FIELD OF THE INVENTION

The present invention relates to inhibitors of kinases. In particular, the invention relates to methods of screening for non-competitive inhibitors of the JAK family of kinases.

BACKGROUND OF THE INVENTION

Maintenance of the haematopoietic system is largely controlled by the secretion of cytokines. Cytokine exposure initiates an intracellular signalling cascade that is driven by activation of a family of receptor-bound tyrosine kinases known as the JAKs (Janus Kinases) (Ihle et al., 1 96; Wilks, 2008). Under physiological conditions, JAK activation and signalling is tightly regulated. However, in certain disease situations JAK is mutated in such a way that renders it constitutively active. In such cases, JAK initiates signaling cascades in the absence of cytokine stimulation, leading to aberrant proliferation and disease.

There are four JAK tyrosine kinases encoded in the human genome, JA ^' Kl -3 and

TYK2. Each are large molecules consisting of four distinct domains (Shuai et al, 1993; Wilks and Harpur, 1994). The PERM domain lies at the N-terminus and binds to the appropriate membrane-bound receptor whilst the kinase (catalytic) domain is at the C- terminus. Between these are a non-canonical SH2 domain and a pseudokinase domain, which is the most distinctive feature of the JAK family. Although the pseudokinase domain is catalytically inactive, it regulates the activity of the catalytic domain. Like most tyrosine kinases, the catalytic domain of all four JAKs contains an "activation loop" that blocks the catalytic cleft. Autophosphorylation of this loop, in trans, causes its translocation away from the catalytic site and allows substrate access, activating the kinase.

Examples of diseases resulting from dysregulated JAK activation include three related myeloproliferative disorders, Polycythemia Vera, Essential Thrombocythemia and Primary Myelofibrosis. The major causative genotype in these three disorders arises from a single point mutation in JAK2 (JAK2v6m) (James et al., 2005; Levine et al., 2005). This mutation results in constitutively active JAK2 and cytokine- independent activation of downstream signalling pathways, resulting in over- proliferation of myeloid cells and disease.

Another example is oncogenic fusion of JAK2 via chromosomal translocation. Expression of the resultant allele results in the production of a chimaeric protein in which the kinase domain of JAK2 is fused with a portion of another protein, usually a transcription factor. Such a fusion can result in oncogenic transformation and disease, for example the TEL-JAK2 chimaeric protein resulting from the t(9; 12)(p24;pl3) translocation, has been intensively studied due to its role in childhood T- and B-cell acute lymphoblastic leukemia (Lacronique, 2000).

Inhibitors of JAK are considered viable therapeutics for the treatment ^' of diseases resulting from aberrant expression and activation of JAK, including those diseases described above. All currently available JAK inhibitors are ATP analogues or competitors and bind to the active site of the enzyme, outcompeting ATP and rendering the enzyme inactive. This approach has two major drawbacks: (1) the high concentration of ATP in the cell competes with inhibitor binding and leads to reduced efficacy in vivo and (2) the ATP binding site of tyrosine kinases are all structurally similar and hence specificity is difficult to achieve. Thus, there remains a need for inhibitors of JAK kinases.

SUMMARY OF THE INVENTION .

The present inventors demonstrate that SOCS1 and SOCS3 inhibit the activity of Janus Kinases (JAK) . and are non-competitive for ATP and JAK substrate. Thus, in light of this finding, it is possible to screen for JAK inhibitors that are non-competitive for ATP and/or JAK substrate.

Accordingly, in one aspect the present inventio provides a method of identifying a non-competitive inhibitor of a Janus Kinase (JAK), the method comprising:

i) determining whether the candidate compound is able to inhibit the activity of the JAK, and

ii) determining if the candidate compound is a non-competitive inhibitor of the

JAK by testing whether the candidate compound inhibits the binding of ATP and/or ADP and the JAK substrate to the JAK,

wherein the non-competitive inhibitor is identified by the ability of the candidate compound to inhibit JAK and not to inhibit the binding of ATP and/or ADP and the JAK substrate to the JAK.

In another aspect, the present invention provides a method of identifying a noncompetitive inhibitor of a Janus Kinase (JAK), the method comprising:

determining if the candidate compound is a non-competitive inhibitor of the JAK by testing whether the candidate compound is able to inhibit the activity of the JAK in the presence of ATP and JAK substrate, wherein the non-competitive inhibitor is identified by the ability of the candidate compound to inhibit JAK activity in the presence of ATP and JAK substrate.

^■ In an embodiment, the method further comprises determining whether the candidate compound inhibits binding of ATP and/or ADP and/or the JAK substrate to the JAK, wherein the non-competitive inhibitor is identified by the ability of the candidate compound to inhibit JAK activity and not to inhibit the binding of ATP and/or ADP and/or the JAK substrate to the JAK.

In one embodiment, the ATP and/or ADP and/or JAK substrate is at a concentration of about 0.1 mM to about 2 mM.

In another embodiment, the concentration of ATP and/or ADP and/or JAK substrate used in the methods of the invention is about the cellular concentration of ATP and/or ADP and/or JAK substrate.

Any suitable JAK substrate as known in the art may be used in the methods of the invention. In one embodiment, the JAK substrate is gpl 30 or a JAK binding fragment thereof, or a STAT polypeptide or a JAK binding fragment thereof. In one particular embodiment, the gpl30 or a JAK binding fragment thereof is a g l 30 cytoplasmic domain, such as provided in SEQ ID NO:28. While the STAT polypeptide or a JAK binding fragment thereof may be any that acts as a JAK substrate, in one particular embodiment, the STAT polypeptide or JAK binding fragment thereof is a STAT 5 derived peptide such as, for example, a peptide comprising the amino acid sequence as provided in SEQ ID NO.27.

In one embodiment, the method comprises:

a) contacting the candidate compound with the JAK in the presence of a JAK substrate and a first concentration of ATP,

b) contacting the candidate compound with the JAK in the presence of the JAK substrate and a second concentration of ATP, and

c) determining if the candidate compound is able to inhibit the activity of the JAK in the first and second concentration of ATP.

In another embodiment, the ability of the compound to inhibit the activity of JAK is substantially the same in the first and second concentration of ATP.

The person skilled in the art can determine suitable first and second concentrations of ATP for use in the methods of the invention. For example, the first concentration of ATP may be about 0.1 mM and the second, concentration of ATP may be about 1 m or alternatively about 2 mM. It may also be desirable to maintain the JAK substrate at a constant concentration, for example at about 1.6 mM.

In yet another embodiment, the method comprises: a) contacting the candidate compound with the JAK in the presence of ATP . and a first concentration of JAK substrate,

b) contacting the candidate compound with the JAK in the presence of ATP and a second concentration of the JAK substrate, and

c) determining if the candidate compound is able to inhibit the activity of the

JAK in the first and second concentration of JAK substrate.

In one particular embodiment, the ability of the compound to inhibit the activity of JAK is substantially the same in the first and second concentration of JAK substrate. Suitable concentrations of JAK substrate cart be determined by the skilled person and include, for example, a first concentration of JAK substrate of about 0.1 mM and asecond concentration of JAK substrate of about 1 mM. The concentration of ATP may be kept constant, for example at about 2 mM.

In another aspect, the present invention provides a method of identifying a noncompetitive inhibitor of a Janus Kinase (JAK), the method comprising:

i) determining if the candidate compound inhibits the ability of a SOCS polypeptide or an active fragment thereof to bind the JAK, and

ii) determining if the candidate compound is a non-competitive inhibitor of the JAK by testing whether the candidate compound is able to inhibit the activity of JAK in the presence of ATP,

wherein the non-competitive inhibitor is identified by the ability of the candidate compound to inhibit binding of the SOCS polypeptide or an active fragment thereof to JAK, and to inhibit JAK activity in the presence of ATP.

In one embodiment, step i) comprises contacting the SOCS polypeptide or active fragment thereof with the JAK in the presence of ATP and/or ADP and/or a JAK substrate.

In another embodiment, the method further comprises determining whether the candidate compound inhibits binding of ATP and/or ADP to the JAK.

In another aspect, the present invention provides a method of identifying a noncompetitive inhibitor of a Janus Kinase (JAK), the method comprising:

i) determining if the candidate compound inhibits the ability of a SOCS polypeptide or an active fragment thereof to bind the JAK, and

ii) determining if the candidate compound is a non-competitive inhibitor of the JAK by testing whether the candidate compound inhibits the binding of ATP and/or ADP to the JAK, wherein the non-competitive inhibitor is identified by the ability of the candidate compound to inhibit binding of the SOCS polypeptide or an active fragment thereof to JAK and not to inhibit the binding of ATP and/or ADP to the JAK.

In another aspect, the present invention provides a method of identifying a non- competitive inhibitor of a Janus Kinase (JAK), the method comprising:

i) obtaining a candidate compound known to bind the JAK, inhibit the ability of a SOCS polypeptide or an active fragment thereof to bind the JAK, and/or inhibit the activity of the JAK, and

ii) determining if the candidate compound is a non-competitive inhibitor of the JAK by testing whether the candidate compound is able to inhibit the activity of JAK in the presence of ATP and/or by testing whether the candidate compound inhibits the binding of ATP and/or ADP to the JAK,

wherein the non-competitive inhibitor is identified by the ability of the candidate compound to inhibit binding of the SOCS polypeptide or an active fragment thereof to bind JAK, and to inhibit JAK activity in the presence of ATP and/or to inhibit the binding of ATP and/or ADP to the JAK.

In one embodiment, step ii) of the methods of the invention comprises contacting the candidate compound with the JAK in the presence of a JAK substrate and ATP and measuring JAK activity.

in another embodiment, the method comprises:

a) contacting the candidate compound with the JAK in the presence of a JAK substrate and a first concentration of ATP,

b) contacting the candidate compound with the JAK in the presence of the JAK substrate and a second concentration of ATP, and

c) determining if the candidate compound is able to inhibit the activity of the

JAK in the first and second concentration of ATP.

In one embodiment, the ability of the compound to inhibit the activity of JAK is substantially the same in the first and second concentration of ATP.

While any suitable JAK substrate may be used in the methods of the invention, in one embodiment the JAK substrate is selected from gpl 30 or JAK binding fragment thereof, and/or a STAT polypeptide or JAK binding fragment thereof.

In an embodiment, the JAK is selected from JAKl , JAK2, and/or TYK2. In one particular embodiment, the JAKl comprises an amino acid sequence selected from SEQ ID NO: l, SEQ ID NO:5, SEQ ID NO:9 or SEQ ID NO: 13, the JAK2 comprises an amino acid sequence selected from SEQ ID NO:2, SEQ ID NO:6, SEQ ID NO: 10, or SEQ ID NO: 14, and/or the TYK2 comprises an amino acid sequence selected from SEQ ID NO:4, SEQ ID N0:8, SEQ ID NO: 12 or SEQ ID NO: 16. In a further embodiment, JAK2 is JAK2 _V 6i 7i ^; . In a particularly preferred embodiment, the JAK comprises the sequence GXM (where X= any amino acid), more preferably GQM, at the C-terminal end of the JAK insertion loop.

In another embodiment, the SOCS polypeptide is a full length SOCS protein such as SOCS1, SOCS3, or an active fragment thereof such as SOCS3 _A PEST(22-225) or SOCS3APEST(22-185), or a chimer thereof such as SOCS 1-3. In one particular embodiment, the SOCS polypeptide or active fragment thereof comprises an amino acid sequence selected from SEQ ID NOs: 18-24.

· The skilled person will appreciate that a full-length JAK polypeptide may be used in the methods of the invention, or alternatively the J AK may be a fragment of the JAK comprising the JAK catalytic domain. By way of example, the fragment of the JAK comprising the JAK catalytic domain may comprise the amino acid sequence as provided in SEQ ID NO:6 or SEQ ID NO: 14.

In one embodiment, the candidate compound is selected from a small molecule, a peptide or mimetic thereof, and/or an antibody or JAK binding fragment thereof.

In another embodiment, the candidate compound identified in the method of the invention is a selective JAK inhibitor. In one particular embodiment, the candidate compound is a selective JAK2 inhibitor.

In yet another aspect, the present invention provides a computer-assisted method of identifying a non-competitive inhibitor of a Janus Kinase (JAK), the method comprising:

i) obtaining atomic coordinates representing the three dimensional structure of JAK in complex with ATP and/or ADP, and a JA substrate, or a, subset of atomic coordinates of thereof, wherein the subset of atomic coordinates at least represents the site to which ATP and/or ADP, and the JAK substrate, binds to JAK,

ii) docking the structure of a candidate compound to a structure defined by the atomic coordinates, or subset thereof, representing the three dimensional structure, wherein the compound should not interfere with the binding of ATP and/or ADP, and the JAK substrate, to the JAK, and

tit) identifying a candidate non-competitive inhibitor which may inhibit JAK activity and which should not interfere with the binding of ATP and/or ADP, and the JAK substrate, to the JAK.

