HIF-1 and HIF-2 inhibitors

Abstract

The present invention relates to inhibitors of HIF-1 and HIF-2 and uses thereof. The present invention further relates to the inhibitors for use in treatment of diseases. An isolated polypeptide is provided, that prevents dimerization of HIF-1α with HIF-1β and HIF-2α with HIF-1β and/or inhibits the activity of HIF-1 and HIF-2, wherein the polypeptide comprises the amino acid sequence C-X1-X2-X3-Z-X4 (SEQ ID NO: 1) and wherein X1, X2, X3 and X4 are any amino acid and wherein Z is leucine, valine of isoleucine or a non-natural derivative or leucine, valine or isoleucine. The isolated polypeptide prevents dimerization of HIF-1α with HIF-1β and HIF-2α with HIF-1β and inhibits the activity of HIF-1 and HIF-2 by binding to HIF-1α or HIF-1β and/or HIF-2α or HIF-1β.

Claims

1. An isolated polypeptide that specifically inhibits or prevents dimerization of HIF-Iα (hypoxia inducible factor-1α) with HIF-Iβ and HIF-2α with HIF-Iβ and/or inhibits the activity of HIF-1 and/or HIF-2, wherein the polypeptide comprises the amino acid sequence CXZZZF (SEQ ID NO: 7), wherein: X is arginine, or lysine; Z is leucine, valine, or isoleucine; and F is phenylalanine; wherein any one or more of X, Z and F may be substituted with a non-natural derivative as follows: X may be a non-natural derivative of arginine or lysine; Z may be a non-natural derivative of leucine, valine, or isoleucine; and F may be a non-natural analogue thereof; and wherein the polypeptide is a cyclic polypeptide.

2. The isolated polypeptide according to claim 1, wherein the polypeptide consists of the sequence CKLIIF (SEQ ID NO: 2), CRLLIF (SEQ ID NO: 3), or CRVIIF (SEQ ID NO: 4).

3. The isolated polypeptide according to claim 1, wherein the polypeptide consists of the sequence CKLIIF (SEQ ID NO: 2).

4. The polypeptide according to claim 1, wherein the polypeptide has a sequence selected from the group consisting of CKLIIF (SEQ ID NO: 2), CRLLIF (SEQ ID NO: 3), and CRVIIF (SEQ ID NO: 4).

5. The polypeptide according to claim 1, wherein the non-natural analogue is a D amino acid.

6. The polypeptide according to claim 1, wherein one or more of the amide bonds in the peptide backbone is modified or replaced with an isostere.

7. The polypeptide according to claim 1 wherein the non-natural analogue of F is selected from the group consisting of 4-Me-F, 4-F—F, 4-Cl—F, 4-CN—F, 4NO-.sub.2-F—F, 4-MeO—F, 3-NO.sub.2—Y, NaI, PaI, and Cha.

8. A pharmaceutical composition comprising the polypeptide of claim 1.

9. The pharmaceutical composition of claim 8, wherein the pharmaceutical composition further comprises a chemotherapeutic agent.

10. A method of inhibiting the activity of HIF-1 and/or HIF-2 in vitro, wherein the method comprises contacting HIF-1 and/or HIF-2 in vitro with the polypeptide according to claim 1.

Description

(1) There now follows by way of example only a detailed description of the present invention with reference to the accompanying drawings, in which;

(2) FIG. 1 shows the thermal shift results for cyclic peptide inhibitors disrupting HIF-1α/HIF-1β and HIF-2α/HIF-1β interactions.

(3) FIGS. 2A and 2B show ELISA results for cyclic peptide inhibitors disrupting HIF-1α/HIF-1β and HIF-2α/HIF-1β interactions. A) inhibitors were assayed for ability to disrupt the HIF-1α/HIF1-β protein-protein interaction. B) inhibitors were assayed for the ability to disrupt the HIF-2α/HIF1-β protein-protein interaction.

(4) FIGS. 3A-3C show MST results for cyclic peptide inhibitors binding to HIF-1α and HIF-2α, and CRVIIF (SEQ ID NO 4) affecting the dimerization constant of HIF-2α/HIF-1β. A) HIF-1α, B) HIF-2α, C) the K.sub.d of the HIF-1α/HIF1-β protein-protein interaction increases in the presence of 50 μM CRVIIF.

