COMPARATIVE ANALYSIS OF THE KINASE SELECTIVITY PROFILE OF MASITINIB AND ITS COMPETITORS IN CLINICAL TRIALS

SUMMARY

The knowledge of the selectivity of protein kinase inhibitors (PKIs) is a critical point for the development of optimal safe and well tolerated compounds in human health, particularly for the treatment of non-lethal inflammatory diseases. Two complementary studies, recently published in Nature Biotechnology, investigated the selectivity of all the kinase inhibitors approved or under clinical development, as well as various investigational ones (Anastassiadis, Deacon et al.; Davis, Hunt et al.). From these studies, it appears that the PKI with the higher kinase selectivity is masitinib from AB Science. Indeed, during clinical trials masitinib has thus far shown an excellent safety record, permitting its clinical development in both oncology and inflammatory diseases.

Masitinib Kinase Selectivity Profile Page 1

INTRODUCTION

The human kinome (Manning, Whyte et al. 2002), which comprises 518 protein kinases, is among the most important pools of therapeutic targets since 150 of them have been shown, or were proposed, to be involved in the etiology and/or the progression of various human diseases including: cancer, inflammatory, metabolic, cardiovascular, or neurodegenerative diseases. Small molecule kinase inhibitors are of considerable therapeutic interest and during the last decade a growing number of therapeutic agents targeting one single kinase (targeted therapy) or several kinases (multi-targeted, polypharmacology) were developed with an impressive success. This is illustrated by the imatinib (Gleevec) story, the first tyrosine kinase inhibitor ever developed, which provided the proof-of-principle that kinase inhibitors can be efficient drugs in human health. To date tens of kinase inhibitors are under clinical development, with the great majority of them targeting the kinase ATP-binding site. Since there are more than 500 different kinases and thousands of ATP-binding proteins, ATP site-targeted compounds have a great potential for cross-reactivity. Indeed, clinical trials have demonstrated that kinase inhibitors are responsible for numerous adverse events, in particular cardiac dysfunctions. It is now well recognized that selectivity is a critical issue for the safety of PKIs and the knowledge of off-target interactions and the assessment of molecular specificity of this class of compound is a key feature for development of the safest drugs.

Masitinib, AB Science

Masitinib (AB1010), developed by AB Science, is a new, orally available, tyrosine kinase inhibitor that primarily targets the proto-oncogene c-Kit and several others kinases. The inhibition of c-Kit permits the development of masitinib both in oncology, by direct inhibition of the activated oncogene, and in inflammatory diseases by the reduction of c-Kit-dependent mast cells, key players of inflammatory processes. In human health, masitinib is currently evaluated in eight phase III clinical trials in both oncology and inflammatory disease fields (Table 1). In most of these indications, masitinib is being evaluated against competitors. In animal health, masitinib (Masivet® in Europe; Kinavet® in USA) was the first veterinary anti-cancer targeted drug approved by EMA in 2008 followed by the FDA in 2010, for treatment of canine mast cell tumors (MCT), a common cutaneous malignant neoplasm in dogs.

Table 1: Clinical development of masitinib in human and veterinary medicine and status of its competitors SPONSOR COMPOUND INDICATION STATUS Human Medicine AB Science Masitinib Smoldering Systemic, Indolent Systemic and Cutaneous Mastocytosis with Handicap Phase III AB Science Masitinib Phase III Novartis Imatinib Gastrointestinal Stromal Tumors Approved Pfizer Sunitinib Approved Novartis Nilotinib Phase III

AB Science Masitinib Phase III AstraZeneca Fostamatinib Rheumatoid Arthritis Phase III Pfizer Tofacitinib Phase III AB Science Masitinib Phase III Melanoma Carrying a Mutation in the Juxta Membrane Domain of c-Kit Novartis Nilotinib Phase III AB Science Masitinib Phase III

Advanced/Metastatic Pancreatic Cancer in Combination with Gemcitabine OSI Pharmaceuticals Erlotinib, Tarceva FDA Approved AB Science Masitinib Multiple Myeloma Phase III

AB Science Masitinib Severe Persistent Asthma Phase III AB Science Masitinib Primary Progressive, Secondary Progressive or Relapse Free Multiple Sclerosis Phase III Veterinary medicine AB Science Masitinib Approved Mast cell tumors Pfizer Toceranib,Palladia Approved

Masitinib Kinase Selectivity Profile Page 2

Masitinib’s competitors under clinical development (Phase III) or already approved

Human medicine

In rheumatoid arthritis, masitinib had two competitors: Fostamatinib (R406), from Rigel Pharmaceuticals and tofacitinib (CP-690550) from Pfizer. Fostamatinib is an orally available compound with potential anti-inflammatory and immunomodulating activities currently developed as an inhibitor of Syk tyrosine kinase. Tofacitinib is developed as a JAK2/JAK3 inhibitor in inflammatory diseases (rheumatoid arthritis but also ulcerative colitis and psoriasis). It acts as a potent immuno-suppressor by suppression of the JAK-STAT signaling pathway. On 16th November 2011, Jakafi® (ruxolitinib, INCB18424) from Incyte Corporation was approved by the FDA for treatment of patients with intermediate or high-risk myelofibrosis (MF), including primary MF, post- polycythemia vera MF and post-essential thrombocythemia MF. Jakafi is an oral JAK1 and JAK2 inhibitor with a mechanism of action similar to that of tofacitinib. This compound is not a direct masitinib competitor but, as a JAKs inhibitor evaluated in inflammatory diseases (psoriasis Phase II), it was included to the comparative analysis of kinase selectivity profiles. In oncology, masitinib is being developed in GIST, melanoma with c-Kit mutation and pancreatic carcinoma. In these indications masitinib’s competitors are imatinib, sunitinib and nilotinib; nilotinib; and erlotinib, respectively (see table 1). Nilotinib (Tasigna™) is developed by Novartis as a second generation BCR-ABL inhibitor, the key cause and driver of Ph+ CML, either in patients with newly diagnosed Ph+ CML or in pretreated imatinib-resistant or -intolerant patients. Because nilotinib equally inhibits c-Kit, this activity had lead Novartis to develop it in c-Kit driven-malignancies, for example, GIST and melanoma with mutated c- Kit. Sunitinib (Pfizer) is a multitargeted tyrosine kinase inhibitor that inhibits primarily VEGFRs 1/2/3 but also numerous RTKs such as platelet derived growth factor receptor (PDGFR), c-Kit, FLT3 or RET. This compound was approved by FDA as antigiogenic compound (anti VEGFRs) for the treatment of renal cell carcinoma (RCC) and for GIST resistant or intolerant to Gleevec. Erlotinib specifically targets the epidermal growth factor receptor (EGFR) tyrosine kinase, which is highly expressed and occasionally mutated in various forms of cancer. In 2004, the FDA approved erlotinib for treatment of patients with locally advanced or metastatic non-small cell lung cancer after failure of at least one prior chemotherapy regimen. In 2005, the FDA approved erlotinib in combination with gemcitabine for treatment of locally advanced, unresectable, or metastatic pancreatic cancer.

Veterinary Medicine

In the treatment of dog mast cell tumors, masitinib has one competitor, toceranib (Pfizer), a sister molecule of sunitinib with which it has identical properties. Toceranib is a multitargeted tyrosine kinase inhibitor that inhibits primarily VEGFRs 1/2/3 but also numerous RTKs including c-Kit.

Masitinib Kinase Selectivity Profile Page 3

METHODS

For several years, the analysis and the quantification of PKIs selectivity was extensively investigated by different complementary in vitro technologies known as chemical proteomics (Kim and Sim 2010; Rix and Superti-Furga 2009). These experimental approaches can be divided in two groups: functional assays and competition assays. The functional assays measure the remaining kinase enzymatic activity in presence of an inhibitory compound, a specific substrate and ATP. The competition assays measure the ability for a compound to compete with an ATP-site specific ligand in absence of ATP and substrate. Several others experimental approaches can be used such as thermal enzyme stabilization by inhibitor binding (thermal shift assays). Davis et al. performed a competition study using the technology developed by Ambit Bioscience (Fabian, Biggs et al. 2005) and known as the KINOMEscan screening platform, which permits the measurement of the binding affinity (Kd) between PKI and a particular kinase. They have extended their previous study of the interaction of 38 PKIs with 317 kinases (Karaman, Herrgard et al. 2008) by including 34 additional inhibitors and 125 additional kinases. The kinase panel comprises most of the human protein kinases and their disease relevant mutants but also lipid kinases and kinases from human pathogens. Briefly, E. coli or mammalian cell-expressed kinases labeled with DNA tag for real-time polymerase chain reaction (qPCR) readout were mixed with the test compound and a known active site binding ligand immobilized on a solid support. Once the binding equilibrium was reached the solid support was washed, in order to remove unbound kinase, and the remaining DNA-tagged bound kinase was quantified by qPCR and compared to control as illustrated in Figure 1.

Figure 1: Schematic representation of the binding study method developed by Ambit Biosciences

A B C D

Compounds that bind the kinase active site prevent kinase binding to the immobilized ligand and then reduce the amount of kinase captured on the solid support (Panels A & B). Conversely, test molecules that do not bind the kinase have no effect on the amount of kinase captured on the solid support (Panel C). Remaining bound kinase is quantified in test versus control samples by using a qPCR method that detects the associated DNA label (Panel D). Dissociation constants (Kds) for test compound-kinase interactions are calculated by measuring the amount of kinase captured on the solid support as a function of the test compound concentration

In their study, Anastassiadis et al. performed a classical functional phospho-transfer enzymatic assay to quantify the inhibitory effect of 178 commercially available PKIs against 300 recombinant kinases. They have quantified the remaining activity of the kinases at given compound and ATP concentrations (0.5 µM and 10 µM, respectively). It is interesting to note that there is an excellent correlation between the results of the in vitro assays used in the Anastassiadis et al. and in the Davis et al. studies by comparison to the direct proteomic approach based on the LC-MS/MS analysis of PKIs protein ligands in whole cell extracts. Most of the previously unsuspected interactions that were detected by one technology were

Masitinib Kinase Selectivity Profile Page 4 confirmed by the other one. For example, the identification of LCK or DDR1 as imatinib targets or TEC kinases as dasatinib targets (Bantscheff, Eberhard et al. 2007; Hantschel, Rix et al. 2007; Rix, Hantschel et al. 2007) and GAK as an erlotinib target (Brehmer, Greff et al. 2005), were all detected independently either by in vitro quantitative binding or enzymatic assays or by proteomic technologies (Karaman, Herrgard et al. 2008).

RESULTS

« Comprehensive analysis of kinase inhibitor selectivity”, Davis MI, Hunt JP, Herrgard S, Ciceri P, Wodicka LM, Pallares G, Hocker M, Treiber DK, Zarrinkar PP. Nat Biotechnol. 2011 Oct 30;29(11):1046-51.