In another aspect, the present invention provides a computer-assisted method of identifying a non-competitive inhibitor of a Janus Kinase (JAK), the method comprising: i) obtaining atomic coordinates representing the three dimensional structure of JAK in complex and ATP and/or ADP, and a JAK substrate, or a subset of atomic coordinates of thereof, wherein the subset of atomic coordinates at least represents the site to which to which ATP and/or ADP, and the JAK substrate, binds to JAK, entered into the computer thereby generating a criteria data set;

ii) comparing, using the computer, the criteria data set to a computer database of chemical structures; «

iii) selecting from the database, using computer methods, chemical structures _, which are complementary to a region of the criteria data set to. identify a candidate non- competitive inhibitor which may inhibit JAK activity and which should not interfere with the binding of ATP and/or ADP, and ^' the JAK substrate, to the JAK and optionally,

iv) outputting, to an output device, the selected chemical structures which are complementary to a region of the criteria data set.

In one particular embodiment, the subset of the atomic coordinates at least represents the JAK catalytic domain.

In yet another aspect, the present invention further provides a computer-assisted method of identifying a non-competitive inhibitor of a Janus Kinase (JAK), the method comprising;

i) obtaining atomic coordinates representing the three dimensional structure of

JAK in complex with a SOCS polypeptide or active fragment thereof, and ATP and/or ADP, or a subset of atomic coordinates of thereof, wherein the subset of atomic coordinates at least represents the site to which the SOCS polypeptide or active fragment thereof binds to JAK,

ii) docking the structure of a candidate compound to a structure defined by the atomic coordinates, or subset thereof, representing the three dimensional structure, and iii) identifying a candidate non-competitive inhibitor which may inhibit binding of the SOCS polypeptide or an active fragment thereof to JAK, and which may inhibit JAK activity in the presence of ATP.

In another aspect, the present invention provides a computer-assisted method of identifying a non-competitive inhibitor of a Janus Kinase (JAK), the method • comprising:

i) obtaining atomic coordinates representing the three dimensional structure of JAK in complex with a SOCS polypeptide or active fragment thereof, and ATP and/or ADP, or a subset of atomic coordinates of thereof, wherein the subset of atomic coordinates at least represents the site to which the SOCS polypeptide or active fragment thereof binds to JAK, entered into the computer thereby generating a criteria data set;

ii) comparing, using the computer, the criteria data set to a computer database of chemical structures;

iii) selecting from the database, using computer methods, chemical structures which are complementary to a region of the criteria data set to identify a candidate noncompetitive inhibitor which may inhibit binding of the SOCS polypeptide or an active fragment thereof to JAK, and which may inhibit JAK activity in the presence of ATP; and optionally,

iv) outputting, to an output device, the selected chemical structures which are complementary to a region of the criteria data set.

In a particularly preferred embodiment of the computer-assisted methods of the invention, the JAK comprises the sequence GXM (where X= any amino acid), more preferably GQM, at the C-terminal end of the JAK insertion loop.

In an embodiment, the atomic coordinates are obtained using nuclear magnetic resonance (NMR).

In one embodiment, the JAK comprises at least the JAK catalytic domain.

In one embodiment, the complex further comprises a JAK substrate.

In another embodiment, the method further comprises synthesising or obtaining an identified candidate compound and performing the method of the invention to identify a non-competitive inhibitor.

In another embodiment, the ATP and/or ADP in complex with the JAK is a non- hydrolyzable ATP or ADP analogue. For example, the non-hydrolyzable ATP analogue may be ATP-yS.

In another aspect, the present invention provides a compound identified by the method of the invention.

In one aspect, the present invention provides use of the compound of the invention in the treatment or prevention of disease. The disease that is treated or prevented may be any disease resulting from dysregulated JAK activity, for example the disease may be cancer and/or a myeloproliferative disorder,

In one embodiment, the myeloproliferative disorder that is treated or prevented is selected from polycythemia vera, essential thrombocytosis, primary or idiopathic myelofibrosis, and chronic myelogenous leukemia.

In yet another aspect, the present invention provides use of the compound of the invention in the manufacture of a medicament for the treatment or prevention of disease. In one embodiment, the disease is cancer or a myeloproliferative disorder. As will be apparent, preferred features and characteristics of one aspect of the invention are applicable to many other aspects of the invention.

Throughout this specification the word "comprise", or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated element, integer or step, or group of elements, integers or steps, but not the exclusion of any other element, integer or step, or group of elements, integers or steps.

The invention is hereinafter described by way of the following non-limiting Examples and with reference to the accompanying figures.

BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS

Figure 1. Top, Autoradiograph and coomassie stained SDS-PAGE gel of a kinase reaction in the presence of serial 10-fold dilutions (10 μΜ-10 nM) of various SOCS/elonginBC complexes. Both SOCS3 and a SOCS l -3 chimaera are effective inhibitors of substrate (gpl 30 cytoplasmic domain) phosphorylation whereas SOCS2 and SOCS4 are not. Bottom left, Quantitation of inhibition by SOCS3 and SOCSl-3. Bottom right, autoradiograph of a kinase assay in the presence of high concentrations of JAK2. Neither SOCS3 nor SOCSl-3, inhibit autophosphorylation

Figure 2. Autoradiograph and coomassie stained SDS-PAGE gel of kinase inhibition reactions in the presence of serial 10-fold dilutions (5 μΜ-200 nM) of various SOCS/elonginBC complexes. A synthetic substrate that consists of the activation loop of JAK2 fused to GST was used as substrate. Both SOCS3 and a SOCS l-3 chimaera are effective inhibitors of substrate phosphorylation whereas SOCS2, SOCS4, SOCS4-3 and SOCS5-3 are not.

Figure 3. Autoradiograph and coomassie stained SDS-PAGE gel of kinase inhibition reactions in the presence of 1 μΜ SOCS3/elonginBC or 1 μΜ SOCS1 - 3/elonginBC or neither. Four different substrates were tested: A synthetic substrate -that consists of the activation loop of JAK2 fused to GST (GST- J), GST alone (GST), the gpl30 cytoplasmic domain (gpl30), or β-catenin (β - cat). Both SOCS3 and a SOCSl -3 chimaera are effective inhibitors of all substrate phosphorylation whereas SOCS2, SOCS4, SOCS4-3 and SOCS5-3 are not.

Figure 4. Autoradiograph of kinase inhibition reactions in the presence of serial

5-fold dilutions (5 μΜ - 200 nM, OnM) of various SOCS mutants. The gpl 30 cytoplasmic domain was used as substrate. Both SOCS3 and a SOCSl -3 chimaera are effective inhibitors of substrate phosphorylation whereas mutations in both the KIR (F25A) and the SH2 domain (R71K) render SOCS3 inactive. As expected, removing the Kinase Inhibtory Region entirely (KIR-) also abrogate the inliibitory effect of SOCS3.

Figure 5. Kinase inhibition reactions in the presence of full-length SOCS3 (closed circles) or the SOCS3 Kinase Inhibitory Region only (open circles) or water (triangles). The SOCS3 KIR alone is unable to inhibit JAK2, even at high concentrations. The assay system used the STATS derived peptide as substrate (1 mM) and 2 mM ATP.

Figure 6. Michaelis-Menten (upper panels) and Lineweaver-Burk (lower panels) analysis of SOCS3 inhibition. ATP (right) and peptide substrate (left) titrations show that SOCS3 alters the V _max but not the K _m of catalysis, indicating noncompetitive inhibition. Inhibition experiments were performed using 1 OnM JAK and, 0-8μΜ SOCS3 with varying concentrations of ATP and substrate for 15 minutes at 25°C. When the STAT peptide was varied (2.7, 1.3, 0.7, 0.3, 0.17mM) ATP was held constant at 2mM. When ATP was varied (1 , 0.5, 0.25, 0.125, 0.06mM) the STAT peptide was held constant at 1.6mM. Linweaver-Burke curves that intersect on the ordinate axis indicates non-competitive inhibition. Error bars represent +/- range from two experiments.

Figure 7. The SOCS3 IC50 is independent of substrate concentration. JAK2 inhibition experiments were performed in the presence of SOCS3 (top panels, 2 independent experiments) or in the presence of ATP competitive inhibitors ADP and CMP-6 (lower panels). The IC50 values of SOCS3 are unchanged, within experimental error, despite a 40-fold difference in substrate concentration whereas the IC50 values for ADP and CMP-6 decreased with decreasing ATP concentration but not with decreasing STAT peptide concentration as expected. Error bars represent +/- range from two experiments. Data is normalized to no-inhibitor controls.

Figure 8. SOCS3 inhibits JAK1 , JAK2 and TYK2 but not JAK 3 due to a three residue motif in the JAK insertion loop. A, Kinase inhibition assays were performed using the kinase domain of all four JAKs expressed as His6 fusions in baculovirus and a STATSb peptide (2mM) as substrate. Only JAK3j _H i could not be inhibited by SOCS3. B, Seven JAK2-JAK3 chimaeric mutants were tested in inhibition assays. Mutants of the GQM motif were not inhibited by SOCS3 whilst all other mutations had no effect. C, Point mutagenesis around the GQM motif shows that Glyl 071 and Metl073 are absolutely required for SOCS3 inhibition whereas Q 1072 and 11074 have a minor effect. D, Sequence alignment of the GQM region of all JAK kinases and the Kinase Inhibitory Region of SOCS3 and SOCSl . Highly conserved residues are shown boxed in grey. The GQM motif is shown. The JAK insertion loop is indicated by an arrow E, The structure of JAK2 is shown as a ribbon diagram. The GQM motif is solvent exposed. Error bars in (A) and (C) represent +/- range from two experiments. Data were normalized to no-inhibitor controls. *Zebrafish JAK2b is grouped with TYK2 in this figure.

Figure 9. Mutational analysis of SOCS3 inhibition of JAK2. (A) IC ₅₀ curves are shown for various fragments and mutants of SOCS3. Inhibition experiments contained lOnM enzyme (JA 2), 1.6mM substrate (STAT peptide), ImM ATP and serial 2-fold dilutions of inhibitor (SOCS). SOCS3 constructs that contained mutations in the KIR (F25A), SH2 domain (R71 ) or that lacked the first eight residues of the KIR (Δ22-29) did not inhibit J AK2. The addition of a high affinity ligand for the SH2 domain (gp! 3075o-764-pY) did not affect inhibition. (B) SOCSl -3 is a. more effective inhibitor of JAK2 than SOCS3. IC50 curves are shown SOCS3, both with and without elonginBC, and SOCSl-3. These experiments were performed under identical conditions ^" to those above with the exception that an 1 1 -point titration of 2.5-fold dilutions of SOCS was performed. Error bars represent +/- range of two experiments. Data is normalized to no-inhibitor controls.

Figure 10. NMR analysis identifies the surface of SOCS3 that interacts with ^' JAK2. A, Ribbon diagram of the structure of SOCS3 ₂₂ -i 85 (PDB ID 2HMH) with important secondary structural motifs indicated. The dashed line indicates the first seven residues of the KI which are unstructured in the absence of JAK. B, Ή- ^{, 5} Ν HSQC and 'H- ¹³ C HSQC analysis of the SOCS3-JAK2 _JH i interaction are shown in upper and lower panels respectively. The apo-SOCS spectra are in red and the SOCS3- JA.K2 _j Hi spectrum are shown in black. Peaks that show a significant chemical shift perturbation in the presence of JAK2 _JH i are labeled. Spectra were recorded at 600MHz in 20mM MES, ImM DTT at 37°C. Samples contained 250μΜ SOCS3 ₂₂ .| _S5 APEST in the presence and absence of 500μΜ JAK2 _JH i. C, Surface diagram of SOCS3 with residues whose resonances shift significantly in the presence of JAK2JHI highlighted. The orientation of SOCS3 on the left is identical to that in (A).