(5) FIGS. 4A-4C exemplify one strategy for improving the potency of the identified molecules. A) Structure of cyclo-CRLLIF (SEQ ID NO 3). B) Potential sites for generation of analogues. X=atom, e.g. S or O; R and R.sup.1=any atom or functional group, e.g. H, Me, Et; R.sup.2=replacement of the LLI motif with any combination of aliphatic amino acids (natural or non-natural) e.g. valine, leucine, homoleucine, isoleucine. R.sup.3=any substituted phenyl, cyclohexyl, naphthyl, pyridyl, or other ring. Amino acids may be L or D in any or all positions, as well as the reverse sequence. The arginine may be replaced with lysine, ornithine or other nitrogen containing non-natural amino acids.

(6) C) The derivatives generated in this study.

(7) FIGS. 5A-5B show MST results for the most potent derivatives of CRLLIF (SEQ ID NO: 3) binding to HIF-1α and HIF-2α. A) Binding affinity of given derivative for the PAS-B domain of HIF-1α. B) Binding affinity of given derivative for the PAS-B domain of HIF-2α.

(8) FIG. 6 shows a fluorescently tagged derivative of CRVIIF (SEQ ID NO 4) binding to the PAS B domain of HIF-2α by ELISA (upper panel) and fluorescence polarization (lower panel).

(9) FIGS. 7A-7B. A) Residues on the PAS B domain of HIF-2α that are affected by CRVIIF (SEQ ID NO 4), as identified by N.sup.15 NMR. B) A ribbon diagram of 2α PASB with CRVIIF (SEQ ID NO 4).

(10) FIG. 8 shows ribbon (left panel) and surface models (middle and right panel) showing the proposed binding site of CRVIIF (SEQ ID NO 4).

(11) FIGS. 9A and 9B show the results of alanine scanning of CRVIIF (SEQ ID NO 4), assessed by ELISA. A) assessed by HIF-1α/HIF1-β ELISA. B) assessed by HIF-2α/HIF1-β ELISA.

(12) FIGS. 10A-D and 10A′-D′ show the results of cell permeability and localisation of fluor-CRVIIF (SEQ ID NO 4) by microscopy. Compound (B and B′) appears to cross cell membrane, and either localises inside vesicles (C and C′), in the cytoplasm, or crosses into the nucleus (D and D′). White areas in panels B and B′ show fluorescently labelled CRVIIF; white areas in panels C and C′ show vesicles; and white areas in panels D and D′ show the nuclei.

(13) FIGS. 11A-11D show the structures of cyclic versions of SEQ ID NO: 5 (A); SEQ ID NO: 3 (B); SEQ ID NO: 4 (C); and SEQ ID NO: 2 (D).

EXPERIMENTAL RESULTS

(14) 4 cyclic peptide hits were identified using a SICLOPPS screen to identify cyclic peptide inhibitors of HIF1 and HIF2.

(15) Thermal shift data of thioredoxin-tagged longer constructs and PASB domains with the identified polypeptides are shown in FIG. 1. Compounds that are effective inhibitors decrease the melting temperatures of the proteins. Compound C (CRVIIF) (SEQ ID NO 4) is most potent as can be seen from the downwards shift in the melting curve.

(16) ELISA with thioredoxin-tagged HIF-1α/2α and GST-tagged HIF-1β (bHLH-PASA-PASB domain constructs) was used to determine whether compounds inhibit dimerisation of HIF heterodimer in vitro. Data is shown in FIG. 2. CKLIIF (SEQ ID NO 2) was found to be the most potent, of the 4 identified cyclic peptides in this assay, for inhibiting the binding of HIF-1β to 1α/2α. IC.sub.50 values for the compound=3.7 μM/8.8 μM for 1α/2α.

(17) TABLE-US-00001 showing the melting temperatures with each of compounds A-D. ΔT.sub.m (° C.) Compound Trx-1α Trx-2α GST-1β 1α PASB 2α PASB 1β PASB 1 mM CRLLIF 0.5 −4 −2 −4 −1 (SEQ ID NO 3) 1 mM CKLIIF −1.5 −2.5 −1 −2.5 −0.5 (SEQ ID NO 2) 1 mM CRVIIF −1 −2 −0.5 −1.5 −2.5 −0.5 (SEQ ID NO 4) 1 mM SGWEMIQRR −2.5 −2 −0.5 −2.5 −0.5 (SEQ ID NO 5)

(18) ELISA assays were conducted using 1α/2α PASB domains. The results, shown in FIG. 2, shows comparable IC.sub.50 values to longer Trx-tagged constructs (5.4 μM vs. 3.7 μM for 1α, 15.2 μM vs. 8.8 μM for 2α). The results in FIG. 2 suggest PASB domain dimerisation is the main target of inhibition by CRVIIF (SEQ ID NO 4).