Based on the determination of the Kd values, Davis et al. have determined an absolute selectivity scores: the selectivity score for the binding interactions with Kd<3 µM, i.e. S(3 µM), that was calculated as follows:

S(3µM)=number of interaction with Kd<3 µM/ number kinases tested

Figure 2: Selectivity scores of masitinib and of its competitors in clinical trials

0.8

0.6

0.4 S(3µM)

0.2

0.0

IMATINIB NILOTINIBERLOTINIB SUNITINIB

MASITINIB AB-1010 FOSTAMATINIB R406 TOFACITINIB CP-690550RUXOLITINIB INCB18424

The graphical representation of S(3µM) (Figure 2) show clearly that masitinib has an equivalent selectivity to that of imatinib and tofacitinib. All others compounds are less selective. This is strikingly illustrated in Figure 3, which represents the kinome interaction map of each of the nine compounds compared here.

Masitinib Kinase Selectivity Profile Page 5

Figure 3: Kinome interaction maps for masitinib and its competitors in clinical development

AB1010 = masitinib; CP-690550 = tofacitinib; INCB18424 = ruxolitinib; R406 = fostamatinib.

Toceranib was not directly evaluated in this study but it can be considered that toceranib is absolutely equivalent to sunitinib. The comparison of the chemical structures of sunitinib and toceranib represented in Figure 4A show that they differ only by one chemical motif (N-diethyl versus pyrrolidine). This minor difference has no impact on their kinase selectivity profile. Indeed, the comparative KINOMEscan analysis of masitinib and toceranib performed at 1 µM concentration (Figure 4B), respectively reveals a pattern identical to that of that of masitinib and sunitinib kinome interaction maps shown in Figure 3. The selectivity scores S(1), S(10) and S(35), used in this case, were calculated using the proportion of control (%Ctrl) as a potency threshold. These selectivity scores are defined as follow:

Masitinib Kinase Selectivity Profile Page 6

S(35) = (number of non-mutant kinases with %Ctrl <35)/(number of non-mutant kinases tested); S(10) = (number of non-mutant kinases with %Ctrl <10)/(number of non-mutant kinases tested); S(1) = (number of non-mutant kinases with %Ctrl <1)/(number of non-mutant kinases tested). As for the previously defined Kd values, they provide a quantitative method to describe the selectivity of the two compounds.

Figure 4: Comparative kinome interaction maps for masitinib and toceranib, its competitor in veterinary medicine

A. B.

Sunitinib

Toceranib

Masitinib Toceranib

C.

A comparison of the selectivity scores demonstrates without any doubt that masitinib is more selective than toceranib (Figure 4C).

CONCLUSION: From the results of Davis et al., and from our unpublished data (toceranib), it appears clear that masitinib is as selective as imatinib and tofacitinib, and that all others competitors are less selective or totally unselective.

However, there are two notable limitations to the study of Davis et al.:

First, the quantification of kinase compound interaction (Kd determination) is not strictly correlated to the inhibition of the catalytic activity as illustrated in figure 5A. Effectively, the comparison of kinase-compound pairs inhibition data from this study with the overlapping ones from large-scale binding studies show that 90% of high affinity kinase-compound interaction (Kd<100 nM) correspond to a functional inhibition (>50%). Moreover, when considering weak affinity kinase- compound interaction (Kd<1000 nM) only 13% showed >50% inhibition. The same finding is made when considering the variation of the melting temperature (Tm) induced by the binding of an

Masitinib Kinase Selectivity Profile Page 7 inhibitor to its target kinase. In general, compounds that induce a Tm variation >4°C are inhibitory compounds but the analysis of disposable data show a significant proportion of both false-positive and false negative (Figure 5B).

Figure 5: Comparison of functional inhibition data from Anastassiadis et al. study with previous kinase-inhibitor interaction profiling studies; (a) binding studies (Kd); (b) thermal shift studies (Tm Shift)

Second, the selectivity score (defined as the number of kinase bound divided by the total number of kinase tested) depends upon an arbitrary hit threshold (xµM), and it appears that selectivity scores generated by the same dataset but using different hit thresholds can produce different ranking of compounds, particularly for compounds of closest selectivity. This point is illustrated in table 2, where masitinib and its competitors are ranked either using the S(300 nM) or the S(3 µM) Kd thresholds.

Table 2: Ranking of PKIs by their selectivity scores at two different Kd thresholds (300 nM and 3 µM).

S(300nM) ranking S(3uM) ranking

Compound S(300nM) Compound S(3uM) CP-690550 0.0207 Imatinib 0.0570 Imatinib 0.0233 AB1010 0.0622 Erlotinib 0.0285 CP-690550 0.0622 AB1010 0.0337 Nilotinib 0.1244 Nilotinib 0.0440 Erlotinib 0.1813 INCB18424 0.0803 INCB18424 0.2565 Sunitinib 0.3109 Sunitinib 0.5959 R406 0.3420 R406 0.6088 AB1010 = masitinib; CP-690550 = tofacitinib; INCB18424 = ruxolitinib; R406 = fostamatinib.

The experimental method and the results analysis used in the second article from Nature Biotechnology (Anastassiadis et al.) permit a response to these two limitations.

Masitinib Kinase Selectivity Profile Page 8

« Comprehensive assay of kinase catalytic activity reveals features of kinase inhibitor selectivity », Anastassiadis T, Deacon SW, Devarajan K, Ma H, Peterson JR. Nat Biotechnol. 2011 Oct 30;29(11):1039-45.

In this study only five of masitinib’s seven competitors were evaluated; those missing being, ruxolitinib and fostamatinib. However, these are all unambiguously less selective than masitinib when considering results from the Davis et al. study, with their selectivity scores higher than that of masitinib at all Kd thresholds considered, i.e. S(300 nM) or S(3µM). Moreover, the internal AB Science functional data (IC50, not shown) have confirmed the weak selectivity of these three compounds. The results obtained from inhibition experiments of each kinase-compound pair were analyzed with a metric for kinase inhibitors based on the Gini coefficient. The application of this statistical method of analyze to PKIs was described in detail by P. Graczyk (Graczyk 2007). In summary, the total inhibition for a given compound is calculated as the sum of magnitudes of inhibition for all kinases tested. After sorting of the kinases in order of increasing inhibition, the cumulative fraction of total inhibition is plotted against the cumulative fraction of kinases. If all kinases are inhibited to an equal extent, the cumulative fraction of total inhibition increases linearly with the cumulative fraction of kinases. This situation of complete lack of selectivity corresponds to the black diagonal line in Figure 6. Conversely, if a compound only strongly inhibits a small number of kinases, the cumulative fraction of total inhibition will initially increase slowly following a Lorenz curve and steeply increase to 1 for the last small fraction of potently inhibited kinases (Figure 6). Considering that the area between the diagonal line and Lorenz curve is A and the area under the Lorenz curve is B, the Gini coefficient (G), is defined as follow: G=A/(A+B), taking account that A+B=0.5, the Gini coefficient equals G=1-2B. The Gini coefficient reflects, on a scale from 0 to 1, the manner in which the degree of aggregate inhibitory activity of a compound is directed against a single target (G=1) or distributed equally against all kinase tested (G=0).

Figure 6: Lorenz Curves Gini scores for masitinib and its competitors at 0.5 µM compound/10 µM ATP.

1.0

0.8

0.6 MASITINIB, G=0.81 NILOTINIB, G=0.78 IMATINIB, G=0.77 0.4 TOFACITINIB, G=0.73 A ERLOTINIB, G=0.70 SUNITINIB, G=0.52 0.2 lack of selectivity, G=0

Cumulative fraction of total inhibition B 0.0 0.0 0.2 0.4 0.6 0.8 1.0 Cumulative fraction of kinases

Masitinib Kinase Selectivity Profile Page 9

From these curves it appears clear that masitinib is the most selective compound of this panel since it had the smaller B area and by consequence the higher Gini coefficient. The full results of the study (178 inhibitors against 300 kinases, available on line http://kir.fccc.edu) are represented in the Figure 7, where the PKIs are ranked by Gini score.

Figure 7: Ranked list of kinase inhibitors sorted by Gini scores as a measure of inhibitor selectivity

Remarkably, when considering the full inhibitor panel (178 compounds), masitinib remains the most selective compound with the highest Gini score.

CONCLUSION

Two independent large scale studies of kinase inhibitor selectivity have been recently published in the journal Nature Biotechnology. From these data it appears clear that masitinib, developed by AB Science, is to date the most selective kinase inhibitor under clinical development or already approved.

Masitinib Kinase Selectivity Profile Page 10

BIBLIOGRAPHY

Anastassiadis, T., S. W. Deacon, et al. "Comprehensive assay of kinase catalytic activity reveals features of kinase inhibitor selectivity." Nat Biotechnol 29(11): 1039-45. Bantscheff, M., D. Eberhard, et al. (2007). "Quantitative chemical proteomics reveals mechanisms of action of clinical ABL kinase inhibitors." Nat Biotechnol 25(9): 1035-44. Brehmer, D., Z. Greff, et al. (2005). "Cellular targets of gefitinib." Cancer Res 65(2): 379-82. Davis, M. I., J. P. Hunt, et al. "Comprehensive analysis of kinase inhibitor selectivity." Nat Biotechnol 29(11): 1046-51. Fabian, M. A., W. H. Biggs, 3rd, et al. (2005). "A small molecule-kinase interaction map for clinical kinase inhibitors." Nat Biotechnol 23(3): 329-36. Graczyk, P. P. (2007). "Gini coefficient: a new way to express selectivity of kinase inhibitors against a family of kinases." J Med Chem 50(23): 5773-9. Hantschel, O., U. Rix, et al. (2007). "The Btk tyrosine kinase is a major target of the Bcr-Abl inhibitor dasatinib." Proc Natl Acad Sci U S A 104(33): 13283-8. Karaman, M. W., S. Herrgard, et al. (2008). "A quantitative analysis of kinase inhibitor selectivity." Nat Biotechnol 26(1): 127-32. Kim, D. H. and T. Sim (2010). "Chemical kinomics: a powerful strategy for target deconvolution." BMB Rep 43(11): 711-9. Manning, G., D. B. Whyte, et al. (2002). "The protein kinase complement of the human genome." Science 298(5600): 1912-34. Rix, U., O. Hantschel, et al. (2007). "Chemical proteomic profiles of the BCR-ABL inhibitors imatinib, nilotinib, and dasatinib reveal novel kinase and nonkinase targets." Blood 110(12): 4055-63. Rix, U. and G. Superti-Furga (2009). "Target profiling of small molecules by chemical proteomics." Nat Chem Biol 5(9): 616-24.