Figure I I. SOCS3 has a slight activating effect on. JAK2 _{J H} i ATPase activity. ATPase assays were performed in the presence of ΙΟΟμΜ ATP, 0.45μΜ JAK2 _JH i and various concentrations of SOCS constructs. Results were analysed by TLC (lower panels) and quantified by phosphorimaging (upper panel). Experiments were performed for 15 minutes at 25°C in the presence of serial two-fold dilutions of SOCS1-3/BC (lower left panel), SOCS3 _22-225 APEST/BC (lower middle panel) or SOCS3(.- ₂ 5 _A /BC (lower right panel). In the absence of SOCS, JAK2 _JH i hydrolyses ATP at the rate of 0.05 s ^"1 under these conditions (right hand lane of each panel) whilst this rate increases to 0.09 s ^"1 in the presence of 15μΜ SOCS3. In the presence of > Ι ΟμΜ SOCS 1 -3/BC or SOCS3/BC this activity is enhanced nearly 2-fold (lanes I and 9 respectively) whereas there was no change in the presence of high concentrations of a SOCS3 mutant that does not bind JAK (SOCS3 _F 2s _A /BC). The activity of SOCS l -3 and SOCS3 titrated with increasing concentration with apparent EC50 values of 1.2μΜ and 2μΜ respectively. Error bars represent +/- range from two experiments. Data is normalized to JAK only controls set to 100%.

Figure 12. SOCS3 catalyses JAK and Receptor ubiquitination. ^' (A) SOCS3/Cullin5/Rbx2 E3 ligase catalyses the ubiquitination of GST-JAK2JHI in the presence of El , E2 (UbcH5a), ATP and ubiquitin (left panel) but does not ubiquitinate GST alone (centre panel). Replacing SOCS3 ₂ 2-22s with the SOCS box domain only does not catalyze JAK ubiquitination (right panel). All panels are Coomassie stained SDS-PAGE gels. *Cullin5 ubiquitination (B) SOCS3 catalyses phosphorylated gpl 30 ubiquitination. Results are visualized by Coomassie staining (upper) and autoradiography (lower). (C) Ubiquitination of JAK requires residues in the KIR and SH2 domain.

Figure 13. A model of the SOCS3 mechanism. The pseudokinase and SH2 domains of JAK are omitted in this schematic for clarity.

KEY TO THE SEQUENCE LISTING

^• SEQ ID NOT - amino acid sequence of human JAK1.

SEQ ID NO:2 - amino acid sequence of human J AK2.

SEQ ID NO:3 - amino acid sequence of human J AK3.

SEQ ID NO:4 - amino acid sequence of human ΊΎΚ2.

SEQ ID NO:5 - amino acid sequence of human JAK1 catalytic domain.

SEQ ID NO:6 - amino acid sequence of human JAK2 catalytic domain.

SEQ ID NO:7 - amino acid sequence of human JAK3 catalytic domain.

SEQ ID NO:8 - amino acid sequence of human TYK2 catalytic domain.

SEQ ID NO:9 - amino acid sequence of murine JAKl . -

SEQ ID NO: 10 - amino acid sequence of murine JAK2.

SEQ ID NO: l 1 - amino acid sequence of murine JAK3.

SEQ ID NO:12 - amino acid sequence of murine ΤΎΚ2.

SEQ ID NO: 13 - amino acid sequence of murine JAKl catalytic domain.

SEQ ID NO: 14 - amino acid sequence of murine JAK2 catalytic domain. SEQ ID N0:15 - amino acid sequence of murine JAK3 catalytic domain.

SEQ ID NO: 16 ^• amino acid sequence of murine TYK2 catalytic domain.

SEQ ID NO: 17 ^• amino acid sequence of murine GST-JAK2 catalytic domain.

SEQ ID NO: 18 - amino acid sequence of human SOCS 1.

SEQ ID NO: 19 ^• amino acid sequence of human SOGS3.

SEQ ID NO:20 amino acid sequence of murine SOCS1.

SEQ ID NO:21 ^■ amino acid sequence of murine SOCS3.

SEQ ID NO:22 amino acid sequence of murine SOCS3 _A PE _S T(22-225).

SEQ ID NO:23 amino acid sequence of murine SOCS3 _APES T(22-185).

SEQ ID NO:24 amino acid sequence of murine SOCS 1 -3.

SEQ ID NO:25 amino acid sequence of murine ElonginB.

SEQ ID NO:26 amino acid sequence of murine ElonginC.

SEQ ID NO:27 STAT5 derived peptide.

SEQ ID NO:28 amino acid sequence of gpl30 cytoplasmic domain.

SEQ ID NO 29 amino acid sequence of J AK loop of human JAK1.

SEQ ID NO 30 amino acid sequence of JAK loop of murine JAK 1.

SEQ ID NO 31 amino acid sequence of JAK loop of chicken JAK I .

SEQ ID NO 32 amino acid sequence of JAK loop of frog JAKl .

SEQ ID NO 33 amino acid sequence of JAK loop of zebrafish JA l .

SEQ ID NO 34 amino acid sequence of JAK loop of human JAK2.

SEQ ID NO 35 amino acid sequence of JAK loop of murine JAK2.

SEQ ID NO 36 amino acid sequence of JAK loop of Xenopus JAK2.

SEQ ID NO 37 amino acid sequence of JAK loop of zebrafish JAK2.

SEQ ID NO 38 amino acid sequence of JAK loop of chicken JAK2.

SEQ ID NO 39 amino acid sequence of JAK loop of human JAK3.

SEQ ID NO 40 amino acid sequence of JAK loop of murine JAK3.

SEQ ID NO 41 amino acid sequence of JAK loop of Xenopus JAK3.

SEQ ID NO 42 amino acid sequence of JAK loop of chicken JAK3.

SEQ ID NO 43 amino acid sequence of JAK loop of zebrafish JAK3.

SEQ ID NO 44 amino acid sequence of JAK loop of human TYK2.

SEQ ID NO 45 amino acid sequence of JAK loop of murine TYK2.

SEQ ID NO 46 amino acid sequence of J AK loop of chicken TYK2.

SEQ ID NO 47 amino acid sequence of JAK loop of Xenopus TYK2.

SEQ ID NO 48 amino acid sequence of JAK loop of zebrafish TYK2.

SEQ ID NO 49 amino acid sequence of J AK loop of Drosophila JAK.

SEQ ID NO 50 amino acid sequence of JAK loop of seaquirt JAK. SEQ ID NO 51 --- amino acid sequence of Kinase Inhibitory Region of human SOCS3. SEQ ID NO 52 - amino acid sequence of Kinase Inhibitory Region of murine SOCS3. SEQ ID NO 53 - amino acid sequence of Kinase inhibitory Region of chicken SOCS3. SEQ ID NO 54 - amino acid sequence of Kinase Inhibitory Region of Xenopus SOCS3.

SEQ ID NO 55 - amino acid sequence of Kinase Inhibitory Region of zebrafish SOCS3.

SEQ ID NO 56 - amino acid sequence of Kinase Inhibitory Region of human SOCS1. SEQ ID NO 57 - amino acid sequence of Kinase Inhibitory Region of murine SOCS1. SEQ ID NO 58 - amino acid sequence of Kinase Inhibitory Region of chicken SOCS 1. SEQ ID NO 59 - amino acid sequence of Kinase Inhibitory Region of Xenopus SOCS1.

SEQ ID NO 60 - amino acid sequence of Kinase Inhibitory Region of zebrafish SOCS1.

SEQ ID NO 61 - amino acid sequence of Kinase Inhibitory Region of seaquirt SOCSl/2/3.

DETAILED DESCRIPTION

General techniques and definitions

Unless specifically defined otherwise, all technical and scientific terms used herein shall be taken to have the same meaning as commonly understood by one of ordinary skill in the art (e.g., in cell biology, molecular genetics, immunology, immunohistochemistryj protein chemistry and biochemistry).

Unless otherwise indicated, the recombinant protein, cell culture, and immunological techniques utilized in the present invention are standard procedures, well known to those skilled in the art. Such techniques are described and explained throughout the literature in sources such as, J. Perbal, A Practical Guide to Molecular Cloning, John Wiley and Sons (1984), J. Sambrook et al., Molecular Cloning: A Laboratory Manual, 3 ^rd edn, Cold Spring Harbour Laboratory Press (2001), T.A. Brown (editor), Essential Molecular Biology: A Practical Approach, Volumes 1 and 2, IRL Press (1991), D.M. Glover and B.D. Hames (editors), DNA Cloning: A Practical Approach, Volumes 1-4, IRL Press (1995 and 1996), and F.M. Ausubel et al. (editors), Current Protocols in Molecular Biology, Greene, Pub. Associates and Wile - Interscience (1988, including all updates until present), Ed Harlow and David Lane (editors) Antibodies: A Laboratory Manual, Cold Spring Harbour Laboratory, (1988), and J.E. Coligan et al. (editors) Current Protocols in Immunology, John Wiley & Sons (including all updates until present).

"JAK" as used herein refers to a polypeptide belonging to the Janus Kinase family of tyrosine kinases and includes reference to a fragment of a JAK comprising a catalytic domain of the polypeptide. Members of the Janus Kinase family of tyrosine kinases include human JAK1 (SEQ ID NO: l), JAK2 (SEQ ID NO:2), JAK3 (SEQ ID NO:3) and TYK2 (SEQ ID NO:4), as well as orthologues thereof, for example murine JAK1 (SEQ ID NO:9), JAK2 (SEQ ID NO: 10), JAK3 (SEQ ID NO: l l) and TYK2 (SEQ ID NO: 12). Also included are JAK mutants, such as the V617F mutant of human JAK2. In a particularly preferred embodiment, the JAK comprises the sequence GXM (where X= any amino acid), more preferably GQM, at the C-terminal end of the JAK insertion loop (see Figure 8D for examples).

JAK "activity" refers to the phosphorylation of a substrate by a JAK.

As used herein, a "JAK substrate" refers to a polypeptide substrate that can be phosphorylated by a JAK exhibiting JAK activity and includes autophosphorylation of the JAK itself.

As used herein "SOCS polypeptide" refers to polypeptides that are members of the family of proteins known as the Suppressor of Cytokine Signalling proteins (Starr et al. _s 1997, Yoshimura et al, 1995, and Naka et al., 1997), and which comprise a Kinase Inhibitory Region (KIR), and also includes reference to active fragments thereof. In one embodiment, the SOCS polypeptide or aetive fragment thereof is human SOCS1 (SEQ ID NO: 18) or SOCS3 (SEQ ID NO: 19) or an orthologue thereof, for example murine SOCS1 (SEQ ID NO:20) or SOCS3 (SEQ ID NO:21). Active fragments of a SOCS polypeptide include any fragments capable of inhibiting JAK activity and include S0CS3 _A PEST(22-225) (SEQ ID NO:22) and SOCS3 _APEST (22-185) (SEQ ID NO:23).

By "inhibits" or "inhibiting" activity of a JAK is meant a decrease or reduction in JAK activity in the presence of a compound when compared to the activity of the JAK in the absence of the compound, such as in a control sample. The degree of decrease or inhibition of JAK activity will vary with the nature and quantity of the compound present, but will be evident e.g., as a detectable decrease in phosphorylation of a JAK substrate; desirably a degree of decrease greater than 10%, 33%, 50%, 75%, 90%, 95% or 99% as compared to phosphorylation of a JAK substrate in the absence of the compound.

By "inhibits" or "inhibiting" the binding of a SOCS polypeptide or active fragment thereof, ATP and/or a JAK substrate to a JAK, is meant a decrease in binding of the SOCS polypeptide or active fragment thereof, ATP and/or the JAK substrate with the JAK in the presence of a compound when compared to the binding of the SOCS polypeptide or active fragment thereof, ATP and/or the JAK substrate with the JAK in the absence of the compound, such as in a control sample. The degree of decrease in the binding will vary with the nature and quantity of the compound present, but will be evident e.g., as a detectable decrease in the binding of the SOCS polypeptide or active fragment thereof, ATP and or a JAK substrate with the JAK; desirably a degree of decrease greater than 10%, 33%, 50%, 75%, 90%, 95% or 99% as compared to the binding of the SOCS polypeptide or active fragment thereof, ATP and/or a JAK substrate with the JAK in the absence of the compound.

By "selective" inhibitor is meant that a compound binds to or inhibits a JAK with greater affinity or potency, respectively, compared to at least one other kinase, for example another tyrosine kinase. In one embodiment, the compound may be a selective inhibitor of JAK2 (for example, over JAK1 , JAK3 and TYK2). Selectivity can be at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 200-fold, at least about 500-fold or at least about 1000-fold. Selectivity can be measured by methods routine in the art. in some embodiments, selectivity can be tested at the K _m ATP or substrate concentration of the enzyme, in some embodiments, selectivity of compounds can be determined at the cellular ATP or substrate concentration. In some embodiments, the selectivity of compounds of the invention can be determined by cellular assays associated with particular JAK kinase activity.