(19) Microscale thermophoresis (MST) was performed to determine K.sub.d of peptide binding to fluorescently labelled HIF-1α/2α PASB domains. The results are shown in FIG. 3.

(20) MST competition assay was used to measure inhibition of heterodimerisation: 1β PASB domain titrated into 2α PASB domain. The binding affinity of these two proteins was determined in absence and presence of 50 μM CRVIIF (SEQ ID NO 4): K.sub.d=378 nm in absence of CRVIIF (SEQ ID NO 4), 3.7 μM in presence of CRVIIF (SEQ ID NO 4).

(21) The data shows that CRVIIF (SEQ ID NO 4) shifts the binding equilibrium between HIF-2α and 1β PASB domains as seen in FIG. 3, right hand panel.

(22) Improving the affinity of the identified cyclic peptides using non-natural amino acids: FIG. 4 exemplifies one strategy that may be used to improve potency of the identified molecules.

(23) FIG. 5 shows the potency of the identified molecules may be significantly improved by substituting one of the amino acids (F) for a natural (W) or non-natural equivalent (e.g. naphthyl). Full list of molecules synthesized, and characterization data by ELISA and MST are shown in table 3.

(24) Validation of fluor-CRVIIF (SEQ ID NO 4): point ELISA shows similar inhibition of dimerisation compared to unlabelled CRVIIF as shown in FIG. 6.

(25) The fluorescent tag appears not to interfere with binding or inhibition of dimerisation. Fluorescence polarisation (FP) assay: titration of 2α PASB into 2 μM fluor-CRVIIF (SEQ ID NO 4) gives K.sub.d of 18.5 μM.

(26) FIGS. 7A-7B show NMR using 15N labelled 2α PASB+/−CRVIIF (SEQ ID NO 4). Slight shifts can be seen from the 15N labelled backbone amides in presence of CRVIIF (SEQ ID NO 4). Backbone amide peaks were assigned to residues using double labelled 13C 15N 2α PASB. Backbone amide peaks that shifted in presence of CRVIIF (SEQ ID NO 4) correspond mainly to residues surrounding the internal water cavity. This suggests CRVIIF (SEQ ID NO 4) induces a conformational change that affects binding of water molecules inside the cavity.

(27) In the absence of detailed NMR data, a potential binding site of CRVIIF (SEQ ID NO 4) was proposed based on the published crystal structures of 2α-1β PASB domains. CRVIIF (SEQ ID NO 4) appears to mimic a region of 1β at the α-β binding interface. 2 isoleucine residues and a phenylalanine point into hydrophobic pockets on the surface of 1α/2α, an arginine residue forms a water-mediated bond to Arg258/260 of 1α/2α, and Leu243/245 of 1α/2α forms hydrophobic protrusion towards the central ring of the CRVIIF (SEQ ID NO 4) motif.

(28) This suggests CRVIIF (SEQ ID NO 4) may act in direct competition with 1β. 1α and 2α are almost identical at the proposed binding site (FIG. 8). The only difference is Val336 in 1α is a methionine in 2α, which may explain slightly better IC.sub.50 and K.sub.d values obtained for 1α vs. 2α.

(29) ELISA using alanine scan peptides with 1α/2α PASB domains and GST-1β identified residues 1, 4, 5 of CRVIIF (SEQ ID NO 4) as critical for inhibiting dimerization (FIG. 9), i.e. inhibitory activity is reduced when these are mutated to alanine. Residues 2, 3 and 6 also important but to a lesser extent. Residues 4 and 5 (both isoleucines) fit well with the proposed binding site model, where 2 Ile residues point into hydrophobic pockets on the surface of 1α/2α.

(30) Site-directed mutagenesis was used to make mutations in 1α/2α PASB domains at the proposed binding site, and corresponding domain-swapped mutations in 1β to still allow heterodimerisation. Mutations should inhibit binding of CRVIIF (SEQ ID NO 4) to 1α/2α if the proposed binding site is correct.