Masitinib Kinase Selectivity Profile Page 11

© 2011 Nature America, Inc. All rights reserved. unoptimized compounds unoptimized libraries large screening from information complement patterns optimized with that can achieved be interaction characteristics the into insight provide fore interest of targets specific against optimized been domains. kinase protein human nonatypical active, against a panel of 442 assays kinase representing > d inhibitor kinase small-molecule approved currently 11 inhib known 72 of total a encompass to set data the DiscoveRx DiscoveRx Corporation, San Diego, California, USA ( NATURE BIOTECHNOLOGY NATURE 3 1 available became molecules small with kinome to approaches other and assays of kinase panels large became only compounds of class this for reactivity e The family. kinase protein the of members multiple inhibito kinase small-molecule of majority vast The Mindy I Davis analysis Comprehensive of kinase inhibitor selectiv a ‘DFG-out’, inactive kinase conformation kinase inactive ‘DFG-out’, a ATP the to adjacent pocket binding a contact which filing data and to quantify selectivity quantify to and data filing ana objective an score’ facilitate to of a ‘selectivity and introduced domains, of >50% human kinase protein ass kinase 317 of panel a against inhibitors kinase compounds. these study to resource common a provided has and kinome the across are currently in clinical trials, has revealed dive profiling of known inhibitors, including compounds We tested the interaction of 72 kinase inhibitors w Michael Hocker guaranteed with a type II binding mode. The compound mode. a II with binding type guaranteed m binding I type a with achieved Thi be may tested. selectivity compounds selective least the among are inhibitors are among the most selective, whereas tw selecti overall in widely vary pocket, this contact loop activation the of conformation ‘DFG-out’ a require inhibito I type contrast, In expected. as selective Received Received 10 May; accepted 30 August; published onli toxicity, as well as for studying the structural ba kinases for which no dedicated inhibitors exist. It a single subfamily of kinases, but selective outsid the compounds tested. Analysis of the interaction p this trend. The data further illustrate that select data show that, as a class, type II inhibitors are direct direct enabling potential of genomics for drug deve These These authors contributed equally to this work. Cor Ambit Biosciences, San Diego, California, USA. The compounds tested here represent mature inhibitors th inhibitors mature represent here tested compounds The We previously described the interaction patterns of 1–3 1 , , Daniel K Treiber , , Jeremy P Hunt 8–10

ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE . We show that most type II inhibitors, inhibitors, II type most Wethat . show 7 . We now update and extend extend and update Wenow . 1–3 2 Present Present addresses: National Institutes of Health, B 1,2 respondence respondence should be addressed to P.P.Z. (pzarrink J.P.H., P.C., L.M.W., G.P., D.K.T.) and Blueprint M 11 , , Sanna Herrgard ne ne 30 October 2011; rse rse interaction patterns , are indeed relatively relatively indeed are , & Patrick P Zarrinkar 1–6 80% 80% of catalytically lysis of kinase pro- of kinase lysis . Systematic kinase kinase . Systematic vity. Several type I type Several vity. o o type II inhibitors rs, which do not not do which rs, . The data there- data The . sis sis of the observed interaction patterns. Our findi ive ive inhibitors have been developed against the majo a set of 38 known more more selective than type I inhibitors, but that the compounds, and compounds, that have been or ays representing representing ays rs interact with with interact rs itors, including including itors, also provides a foundation for further exploring k interrogate the the interrogate and selectivity selectivity and e e that subfamily. The data set suggests compounds t ode and is not not is and ode site and prefer prefer and site apparent once once apparent xtent of cross- of xtent ith ith 442 kinases covering >80% of the human catalyti lopment lopment and basic research about cellular signaling s shows that that shows s atterns atterns reveals a class of inhibi ‘group-selective’ set contains contains set comprising comprising the concept the concept and do not not do and rugs, tested tested rugs, further further at have have at doi:10.1038/nbt.1990 1,3 , , Pietro Ciceri iae, 1 ii knss nt onig uat vari mutant counting (not from kinases kinases lipid 11 kinases, includes also panel The variants). mutant d counting kinase distinct (363 domains kinase protein human catalyt of >80% representing assays 442 yielded has doma kinase noncatalytic and variants mutant linked such as mTOR, and kinases from human pathogens, as atypical several kinases, lipid PI3K-family for assays relevance, biological and therapeutic high their to tuberculosis We assays binding competition used kinases protein for set assay A comprehensive RESULTS globa address reliably not does do selectivity, compound been frequently has as target, intended primary, closely kinases against compounds testing Therefore, family or group wit is therefore not always predictive o Selectivity group. targeted the of outside tive compounds that interact broadly with one kinase group a o class reveals or groups subfamilies major kinase of selec analysis a quantitative Furthermore, kinases. inhibito selective develop to possible indeed is it intended, targets primary of the compounds tested. T selective inhibitors for the majority of the 28 kin kinase domains in the eight major ‘typical’ groups, as groups, ‘typical’ major eight the in domains kinase emphasiskinome. was on buildingthe assays for catalytically across a compounds of testing direct and enable would that panel assay a biochemical develop ains n to iae oan blee t b nonc be to believed domains kinase ( two and variants h asy ae t gnrt a out vriw f what of overview robust a generate to panel assay the Supplementary Table 1Supplementary 1,2 We screened a diverse set of 72 known kinase inhibi kinase known 72 of set diverse a Wescreened , 7 activation-state variants, 49 disease-relevant disease-relevant 49 variants, activation-state 7 , ethesda, ethesda, Maryland, USA (M.I.D.), KINOME Plasmodium falciparum Plasmodium 1–3 edicines edicines Corporation, San Diego, California, USA (P [email protected]) or [email protected]) D.K.T. ([email protected]). , Lisa , M Lisa Wodicka ). ngs ngs will help to realize the re re are important exceptions to inase inase inhibitor biology and tors tors broadly active against rity rity of kinases targeted by 3 to undertake a 10-year effort to a effort to 10-year undertake and one from from one and o o use as tools to study 1,2 . c c protein kinome. Our , , Gabriel Pallares ases ases that represent the E C R U O S E R ctive ctive human protein scan rs for a diversity of of diversity a for rs f f overall selectivity. hin a kinase sub- kinase a hin f f ‘group-selective’ we also included included also we his his suggests that tivity across the across tivity Mycobacterium Mycobacterium protein kinases kinases protein Division of comprehensive comprehensive ic, nonatypical nonatypical ic, defined well as disease- l selectivity.l ne to estimate estimate to ne seven atypical atypical seven ins. The effort effort The ins. , , but are selec- related to the the to related The primary primary The omains, not not omains, tors against against tors .P.Z.). ity ants), two two ants), types of of types atalytic atalytic mutant mutant 12 1,2 . Due . Due , 1 © 2011 Nature America, Inc. All rights reserved. affinity for phosphorylated and nonphosphorylated v nonphosphorylated and phosphorylated for affinity variants of ABL1 ( ABL1 of variants ( as described in the text, and compounds were binned were compounds and text, the in described as 2 2 o be can patterns ( interaction molecule–kinome small 1Figure E C R U O S E R panel at a single concentration (10 (10 concentration single a at panel We a of variety each com kinases. screened initially fo optimized of inhibitors a cross-section represents ( compounds ta intended primary, the among are subfamilies, kinase as as th well groups kinase ‘typical’ of six eight the have been published previously published been have mined using biochemical enzyme activity assays (Table 3 assays activity enzyme biochemical using mined stants measured here generally agree well with publ compound/kinase combinations queried. As before num the by assessed results, published previously the to points pared data of number the in increase 2.5-fold a to set data extended and updated The analyses. and ies a resource as used and accessed be readily may that an systematic single, a presenting of interest the in targets, and determined a quantitative dissociation c dissociation a quantitative determined and targets, each interaction observed in this primary screen (Table4 screen primary this in observed interaction each kinases tested ( tested kinases previouslydescribed (386, after queried excluding mutant and activation bound bound with numb the dividing by compound each for scores tivity kinome selectivity, interaction maps kinome ( global of overview an provide To inhibitorsII type I and type of Selectivity of 72, or 64%), S(3 S(3 64%), or 72, of ATP site itself from the pocket exploited by type II inhibitors, wi ATP the to adjacent pocket allosteric an bind inhibitors from most k other kinase the target may distinguish a structu exploits compounds selective highly of these (Tyk lapatinib inhibitor ERBB2/EGFR the and GW-2580 CSF the SGX-523, inhibitor MET the CI-1040, and 886 AZD-62 inhibitors MEK for the observed were selectivity, ther and scores, selectivity lowest The kinases. of fr large a with interact to known is which itself, staurosporine analog staurosporine the are outliers The >0.8. scores out promiscuous highly two of exception the with 0.7, and most of the remaining compounds are broadly distributed between 0.2 b a Supplementary Table 2Supplementary

) Distribution of selectivity for 37 compounds clas 37 compounds for of selectivity ) Distribution Percent of all compounds 10 20 30 40 50 60 0 0–0.1 and and ). Data for 40 of the compounds against smaller assay assay smaller against compounds the of 40 for Data ). Quantitative distribution of kinome-wide selectivi kinome-wide of distribution Quantitative 0.1–0.2 K

Supplementary Fig. 1 Fig. Supplementary 0.2–0.3 Supplementary Table 3 Supplementary d 14 < 3 ; SGX-523 ; requires SGX-523 a conformation unique inactive of Fig. 1a Fig. 0.3–0.4 Supplementary Table 5Supplementary M M M M by the total number of kinase distinct domains 0.4–0.5 M) < 0.2, indicating that they bind <20% of the the of <20% bind they that indicating 0.2, < M) 7 (S(3 (S(3 and and S(3 0.5–0.6 ). At least 28 different kinases, representing representing kinases, Atdifferent ). 28 least Supplementary Fig. 2 µ M) M Supplementary Table 5 Supplementary

M)). For the majority of compounds (46 of For M)). compounds majority the 0.6–0.7

7,13 0.7–0.8 and are included once more here here more once included are and ). M 0.8–0.9 M) to identify candidate kinase kinase candidate identify to M) ). The compound set therefore therefore set compound The ). ). (). c

) Distribution of selectivity for 13 compounds clas compounds 13 for selectivity of ) Distribution 0.9–1 ) and calculated selec- sified as type I inhibitors based on relative bindi on relative based I as inhibitors type sified b ariants of ABL1 ( ABL1 of ariants thout contacting the

according to their scores. ( scores. their to according Percent of type I inhibitors e atypical and lipid e atypical 10 20 30 40 50 60 ished ished values deter- 7 pound against the pound efore the greatest greatest the efore d unified data set set data unified d 0 , , the binding con- state variants), state as variants), for further stud- further for inases. The MEK inases. r activity against against r activity ). The scores for for scores The ). Supplementary Supplementary Supplementary Supplementary 0–0.1 represents close close represents onstant (onstant available com- available ral feature that feature ral ty of compounds. A kinome selectivity score (S(3 (S(3 score selectivity A kinome compounds. of ty CEP-701 and and CEP-701

rgets of these these of rgets 0.1–0.2 er of kinases kinases of er we prepared prepared we site, distinct distinct site, 1R inhibitor inhibitor 1R erb). Each Each erb). 44/ARRY-

liers with with liers 0.2–0.3 bserved bserved K panels panels action action ber of ber d