As used herein, a "non-competitive inhibitor" has about the same affinit for both free enzyme and the enzyme-substrate complex. By virtue of its non-competitive mechanism, the inhibitor}' function of SOCS3 is unaffected by high intracellular ATP concentration or high concentrations of protein substrates. Although JAK substrates (STATs and cytokine receptors) are not present at high concentration in the cytoplasm, JAK is tethered directly to its receptor substrate and through it, to STAT, thus the local concentration of substrate can be high, A JAK inhibitor that is non-competitive toward substrate will be unaffected by the high local substrate concentration and therefore be more effective in vivo.

"STAT polypeptide" as used herein refers to proteins that are members of the Signal Transducers and Activation of Transcription (STAT) protein family and which are substrates for a JAK.

As used herein, the terms "treating", "treat" or "treatment" include administering a therapeutically effective amount of a compound(s) described herein sufficient to reduce or eliminate at least one symptom of a disease. As used herein, the terms "preventing", "prevent" or "prevention" include administering a therapeutically effective amount of a compound(s) described herein sufficient to stop or hinder the development of at least one symptom of a disease.

The term "about" as used herein refers to a range of +1-5% of the specified value.

Candidate compounds

By a "candidate compound" is meant an agent to be evaluated as a noncompetitive inhibitor of JAK activity. Candidate compounds may include, for example, small molecules, peptides or mimetics thereof, polypeptides, antibodies, nucleic acid molecules such as aptamers, peptide nucleic acid molecules, and components and derivatives thereof.

Small molecules

The candidate compounds and/or "compounds identified or designed using a method of the present invention may be any suitable compound, synthetic or naturally occurring. The compounds may encompass numerous chemical classes though typically they are organic molecules, preferably small organic compounds. In one embodiment, a synthetic compound identified or designed by the methods of the invention has a molecular weight equal to or less than about 5000, 4000, 3000, 2000. 1000 or 500 daltons. A compound of the present invention is preferably soluble under physiological conditions.

Such compounds can comprise functional groups necessary for structural interaction with proteins, for example hydrogen bonding, and may include at least an amine, carbonyl, hydroxy 1 or carboxyl group, or at least two of the functional chemical groups. The compound may comprise cyclical carbon or heterocyclic structures and/or aromatic or polyaromatic structures substituted with one or more of the above functional groups. Compounds can also comprise biomolecules including peptides, saccharides, fatty acids, steroids, purines, pyrimidines, derivatives, structural analogs, or combinations thereof.

Candidate compounds can be obtained from a wide variety of sources including libraries of synthetic or natural compounds. Synthetic compound libraries are commercially available from, for example, Maybridge Chemical Co. (Tintagel, Cornwall, UK), AMRI (Budapest, Hungary) and ChemDiv (San Diego, CA), Specs (Delft, The Netherlands). Natural compound libraries comprising bacterial, fungal, plant or animal extracts are available from, for example, Pan Laboratories (Bothell, WA), TimTec (Newark, DE). In addition, numerous means are available for random and directed synthesis of a wide variety of organic compounds.

Alternatively, libraries of natural compounds in the form of bacterial, fungal, plant and animal extracts can be readily produced. Methods for the synthesis of molecular libraries are readily available (see, for example, eWitt et al., 1993; Erb et al., 1994; Zuckermann et al., 1994; Cho et al., 1993; and Gallop et al., 1994). In addition, natural or synthetic compound libraries and compounds can be readily modified through conventional chemical, physical and biochemical means (see, for example, Blondelle and Houghton, 1996), and may be used to produce combinatorial libraries. In another approach, previously identified pharmacological agents can be subjected to directed or random chemical modifications, such as acylation, alkylation, eslerification, amidification, and the analogs can be screened for a desired activity.

A combinatorial chemical library is a collection of diverse chemical compounds generated by either chemical synthesis or biological synthesis by combining a number of chemical "building blocks" such as reagents. Millions of chemical compounds can be synthesized through such combinatorial mixing of chemical building blocks (Gallop et al, 1994). Preparation and screening of combinatorial chemical libraries is well known to those of skill in the art. Such combinatorial chemical libraries include, but are not limited to, small organic molecule libraries and carbohydrate libraries.

Compounds also include those that may be synthesized from leads generated by fragment-based drug design, wherein the binding of such chemical fragments is assessed by soaking or co-crystallizing such screen fragments into crystals comprising a JAK and then subjecting these to an X-ray beam and obtaining diffraction data. Difference Fourier techniques are readily applied by those skilled in the art to determine the location within the JAK structure at which these fragments bind, and such fragments can then be assembled by synthetic chemistry into larger compounds with increased affinity for the JAK.

Peptides and mimetics thereof

In another embodiment, the candidate compound used in the methods of the invention is a peptide or a mimetic thereof. The term "peptide" as used herein is typically used to refer to chains of amino acids which are not large, for instance 100 or less residues in length. In one embodiment, candidate compounds are peptides of from about 5 to about 30 amino acids, or from about 5 to about 20 amino acids, or from about 7 to about 15 amino acids. In one embodiment, peptides are chemically or recombinantly synthesized as oligopeptides.

The terms "mimetic", "peptidomimetic" and "mimic" as used herein refer to a synthetic chemical compound that has substantially the same structural and/or functional characteristics of the peptides. The mimetic can be entirely composed of synthetic, non-natural analogues of amino acids, or, may be a chimeric molecule of partly natural amino acid residues and partly non-natural analogs of amino acids.

A peptide may be characterized as a mimetic when all or some of its residues are joined by chemical means other than natural peptide bonds. Individual mimetic residues can be joined by peptide bonds, other chemical bonds or coupling means, such as, e.g., glutaraldehyde, N-hydroxysuccinimide esters, bifunctional maleimides, Ν,Ν'- dicyclohexylcarbodiimide (DCC) or Ν,Ν'-diisopropylcarbodiimide (DIC). Linking groups that can be an alternative to the traditional amide bond ("peptide bond") linkages include, but not limited.to, ketomethylene (e.g., ~C(=0)~CH2~ for -C(=0)~ NH-), aminomethylene (CH ₂ -NH), ethylene, olefin (CH=CH), ether (CH ₂ -0), thioether (CH ₂ -S), tetrazole (CN ₄ -), thiazole, retroamide, thioamide, or ester (see, e.g., Spatola (1983) In: Chemistry and Biochemistry ol ^' Amino Acids, Peptides and Proteins, Vol. 7, pp 267-357, "Peptide Backbone Modifications," Marcell Dekker, NY).

A mimetic also can be a peptide-like molecule which contains, for example, an amide bond isostere such as a retro-inverso modification; reduced amide bond; methylenethioether or methylene-sulfoxide bond; methylene ether bond; ethylene bond; thioamide bond; trans-olefin or fluoroolefin bond; 1 ,5-di substituted tetrazole ring; ketomethylene or fluoroketomethylene bond or another amide isostere. Retro-inverso modification of naturally occurring peptides involves the synthetic assembly of amino acids with a-carbon stereochemistry opposite to that of the corresponding L-amino acids, i.e., D- or D-allo-amino acids in inverse order to the native peptide sequence. A rerto-inverso analogue, thus, has reversed termini and reversed direction of peptide bonds, while essentially maintaining the topology of the side chains as in the native peptide sequence. One skilled in the art understands that these and other mimetics are encompassed within the meaning of the term "mimetic" as used herein.

Antibodies and fragments thereof

In one embodiment, the candidate compound is an antibody or antigen binding fragment thereof, for example, such as an antibody or fragment thereof that binds to a JA and inhibits binding of a SOCS polypeptide to the JAK, or an antibody or fragment thereof that binds to JAK and inhibits JAK activity without inhibiting binding of ATP and/or JAK substrate to the JAK.

The term ".antibody" as used in this invention includes intact molecules as weil as molecules comprising or consisting of fragments thereof, such as Fab, F(ab')2, and Fv which are capable of binding an epitopic determinant. Thus, antibodies may exist as intact immunoglobulins, or as modifications in a variety of forms including, for example, but not limited to, domain antibodies including either the VH or VL domain, a dimer of the heavy chain variable region (VF1H, as described for a camelid), a dimer of the light chain variable region (VLL), Fv fragments containing only the light and heavy chain variable regions, or Fd fragments containing the heavy chain variable region and the CHI domain. A scFv consisting of the variable regions of the heavy and light chains linked together to form a single-chain antibody and oligomers of scFvs such as diabodies and triabodies are also encompassed by the term "antibody". As outlined above, also encompassed are fragments of antibodies such as Fab, (Fab') ₂ and FabFcj fragments which contain the variable regions and parts of the constant regions. CDR-grafted antibody fragments and oligomers of antibody fragments are also encompassed. The heavy and light chain components of an Fv may be derived from the same antibody or different antibodies thereby producing a chimeric Fv region. The antibody may be of animal (for example mouse, rabbit or rat) or human origin or may be chimeric or humanized.

The antibodies may be Fv regions comprising a variable light (V _L ) and a variable heavy (V _H ) chain. The light and heavy chains may be joined directly or through a linker. As used herein a linker refers to a molecule that is covalently linked to the light and heavy chain and provides enough spacing and flexibility between the two chains such that they are able to achieve a conformation in which they are capable of specifically binding the epitope to which they are directed. Protein linkers are particularly preferred as they may be expressed as an intrinsic component of the Ig portion of the fusion polypeptide.

The antibody or fragment thereof used in the methods of the invention may be an internalizing antibody. An "internalizing antibody" is an antibody that is capable of being transported into a ceil. Methods for producing and/or selecting internalizing antibodies are known in the art, such as those methods described in Poul et al. (2000) and Becerril et al. ( 1999).

In another embodiment, recombinantly produced single chain scFv antibody, preferably a humanized scFv, is used in the methods of the invention. A variety of immunoassay formats may be used to select antibodies specifically immunoreactive with a JAK and which inhibit JAK activity, including those that may inhibit binding of a SOCS polypeptide to the JAK. For example, surface labelling and flow cytometric analysis or solid-phase EL1SA immunoassays are routinely used to select antibodies specifically immunoreactive with a protein or carbohydrate. See Harlow & Lane (supra) for a description of immunoassay formats and conditions that can be used to determine specific immunoreactivity.

Identification of potential inhibitors of JAK activity

Inhibitors of JAK activity are screened using assays and techniques useful for identifying compounds capable of reducing interaction of a SOCS polypeptide or fragment thereof with a JAK, and for identifying compounds able to inhibit the activity of JAK in the presence of ATP and/or a JAK substrate. Such assays include those described in the Examples section.

Screening assays

in certain embodiments candidate compounds will be screened for an ability to reduce interaction of a SOCS polypeptide or active fragment thereof with a JAK, and/or to inhibit JAK activity. Conventionally, new chemical entities with useful properties are generated by identifying a chemical compound (called a "lead compound") with some desirable property or activity, for example, reducing binding of a SOCS polypeptide to JAK, creating variants of the lead compound, and evaluating the property and activity of those variant compounds. Often, high throughput screening (HTS) methods are employed for such an analysis.

in one embodiment, high throughput screening methods involve providing a library containing a large number of candidate compounds. Such "combinatorial chemical libraries" are then screened in one or more assays to identify those library members (particular chemical species or subclasses) that display a desired characteristic activity. The compounds identified can serve as conventional lead compounds or can be used as potential or actual therapeutics.

High throughput screening systems are commercially available and typically automate entire procedures, including all sample and reagent pipetting, liquid dispensing, timed incubations, and final readings of the microplate in detectors appropriate for the assay. These configurable systems provide high throughput and rapid start up as well as a high degree of flexibility and customization. The manufacturers of such systems provide detailed protocols for various high throughput systems. Thus, for example, Zymark Corp. provides technical bulletins describing screening systems for detecting the modulation of gene transcription, ligand binding, and the like.

Alternatively, compounds which reduce binding of a SOCS polypeptide or active fragment thereof with a JAK and/or which are capable of inhibiting JAK activity may be identified and isolated by other methods known to those of skill in the art. Examples of such methods that may be used include yeast-2-hybrid screening. Surface Plasmon Resonance, high-resolution NMR, phage display, affinity chromatography, Isothermal Titration Calorimetry (ITC), immunoprecipitation and GST pull downs coupled with mass spectroscopy,

Surface Plasmon Resonance (SPR) or Biomolecular Interaction Analysis (BIA; e.g., Biacore) detects biospecific interactions in real time, without labeling any of the interactants. Changes in the mass at the binding surface (indicative of a binding event) of the BIA chip result in alterations of the refractive index of light near the surface. The changes in the refractivity generate a detectable signal, which are measured as an indication of real-time reactions between biological molecules. For example, structural modelling of a SOCS polypeptide bound to a JAK can be used to guide the design of compounds that bind JAK and inhibit the binding of the SOCS polypeptide to the JAK.

