particular, the analysis described predicts that a percentage of
`
`about 10% of the antibodies present in the Kabat database are intracellular antibodies.
`
`The ICS can be employed for constructing antibody libraries that are greatly enriched in
`
`functional intracellular antibodies. The libraries will preferably express scFv fragments
`
`based on ICS.
`
`According to the above aspect of the invention, the term ICS denotes ‘intracellular
`
`consensus sequence’ and is a consensus sequence for an immunoglobulin molecule
`
`capable of bindingto its ligand within an intracellular environment. ICS’s as herein
`
`described are generated using the methodsof the present invention. One skilled in the art
`
`will appreciate that the amino acid residues and their sequence comprising each ICS will
`
`depend upon the number of sequences comparedin order to generate the ICS, the nature
`
`of the sequences compared and the frequency threshold value selected.
`
`Asherein described a ‘VIDA’ denotes a ‘validated intracellularly binding antibody’. That
`
`is, it denotes an antibody which has been shownbyfunctional studies to bind specifically
`
`to one or more ligands within an intracellular environment. VIDAsas herein defined
`
`include those antibodies which have been shown by the present inventors to function
`
`within an intracellular environment as well as those antibody molecules which are
`
`reported in the literature as binding to one or moreligands specifically within an
`
`intracellular environment.
`
`Accordingly, a “VIDA database’ includes the sequences of those antibodies which have
`
`been shown by the present inventors to function within an intracellular environment as
`
`well as those antibody molecules which are reported in the literature as binding to one or
`
`more ligands specifically within an intracellular environment.
`
`As used herein the term ‘frequency distribution’ refers to a representation of the
`
`relationship between two or more characteristics. Advantageously, the representation is a
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`11
`
`graphical representation. Specifically, the term ‘VIDA distribution’ of a particular ICS
`
`refers to a representation of the relationship between the numberof identical amino acids
`
`between that particular ICS and each of the antibodies making up the VIDA databaseas
`
`herein defined.
`
`Likewise the term ‘Kabat’ distribution of a particular ICS refers to a representation,
`
`preferably a graphical representation of the numberof identical amino acids between that
`
`particular ICS and each of the antibodies which make up the Kabat database.
`
`Asdefined herein, the “D distance’ is that graphically defined distance between a given
`
`VIDA(validated intracellular antibody) distribution value and a given Kabat distribution
`
`value for a given LP value (threshold value), as herein defined.
`
`Accordingto the present invention, the term ‘optimum ICS’ refers to the ICS
`
`(intracellularly binding antibody consensus sequence) correspondingto the LP (threshold
`
`value) for which the calculated distance D as herein defined is a maximum.
`
`Advantageously, the consensus sequence is one of the consensus sequences for VH
`
`and/or VL comprising:
`
`a) for a VH consensus sequence,at least the following amino acids in the positions
`indicated according to Chothia numbering (Chothia and Lesk, (1987) J. Mol. Biol.
`
`196:910-917):
`
`8-21, C-22, S-25, G-26, M-32, W-36, P-41, L-45, E-46, D-72, Q81, L-82c, E-85, D-86,
`
`A-88, Y-90, C-92, W-103, G-104, G-106, T-107, T-110, V-111, S-112;
`
`b) for a VL consensus sequence, at least the following aminoacids in the positions
`
`indicated according to Chothia numbering:
`
`G-16, C-23, W-35, G-57, G-64, S-65, S-67, I-75, D-82, Y-86, C-88, T-102, K-103.
`
`In a preferred embodimentof this aspect of the invention, the consensus sequenceis for
`
`human VH and essentially comprises the following amino acids in the positions indicated
`
`according to Chothia numbering:
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`12
`
`Q-1, V-2, Q-3, L-4, 8-7, G-8, G-9, G10, V-12, P-14, G-15, S-17, L-18, R-19, L-20, S-21,
`
`C-22, A-24, S-25, G-26, F-27, T-28, F-29, Y-31a, M-32, W-36, R-38, Q-39, A-40, P-41,
`
`G-42, K-43, G-44, L-45, E-46, W-47, V-48, S-52, G-54, Y-58, Y-59, A-60, D-61, S-62,
`
`V-63, K-64, G-65, R-66, F-67, T-68, I-69, S-70, R-71, D-72, N-73, S-74, N-76, T-77, L-
`
`80, Q81, M-82, L-82c, R-83, A-84, E-85, D-86, I-87, A-88, Y-90, C-92, A-93, W-103,
`
`G-104, G-106, T-107, L-108, V-109, T-110, V-111, 8-112, S-113:
`
`In a further preferred embodiment, the consensus sequence represents human VL, and
`
`essentially comprises the following amino acids in the positions indicated according to
`
`Chothia numbering:
`
`T-5, P-8, G-16, I-21, C-23, W-35, Y-36, Q-37, P-40, G-41, P-44, I-48, 8-56, G-57, S-63,
`
`G-64, S-65, 8-67, G-68, L-73, T-74, I-75, D-82, A-84, Y-86, C-88, T-102, K-103.