Supplementary Table 5Supplementary 0.3–0.4 ) ) for

0.4–0.5 S(3 0.5–0.6 te tp I ihbtr wl lkl rqie struct require likely will inhibitors II type other scores. Understanding why these compounds behave dif cal for cutoffs affinity of range a with t repeated is when outliers remain compounds two These kinases. a with interact which respectively, 0.49, and 0.44 inhibitors, EXEL-2880/GSK1363089 and AST-487, with with AST-487, and EXEL-2880/GSK1363089 inhibitors, compound set as set a (compound whole for responsible the ob compounds bias toward selective 3 Fig. Supplementary on a 300 nM cutoff affinity (S(300 nM)) ( observed when the analysis was repeated using si selectivity A scores based <0.2. scores had inhibitors II type the of two con In whole. a as set compound the for observed range The scores for I type inhibitors are fairly d evenly mation specific to a target kinase target a to specific mation inactiv the for selectivity optimizing for opportunities active canonical the in than states kinase inactive tors is that there may be greater he conformational unusual displacement of the alpha-C helix alpha-C the of displacement inhib unusual II type typical a not although lapatinib, and the the paired ABL1 variants with Of ing the modes. 72 50 tested, bin compounds (69%) rule out the possibility that some inhibitors have ki can we robust, is method classification this Although our in included variants mutant ABL1 five and ABL1 forms nonphosphorylated and phosphorylated the for affi binding used therefore we study current the in variant of ABL1 (ref. 18). To classify functionally for affinity modest least at exhibit but inhibitors prim not are that compounds for even inhibitors), I (type th those from inhibitors) II (type conformation kinase out’ an prefer that compounds differentiate functionally o forms nonphosphorylated and phosphorylated to ing diff that shown We however. recently have performed, II inh I and of type type set a diverse across tivity ( phosphorylated state and therefore were classified f preference marked a have 13 remaining The inhibitors. class were and state nonphosphorylated the for ence little exhibit 37 compounds, 50 these Of data. these on the the kinase activation loop tivity scores separately for type I and type II inh II type and I type for separately scores tivity ized ized by an requiring ‘DFG-out’ inactive conformation Supplementary Table 5Supplementary µ M) One One of the advantages potential frequently noted for a

) Distribution of selectivity for the entire set of set entire the for selectivity of ) Distribution 0.6–0.7

ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE 0.7–0.8

0.8–0.9 ).

sified as type II inhibitors based on relative bind relative on based inhibitors II type as sified 0.9–1 ng affinity for phosphorylated and nonphosphorylate and phosphorylated for ng affinity c ). The type II inhibitors therefore are largely largely are therefore inhibitors II type The ). Percent of type II inhibitors 10 20 30 40 50 60 ). We then plotted the distribution of of selec- distribution Wethe ). plotted then 15 0 Fig. 1a Fig.

; ; GW-2580 is a type II inhibitor, character- 0–0.1 M M)) was calculated for each compound compound each for calculated was M))

K 0.1–0.2 d < 3 ). ). Importantly, are II there two type 11,17 0.2–0.3 M . A broad comparison of selec- of comparison broad A . M M and could be classified based 0.3–0.4 Supplementary TableSupplementary 5

NATURE BIOTECHNOLOGY NATURE 0.4–0.5 16 . S(3

72 compounds. compounds. 72 0.5–0.6 µ istributed across the istributed the the inhibitors tested M) ibitors has not been not has been ibitors ABL1 or a mutant mutant a or ABL1 as as type II inhibitors culating selectivity selectivity culating nase-specific bind- nase-specific ibitors ( ibitors terogeneity terogeneity among 0.6–0.7 milar pattern was was pattern milar ural studies, and and studies, ural state, providing providing state, large number of of number large itor, requires an an requires itor, nities measured measured nities inactive, ‘DFG- inactive, of the kinase d d at one least of not completely completely not

ified as type I type as ified 0.7–0.8 type type II inhibi- e-like confor- e-like erential bind- erential served for the served ferently ferently from or no prefer- no or of wild-type wild-type of assay panel. panel. assay trast, all but but all trast, f ABL1 can can ABL1 f or the non- the or he analysis analysis he arily ABL1 ABL1 arily S(3 Fig. 1b Fig. 0.8–0.9 at do not not do at ing ing M

M) of of M) 0.9–1 and , c 11 ). ). d ; © 2011 Nature America, Inc. All rights reserved. of the the of target is shown for the most selective compound tar compound selective most the for shown is target Table 5 for selectivity absolute greatest the with compound Table 1 a a 300 nM cutoff to Saffinity calculate was the analysis when was observed pattern A similar fr and are most the tested, of compounds the >60% with LCK and Y hit, whereas and are frequently least the compound with the greatest relative selectivity for selectivity relative greatest the with compound NATURE BIOTECHNOLOGY NATURE 3Figure with bound the pounds dividing by kinase each for scores selectivity observation these To two. quantify or one only with ot whereas many tested, with of compounds the interact some are there as reactive, broadly are others and just selective that set data the from apparent is It kinasesof Selectivity t compounds selective least the among are AST-487 and EXEL-288 inhibitors II type the whereas inhibitors, an as selective as are BIBW-2992, and (tofacitinib) inclu inhibitors, I type Several selectivity. achieve to n selectivity, high guarantee not does mode binding t a that highlight data However, the inhibitors. I type than tive confirms that, in general, type II inhibitors are m ing mode, kinase conformation and selectivity. Our interplay the into insights important provide should 2Figure screened (S screened ERK2 and TRPM6, bind only one compound each with with each compound one only bind TRPM6, and ERK2 at with one Three least compound. interacting kinase of kinases interacting with 10–40% of the compounds is fairly selectivities of kinase distribution overall their scores.their binned kinases and binds, kinase each compounds of text the in described as panel assay the in kinases A selectivity score (S score A selectivity here (here a Primary target p38-alpha VEGFR2 IKK-beta PIK3CG PIK3CA PRKCH AURKB AURKA ERBB2 CSF1R Supplementary Table 3Supplementary IGF1R MTOR EGFR MEK1 CDK2 BRAF K AKT1 ABL1 PLK1 JAK3 JAK2 FLT3 MET ). The S(3 S(3 The ). SYK RET ALK d ). There is generally good agreement between our re between agreement good is generally ). There Selective inhibitors for primary targets. Each kin Each targets. primary for inhibitors Selective selectivity compound of distribution Quantitative KIT for the primary target for each compound targeting compound each for target primary the for 0 kinase (3 (3 M M) is shown for the most selective compound targeti compound selective most the for shown is M) BMS-540215 GSK-461364A TG-100-115 GDC-0941 VX-745 SGX-523 AZD-6244/ARRY-886 CP-690550 INCB018424 MLN-120B CHIR-258/TKI-258 Lapatinib BIBW-2992 GW-2580 GDC-0879 MLN-8054 Imatinib kinase AZD-1152HQPA R547 M Sunitinib Crizotinib Number ofofftargetswith M)) ( M)) K 03 050 40 30 10 d < 3 3 < (3 (3 M GSK-1838705A PP-242 Fig. 2 Fig. ). The most selective compound for each target is s is target each for compound selective most The ).

GSK-690693 M)) was calculated for each of the 442 442 the of each for calculated was M)) M ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE M by the total number of compounds compounds of number total the by M and and 20 kinase Supplementary Table 1Table Supplementary K d (300 (300 nM) (

≤ 10 that target was identified by counting the number number the counting by identified was target that , based on the number number the on , based × that target was identified, using the S(3 S(3 the using identified, was target that

geting each of the kinases shown. ( shown. kinases the of each geting K ore likely to be selec- d primarytarget narrow, >60% with according to to according some kinases that that kinases some SK4 each interact SK4 interact each analysis analysis therefore number of com- of number ding CP-690550 CP-690550 ding Supplementary Supplementary tested, and each compounds are are compounds y of the type II II type the of y s, we calculated calculated we s, or is it required required it is or kinases, ERK1, ERK1, kinases, ase on the the on ase 0/GSK1363089 0/GSK1363089 between bind- between repeated using repeated of kinases. kinases. of Vandetanib Staurosporine R406 the kinase. The number of kinases bound with with bound kinases of number The kinase. the equently hit. equently hers interact interact hers ested here.ested K d sults and sults < 3 3 < ). The The ). ype II II ype y axis is a primary, intended target of one or more more or one of target intended a primary, is axis M M, M, ng each of the kinases shown. kinases the of each ng 110 184 here (Primary Target 1 in Target in 1 (Primary here compou the of targets primary the represent to sidered collect that kinases distinct 28 are There question. begi to suited well are they such, As target. primary agai optimization significant of result the are cases inhibitor mature staurosporine, of exception the with u compounds The inhibitors. selective identify to ficult f kinases of subset significant a is there whether or fo inhibitors selective reasonably develop to possible address to difficult been has that question major One kinases many for inhibitors Selective experiments. the of scale compound the of nature different the reflect i likely hit are kinases individual which with frequencies pounds against much smaller panels of kinasesof panels smaller much against pounds un of collections different very of screens primary sing in previously observed selectivities kinase the b

hown. (hown. Primary target

p38-alpha Percent of kinases VEGFR2 10 12 14 16 18 20 IKK-beta PIK3CG PIK3CA PRKCH AURKB AURKA 0 2 4 6 8 ERBB2 CSF1R IGF1R MTOR EGFR

MEK1 0–0.05 CDK2 BRAF AKT1 PLK1 ABL1 JAK3 JAK2 FLT3 MET SYK RET ALK KIT b a ) Absolute selectivity. For each of the primary tar primary the of each For selectivity. ) Absolute ) Relative selectivity. For each of these primary t primary these of each For selectivity. ) Relative

. . . . 1 0.8 0.6 0.4 0.2 0 0.1–0.15 M M) as a measure of absolute selectivity ( selectivity absolute of a measure as M)

0.2–0.25 Imatinib GSK-461364A PI-103 VX-745 SGX-523 AZD-6244/ARRY-886 AB-1010 CP-690550 MLN-120B MLN-518 Lapatinib BIBW-2992 GW-2580 AT-7519 GDC-0879 AZD-1152HQPA MLN-8054 PTK-787 TG-100-115 GSK-1838705A GSK-690693 of kinases bound with a with bound kinases of

0.3–0.35 Vandetanib Supplementary Table 3 Supplementary INCB018424 K d within tenfold of that for the primary primary the for that of tenfold within

Crizotinib 0.4–0.45 S kinase

0.5–0.55 S(3 PP-242 (3 compounds in the set tested tested set the in compounds µ µ M) M) 0.6–0.65 K d within tenfold or better better or tenfold within R406 8,9 E C R U O S E R

Staurosporine 0.7–0.75 . Differences in the the in . Differences ). Of these, 27 are are 27 these, Of ). ively may be con- be may ively or which it is dif- is it which or Supplementary Supplementary le-concentration le-concentration nst an intended, intended, an nst n to address this this address to n optimized com- optimized s tested and the the and tested s n these studies studies these n r most kinases, kinases, most r gets the the gets is whether it is whether is argets, the the argets, s that in most most in that s 0.8–0.85 nd set tested tested set nd sed here are, are, here sed