(31) FIGS. 10A-D and 10A′-D′ show cell permeability tested using fluorescently labeled-CRVIIF (SEQ ID NO 4) in MCF-7 cells. Compound (B and B′) appears to cross cell membrane, and either localises inside vesicles (C and C′), in the cytoplasm, or crosses into the nucleus (D and D′).

(32) FIGS. 11A-11D show the structure of the identified cyclic peptides.

(33) Materials and Methods

(34) Reagents from Fisher/Sigma unless specified otherwise.

(35) Constructs for Protein Expression

(36) The expression constructs used to produce proteins for this work are listed in Table 2.

(37) TABLE-US-00002 TABLE 2 Constructs for protein expression Protein ID Domain Vector Cloning sites Construct 1α PASB 238-349 pET28 BamH1/Sac1 His.sub.6-Met[Protein] 2α PASB 240-350 pET28 BamH1/Sac1 His.sub.6-Met[Protein] 1β PASB 356-474 pET28 Nde1/Sac1 His.sub.6-[Protein] Trx-1α  1-350 pET32 BamH1/Eag1 Thioredoxin-His.sub.6-[Protein] Trx-2α  1-360 pET32 BamH1/Xho1 Thioredoxin-His.sub.6-[Protein] GST-1β   1-474* pGEX-2TK BamH1/EcoR1 GST-[Protein] 2α PASB .sup.R260A 240-350 pET28 BamH1/Sac1 His.sub.6-Met[Protein] *Isoform 2, Δ77-91
Site-Directed Mutagenesis

(38) Site-directed mutagenesis was performed following a QuikChange protocol (Agilent) using either 2× Phusion HF Master Mix (NEB) or Pfu polymerase (Promega) following the manufacturers' recommended conditions for thermal cycling. PCR products were treated with 1 μL Dpn1 (NEB) for 1 hr at 37° C., then 5 μL of each reaction were used to transform 50 μL chemically competent DH5α via heat shock (42° C., 45 sec). Recovered cells were plated on LB agar plates supplemented with 50 μg/mL kanamycin. Individual colonies were picked and sequenced (Eurofins) to confirm the desired mutation.

(39) Protein Expression

(40) The proteins used in this study were produced and purified per previously published protocols (e.g. Miranda E. et al., J. Am. Chem. Soc., 2013, 135 (28), 10418-10425; and Scheuermann T. H. et al., Nat. Chem. Bio., 2013, 9 (4), 271-276).

(41) Thermal Shift Assay

(42) Thermal denaturation of HIF proteins was performed in 20 μL reactions in triplicate using a Bio-Rad CFX Connect Real-Time System. Proteins were used at either 5 μM (Trx-1α, Trx-2α and GST-1β) or 15 μM (His-tagged PASB domains) in HIF assay buffer, with 10 mM DTT and SYPRO orange dye (Life Technologies) at a final concentration of 8× (relative to 5000× stock), and supplemented with either DMSO or compound dissolved in DMSO at the stated concentrations. The reaction plate was assembled on ice, then transferred to the pre-chilled thermal cycler, held at 10° C. for 5 min (heated lid, 105° C.), then the temperature ramped to 95° C. Fluorescence was recorded at 0.5° C. increments, holding for 5 sec at each temperature. Data were plotted using GraphPad Prism 6, and melting temperatures calculated using the first derivative method in which the peak (i.e. the steepest transition between data points) represents the point at which 50% of the protein sample is denatured. Errors are quoted as the inherent error associated with the first derivative of the data, i.e. ±0.25° C. for T.sub.m values and ±0.5° C. for ΔT.sub.m values.