`
`In an especially preferred embodimentof this aspect of the invention an immunoglobulin
`
`of the present invention comprises a consensus sequence which comprises the following
`
`aminoacidsin the positions indicated according to Chothia numbering:
`
`Q-1, V-2, Q-3, L-4, S-7, G-8, G-9, G10, V-12, P-14, G-15, S-17, L-18, R-19, L-20, S-21,
`
`C-22, A-24, S-25, G-26, F-27, T-28, F-29, Y-31a, M-32, W-36, R-38, Q-39, A-40, P-41,
`
`G-42, K-43, G-44, L-45, E-46, W-47, V-48, S-52, G-54, Y-58, Y-59, A-60, D-61, S-62,
`
`V-63, K-64, G-65, R-66, F-67, T-68, I-69, S-70, R-71, D-72, N-73, S-74, N-76, T-77, L-
`
`80, Q81, M-82, L-82c, R-83, A-84, E-85, D-86, T-87, A-88, Y-90, C-92, A-93, W-103,
`
`G-104, G-106, T-107, L-108, V-109, T-110, V-111, 8-112, S-113 and a variable light
`
`chain which comprises the following amino acids in the positions indicated according to
`
`Chothia numbering:
`
`T-5, P-8, G-16, I-21, C-23, W-35, Y-36, Q-37, P-40, G-41, P-44, I-48, 8-56, G-57, S-63,
`
`G-64, S-65, S-67, G-68, L-73, T-74, I-75, D-82, A-84, Y-86, C-88, T-102, K-103.
`
`In a further aspect the present invention provides an intracellularly binding
`
`immunoglobulin molecule comprising at least one variable chain which is described by
`
`at least one of the consensus sequences described in figs 11a and 11b and depicted SEQ
`
`ID no 41 and 42 respectively.
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`13
`
`Advantageously, an immunoglobulin molecule of the present invention comprises,
`(a) for the heavy chain,at least the following amino acidsin the positions indicated
`according to Chothia numbering (Chothia and Lesk, (1987) J. Mol. Biol. 196:910-917):
`§-21, C-22, S-25, G-26, M-32, W-36, P-41, L-45, E-46, D-72, Q81, L-82c, E-85, D-86,
`A-88, Y-90, C-92, W-103, G-104, G-106, T-107, T-110, V-111, S-112; and
`(b) for the light chain at least the following aminoacidsin the positions indicated
`according to Chothia numbering:
`G-16, C-23, W-35, G-57, G-64, S-65, S-67, I-75, D-82, Y-86, C-88, T-102, K-103.
`
`In a further aspect still, the present invention provides the use of an immunoglobulin
`molecule comprising at least one consensus sequence described in fig 11a and/or 11b and
`depicted SEQ 41 and SEQ 42respectively in the selective binding of a ligand within an
`
`intracellular environment.
`
`Asherein defined, the term ‘selective binding’ (of a ligand within an intracellular
`environment) meansthat the interaction between the immunoglobulin andthe ligand are
`specific, that is, in the event that a number of molecules are presented to the
`immunoglobulin, the latter will only bind to one or a few of those molecules presented.
`Advantageously, the immunoglobulin ligand interaction will be of high affinity. The
`interaction between immunoglobulin and ligand will be mediated by non-covalent
`interactions such as hydrogen bonding and Van der Waals interactions. Generally, the
`interaction will occur in the cleft between the heavy and the light chains of the
`
`immunoglobulin.
`
`In a further aspect still, the present invention provides a method for predicting whether an
`antibody is a functioning intracellular antibody comprising the stepsof:
`a) aligning the sequence of the antibody with the sequences of the antibodies ofthe
`
`reference VIDA database;
`b) aligning the sequence of the antibody with the optimum ICS sequenceofthe reference
`
`VIDA database;
`
`

`

`WO 03/014960
`
`PCT/GB02/03512
`
`14
`
`c) constructing VIDA and KABATdistributions corresponding to the optimum value of
`
`LP (frequency threshold value);
`
`d) determining the corresponding distance D;
`
`e) determining the identity number N between the sequence of the antibody and the ICS
`
`reference sequence;
`
`f) calculating the difference between the mean value of the VIDAdistribution and the
`
`product between D andthe standard deviation of the VIDA distribution, obtaining the
`
`parameter Sintra’
`
`g) if the identity numberN is greater than or equal to Sintra, identifying the antibody as
`
`intracellular antibody.
`
`In a further aspect the present invention provides a method for conferring upon an
`
`immunoglobulin molecule the ability to function within an intracellular environment,
`
`comprising the steps of:
`
`a)
`
`identifying the optimum ICS reference sequence
`
`b) optionally, modifying, by site-specific mutagenesis, the amino acid residues that are
`
`located in the positions defined by the optimum ICS,or a subset of these residues, in such
`
`a way that they are those identified by the optimum ICS.
`
`According to this aspect of the invention, advantageously the aligning step is performed
`
`as described herein. Advantageously, the ICS generation is performed using the methods
`
`herein described.
`
`The present inventors have foundthat by performing a functional binding assay (that is by
`
`performing a yeast two-hybrid based [ACTassay) to all of the antibodies proposed to be
`
`included in a database, and only using those antibodies which are found using to IACTto
`
`b