0.9–0.95 3 © 2011 Nature America, Inc. All rights reserved. were generated using TREE using generated were from the S(3 S(3 the from one group-selective and one non-group-selective com non-group-selective one and group-selective one lip atypical, excluding (thereby panel assay the in (S(3 scores selectivity and >0.4), 4 4 with bound kinases First, we two approaches. for counted ou compound each in compounds by selectively which targeted are determine kinases To panel. assay the in represented 4Figure E C R U O S E R domain sequence. The bars in each bar graph indicat graph bar each in bars The sequence. domain target the compound with the greatest absolute selectiv absolute greatest the with compound the target selectivity score (S(3 (S(3 score selectivity low the with compound the target primary for each mined comparable to that for the intended, primary target. kina other five than fewer bound that set our tor in at was there kinases, target For primary 27 of 17 the compoundpri with the each greatest relativefor selectivity for identified thereby and target, primary comparable to that for the primary target, and an S an and target, primary target) the for that primary to the comparable than other (kinases off-targets fe both featured that compound one least at was there in in our S(3 with set For 16 of the 27 primary target kinases, there was primary target for each compound. Dashed lines sign lines Dashed compound. each for target primary d c b a XL28/S-338 KW-2449 EXEL-2880/GSK-1363089 Group-selective compounds. Compounds were divided divided were Compounds compounds. Group-selective T71 SB-203580 AT-7519 adtnbCHIR-258/TKI-258 Vandetanib ioii GSK-1838705A Nilotinib M M) = 0.1–0.2 bin. ( bin. = 0.1–0.2 M) M M) M) < For 0.1. 15 kinases, of target the 27 primary K M d M)), and thereby identified for each primary primary each for identified thereby and M)), within tenfold of the the of tenfold within spot software ( software M c M)) were calculated for each compound for the kinas the for compound each for calculated were M)) ) Compounds from the S(3 S(3 the from ) Compounds http://www.kinomescan.com/ id and CK1 kinases). Shown here are the kinase inte kinase the are here Shown kinases). CK1 and id K d for the compound’s the for at least one inhibitor that target ( ify the overall S(3 S(3 overall the ify pound from each selectivity bin. ( bin. selectivity each from pound e S(3 e S(3 ses with affinities affinities with ses Second, Second, we deter- least one inhibi- least mary target the the target mary S(3 µM) S(3 µM) S(3 µM) S(3 µM) the number of the number r set, we used used we set, r 0.2 0.4 0.6 0.8 1.0 (3 (3 0.2 0.4 0.6 0.8 1.0 with affinity affinity with M 0.1 0.2 0.3 0.4 0.5 0.6 0.1 0.2 0.3 0.4 0.5 0.6 wer than five five than wer 0 M) for the individual kinase groups. Red bars indic bars Red groups. kinase individual the for M) 0 ity ( ity 0 0 of these 27 27 these of M M est overall overall est G AKCG te T KTKL TK STE Other CMGC CAMK AGC G AKCG te T KTKL TK STE Other CMGC CAMK AGC G AKCG te T KTKL TK STE Other CMGC CAMK AGC M) = 0.2–0.4 bin. ( bin. = 0.2–0.4 M) M) < 0.1. 0.1. < M) into selectivity bins based on their overall select overall their on based bins selectivity into G AKMCOhrSET TKL TK STE Other CAMKCMGC AGC Fig. 3b Fig. Fig. 3a Fig. ) and display a circular representation of the kina the of representation a circular display ) and M EXEL-2880/GSK-1363089 M) for each compound. each for M) ). ). ). Vandetanib compounds’ kinase group preferences that is difficu is that preferences group de kinase objective an compounds’ provides approach quantitative This ac distributed is pattern interaction the kinase whether or more or one targets r preferentially scores compound selectivity group-specific these of print, among compounds that share a primary target w primary vary a share can that subfamilies compounds or among groups kinase intera various of the pattern the that noted previously have We patterns interaction of fingerprint A quantitative targets. kinase of diversity a for results s reasonably these develop to possible kinases, is it that of suggest selection unbiased an her not examined targets primary 27 of set the Although each of the major kinase groups kinase major the of each individual selectivity scores, again using a 3 3 a using again scores, w selectivity quantitatively, individual compounds of patterns interaction Nilotinib AT-7519 d ) Compounds from the S(3 S(3 the from ) Compounds ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE a ) Compounds from the S(3 S(3 the from ) Compounds e groups for which more than fifteen kinases are re are kinases fifteen than more which for e groups raction maps and kinase group fingerprints for for fingerprints group kinase and maps raction 0.2 0.4 0.6 0.8 1.0 0.1 0.2 0.3 0.4 0.5 0.6 0.2 0.4 0.6 0.8 1.0 0.1 0.2 0.3 0.4 0.5 0.6 0 0 0 0 G AKCG te T KTKL TK STE Other CMGC CAMK AGC G AKCG te T KTKL TK STE Other CMGC CAMK AGC G AKCG te T KTKL TK STE Other CMGC CAMK AGC G CAMK AGC ate the kinase group containing the the containing group kinase the ate M ivity (S(3 (S(3 ivity M) > 0.4 bin. The interaction maps maps interaction The bin. > 0.4 M) 12 M M) = 0–0.1 bin. ( bin. = 0–0.1 M) . The relative pattern, or finger- or pattern, relative The . se family tree based on kinase kinase on based tree family se CMGC M CHIR-258/TKI-258 NATURE BIOTECHNOLOGY NATURE M) 0–0.1, 0.1–0.2, 0.2–0.4, 0.2–0.4, 0.1–0.2, 0–0.1, M) GSK-1838705A SB-203580 KW-2449 te T KTKL TK STE Other M M affinity cutoff, for for cutoff, affinity M 7 elective inhibitors inhibitors elective . To describe the the describe To . eveals whether a whether eveals b ross the kinome. kinome. the ross ) Compounds ) Compounds nevertheless nevertheless ctions across across ctions e certainly is is certainly e presented presented subfamilies, subfamilies, e calculated calculated e scription of of scription lt to obtain obtain to lt idely, even even idely, © 2011 Nature America, Inc. All rights reserved. selective (selective strong were BMS-387032/SNS-032) R547, (AT-7519, tested ( kinase of Three four the cyclin-dependent kinases. interact to propensity global or scaffold chemical gros by governed not is selectivity group that gest compounds with diverse chemical structures. These o NATURE BIOTECHNOLOGY NATURE groups, and than the overall score ( that than both higher substantially is target) mary com the includes that group the (generally group kinase the however, compounds, of subset a For score. wide t to close and similar, relatively are groups kinase kinases. of number limited a screening from or intera of assessment visual qualitative more a from of known and available inhibitors, along with the o the with along inhibitors, available and known of Our data set reveals a large number of previously unde inhibitors are available for only a very small frac Nonethe kinases. of specific potential therapeutic and kinases.of majority a for inhibitors d to possible be may it that suggestion encouraging nevertheles results these limited, still is assessed tested here. the number Although of targets a primary kinome and selectivity relative to the primary targ by as absolute both inhibitors, assessed selective for and tested that the compounds are i represented primary are the that kinases 27 of the 15 for at least the tinguish target from kinase The other kinases. t conformations or features kinase structural exploit typ inhibitor of regardless compounds, selective most comm A selectivity. dictate not does type inhibitor selectiv quite are inhibitors I type of number a whereas also demonstrate that several type II inhibitors exhibit poor selectivity tion that can accommodate typical type I inhibitors by than II type inhibitors is more kinase-specific with the general assumption that the confo inactive observat This selectivities. of range wide a have can tha and inhibitors, I type than selective be to likely in II type class, a as that, shows class compound by int inhibitors kinome. the small-molecule a optimized yields how of tested, view collection compound the primary by of and sented scaffolds chemical of diversity the appro panel t and, kinome protein assay human of the coverage The complete date. to published kinome the kinase comprehen clinical and of known reactivity of the ment detailed most the represents set data Our DISCUSSION each that groups other exploits.compounds in kinases from kinases CMGC tha feature structural fundamental more a signal may non-CMG over CDKs for selectivity optimize to taken of overall selectivity ( selectivity overall of with compounds included inhibitors Group-selective o their of assessment accurate an yield to unlikely t primary the to related closely kinases targeted testing and primary the of members against reactivity inhibito Group-selective target. specific their for be considered group‘kinase selective’, but are not For many compounds, the selectivity scores for the the for scores selectivity the compounds, many For Small-molecule inhibitors are valuable tools to stud to tools valuable are inhibitors Small-molecule t compounds the of selectivity overall of assessment An Supplementary Fig. 2 Fig. Supplementary Fig. 4 Fig.