Information from SPR can be used to provide an accurate and quantitative measure of the equilibrium dissociation constant (Kd), and kinetic parameters, including k _on and r, for the binding of a molecule to a target. Such data can be used to compare different molecules. Information from SPR can also be used to develop structure-activity relationships (SAR). For example, the kinetic and equilibrium binding parameters of different compounds can be evaluated. Variants at given positions in the compound can be identified that correlate with particular binding parameters, e.g., high affinity and slow k _oi r. This information can be combined with structural modelling (e.g., using homology modeling, energy minimization, or structure determination by x- ray crystallography or NMR). As a result, an understanding of the physical interaction between the compound and its target can be formulated and used to guide other design processes.

Standard solid-phase ELISA assay formats are also useful for identifying antagonists of protein-protein interaction. In accordance with this embodiment, one of the binding partners, for example, the JAK is immobilized on a solid matrix, such as, for example an array of polymeric pins or a glass support. Conveniently, the immobilized binding partner may be a fusion polypeptide comprising, for example, Glutathione-S-transferase, wherein the GST moiety facilitates immobilization of the protein to the solid phase support. The second binding partner (for example, the SOCS polypeptide or active fragment thereof) in solution is brought into physical relation with the immobilized protein to form a protein complex in the presence of a candidate compound, which complex is then detected by methods known in the art. If a candidate compound is capable of inhibiting SOCS binding to JAK, this will be identified by a reduction in the formation of the polypeptide complex when compared to the same assay conditions in the absence of the candidate compound.

The skilled person will understand that when performing an assay to identify a candidate non-competitive inhibitor of a JAK, the SOCS polypeptide and/or JAK may be either full-length polypeptides or alternatively active fragments thereof. The JAK used in the methods of the invention may be a fragment of the JAK that maintains JAK activity, for example the fragment may comprise or consist of the JAK catalytic domain. The SOCS polypeptide fragment used in the methods of the invention is capable of binding to at least a catalytic domain of a JAK. As would be understood in the art the SOCS polypeptide fragment will comprise at least the Kinase Inhibitory Region (KIR), and preferably comprises at least the first 71 amino acids of the SOCS polypeptide. In one embodiment, the SOCS polypeptide fragment comprises both the KIR and the SH2 domain. The person skilled in the art will able to determine fragments of a SOCS polypeptide that bind a JAK and inhibit JAK activity using methods known in the art.

It may also be desirable in some instances to modify one or more amino acids in the SOCS polypeptide or active fragment thereof to facilitate the production of recombinant protein. For example, in order to produce sufficient quantities of a SOCS polypeptide or active fragment thereof, it may be desirable to produce a chimeric SOCS polypeptide comprising at least one domain from a first SOCS polypeptide that is difficult to produce recombinantly in sufficient quantities and at least one domain from a second SOCS polypeptide which is able to be produced recombinantly in sufficient quantities. By way of non-limiting example, when testing SOCS 1 for JAK inhibitory activity, it may be desirable to express a SOCSl-3 chimeric protein comprising the KIR of SOCS1 and the SH2 domain of SOCS3. Such a chimeric protein can be used, for example, to investigate the interaction of SOCS 1 with a JAK. The person skilled in the art will be able to produce modified SOCS polypeptides or active fragments thereof that maintain the ability to bind a JAK and inhibit JAK activity and which are suitable for use in the methods of the invention using known techniques.

When performing the methods of the invention, it may also be desirable to express and purify SOCS polypeptides and/or active fragments thereof in complex with a physiological ligand, for example ElonginB and/or ElonginC in order to increase the solubility of the expressed polypeptide.

A screening assay as described herein to identify candidate compounds that inhibit the binding of a SOCS polypeptide or active fragment thereof to a JAK may also be performed in the presence of ATP and/or ADP. In one embodiment, the concentration of ATP or ADP used in the methods of the invention is about the cellular concentration of ATP, for example about 2 mM ATP.

To test for JAK activity in the methods of the invention, the person skilled in the art will be able to determine suitable conditions for performing an assay to test for kinase activity. Typically, an assay to test for JAK activity will comprise the JAK together with a polypeptide substrate, for example gpl30 or a STAT polypeptide, ATP and magnesium as a cofactor. As would be understood in the art, the assay is performed in a suitable buffer, such as for example HEPES (pH 7.5) or Tris (pH 8.0) containing other components as required, for example containing about 100 mM NaCl and about 1 mM dithiothreitol. Magnesium in the buffer may be MgCl ₂ at a concentration of about 4 mM. The person skilled in the art will be able to determine suitable conditions for performing an assay to test kinase activity that vary from the conditions provided herein. Assay formats for testing kinase activity are commercially available and include the Alphascreen Protein Tyrosine Kinase PI 00 detection kit (Perkin Elmer) which can be used in conjunction with a Perkin Elmer Fusion Alpha Instrument such as described in Williams et al. (2009). Examples of other suitable kinase assays are described in Jia et al. (2008).

Labelled ATP may be incorporated into the assay to allow for the detection of phosphorylated substrates as a measure of kinase activity. One example of a suitable labelled ATP is radiolabeled ATP, such as ² Ρ-γ-ΑΤΡ, although other labels known in the art, for example fluorescent labels, may be used in the methods of the invention. The labelled ATP may be detected using known means. For example, a substrate phosphorylated with ³² Ρ-γ-ΑΤΡ may be detected by autoradiography or phosphorimaging by first subjecting the substrate to SDS-PAGE or by spotting the substrate onto nitrocellulose membrane and exposing the gel or membrane to film or a phosphorimager plate. The level of phosphorylation of the substrate can then be quantified using known techniques.

In addition, to determine whether a candidate compound is a non-competitive inhibitor of a JAK, the candidate compound is tested for JAK inhibitory activity in the presence of ATP and/or JAK substrate. The concentration of ATP or JAK substrate may be any suitable concentration, for example such as would be found within a cell. By way of example, the concentration of ATP and/or JAK substrate in the assay may be about 0.1 mM to about 2 mM, or about 0.5 mM, 1.0 m or about 1.5 mM.

Typically, the ability of a non-competitive inhibitor to inhibit JAK activity will be independent of the concentration of ATP and/or JAK substrate. Thus, in one embodiment, the candidate compound is tested for JAK inhibitory activity in the presence of ATP and/or JAK substrate in a first concentration of ATP and/or substrate and separately in a second concentration of ATP and/or substrate. For example, the first concentration of ATP may be relatively low, for example about 0.1 mM ATP, and the second concentration may be higher such as may be found within a cell, for example about 2 mM ATP. In one embodiment, the compound is a non-ATP-competitive inhibitor, the ability of which to reduce the activity of JAK is substantially the same in ^■ the first and second concentration of ATP. By "substantially the same" is meant that the compound is able to inhibit the JAK in both the first and second concentrations of ATP and/or JAK substrate and preferably maintains at least about 50%, more preferably about 60%, 70%, 75%, 80%, 85%, 90% or about 95% or more activity in a second higher concentration of ATP and/or substrate compared to a lower first concentration of ATP and/or substrate. In another embodiment, the compound is a non-ATP and non- substrate-competitive inhibitor. The compound may also be a selective inhibitor of a JAK kinase.

In an alternate embodiment, the screening assay for detecting the inhibition of

JAK includes testing to determine if the candidate compound decreases the degradation of JAK. In this embodiment, the candidate compound is screened to determine if it is blocking SOCS binding JAK. In an example, JAK is tested for a decrease in ubiquitination using standard assays such as those outlined in Example 9.

Computer modelling and structural determination

Computer modelling and searching technologies permit identification of compounds that can reduce the interaction of a SOCS polypeptide or active fragment thereof with a JAK and/or which are able to inhibit JAK activity. The three dimensional geometric structure of the SOCS polypeptide or active fragment thereof in complex with the JAK can be determined. This can be done by known methods, including X-ray crystallography, which can determine a complete molecular structure. For example, the crystal structure at 2.0 A resolution of the N-termina ly extended SH2 domain of SOCS3 in complex with its phosphopeptide target on the cytokine receptor gpl 30 is described by Bergamin et al. (201 1). Methods for the determination of the three-dimensional structure of proteins are known in the art and include protein crystallography and NMR. For example, a JA , or a SOCS polypeptide or active fragment thereof complexed with a JAK, may be subject to a screen against a library of crystallisation conditions. Commercially available crystallisation screens include Hampton Research Crystal Screen I, Crystal Screen II, Crystal Screen I-Lite, and from Emerald Biostructures, Inc. (Bainbridge Island, Washington), Wizard I, Wizard II, Cryo I and Cryo II. Alternatively, other conditions known to those of skill in the art, including those provided in screening kits available from other companies, may also be tested.

Conditions are typically tested at multiple protein concentrations and at different temperatures (for example, 4° and 20°C). Crystal setups may be performed manually or by a liquid handling robot appropriately programmed for sitting drop experiments. Crystallization plates are observed are various time points such as two days, two weeks, and 1 month after being set. Having identified conditions that are best suited for further crystal refinement, subsequent crystallisation drops are set up to explore the affects of variables such as temperature, pH, salt or PEG concentration on crystal size and form, with the intent of establishing conditions where the protein is able to form crystals of suitable size and morphology for diffraction analysis.

Methods of computer based numerical modelling can be used to complete the structure (e.g., when an incomplete or insufficiently accurate structure is determined) or to improve its accuracy. Any method recognized in the art may be used, including, but not limited to, parameterized models specific to particular biopolymers such as proteins, molecular dynamics models based on computing molecular motions, statistical mechanics models based on thermal ensembles, or combined models.

The present inventors have identified that SOCS3 binds JA 2 with higher affinity in the presence of ATP or ADP. Thus, in one embodiment, the structure of the SOCS polypeptide or active fragment thereof in complex with the JAK is determined in the presence of ATP and/or ADP. In addition, it may be desirable that the ATP is an ATP analogue, for example ATP-yS to prevent catalysis.

The three-dimensional structure of a SOCS polypeptide or active fragment thereof in complex with a JAK can be used to identify compounds that reduce interaction of the SOCS polypeptide and the JAK through the use of computer modelling using a docking program such as GRAM, DOCK, or AUTODOCK (Dunbrack et al., 1997). Computer programs can also be employed to estimate the attraction, repulsion, and steric hindrance of a candidate compound to the JAK. Generally the tighter the fit (e.g., the lower the steric hindrance, and/or the greater the attractive force) the more potent the potential agonist or antagonist will be since these properties are consistent with a tighter binding constant. Furthermore, the more specificity in the design of a potential inhibitor the more likely that it will not interfere with other proteins.

Initially a potential compound could be obtained, for example, using methods of the invention such as by screening a random peptide library produced by a recombinant bacteriophage or a chemical library. A compound selected in this manner could then be systematically modified by computer modelling programs until one or more promising potential compounds are identified.

Such computer modelling allows the selection of a finite number of rational chemical modifications, as opposed to the countless number of essentially random chemical modifications that could be made, and of which any one might lead to a useful compound. Each chemical modification requires additional chemical steps, which while being reasonable for the synthesis of a finite number of compounds, quickly becomes overwhelming if all possible modifications needed to be synthesized. Thus through the use of the three-dimensional structure and computer modelling, a large number of these compounds can be rapidly screened on the computer monitor screen, and a few likely candidates can be determined without the laborious synthesis of untold numbers of compounds.

For most types of models, standard molecular force fields, representing the forces between constituent atoms and groups, are necessary, and can be selected from force fields known in physical chemistry. Exemplary forcefields that are known in the art and can be used in such methods include, but are not limited to, the Constant Valence Force Field (CVFF), the AMBER force field and the CHARM force field. The incomplete or less accurate experimental structures can serve as constraints on the complete and more accurate structures computed by these modelling methods.

Further examples of molecular modelling systems are the CHARMm and QUANTA programs (Polygen Corporation, Waltham, MA). CHARMm performs the energy minimization and molecular dynamics functions. QUANTA performs the construction, graphic modelling and analysis of molecular structure. QUANTA allows interactive construction, modification, visualization, and analysis of the behaviour of molecules with each other.