(43) ELISA

(44) 25 μL of His-tagged HIF-α protein (either Trx-1α, Trx-2α, 1α PASB, or 2α PASB) at 0.1 μM in HIF assay buffer were added to individual wells of a Pierce Ni.sup.2+-coated 96-well plate (Thermo) in triplicate. The plate was rocked gently at room temperature for 1 hr, before washing with 3×200 μL assay buffer (5 min incubation per wash). 25 μL of compounds and/or DMSO at the stated concentrations were added to the plate and allowed to equilibrate for 30 min prior to the addition of 25 μL of GST-1β at 0.2 μM in assay buffer for 1 hr. For conditions in which no compound or DMSO were used, GST-1β was added as a 0.1 μM solution. After incubation with GST-1β, the wells were washed with 3×200 μL assay buffer then 1×200 μL PBS. The wells were then incubated with 100 μL mouse α-GST antibody (Thermo, MS-707-P) at a 1/1000 dilution in PBS supplemented with 0.1% Tween-20 (PBST) and 2% non-fat powdered milk (Marvel) for 1 hr, followed by 3×200 μL washed with PBS. 100 μL of secondary HRP-conjugated α-mouse antibody (GE Healthcare) was added at a 1/6000 dilution in 2% milk/PBST for 1 hr, then washed with 3×200 μL PBS. Wells were incubated with 100 μL of 1-Step Ultra TMB-ELISA solution (Thermo) until a blue colour developed (typically 10-15 min) and the reaction was stopped by the addition of 100 μL of a 1/5 dilution of conc. H.sub.2SO.sub.4, which turned the solution yellow. Absorbance was read at 450 nm using a Tecan Infinite M200 Pro plate reader. Data were fitted and IC.sub.50 values determined by non-linear regression using GraphPad Prism 6.

(45) Microscale Thermophoresis

(46) 0.5 mL of 1α or 2α PASB protein stocks in HIF storage/assay buffer were dialysed against 1 L of MST labelling buffer (100 mM NaHCO.sub.3 pH 8, 150 mM NaCl and 5% glycerol) for 1.5 hr at 4° C., then against 1 L of fresh labelling buffer at 4° C. overnight. 100 μL of each dialysed protein diluted to 10 μM were incubated with 100 μL of 30 μM NT-647-NHS dye (NanoTemper protein labelling kit) for 20 min in the dark, then exchanged into 650 μL HIF assay buffer supplemented with 0.05% Tween-20 using the supplied gravity flow purification columns. A dilution series of cyclo-CRVIIF (SEQ ID NO 4) or alanine scan analogues (in HIF assay buffer, 10% DMSO) was mixed with equal volumes of labelled 1α PASB or 2α PASB at 100 nM and allowed to equilibrate for 10 min. Samples were loaded into hydrophilic capillaries and, after optimising the power, MST was performed using a Monolith NT.155 instrument (NanoTemper). Data were analysed using NT Analysis Software using the Temperature Jump analysis mode.

(47) For competition assays, NT-Red-labelled 2α PASB (100 nM) was mixed with equal volumes of a dilution series of unlabelled 113 PASB (67 μM stock), then each 9 μL sample was supplemented with 1 μL 500 μM cyclo-CRVIIF (SEQ ID NO 4) (50% DMSO) to give final concentrations of 45 nM 2α PASB, 50 μM cyclo-CRVIIF (SEQ ID NO 4), 5% DMSO, and variable 1β PASB (30 μM highest concentration). MST was performed and data analysed as described above.

(48) Fluorescence Polarisation

(49) K.sub.d values were determined by fluorescence polarisation (FP) using fluorescently labelled cyclo-CRVIIF (fluor-CRVIIF) (SEQ ID NO 4) dissolved in DMSO as the probe. In each condition, the concentration of the probe was kept constant (2 μM, 1% DMSO) whilst proteins were titrated. Reactions were performed in HIF assay buffer with the addition of 0.1% Triton X-100. Samples were assembled in black 96-well plates (40 μL/well, in triplicate) and allowed to equilibrate for 2 hr in the dark. Plates were read using a POLARstar Omega plate reader (BMG Labtech) at Ex: 490 nm, Em: 520 nm, with target gain set to 300 mP. Polarisation data were plotted and K.sub.d values calculated by non-linear regression using GraphPad Prism 6.

(50) Competition Assay

(51) A competition assay may be employed for the identification of other molecules that competitively bind to HIF-1α or HIF-2α by measuring displacement of the fluor-CRVIIF (SEQ ID NO 4) from the protein.