), both type I and type II inhibitors, and and inhibitors, II type and I type both ), ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE ). This may reflect similar strategies strategies similar reflect may This ). Fig. Fig. 4 ). ). These compounds can thus tion of protein kinases. selectivity across the selectivity necessarily selective selective necessarily an active an conforma- active et, et, are among the 72 s provide an initial initial an provide s he overall kinome- overall he data also data also show that rs may have broad broad have may rs . . However, the data for the remaining remaining the for rmation rmation preferred t type I inhibitors inhibitors I type t , intended targets targets , intended hibitors are more more are hibitors across inhibitors n n the assay panel, arget is therefore therefore is arget verall selectivity. selectivity. verall on theme for the the for theme on s binding mode, mode, binding s CDK) inhibitors inhibitors CDK) verall selectivity selectivity verall hat hat can help dis- ion is consistent consistent is ion scribed activities bservations bservations sug- with a range of of range a with evelop selective selective evelop ction patterns, patterns, ction nd nd compounds t distinguishes distinguishes t less, dedicated dedicated less, targets repre- targets score for one one for score kinase group, group, kinase e, is that they they that is e, y the biology biology the y a wide range range wide a C kinases, or or kinases, C ogether with with ogether e. Therefore, Therefore, e. ra over- broad pounds’ pri- pounds’ aches near- aches sive assess- sive eract with with eract individual individual ested here here ested of these these of ly group group ly , inhibitors has which of in resulted kinases, the sever discovery ag compounds of libraries entire screening on based inhi kinase to approach novel a enabled have further in clinical trials and has exhibited efficacy in pa in efficacy exhibited has and trials clinical in of the paper at http://www.nature.com/naturebiotech Methods and any associated references are available METHODS th to health.up human improving living is sequence genome the which in way drugs existing of understanding greater a and drugs to pro kinase kinome from and from discovery, drug kinome, profiling–based to sequence genome from path hs cmons a afc biology affect may compounds these underst complete more a yielded has and activities, u previously many revealed has panels assay the across were the representative and how complete for understanding small with kinome the interrogate to assays kinase systematically to essential been has turn in which sequence was essential to enumerate the human prote human the enumerate to essential was discovery. sequence drug to profiling kinase high-throughput by exemplified is path This discovery. target than drug facilitate directly that technologies of ment enabling path through is discovery alternative drug affect An can genome targets. drug new potential of gene-disease novel of identification the through is from path geno assumed The commonly most discovery. an facilitate would sequence genome the that promise interaction map images.map interaction numbers large examine manually to need the without inter specific with compounds identify to sets data systematical to possible it makes fingerprint group but their detailed interaction pattern across kinase groups. This kinase a of selectivity kinome overall of the not only quantitative description a generating by handle a such provides group kinase major the of each for scores selectivity here describe we approach The hits. interesting and the m and compounds revealing for handles classifying require application this by generated sets data large discoverystarti and inhibitors novel identify efficiently to compo large of screening the is panels assay kinase (Zelboraf). vemurafenib drug approved t related closely compound a PLX-4720, with parasite currently available; and the interaction of PFCDPK1 f whic against and cancer lung resistance in inhibitors major EGFR for a is which EGFR(T790M), with ment, the interaction of PKC-412, a compound in late-stage clinical develop- high affinity by the approved RET inhibitor vandetanib ( vandetanib by approved inhibitor RET the affinity high b not are which (RET(V804L/M)), mutations gatekeeper vities include the high affinity of sunitinib (Sutent) Intere kinases. novel targeting inhibitors optimized for de the points or starting drugs for known to explore n possible suggest may it cases, some In inhibito available. targeted specifically no here which studied for set kinases the in compounds of use the enable info The compound. each for pattern interaction and mjr ainl fr eunig h hmn eoe w genome human the sequencing for rationale major A comprehe of applications compelling most the of One 2–4,6 . Profiling of known inhibitors, including approved d approved including inhibitors, known of Profiling . 8–10,20–22 8,9,13,20–22,28,29 . Identifying compounds of interest from the the from interest of compounds Identifying . . At least one of these inhibitors is currently currently is inhibitors these of one Atleast . 2–4,6,7,23–27 E C R U O S E R in the online version tients . The assay panels panels assay The . ng points for drug drug for points ng for RET harboring the application of of application the discovery, rather rather discovery, al al promising new ovel applications applications ovel associations and and associations nology/. sting novel acti- novel sting s computational computational s rs are currently currently are rs s or subfamilies subfamilies or s rom the malaria und collections collections und bitor discovery, discovery, bitor build panels of of panels build action patterns patterns action molecules, and and molecules, ly search large large search ly anding of how how of anding illustrates one one illustrates h no drugs are are drugs no h ainst panels of of panels ainst of calculating calculating of filing to novel novel to filing 30 o the recently recently the o d enable drug drug enable d by which the the which by The genome genome The ost ost promising the develop- the of qualitative qualitative of rmation may may rmation velopment of velopment e promise of of promise e in kinome in as tools for for tools as nrecognized nrecognized compounds, compounds, . This direct direct This . nd numeric numeric nd assay panels panels assay me to drugs me to drugs Zactima) ound with with ound mutation mutation kinase kinase as the the as nsive nsive rugs, rugs, 12 19 5 , ; © 2011 Nature America, Inc. All rights reserved. 6. Fedorov, O. Fedorov, 6. 6 6 P.Cohen, kinase & protein M. of Elliott, specificities H., The McLauchlan, J., Bain, 1. reprints/index.html. o available is information permissions and Reprints Published online at http://www.nature.com/nbt/index Note: Supplementary information is available on the E C R U O S E R . Bantscheff, M. 2. development, G.P. coordinated and executed the meas J.P.H., P.C. and L.M.W. developed binding assay tec to of enhance the compoundefficiency screening, S. M.I.D. coordinated development of the assay panel, synthesis synthesis and D. Jones for assistance preparing Fig compound screening assistance,technical R. Faraoni molecular assistance,technical biology C. Shewmake and M. forCosta compound management, M. Floyd and We thank W. Wierenga for reading critical of the ma 10. Metz, J.T.10. Metz, . Posy,S.L. 9. 8. Bamborough, P., Drewry, D., Harper, G., Smith, G.K. & Schneider, K. Assessment K. Schneider, & G.K. Smith, G., Harper, P.,D., Drewry, Bamborough, 8. . ain M.A. Fabian, 3. . aaa, M.W. Karaman, 7. HTML HTML version of the paper at The authors declare competing interests:financial supervised the project, analyzed supervised data and wrote the and D.K.T. contributed to preparation of the manusc development strategies, a technology and supervised M.H. synthesized compounds, D.K.T. conceived the te 5. Patricelli, M.P. Patricelli, 5. 4. Melnick, J.S. Melnick, 4. AUTHOR CONTRIBUTIONS ACKNOWLEDGMENTS COMPETING FINANCIAL INTERESTSCOMPETING phosphates. Ser/Thr kinases. Ser/Thr inhibitors: an update. an inhibitors: of clinical ABL kinase inhibitors. kinase ABL clinical of inhibitors. screening. J. Med. Chem.Med. J. of chemical coverage of kinome space and its implications for kinase drug discovery. Nat. Biotechnol.Nat. molecular libraries. molecular et al.et et al.et Nat. Biotechnol.Nat. J. Med. Chem.Med. J. et al.et Biochemistry et al.et t al.et

et al. Trends in kinase selectivity: insights for target class-focused library class-focused Trendstarget for insights selectivity: kinase in et al.et 51 Navigating the kinome. the Navigating

Proc. Natl. Acad. Sci. USASci. Acad. Natl. Proc. 26 A systematic interaction map of validated kinase inhibitors with inhibitors kinase validated of map interaction systematic A t al. et , 7898–7914 (2008).7898–7914 , An efficient rapid system for profiling the cellular activities of activities cellular the profiling for system rapid efficient An sal oeuekns itrcin a fr lncl kinase clinical for map interaction molecule-kinase small A Quantitative chemical proteomics reveals mechanisms of action , 127–132 (2008).127–132 , Proc. Natl. Acad. Sci. USASci. Acad. Natl. Proc. Functional interrogation of the kinome using nucleotide acyl nucleotide using kinome the of interrogation Functional Biochem. J.Biochem. qatttv aayi o kns ihbtr selectivity. inhibitor kinase of analysis quantitative A

46 54

http://www.nature.com/nbt/index.html 23 , 350–358 (2007).350–358 , , 54–66 (2011).54–66 , , 329–336 (2005).329–336 , Nat. Biotechnol.Nat.

371 , 199–204 (2003).199–204 , Nat. Chem. Biol.Chem. Nat.

104

ure 4. 103 details details accompany the full-text

hnology and hnology assayperformed manuscript. , 20523–20528 (2007).20523–20528 , Nature Biotechnology nline at http://www.nature.com/ at nline J.P.H. developed technology hm. .html. H. H. analyzed data, M.I.D., nuscript, A. Torres, G. Riggs

ript, P.P.Z. the designed study, nd assay development, S.H. for advice on compound r r and J. Lowe for expert 25 urement of K chnology, assaydesigned , 3153–3158 (2006).3153–3158 , , 1035–1044 (2007).1035–1044 , L. L. Ramos for expert

7 , 200–202 (2011).200–202 , d values, . website. 29. Miduturu, C.V. 29. Miduturu, 8 ngt ZA, i, . Soa, .. agtn te acr ioe through kinome cancer the Targeting K.M. Shokat, & H. Lin, Z.A., 28. Knight, 30. 30. Cortes, J. 7 esn Rx L.L. Rix, 27. Remsing 6 ag X. 26. Yang, 23. 23. Hasinoff, B.B. & Patel, D. The lack of target specificity of small molecule anticancer 5 lhrk, A.J. 25. Olaharski, 4 un, . Zo, . Lflu, . Nvd, . Cflsh A Kns selectivity Kinase A. Caflisch, & C. Nevado, K., Lafleur, T., Zhou, D., 24. Huang, 22. Deng, X. 22. Deng, 7 iad J.R. 17. Simard, 18. 18. L.M. Wodicka, 21. Kwiatkowski, N. 21. Kwiatkowski, 20. Goldstein, D.M., Gray, N.S. & Zarrinkar, P.P. High-throughput kinase profiling as a as P.P.profiling Zarrinkar, kinase & High-throughput N.S. Gray, 20. D.M., Goldstein, 15. 15. S.G. Buchanan, 16. 16. E.R. Wood, 4 he, J.F. 14. Ohren, 19. 19. Carlomagno, F. 13. 13. Zarrinkar,P.P. 12. Manning, G., Whyte, D.B., Martinez, R., Hunter, T. & Sudarsanam, S. The protein The S. Sudarsanam, T.& Hunter, R., Martinez, D.B., Whyte, 12. G., Manning, 1 i, . Ga, .. ainl ein f niios ht id o ncie kinase inactive to bind that inhibitors of design Rational N.S. Gray, & Y. 11. Liu, toward the discovery of new kinase inhibitors. inhibitors. kinase new of discovery the toward polypharmacology. Annual Meeting Abstracts)Meeting Annual trial. clinical I phase (FIH) first-in-human a in inhibitor,(RTK) kinase (2011). primary chronic myeloid leukemia cells. leukemia myeloid chronic primary kinome and clinical trial data. trial clinical and kinome (2009).e1000446 Appl. Pharmacol.Appl. kinase inhibitors is correlated with their ability to damage myocytes in vitro. aae nue b sal oeue iae inhibitors. kinase molecule small by induced damage oeta fr niios agtn te T bnig ie a ewr analysis. network a site: binding ATP the targeting Bioinformatics inhibitors for potential kinase LRRK2. kinase Chem. Soc.Chem. conformations. kinase inactive and active to bind that inhibitors identifying niios srcua isgt fo biochemistry. from insights structural inhibitors: cell cycle function. cycle cell platform for drug discovery. drug for platform (2004).6056–6063 f h MT eetr yoie iae ih niuo atvt i vivo. in activity Ther. antitumor with kinase tyrosine receptor MET the of (2004).1192–1197 rate, and receptor activity in tumor cells. tumor in activity off-receptor and inhibitor rate, conformation, protein among relationships (Lapatinib): GW572016 MEK2 describe novel noncompetitive kinase inhibition. kinase noncompetitive novel describe MEK2 yoie iae ofr eitne o eetv kns inhibitors. kinase selective to resistance confer kinase tyrosine (2010). the treatment of acute myeloid leukemia (AML). leukemia myeloid acute of treatment the kinase complement of the human genome. human the of complement kinase conformations.

ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE 8 , 3181–3190 (2009).3181–3190 , et al.et et al. t al.et

et al.et 132 t al. et t al.et

et al.et AC220, a potent, selective, second generation FLT3 receptor tyrosine Characterization of a selective inhibitor of the Parkinson’s disease Parkinson’s the of inhibitor selective a of Characterization et al. 26 et al.et iae niiinrltd des eet peitd rm n vitro in from predicted events adverse inhibition-related Kinase Nat. Chem. Biol.Chem. Nat. et al. et t al.et Nat. Chem. Biol.Chem. Nat. et al.et , 4152–4160 (2010).4152–4160 , et al.et A unique structure for epidermal growth factor receptor bound to bound receptor factor growth epidermal for structure unique A

, 198–204 (2010).198–204 , 249 Nat. Rev. CancerRev. Nat. tutrs f ua MP iae iae (E1 and (MEK1) 1 kinase kinase MAP human of Structures loohr lbln o te lcn-ih op s mto of method a as loop glycine-rich the of labeling Fluorophore t al.et AC220 is a uniquely potent and selective inhibitor of FLT3of for inhibitor selective and potent uniquely a is AC220 Disease associated mutations at valine 804 in the RET receptor Nat. Chem. Biol.Chem. Nat. Activation state-dependent binding of small molecule kinase molecule small of binding state-dependent Activation High-throughput kinase profiling: a more efficient a efficient more a profiling: kinase High-throughput dnicto o a iae rfie ht rdcs chromosome predicts that profile kinase a of Identification SGX523 is an exquisitely selective, ATP-competitiveselective, exquisitely an is SGX523 inhibitor Small-molecule kinase inhibitors provide insight into Mps1 into insight provide inhibitors kinase Small-molecule , 132–139 (2010).132–139 , lbl agt rfie f h kns ihbtr ouii in bosutinib inhibitor kinase the of profile target Global

114 Nat. Rev. Drug Discov.Drug Rev. Nat. J. Biomed. Inform.Biomed. J. , Abstract 636 (2009).636 Abstract ,

2 7 , 358–364 (2006).358–364 , , 203–205 (2011).203–205 ,

10

6 , 359–368 (2010).359–368 , , 130–137 (2010).130–137 , Leukemia Cancer Res.Cancer Science NATURE BIOTECHNOLOGY NATURE Blood

43

23 hm Biol. Chem.

7 298 hm Biol. Chem. , 391–397 (2008).391–397 , , 376–384 (2010).376–384 , , 477–485 (2009).477–485 ,

64 Nat. Struct. Mol. Biol.Mol. Struct. Nat. 114 , 1912–1934 (2002).1912–1934 , LS opt Biol. Comput. PLOS , 6652–6659 (2004).6652–6659 , , 2984–2992 (2009).2984–2992 ,

17

18 Oncogene 1241–1249 , o. Cancer Mol. , 868–879 868–879 , Blood (ASH Blood pproach pproach Toxicol. J. Am. J.

23 11

5 , , , © 2011 Nature America, Inc. All rights reserved. validated and performed as described previously described as performed and validated assays. binding Competition ApBiosciences. at Ambit synthesized or SynChem WuXi Technologies, SynCores Shangai Qventas, CiVentiChem, by synthesized custom SYNthesismedchem, or Calbiochem/EMD Chemicals, Tocris Bioscience, Archer doi:10.1038/nbt.1990 Compounds. METHODSONLINE otd antc ed wr tetd ih biotinylate with treated were beads magnetic coated bind the for Briefly, kinases. multidomain for used f appropriate including constructs domain catalytic general, constructs full-length were used for small, phage T7 to fusions as either HEK-293 cells and subsequently tagged with DNA for for DNA with tagged subsequently and cells HEK-293 niios ee ihr ucae fo AG Scienti A.G. from purchased either were Inhibitors Competition binding assays were developed, developed, were assays binding Competition 3 , or were expressed as fusions to NF- to fusions as expressed were or , 3,18 single-domain single-domain kinases, and ing assays, streptavidin- streptavidin- assays, ing . Kinases were produced produced were Kinases . lanking sequences were were sequences lanking chem, Axon Medchem d affinity ligands to to ligands affinity d PCR detection PCR SAI Advantium, Advantium, SAI pTec, BioDuro, BioDuro, pTec, K 18 B in in B . In . fic, fic, plates, whereas whereas plates, ing were a interactions screen test compound. Primary c to added was DMSO DMSO. in stocks 100× as prepared Bindin ligand. bead affinity liganded unbound kinase, combining by assembled remove to washed and binding generate affinity resins. The liganded beads were b PCR. PCR. shaking. The kinase concentration in the eluates wa ligands affinity nonbiotinylated of presence the in were washed extensively h to remove 1 unbound protein. for Bound shaking with °C 25 at incubated were plates ated using TREE using ated and a DMSO control. Kinase interaction maps and shown i a interaction DMSO control. Kinase screens have been previously determined previously been have screens False-positive and false-negative rates for single- for rates false-negative and False-positive K d s s were determined using 11 serial threefold dilutio K spot d determinations were performed in 96-well plates. A plates. 96-well in performed were determinations software ( software http://www.kinomescan.com/ 7 . NATURE BIOTECHNOLOGY NATURE locked locked to reduce nonspecific s s measured by quantitative for 30 min at 25 °C with with °C 25 at min 30 for , and the affinity beads beads affinity the and , concentration primary primary concentration n n performed in 384-well performed s and test compounds compounds test and s Figure 4 Figure ns of test compound ontrol assays lack- assays ontrol g reactions were were reactions g kinase kinase was eluted ). were gener- ssay ssay Supplementary Figure 1.

1000

(nM) 100 d /K i

/K 10 50 IC ! 1

0.1 Published

0.1 1 10 100 1000 Measured !K (nM) d !

Supplementary Figure 1. Comparison of K d values reported here and published

biochemical IC 50 /K i/K d values for interactions between inhibitors and their primary, intended targets. Values used to create this plot are reported in Supplementary Table 3. Comparative data were available for 66 of the 72 compounds addressed in this study. The line of equivalence is indicated in solid red, and dashed red lines indicate 10-fold offsets.

Nature Biotechnology: doi:10.1038/nbt.1990 Supplementary Figure 2.

Nature Biotechnology: doi:10.1038/nbt.1990 Supplementary Figure 2.

Nature Biotechnology: doi:10.1038/nbt.1990 Supplementary Figure 2.

Nature Biotechnology: doi:10.1038/nbt.1990 Supplementary Figure 2.

Nature Biotechnology: doi:10.1038/nbt.1990 Supplementary Figure 2.

Nature Biotechnology: doi:10.1038/nbt.1990 Supplementary Figure 2.

Nature Biotechnology: doi:10.1038/nbt.1990 Supplementary Figure 2.

Supplementary Figure 2. Kinome interaction maps for the 72 compounds tested. Each red circle indicates a kinase found to bind to a compound. Larger circles indicate higher affinity

interactions. Interactions with K d < 3 µM are shown.

Nature Biotechnology: doi:10.1038/nbt.1990 Supplementary Figure 3. a All compounds

70

60

50

40

30

20

10

Percent of compounds of Percent 0

S(300 nM)

b Type I inhibitors

70

60

50

40

30

20

10

Percent of compounds of Percent 0

S(300 nM)

c Type II inhibitors

70

60

50

40

30

20

10

Percent of compounds of Percent 0

S(300 nM) !

Supplementary Figure 3. Quantitative distribution of kinome-wide selectivity of compounds. An alternative analysis of the data shown in Figure 1 using a cutoff of S(300 nM) instead of S(3 !M) and reduced bin sizes (0.05 units instead of 0.1). (a), (b), and (c) are otherwise as described in Figure 1.

Nature Biotechnology: doi:10.1038/nbt.1990 Supplementary !Table !2. !! Compound !structures !and !highest !affinity !target !kinases !(excluding !mutants)

Compound Structure Kinase Kd, !nM CLK2 0.51 CLK4 0.53

NH CLK1 1.4 NH 2 N DYRK1A 2.1 O DYRK1B 3.9 AR674563 PRKCH 9.5 N PRKCE 11 PRKC Q 12 CIT 13 PRKG2 19 ABL1 Rnonphosphorylated 2.1 CSF1R 7.6 KIT 8.1 O S N PDGFRB 8.4 DDR1 8.7 AB R1010 N NH NH N PDGFRA 25 N DDR2 26 LCK 31 ABL1 Rphosphorylated 55 LYN 61 FLT3 0.63 PDGFRB 1.9 F KIT 2 NH 2 NH NH CSF1R 3.4 N PDGFRA 4.2 ABT R869 HN O FLT1 7.5 VEGFR2 8.1 MUSK 10 FLT4 16 EPHB6 33 FLT3 1.3 KIT 4.8 O N O PDGFRB 7.7 RET 8 N CSF1R 10 AC220 S O N PDGFRA 11 O FLT1 41 N NH NH FLT4 41 DDR1 81 VEGFR2 87 O NH PDGFRA 0.51 PDGFRB 0.57 S NH AURKC 1.3 N KIT 3.2 FLT1 5.8 AG R013736 VEGFR2 5.9 AURKB 11

N PLK4 16 CSF1R 21 ABL1 Rphosphorylated 36 KIT 3.7 O CSF1R 5.6 ZAK 8 NH NH PDGFRB 9.1 FLT4 9.7 AMG R706 N NH PDGFRA 10 FLT1 12 RET 14 N VEGFR2 26 FLT3 71 TIE1 0.29 N N CDKL2 0.52 F F ABL1 Rnonphosphorylated 0.68 F DDR1 0.69 FLT3 0.79 AST R487 O NH LOK 0.92 NH CDC2L5 0.94 NH N CDK8 1.4 N O CDK11 1.5 TAK1 1.5

Nature Biotechnology: doi:10.1038/nbt.1990 1 Supplementary !Table !2. !! Compound !structures !and !highest !affinity !target !kinases !(excluding !mutants)

Compound Structure Kinase Kd, !nM PCTK2 0.95 PCTK1 1.1 Cl O CDKL5 2.6 NH CDK7 2.8 NH CDC2L5 3.2 AT R7519 N NH CDC2L2 5.2 Cl O NH CDK9 5.8 ICK 8.3 CDC2L1 8.4 GSK3B 11 MEK5 2.8 AURKC 4.4 N NH AURKB 4.6

HN NH FLT3 8.2 O KIT 17 AZD R1152HQPA N HO F N O N PDGFRA 38 PDGFRB 41 EPHB6 50 RET 80 HIPK4 97 NH PDGFRB 0.32 KIT 0.38 O PDGFRA 0.41 O F N FLT1 0.74 VEGFR2 1.1 AZD R2171 O N DDR1 1.7 FLT4 4.3 STK35 5.4 N RET 6.1 CSF1R 13 MEK1 99 MEK2 530 HO NH O O Cl EGFR 7000 NH AZD R6244/ARRY R886

N F Br N

PLK1 0.19

N O PLK2 0.81 PLK3 4 N O NH N RPS6KA4(Kin.Dom.2 RCRterminal) 12 CAMKK1 22 BI R2536 NH N N CAMKK2 23 O MYLK 97 PIP5K2C 110 DAPK3 130 FAK 150 O MEK5 1.8

N N BIKE 2.2 N VEGFR2 2.9 PKNB(M.tuberculosis) 3.6 FLT3 3.8 BIBF R1120 !(derivative) NH TAK1 4.1 TRKA 4.5 JAK1(JH2domain Rpseudokinase) 4.8 O NH MELK 4.9 YSK4 5.2 EGFR 0.25 F ERBB2 5 ERBB4 6.3 Cl NH GAK 79 NH BLK 220 BIBW R2992 N N IRAK1 240 O N O EPHA6 340 HIPK4 360 PHKG2 470 O ABL1 Rphosphorylated 570

Nature Biotechnology: doi:10.1038/nbt.1990 2 Supplementary !Table !2. !! Compound !structures !and !highest !affinity !target !kinases !(excluding !mutants)