A compound designed or selected as binding to a JAK may be further computationally optimised so that in its bound state it would preferably lack repulsive electrostatic interaction with the target protein. Such non-complementary (e.g., electrostatic) interactions include repulsive charge-charge, dipole-dipole and charge- dipole interactions. Specifically, the sum of all electrostatic interactions between the compound and the protein when the compound is bound to JAK, preferably make a neutral or favourable contribution to the enthalpy of binding.

Once a candidate compound has been optimally selected or designed, as described above, substitutions may then be made in some of its atoms or side groups to improve or modify its binding properties. Generally, initial substitutions are conservative, i.e., the replacement group will have approximately the same size, shape, hydrophobicity and charge as the original group, it should, of course, be understood that components known in the art to alter conformation should be avoided. Such substituted chemical compounds may then be analysed for efficiency of fit to JAK by the same computer methods described above. Specific computer software is available in the art to evaluate compound deformation energy and electrostatic interaction. Examples of programs designed for such uses include: Gaussian 03, GAMESS; Jaguar; AMBER, version 9.0; CHARMM and GROMACS version 4.0.

Candidate compounds identified or designed using the techniques described above may be further tested in JAK inhibition and activity assays as described herein.

EXAMPLES

Example 1. Materials and methods

Cloning and expression

DNA encoding murine SOCS3 _AI¾S T(22-225) (SEQ ID NO:22), was cloned as a

GST-fusion protein in pGEX-4T (Amersham) modified to contain an Ascl site immediately following the Thrombin cleavage site. The vector was further modified to contain DNA encoding mouse ElonginB (SEQ ID NO:25) preceded by an internal ribosome entry site, located downstream of SOCS3. This vector was used to express SOCSl-3 (SEQ ID NO:24), SOCS2, SOCS4 and SOCS5 by inserting DNA encoding these proteins into the same sites as SOCS3, once it had been excised using Ascl and EcoRI. These vectors were co-transformed with pBB75(mouseElonginCi .n2) into BL21(DE3) cells to produce the ternary SOCS/elonginB/elonginC complexes. Cells were grown in the presence of ampicillin and kanamycin and were harvested 8 h after IPTG induction by centrifugation at 6200 g at 4 °C for 30 min. Cells were resuspended in PBS and lysed by french press. The lysates were centrifuged at 20,000 g, and purified using glutathione-Sepharose chromatography. Thrombin cleavage was used to remove the GST followed by size exclusion chromatography on a Superdex 75 or Superdex 200 16/60 column (GE Healthcare) run in Tris buffered saline. D A encoding the cytoplasmic domain of the g l30 receptor was also cloned as a GST-fusion protein in pGEX-4T (Amersham) modified to contain an Ascl site immediately following the Thrombin cleavage site and then purified under identical conditions as SOCS3/elonginBC.

JAK1 (SEQ ID NO: 13), JA 2 (SEQ ID NO: 14), JAK3 (SEQ ID NO: 15) and

TYK2 (SEQ ID NO: 16) (kinase domains) were cloned into pFASTBAC (Invitrogen) and expressed as a 6 * HIS tagged proteins. In addition JAK2 was also expressed as a GST fusion (SEQ ID NO: 17) by cloning into pDEST-20 (Invitrogen). All JA constructs were expressed in Spodoptera frugiperda Sf21 insect cells using the baculovirus system. High titre virus stocks were produced using standard protocols and typically used to infect 1-2 L of insect cells at 2 * 10 ⁶ cells/ml. Following infection, the proteins were allowed to express for 60 hours at 25°C in a humidified shaking incubator containing 10% C0 ₂ .

If the kinases were not to be used for inhibition assays the JAK inhibitor 2-(l,l- Dimethylethyl)-9-fluoro-3,6-dihydro-7H-benz[h]-imidaz[4,5-f] isoquinolin-7-one

(EMD biosciences) was added to a final concentration of 0.4 uM to increase the yield. Cells were resuspended into a buffer consisting of 20 mM Hepes, pH 7,4, 150 mM NaCl, 1% Tween 20, 1 mM PMSF, and 1 mM DTT, lysed by incubation on ice for 30 minutes followed by sonication, and centrifuged at 20,000g for 10 minutes. Depending on the construct, the supernatant was loaded onto either GST resin (Amersham) or Ni- NTA resin (Qiagen). After extensive washes, the fusion protein was eluted, and fractions of the kinase were pooled and concentrated before being loaded onto a Superdex 200 (16/60) column (Amersham) and purified by gel filtration. If the JAK inhibitor 2-(l,l-Dimethylethyl)-9-fluoro-3,6-dihydro-7H-benz[h]-imidaz [4,5- fJisoquinolin-7-one (EMD biosciences) was added during expression it was displaced by incubation of the kinase with 2 mM ATP/4 mM MgCl ₂ overnight prior to gel filtration.

Kinase inhibition assays

Kinase inhibition assays using protein substrates were performed by incubating lmg/mL substrate with 50nM JAK2 _JH i at 25°C for 30 minutes in the presence of 20mM Tris pH 8.0, l OOmM NaCl, ImM DTT, 2mM ATP and 4mM MgCl ₂ and various concentrations of SOCS/elonginBC complexes ^'32 Ρ-γ-ΑΤΡ was included to allow visualization of phosphorylation through autoradiography and phosphorimaging. Following incubation, the reactions were either boiled and subjected to analysis by SDS-PAGE or, if quantification was required, terminated with 50mM EDTA and spotted onto a Nitrocellulose membrane. , Membranes were washed extensively (4x200tnl, 15 minutes each) with PBS to remove unincorporated ³² P-y-ATP and subsequently exposed to a phosphorimager plate (Fuji). Results were plotted and analyzed using SigmaPlot. Various SOCS/elonginBC complexes were tested for inhibition at a range of concentrations, typically 0-10μΜ. ElonginBC is the physiological ligand for all SOCS proteins and is present to enhance solubility.

Kinase inhibition assays using peptide substrates were performed by incubating 0-2mM substrate with lOnM JAK2 _JH i at 25°C for 10-20 minutes in the presence of 20mM Tris pH 8.0, lOOmM NaCl, ImM DTT, 2mM ATP and 4mM MgCl ₂ and ³² Ρ-γ- ATP. Following incubation, the reactions were spotted onto P81 phosphoceHuiose paper and quenched in 5% H3PO4. The paper was washed extensively (4x200ml, 15 minutes each) with 5% H3PO4 to remove unincorporated ³² Ρ-γ-ΑΤΡ and subsequently exposed to a phosphorimager plate (Fuji). Results were plotted and analyzed using SigmaPlot. ·

Steady state kinetic analysis

Michaelis Menten analysis requires the use of low enzyme concentrations and/or incubation times in order to determine initial reaction velocities, in addition the substrate concentrations must exceed their Km values in order to determine the maximum velocity (Vmax). Therefore significantly lower concentrations of JAK2 (kinase domain) were used (2 nM) and a STAT5 derived peptide (GL Biosciences, China (SEQ ID NO:27)) was used as substrate in order to enable saturating concentrations to be achieved. Inhibitor (SOCS3 or SOCS l-3)/elonginBC was included at 0-10 μΜ fmai concentration. Reactions were performed in 20mM Tris pH 8.0, lOOmM NaCl, ImM DTT, and 4mM MgCi ₂ at 30°C for 5-10 minutes. ³² Ρ-γ-ΑΤΡ was included at 0.5 μΜ to allow visualization. These reactions contained two substrates: ATP (the phosphate donor) and the STATS derived peptide (the phosphate acceptor). Both of these substrates were titrated independently. If ATP was to be titrated the STAT5 concentration was kept at 1.2 mM (~Km) whilst if the STAT5 peptide was titrated the ATP concentration was fixed at 2 mM (-50 * Km). Following incubation, the reactions were spotted onto P81 phosphoceHuiose paper (Millipore), washed extensively (4 * 200 ml, 15 minutes each) with 5% Orthophosphoric acid to remove unincorporated ³² Ρ-γ-ΑΤΡ, the paper was then air-dried and exposed to a phosphorimager plate (Fuji). Results were plotted and analyzed using Sigma Plot. ATPase assays ATPase assays were performed by incubating 0.25 μΜ JA 2 _JH i with 0.1 mM ATP for 30 minutes at 25°C. Reactions were performed in 20mM Tris pH 8.0, lOOmM NaCl, ImM DTT, 2mM MgCl ₂ and contained ϊ μθϊ ³² Ρ-γ-ΑΤΡ. SOCS3/elonginBC or mutants thereof were present in the reaction at 0-1 ΟμΜ concentration. Reactions were stopped by the addition of EDTA to 25mM and spotted onto PEI-cellulose thin-layer chromatography plates. TLC plates were developed in I ^' M LiCl, 1M Formic acid for approximately 45 minutes, air-dried and exposed to a phosphorimager plate. Results were plotted and analyzed using Sigma Plot. Surface Piasmon Resonance

GST-SOCS3 ₂ 2- ₂ 2 ₅ APEST/elonginBC was immobilized onto a Biacore CM5 sensor chip (GE Healthcare) that had been previously coated with anti-GST antibody (GE Healthcare) according to manufacturers instructions. 15 μΐ of 10 μg/mL GST- SOCS3 ₂ 2. ₂ 2sAPEST/elonginBC (sample lane) or GST-SOCS3 ₂ 9- ₂₂ sAPEST/elonginBC alone (control lane) was loaded onto the chip to give a surface density of approximately 1000 response units. The binding of JAK2JHI to SOCS3 was then analysed by passing 0-2 μΜ JAK2JHI over the chip at 20 μΐ min ^"1 in Hepes buffered saline (HBS) containing 0.01 % (v/v) Tween-20. The chip was regenerated by the addition of l OmM glycine, pH 2,0 after each passage. The kinetic data were analyzed using the BIAevaluation software (GE Healthcare) and fitting to a 1 : 1 Langmuir binding model.

Cloning and expression of the SOCS/elonginBC/cullin5/Rbx2 E3 ligase complex

Mouse CullinS was co-expressed as two domains, the N-terminal domain (1 - 384) and C-terminal domain (385-780) analogous to that performed previously for Cullinl (Zheng et al., 2002). The N-terminal domain was GST-tagged and the C- terminal domain untagged. These were co-expressed with His6 Rbx2 in BL21 (DE3) cells to yield a ternary GST-Cul5( TD)/HIS ₆ -Rbx2/Cul(CTD) complex. Expression was at 18°C overnight and purifications were performed using standard protocols. Once GST was removed with thrombin, Cul5/Rbx2 as added to purified SOCS3/elonginBC to form the full E3 ligase. Gel filtration on a Superdex200 column (GE Healthcare) was used as a final purification step and to ensure the complex was of the expected size and stoichiometry.

Ubiquitin cascade components and ubiquitination assays

Murine UbcH5a (E2) was expressed as a GST-fusion protein and purified using standard protocols. Human El (His ₆ tagged) and a panel of human E2 enzymes (Hise T/AU2011/000710

32 tagged), were purchased from Biomol International. Bovine ubiquitin was purchased from Sigma-Aldrich. Ubiquitination assays were performed in 20 μΐ in 20mM Tris- HCI, SOmM NaCl, 5mM MgCl ₂ , 2.5mM ATP, 0.1 mM DTT at 37°C. Reactions were stopped by the addition of 2x SDS PAGE loading buffer and heating at 95°C for 5 minutes. Typically reactions contained 0.1 μΜ El , 2.5 μΜ E2, 2.5 μΜ E3, 50 μ ubiquitin and 5 μΜ JAK.

Example 2. SOCS1 and SOCS3 inhibit substrate phosphorylation by JAK2

In order to determine the mechanism of SOCS1 and SOCS3 suppression of JAK activity, an in vitro kinase assay was developed that consisted of purified, recombinant enzyme (JAK2 catalytic domain), substrate (gpl30 cytoplasmic domain) and inhibitor (SOCS3). Expression of SOCS1 in the quantities required for rigorous assaying is not feasible, and hence we constructed a chimaeric version of SOCS3, termed SOCSl-3, which contains the KIR of SOCS1, rather than its own. This chimaera was used to address characteristics of SOCS1 based inhibition. SOCS2 and SOCS4, which do not contain a kinase inhibitory region, were purified for use as negative controls. All SOCS proteins were expressed and purified in complex with elonginBC, their physiological ligand, due to increased solubility.