(52) Cell Culture

(53) For initial compound permeability experiments, 1.5 mL MCF-7 cells were plated on 3 cm petri dishes (200,000 cells/plate) and grown at 37° C. After 24 hr, cells were incubated with fresh media containing the stated concentrations of fluor-CRVIIF (SEQ ID NO 4) (0.5% DMSO) or 66 μM CP-61 control compound (0.66% DMSO) for a further 24 hr. Cells were washed twice with PBS and imaged live in plates. For co-localisation experiments, an 8-well Lab-Tek Chamber Slide (Nunc) was treated with collagen prior to plating 25,000 MCF-7 cells/well. After 24 hr, cells were dosed with 300 μL fresh media containing the stated concentrations of fluor-CRVIIF (SEQ ID NO 4) (0.5% DMSO) and/or 5 μg/mL FM 4-64FX Membrane Stain (Life Technologies) and incubated for a further 24 hr. Cells were washed with 2×300 μL PBS, fixed with 150 μL 10% formalin in PBS for 5-10 min, then washed again with PBS. Growth chambers were removed and nuclear staining was achieved using Vectashield Mounting Medium with DAPI (Vector Laboratories) prior to sealing the slide with a coverslip. Cells were imaged using a fluorescent microscope.

(54) Peptide Synthesis

(55) Cyclic peptides were synthesized by solid phase peptide synthesis, as detailed in Miranda et al. J. Am. Chem. Soc., 2013, 10418.

(56) Derivatives

(57) Derivatives of the above compounds may also be used in the present invention. These are as detailed below with cyclo-CRLLIF (SEQ ID NO 4) as an example but similar derivatives may be made of any of the compounds described herein provided that the derivatives inhibit the activity of HIF-1 and HIF-2.

(58) The derivatives may be derivatives of:

(59) C-X1-X2-X3-Z-X4 (SEQ ID NO 1) wherein X1, X2, X3 and X4 are any amino acid;

(60) CRVIIF (SEQ ID NO: 2);

(61) CRLLIF (SEQ ID NO: 3);

(62) CKLIIF (SEQ ID NO: 4); or

(63) SGWEMIQRR (SEQ ID NO: 5), wherein the L amino acids are replaced with D amino acids and the sequences are reversed.

(64) The derivatives may comprise the same motif of CXZZZF (SEQ ID NO 7) where X=arginine or lysine (and so can be replaced with any natural or non-natural derivative of these two amino acids), Z=leucine, valine, or isoleucine and these amino acids or non-natural derivatives, as well as aliphatic derivatives can be placed in any of these 3 positions in any combination. There may also be derivatisation at the phenylalanine with non-natural analogues and/or attachment of a moiety to the sulphur. The cysteine may also be replaced with natural or non-natural amino acids.

(65) Data showing activity of the above molecules (FIGS. 10A-D and 10A′-D) in HIF-1 and HIF-2 ELISA (showing the compounds disrupt dimerization of these protein-protein interactions) as well as the binding constants determined by microscale thermophoresis (MST) for binding to the PASB domain of HIF-1α and HIF-2α.

(66) TABLE-US-00003 TABLE 3 HIF-1 and HIF-2 inhibitory activity of CRLLIF (SEQ ID NO 3) and its analogues shown in FIG. 4. CRLLIX IC.sub.50 (μM) IC.sub.50 (μM) K.sub.d (μM) K.sub.d (μM) where in HIF-1 in HIF-2 for HIF-1α for HIF-2α X= ELISA ELISA by MST by MST F 71.32 98.27 36.93 15.43 4-Me-F 11.9 6.07 9.61 5.27 4-F-F 37.46 11.47 6.84 7.41 4-Cl-F 13.16 4.88 2.8 4.15 4-CN-F 44.24 6.27 10.01 9.46 4-NO2-F 27.76 12.04 3.73 5.76 4-MeO-F 36.69 15.34 10.79 4.5 3-NO2-Y 19.5 22.74 90.66 49.03 Nal 15.11 11.03 3.6 3.73 Pal 293.8 212.4 69.23 45.33 Cha 25.49 19.83 1.37 5.63 CRLLIW 7.7 9.6

(67) Activity of the molecules improves with a fluorophore attached to the cysteine (in this case FITC-maleimide bound to the cysteine). But this can be extended to any bulky aromatic or aliphatic residue in that position.

(68) Activity as measured by a HIF-2 ELISA is as follows for fluorescently tagged molecule:

(69) FITC-CRLLIF—0.47 μM (SEQ ID NO 3)

(70) FITC-CKLIIF—0.77 μM (SEQ ID NO 2)

(71) FITC-CRVIIF—1.46 μM (SEQ ID NO 4)

(72) The above molecules, or derivatives with non-natural amino acids may be used for the discovery of new HIF-1 and HIF-2 inhibitors using a fluorescence polarization assay.