Kinase assays were performed with an enzyme:substrate ratio of > 1 : 100 at 30° C for 10 minutes in the presence of γ- ³² Ρ-ΑΤΡ and the results visualized by autoradiography. Figure 1A shows that both SOCS3 and SOCSl-3 are potent inhibitors of s tbstrate phosphorylation whereas SOCS-2 and -4 have no effect. The assay was modified to incorporate a filter-binding step, rather than SDS-PAGE and the inhibition quantified using phosphorimaging (Figure I B). SOCS3 and SOCS l-3 inhibit the phosphorylation of the gp! 30 cytoplasmic domain by JAK2 with IC50 values of 240 and 80 nM, respectively. Interestingly, SOCS3 and SOCSl-3 had no effect on JAK autophosphorylation (Figure 1C).

In order to determine whether this effect was specific for gpl30 phosphorylation, other substrates were tested. Figures 2 and 3 show that SOCS3 and SOCSl-3 are able to inhibit the phosphorylation of every exogenous substrate tested. Although the kinase inhibitor region is necessary for inhibition it is not sufficient, as mutations outside this region, for example Arg71 , render SOCS3 inactive (Figure 4). In support of this, Figure 5 shows that whilst full length SOCS3 has inhibitory activity towards JAK2, the KIR alone (as a synthetic peptide), does not. Thus regions outside the KIR are required for activity.

Example 3. SOCS is not a competitive inhibitor of JAK2

The prevailing model regarding the mechanism of JAK inhibition by SOCS3 is that the KIR acts as a "pseudosubstrate" and blocks substrate binding. Kinases are two substrate enzymes: ATP (the phosphate donor) and the tyrosine-containing substrate (the phosphate acceptor). If SOCS3 acts as a pseudosubstrate then it must compete with the binding of one or both of these substrates. Therefore. Michaelis-Menten kinetic experiments were performed in the presence of SOCS3. Such experiments have been widely used to assess the nature of enzyme inhibitors. These experiments used a STAT5 derived peptide as a substrate and the results were quantified using scintillation counting.

To determine whether SOCS3 was a competitive inhibitor of ATP binding, the STAT substrate concentration was fixed at 1.6 mM and the reactions were performed using a range of different ATP concentrations (0-2 mM). Conversely, to determine whether SOCS3 was a competitive inhibitor of tyrosine-substrate binding, the ATP concentration was fixed at 2 mM and the reactions were performed using a range of different STAT peptide substrate concentrations (0-2 mM). In both cases, the initial reaction velocity was plotted against substrate concentration. Figure 6 shows that SOCS3 is a non-competitive inhibitor of JAK2 because V _max but not K _m is affected and inhibition cannot be overcome by high substrate concentrations. Experiments were performed at ATP concentrations of up to 50 x K _lt) but STAT concentrations of up to only 3 x K _m due to solubility issues. Therefore, reciprocal analyses (Lineweaver-Burk) were performed to confirm that SOCS3 is also non-competitive with regard to peptide substrate (Figure 6, lower panels). Since K _m is not affected by SOCS3 the analyses also show that SOCS3 is not an allosteric inhibitor of ATP or substrate binding and therefore represents a novel class of kinase inhibitor.

One attribute of a non-competitive inhibitor is that inhibition is not reduced by the presence of high substrate concentration. This can be illustrated by performing standard inhibition experiments in the presence of high and low substrate concentrations and measuring IC50. As shown in Figure 7, SOCS3 inhibited JAK with identical IC50 values at ATP and substrate concentrations that varied 40-fold in each case. As a control, ADP and CMP-6, ATP competitive inhibitors, were tested. The IC50 values for these inhibitors increased in the presence of high ATP concentration, but not in the presence of high substrate concentration, as expected. Example 4. SOCS3 inhibits JAK1, JAK2 and TYK2 but not JAK3 due to the presence of a tri-residue (GQM) motif in the JAK insertion loop

The data show that SOCS3 can directly inhibit the kinase activity of JAK2. However, it is was not known whether it acts similarly on the other three members of the JAK family (JAK1 , JAK3, TYK2). The ability of SOCS3 to inhibit the three other members of the JAK family was determined using a STAT derived peptide and identical assay conditions to those described above. All four JAKs (kinase domain only) were expressed using the baculovirus system. SOCS3 was used at concentrations between 0-5 μΜ and the results visualized and quantified by phosphorimaging. As shown in Figure 8A, SOCS3 was able to inhibit JAK1 , JAK2 and TYK2 with IC ₅₀ values of 1.5 to 7 μΜ, but had no inhibitory effect on JAK3.

As the sequence of all four JAKs is highly conserved the data suggested that particular residue(s) on JAK3 are responsible for its insensitivity toward SOCS3. A comparison of the sequence of all four kinases domains highlighted a number of residues that were conserved within JAKl , JAK2 and TYK2 but not JAK3. Therefore, a series of seven mutant JAK2 molecules were produced that replaced residue(s) of JAK2 with the corresponding residues of JAK3. All mutant kinases were active. As shown in Figure 8B, the only mutants to have an effect on SOCS3 -mediated inhibition were two mutants that mutated a three resi due motif ^{Km" 1073} GQM. Mutating this tri- residue motif completely abolished the ability of SOCS3 to inhibit JAK2. Depending on the method of sequence alignment, these three residues correspond to either ^{i043" )045} DVP or ^{I 044} - ^{,0 6} VPA in JAK3. Both the QGM-DVP and QGM-VPA mutants were not inhibited by SOCS3.

The GQM motif in JAK2 is located in the JAK insertion loop (residues 1052-

1073, Figure 8D) as described by Lucet et al. (2006). A similar region (residues 1056- 1070) has been termed the JAK specific insertion (JSI) by Haan and colleagues (2009). The JAK insertion loop lies between helix ctG and ocl and the GQM motif forms the last three residues. More detailed mutagenesis around the GQM motif illustrated that the first and third of these residues, Gly 1071 and Met 1073 , were absolutely required for SOCS3 inhibition, whilst mutation of the central glutamine, Gin 1072, had only a minor effect (Figure 8C). Mutation of He 1074 also had a small effect on the IC _J0 of SOCS3 whilst mutation of Vail 075 did not. All these point mutants were active and there was no significant difference in their specific activity. Examination of the structure of JAK2 (PDB ID: 2B7A, Figure 8E) shows that the GQM motif is solvent exposed, suggesting it may form a direct contact with SOCS3. The GQM motif is unlikely to represent the entire binding surface on JAK, but is rather an essential motif within this surface.

All organisms from insects to mammals contain at least one JAK kinase. Within this set, vertebrates show most complexity in the JAK system, typically containing four JAKs whilst insects such as Drosophila and urochordates such as Ciona contain only one. A sequence alignment of JAKs from within these organisms (Figure 8D) shows that the GQM motif is completely conserved in JAK1, JAK 2 and TYK2 from all vertebrate species listed with the exception of zebrafish JAK2b which contains GQT at this position. Conversely, none of these organisms contained this motif in JAK3 and in fact no conserved sequence within this region of JAK3 at all. The JAK sequences from Drosophila and Ciona show no GQM motif. A sequence comparison of SOCSl and SOCS3 homology mirrors this phenomenon. Once again* vertebrates have SOCS l and SOCS3 homologues and these all have highly similar kinase inhibitory regions. In contrast, the urochordate Ciona has an annotated SOCS 1/2/3, however this is more similar in sequence to SOCS5 and insects contain only SOCS4-7 homologues. Therefore, the GQM motif is specific to JAK1 , JAK2 and TYK2 and is present in all organisms that contain a SOCS protein with a functional KIR.

Example 5. The SOCSl KIR is » more potent inhibitor than the SOCS3 _K m

The effectiveness of the SOCS3 KIR and SOCSl KIR were compared by performing inhibition experiments using SOCS3 and SOCS 1-3. As shown in Figure 9A, SOCS3 and SOCS 1 -3 inhibited the phosphorylation of the gp 130 cytoplasmic domain with IC ₅₀ values of 1.2 +/- 0.1 μΜ and 150 +/- 20 nM, respectively, suggesting that the KIR of SOCSl is a more potent inliibitor of JAK2 than the KIR of SOCS3. A truncated construct of SOCS3, SOCS3 ₂₂ -i85 that lacks the SOCS box, and hence e!onginBC binding, was equally effective at inhibiting JAK2, showing the kinase inhibitory activity of SOCS3 relies solely upon the KIR and extended SH2 domain of SOCS3.

Example 6. Both the KIR and SH2 domain of SOCS3 are required for kinase inhibition whilst phosphotyrosine binding is not

To determine in more detail the regions of SOCS3 required for kinase inhibition a number of mutant SOCS3 constructs were examined. Previous work had highlighted the importance of F25A within the KIR and R71 within the SH2 domain (Nicholson et al., 1999; Sasaki et al., 1999; Yasukawa et al., 1999). In vitro, deletion of the first eight residues in the KIR (residues 22-29), or mutagenesis of F25 and R71 completely abrogated inhibition (Figure 9B). The SOCS3 KIR in isolation, as a synthetic peptide, could not inhibit JAK2, even at concentrations that were lOOx the IC ₅₀ values of the full length protein (data not shown). This indicates that both the kinase inhibitory region and a functional SH2 domain are necessary for inhibition. The requirement for R71, which directly binds phosphotyrosine and is conserved in almost every SH2 domain in the human genome, implies that SOCS3 may bind the phosphor lated activation loop of JAK2 as part of its inhibitory mechanism. However, the addition of a known high affinity ligand (lOOnM) for the SOCS3 SH2 domain, murine g l 307 ₅₀ .7 ₆₄ (Nicholson et al, 2000), even at a 5-fold molar excess, had no effect on JAK inhibition by SOCS3. In addition, the inventors were able to form a ternary complex of JAK2:SOCS3:gpl30 ₇ 5o-764 that eluted as a single peak from a Gel filtration column and contained all three components at a stoichiometric ratio as analysed by rpHPLC. Therefore, whilst R71 may contact JAK2 when bound, phosphopeptide binding is undisturbed.

Example 7. NMR analysis reveals a surface on the SH2 domain of SOCS3, adjacent to the pY binding groove, that interacts with JAK2

NMR was used to map the surface of SOCS3 that binds to JAK2. In order to maximize signal-to-noise SOCS3 ₂₂ . i85 was used rather than the full length protein. This construct has been assigned previously (Babon et al., 2006). The gpl30 phosphopeptide (gpl 30 ₇ 5o- ₇₆ 4-pTyr757), was included as it increases solubility but does not interfere with JAK inhibition. Both 'H-'^ -HMQC and 'H- ^,3 C-HMQC spectra were recorded at 37°C using 250μΜ SOCS +/- 500μΜ unlabelled JAK2 kinase domain in order to map the JAK2 binding surface of SOCS3 by chemical shift perturbation.

In the absence of JAK2, the ^I5 N-HMQC spectrum of SOCS3 was reasonably well dispersed and displayed narrow line- widths, as expected for a 14kDa protein. The addition of a 2-fold molar excess of unlabelled JAK2 resulted in intense line broadening of this spectrum and widespread chemical shift perturbation (Figure 10B), consistent with the formation of ca. 50kDa SOCS3-JAK2 complex. As a negative control, no line broadening or shifting of the SOCS3 spectrum was seen upon addition of JAK2 _G QM-DV _P which is not susceptible to SOCS inhibition. Likewise, SOCS32 -i 8s, which lacks the KIR, showed no interaction with wild-type JAK2. Although the spectrum of the SOCS3/JAK2 complex could not be assigned, due to its large size (>50 kDa) and the fact that it is in slow exchange, the surface of SOCS3 that interacts with JAK2 was mapped by identifying amide resonances from the apo spectra that shift when in complex with JAK2. In order to avoid false positives, only non-overlapped peaks that shift by more than one peak-width (0.1 and 0.6 ppm in the Ή and ^l5 N dimensions respectively), were considered in this analysis. Lys22-Ser29, the first eight residues of the kinase inhibitory region had to be excluded from analysis on this basis (they are unstructured in the absence of JAK2 and hence their resonances reside in the heavily overlapped, random coil region of the spectrum). Despite this limitation, 21 backbone and two sidechain amides were identified that shifted significantly in the presence of JAK2. Several of these shifts were extremely large, for example Ser74 has a chemical shift perturbation of at least: + (Δδι ₅ Ν/5) ² /2)' ^/! . Repeating this analysis on the well-dispersed methyl region of 'H- ^I C-HMQC spectra in the presence and absence of JAK2 identified a further 10 hydrophobic residues whose methyl groups shifted significantly. The combination of these yielded 30 residues whose resonances shift upon interaction with JA 2 (Table 1 ). When mapped onto the crystal structure of SOCS3 (PDB ID 2HMH) they can be seen to form a large surface on SOCS3 that is centred around the extended SH2 subdomain (ESS) helix. In addition to the helix itself, this includes its junction with the SH2 domain proper and three secondary structural elements that are located in close spatial proximity: the N- terminal portion of helix aA, the BC loop and the DE loop.

Table 1 - List of residues of human SOCS3 whose chemical shift was perturbed in the presence of JAK2JH1.

Of the 10 methyl containing residues identified, six are within the ESS helix and five of these have solvent exposed sidechains in the unbound state, making it likely they represent par of the true binding surface. The other residue, Leu41 , forms the junction with the SH2 domain and appears to anchor the ESS helix at a set geometry in relation to the core of the SH2 domain by a number of hydrophobic interactions. This residue contains the most upfield shifted resonance in the SOCS3 spectra, a methyl group at -0.4 ppm at 37°C, due to ring current effects from nearby Tyr47, Phe80 and Phel02. This resonance moves even further upfield in the presence of JAK2, suggesting that a subtle change in conformation in this region moves the Leu sidechain closer to one of these three aromatic groups.

Interestingly, the mapped interaction surface is directly adjacent to one end of the phosphopeptide binding groove on the SH2 domain. This is a long, largely hydrophobic groove, which extends along one entire face of the SH2 domain. Residues that shift upon JAK binding include several (T42-G54) adjacent to the pY-2 pocket as well as several in the BC loop (R71 -D75). This includes R71, the arginine that forms a sal bridge with phosphotyrosine. Nevertheless, it is clear from chromatographic analysis that the gpl 30 phosphopeptide remains bound in the presence of JAK2. In addition, residues that display characteristic chemical shift perturbations when the gpl30 peptide is bound (T89, D107, R94, LI 78) (Babon et al., 2006), maintain these characteristic chemical shift positions in the presence of JAK2. Finally, examination of the JAK2 binding surface (Figure 10) shows that it borders, rather than overlaps, the phosphotyrosine binding groove. In summary, the JAK interaction surface of SOCS3 is broadly defined as containing the KIR, ESS helix and the edge of the phosphotyrosine binding region.

Example 8. The mechanism of SOCS mediated inhibition of JAK kinase activity

As SOCS3 is a non-competitive inhibitor of JAK it cannot function by acting as a pseudosubstrate. As K _M ^ATP and K _M ^pepl,de are unaffected by the presence of SOCS3 it was concluded that SOCS3 is also not an allosteric inhibitor of ATP or substrate binding. Two possibilities were considered as the mechanism of JAK inhibition by SOCS3: (A) SOCS3 prevents product (ADP) release and (B) SOCS3 inhibits phosphate transfer. Mechanism (A) would involve SOCS3 stabilising a JAK-ADP (Enzyme- Product) complex and thus preventing ATP access to the active site. This mechanism is attractive as ADP release is considered the rate limiting step for a number of kinases (Adams, 2001). This mechanism can be addressed by performing single-turnover kinase reactions where [Substrate]«[Enzyme]. In these experiments, each molecule of JAK catalyses only a single phosphorylation event. As ADP is not present at t=0, the enzyme would be unaffected by the presence of SOCS3 during its first (and only) round of catalysis. However, clear inhibition of JAK2 by SOCS3 was observed under these conditions (data not shown) suggesting that SOCS3 does not inhibit via this mechanism. In addition, no synergistic effect was observed when a combination of SOCS3 and ADP were used in standard kinase inhibition experiments (data not shown).

Mechanism (B) was addressed by measuring the ATPase activity of JAK2 in the presence of SOCS3. An ATPase reaction can be considered a phosphorylation event where water, rather than tyrosine, is the phosphate acceptor. If phosphate transfer is prevented by SOCS3 then the ATPase activity should likewise be inhibited. Therefore an ATPase assay was developed using thin-layer chromatography (TLC). Surprisingly, SOCS3 had a small, but reproducible, activating effect on the ATPase activity of JAK2. As shown in Figure 1 1 , in the presence of excess SOCS3, JAK2 catalyses the hydrolysis of ATP at a nearly 2-fold faster rate than in its absence. This activity titrated out with decreasing SOCS3 with an apparent EC50 of 2μΜ. SOCSl -3 could also activate the ATPase activity of JAK2 whilst the F25A mutant, which does not bind to JAK, had no effect. These results indicate that phosphate transfer is not inhibited by SOCS3. Example 9. SOCS3 catalyses ubiquitination of JAK and receptor-a complete model of SOCS3 activity

Although direct inhibition of kinase activity is the dominant mode-of-action of SOCS3 in vivo (Babon et ^' al., 2009; Boyle et al., 2007), the inventors wanted to determine if the same mutations in SOCS3 that prevent inhibition of JAK activity also prevent ubiquitination of JAK by a SOCS3 based E3 ligase. A fully in vitro Cullin5/Rbx2-based ubiquitination system using purified recombinant components was developed. As shown in Figure 12A, SOCS3 induced ubiquitination of GST-JAK2JH1 in vitro in a process that was El, E2, E3 and ATP dependent (left panel). As negative controls GST alone was not ubiquitinated (centre panel) and neither was GST-JAK2JH _I when incubated with an E3 ligase containing only the SOCS3 SOCS box, rather than full length SOCS3 (right panel). CullinS was specifically ubiquitinated in this system, on K724, as described previously (Duda et al., 2008).

Both the KIR, including F25, and the SH2 domain, including R71 , were required for SOCS3 to induce JAKJHI ubiquitination. Neither SOCS3 ^F25A , SOCS3 ^R7,A nor SOCS3 ^R71K could catalyze JAK2JH I ubiquitination (Figure 12C). This suggests that the KIR, in addition to its ability to inhibit JAK kinase activity, is required to bind JAK, forming an integral part of the JAK binding surface. The gp! 30 phosphopeptide did not inhibit SOCS3 ubiquitination of JAK, again illustrating that the pTyr binding groove of the SH2 domain is not required for J AK interaction.

SOCS3 is known to bind to pTyr ⁷⁵⁷ of the gpl 30 receptor cytoplasmic domain. In agreement with this the present inventors showed that SOCS3 catalyzed the ubiquitination gpl 30 _cyt once it was phosphorylated (Figure 12B). gpl 30 ^Y757F was not ubiquitinated under identical conditions confirming that the interaction is mediated by p ^Y757 . As expected, SOCS3 KIR mutants were able to ubiquitinate gpl 30 _cy) as efficiently as wild-type SOCS3.

Conclusions

The discovery that a range of myeloproliferative disorders are caused by dysregulated JAK2 has highlighted the importance of JAK/STAT signalling in homeostasis and pathology. The development of kinase inhibitors is a major component of current drug discovery efforts worldwide. Structure-guided drug design has historically targeted the ATP binding site as the most amenable for inhibitor development however structural conservation of this site across the >500 kinases, including >90 tyrosine kinases, in the human genome makes targeting a single kinase a significant challenge. In addition, intracellular ATP concentrations can be as high as 2 mM, thus the potency of ATP-competitive inhibitors (as well as most allosteric inhibitors) is reduced in vivo. In contrast, the SOCS polypeptides and active fragments thereof described herein are non-competitive kinase inhibitors whose function is to selectively inhibit JAK. For example, SOCS3 inhibits JAK with an IC ₅₀ of 240 nM, even in the presence of 2 mM ATP. Our data shows that SOCS polypeptides and active fragments thereof inhibit JAK2 via a mechanism that is non-competitive towards ATP and substrate. As such it is unaffected by the high intracellular ATP concentrations that hinder the efficacy of current JAK inhibitors and also unaffected by the presence of high concentrations of protein substrates such as members of the STAT and cytokine receptor families. In addition, the fact that SOCS polypeptides and active fragments thereof do not hinder ATP or substrate binding shows that they targe a region that is distal to the ATP and substrate binding sites of JAK1 , JAK2 and TYK2. By binding outside these sites, SOCS polypeptides and active fragments thereof target a region that is likely to show greater structural heterogeneity amongst kinases and thus allow inhibitors that target this site to achieve greater specificity than current drugs. Although steady state kinetics show SOCS3 to act non-competitively, at a molecular level the precise mechanism remains to be determined. Truly noncompetitive inhibitors of single substrate enzymes are rare, indeed, a mechanism that prevents catalysis yet leaves the structure of the active site intact is difficult to envisage. However, in a two substrate enzyme system, an inhibitor that alters the distance between the two substrate binding sites or that alters their distance and/or relative geometry, without changing the structure of either, will be non-competitive. Such a mechanism is the favored model for SOCS3 action as JAK retains its ATPase activity in the presence of SOCS3, showing that phosphate transfer per se, is not inhibited. In fact JAK catalyses the transfer of phosphate to water with increased efficiency in the presence of SOCS3, suggesting water has increased access to the active site. An opening of the active site to allow water in would be consistent with the ATP and substrate binding sites being moved apart. Even a small shift in their relative placements could dramatically effect phosphate transfer from ATP to the tyrosine hydroxyl as these moieties need to be positioned within 3 A to allow nucleophilic attack within the developing transition state.

In summary, SOCS3 binds to JAK and the receptor to which it is attached simultaneously. SOCS3 can then catalyze the ubiquitination of both JAK and receptor but its dominant mode-of-action is to directly inhibit JAK catalytic activity via a non- competitive mechanism (Figure 13): Surfaces on both SOCS3 and JAK responsible for this interaction, as identified by this study, are shown bounded, by a half-circle. SOCS3 can -inhibit JAKi ; JAK2 and TYK2 in isolation but is a much more effective inhibitor when it is tethered to the same receptor as the kinase.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

All publications discussed and/or referenced herein are incorporated herein in , their entirety. \

The present application claims priority from US 61/352,913 filed 9 June 2010 and US 61 /450,593 filed 8 March 201 1, both of which are incorporated herein by reference. Some figures contain coloured representations or entities. Coloured versions of the figures are available from the Patentee upon request or from an appropriate Patent Office. A fee may be imposed if obtained from a Patent Offiee. Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification is solely for the purpose of providing a context for the present invention. It is not to be taken as an admission that any or all of these matters form part of the prior art base or were common general knowledge in the field relevant to the present invention as it existed before the priority date of each claim of this application.

REFERENCES

Adams (2001) Chem. Rev. 101 :2271 -2290.

Babon et al. (2006). Mol. Cell. 22:205-216.

Babon et al. (2009) J. Mol. Biol. 387: 162- 174.Becerril et ai. (1999) Biochem. Biophys. Res. Commun. 255:386-393.

Bergamin et al. (2006) Structure 14: 1285-1292. .

Blondelle and Houghten (1996) Trends Biotechnol 14:60-65.

Boyle et al. (2007) Blood 1 10: 1466-1474.Cho et al. (1993) Science 261 : 1303.

DeWitt et al. (1 93) Proc. Natl. Acad. Sci. USA, 90:6909-6913.

Duda et al. (2008). Cell 134:995-1006.Dunbrack et al. (1997) Folding and Design, 2.-R27-42.

Erb et al. ( 1994) Proc. Natl. Acad. Sci. USA 91 : 1 1422-1 1426.

Gallop et al. ( 1994) J . Med. Chem. 37: 1233- 1251.

Haan et al. (2009) J. Immunol. 182:2969-2977.

Ihle et al. (3996) Philos. Trans. R. Soc. Lond. B. Biol. Sci. 351 : 159-166.

James et al. (2005) Nature 434: 1 144-1 148.

Jia et al. (2008) Curr. Drug Discoy. Technol. 5:59-69.

Lacromque et al. (2000) Blood 95 :2076-2083.

Levine et al. (2005) Cancer Cell 7:387-397.

Lucet ct al. (2006) Blood 107: 176-183.

Naka et al. (1997) Nature 387:924-929. .

Nicholson et ai. (1999) Embo. J. 18:375-385.

Nicholson et a], (2000) Proc. Natl. Acad. Sci. USA 97:6493-6498.

Poul et al. (2000) J. Mol. Biol. 301 : 1 149-1 161.

Sasaki et al. (1999) Genes Cells 4:339-351.

Shuai et al. (1 93) Nature 366:580-583.

Starr et al. ( 1997) Nature 387:917-921.

Wilks and Harpur ( 1 94) Bioessays 16:313-320.

Wilks (2008) Semin. Cell. Dev. Biol. 19:3 19-328.

Williams et al. (2009) J. Mol. Biol. 387:219-232.

Yasukawa et al. (1999) Embo. J. 18:1309-1320.

Yoshimura et al. (1995) EMBO J. 14:2816-2826.

Zheng et al. (2002) Nature 416:703-709.

Zuckermann et al. (1994) J. Med. Chem, 37:2678-2685.