Compound Structure Kinase Kd, !nM p38 Ralpha 0.45 DDR1 1.9 O O N p38 Rgamma 2.9 p38 Rbeta 7.2 N O JNK2 7.3 BIRB R796 N NH NH TIE1 8.3 LOK 12 TIE2 20 DDR2 33 p38 Rdelta 78 IKK Rbeta 130 YSK4 260 CDC2L2 390 CDC2L1 420 N N ERK5 620 BMS R345541 CDK7 680 NH 2 N NH MYLK4 700 CDC2L5 800 PCTK1 890 ERN1 1000 CDKL5 1.7 PCTK1 7.1 PCTK2 13 N O O S CDC2L5 23 GSK3A 28 BMS R387032/SNS R032 S NH CDK7 31 N NH GSK3B 37 CDKL2 41 PCTK3 44 CDC2L2 48 VEGFR2 5 NH FLT1 10 PDGFRA 11 STK35 26 O KIT 36 BMS R540215 F PDGFRB 50 N FLT4 56 HO O N FGFR1 99 N FGFR2 110 DDR1 160

NH PHKG1 0.39 O YSK4 0.52 LATS2 1 SNARK 1 MKNK2 1.4 CEP R701 N N PLK4 1.5 O IRAK4 1.7 PHKG2 1.7 HO OH PKN2 1.8 JAK3(JH1domain Rcatalytic) 2.3 FLT3 0.64 MLCK 2 PDGFRB 3.8 N N F NH 2 N KIT 7.5 MINK 9.1 CHIR R258/TKI R258 NH YSK4 12 NH O TNIK 24 MEK5 32 MAST1 40 HPK1 44 DDR1 13 EPHB6 43 N YSK4 54 O CF 3 GCN2(Kin.Dom.2,S808G) 55 N NH NH LOK 60 CHIR R265/RAF R265 N N ZAK 63 CIT 87 TAOK2 140 CF 3 RET 150 TIE1 150

Nature Biotechnology: doi:10.1038/nbt.1990 3 Supplementary !Table !2. !! Compound !structures !and !highest !affinity !target !kinases !(excluding !mutants)

Compound Structure Kinase Kd, !nM EGFR 0.19 F ERBB4 6.5

O ABL1 Rphosphorylated 30 HN Cl BLK 45 HN ERBB2 56 CI R1033 N MEK5 60

N O N GAK 100 MKK7 110 O ABL1 Rnon phos p hor y lated 210 ERBB3 210

O MEK1 120 MEK2 370 O NH AURKC 1800 PDGFRB 3100 F NH CAMK2A 3400 CI R1040 F Cl

I JAK3(JH1domain Rcatalytic) 0.16 JAK2(JH1domain Rcatalytic) 0.58 JAK1(JH1domain Rcatalytic) 1.6 N TYK2(JH1domain Rcatalytic) 4.8 N N DCAMKL3 12 CP R690550 O N TNK1 120 PKN1 170 N NH SNARK 240 ROCK2 420 LCK 460

NH MET 2.1 ALK 3.3 MERTK 3.6 NN ROS1 4.1 EPHB6 6 Crizotinib Cl AXL 7.8 N LTK 12 O SLK 18 NH 2 Cl MST1R 25 F LCK 30 ABL1 Rnonphosphorylated 0.029 EPHB6 0.039 N OH ABL1 Rphosphorylated 0.046 N N N EPHA3 0.093 N ABL2 0.17 Dasatinib NH NH LCK 0.2 S BLK 0.21 O Cl SRC 0.21 EPHA5 0.24 EPHA8 0.24 EGFR 0.67 GAK 3.1 LOK 19

HN YSK4 25 SLK 26 Erlotinib O O N ABL1 Rphosphorylated 76 O MEK5 96 O N BLK 190 ABL2 200 ERBB4 230 AXL 0.093 DDR1 0.2 F NH NH MERTK 0.27

O O HIPK4 0.51 O F LOK 0.53 EXEL R2880/GSK R1363089 O RET 0.74

N O N TIE1 0.79 O FLT3 0.9 PDGFRB 0.96 EPHA3 1

Nature Biotechnology: doi:10.1038/nbt.1990 4 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM ICK 0.69 CDK4 RcyclinD3 3.3 N CDK9 6.4 CDKL5 7.1 Flavopiridol OH CDK4 RcyclinD1 9 HO O CDK7 23 MAK 28 Cl TYK2 (JH2domain Rpseudokinase ) 35 TNNI3K 55 OH O CDK11 57 OH BRAF 0.19 RAF1 6.6 NN CSNK1E 130 YSK4 910 MINK 1000 GDC R0879 N RIOK2 1200 LOK 1300 SLK 1300 N CSNK1D 1400 OH BMPR1B 1800 O PIK3CA 1.1 PIK3CD 5 N PIK3CB 16 PIK3CG 48 S N N PIK3C2B 130 GDC R0941 NH N N MTOR 200 PIK3C2G 300 O N JAK1 (JH2domain Rpseudokinase ) 430 S HIPK2 520 O JNK3 560 EGFR 1 F GAK 13 IRAK1 69 YSK4 240 O HN Cl MKNK1 290 Gefitinib N O N HIPK4 310 ERBB4 410

O N CSNK1E 430 LOK 470 ABL1 Rphos phor ylated 480 O F ALK 0.55 LTK 1.1 NH INSR 1.7 NH IGF1R 7 GSK R1838705A N INSRR 8.6 NH N NH FER 9.3 O MYLK 11 O ROS1 15 N N CLK2 16 CLK1 21 O PLK1 0.094 SNARK 23 S NH 2 N N LOK 69 O RSK2 (Kin.Dom.1 RNRterminal ) 190 GSK R461364A PIM1 250 F NEK2 260 RSK4 (Kin.Dom.1 RNRterminal ) 350 N FF BIKE 470 N PLK2 500 CAMK2D 620 AKT2 2.1

HO AKT1 2.2 PRKCH 2.4 AKT3 3 GSK R690693 H2N PRKG2 3.1 N PKNB (M.tuberculosis ) 3.2 N N PRKCE 5.3 O N N PRKX 7.2

HN O PAK7 8.9 PKAC Rbeta 13

Nature Biotechnology: doi:10.1038/nbt.1990 5 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM CSF1R 2.2

NH TRKB 36 2 TRKC 120 N TRKA 630 GW R2580 H2N N O O O

MAP4K5 0.65 EGFR 1.1 O ERBB4 2.4 N ERBB2 6 O MST3 6.5 HKI R272 N HN Cl MST4 7.4 HN N ERBB3 7.7 MAP4K3 7.7 O N YSK1 12 LOK 13 DDR1 0.7

N ABL1 Rnon phos phor ylated 1.1 ABL2 10 N NH NH N CSF1R 11 KIT 13 Imatinib N O PDGFRB 14 DDR2 15 ABL1 Rphos phor ylated 21 N PDGFRA 31 LCK 40 JAK2(JH1domain Rcatalytic) 0. 036 TYK2 (JH1domain Rcatal ytic ) 0.9 JAK3 (JH1domain Rcatal ytic ) 2 N JAK1 (JH1domain Rcatal ytic ) 3.4 N INCB018424 HN MAP3K2 41 CAMK2A 46 ROCK2 52 N N ROCK1 60 N DCAMKL1 68 DAPK1 72

NH CSF1R 3.2 KIT 3.6 O N N AMPK Ral pha2 4.1 NH AXL 5.3 JNJ R28312141 TYK2 (JH1domain Rcatal ytic ) 5.8 TAK1 7.2 RET 9.2 N N DDR1 9.4 CHEK1 9.9 O JAK2 (JH1domain Rcatal ytic ) 9.9 PDGFRB 0.29 O N PDGFRA 0.49 NH NH KIT 0.69 S CSF1R 0.83 O EPHB6 5 Ki R20227 O MEK5 7.4 O DDR1 12 VEGFR2 18 O N LOK 22 SLK 60 NH DRAK1 2.9 MLCK 4.3 N O DRAK2 4.8 YSK4 5.2 KW R2449 BIKE 9.6 MAP4K2 11 LOK 13 SLK 13 N FLT3 15 NH SRPK2 15

Nature Biotechnology: doi:10.1038/nbt.1990 6 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM EGFR 2.4 ERBB2 7 ERBB4 54 O F PIK3C2B 670 O PIK4CB 940 Lapatinib S O NH HN Cl MEK5 1100 O N SLK 3300 RIPK2 3600 N LOK 4400 MKK7 4400 GSK3B 8.3 O NH O PRKCE 8.9 PRKCD 25

N RSK4 (Kin.Dom.1 RNRterminal ) 25 LY R317615 N PRKCQ 36 PRKCH 46 N ERK8 76 N RSK2 (Kin.Dom.1 RNRterminal ) 87 DYRK1A 160 PRKG2 170 PRKCQ 2.5 NH O O PRKCD 3.6 PRKCE 11 PIM3 12 LY R333531 YSK4 48 N N HIPK3 77 ERK8 88 TAOK3 97 O N HIPK2 140 MAP3K3 170 IKK Rbeta 19 Cl N PIK4CB 270 TYK2 (JH2domain Rpseudokinase ) 500 CLK1 640 O NH MLN R120B CLK4 1000 O NH IKK Ral pha 1000 DYRK1A 2100 CLK2 2300 N PDGFRA 2.4 KIT 2.7 O NH FLT3 3 N O PDGFRB 4.5 CSF1R 4.9 MLN R518 N O DDR2 120 N EGFR 410 N O N TRKA 450 CLK1 630 IRAK3 730 AURKA 6.5 DRAK2 8.1 O F AURKC 26 N HO N AURKB 43 MLN R8054 BLK 68 NH N DRAK1 190 F FGR 220 YES 260 Cl TIE2 300 EPHB1 330 DDR1 1.1 ABL1 Rnon phos phor ylated 10 F N F ZAK 11 NH NH N ABL1 Rphos phor ylated 13 F ABL2 26 Nilotinib O N KIT 29 N DDR2 33 p38 Rbeta 36 N EPHA8 37 CSF1R 45

Nature Biotechnology: doi:10.1038/nbt.1990 7 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM PDGFRB 2 KIT 2.8 NHN N N PDGFRA 4.9 N CSF1R 7.9 N Pazopanib FLT1 14 O NH 2 VEGFR2 14 S FLT4 27 O TAOK3 45 DDR1 57 EPHB6 81 ABL1 Rphos phor ylated 0.58 Cl CSF1R 0.67 ABL2 0.69 SRC 0.71 PD R173955 N LCK 1.1 PDGFRB 1.4 Cl ABL1 Rnon phos phor ylated 1.5 S NH N N O BLK 1.5 KIT 1.8 EPHB6 2 MET 0.27 SRPK1 41

Cl TAOK3 43 NH N BIKE 47 N PHA R665752 GRK7 61 Cl O HIPK3 63 O S CAMKK1 64 O O DDR1 70 NH CAMKK2 73 YSK4 78 PIK3CA 1.5 O PIK3CB 1.7 PIK3C2B 10 N MTOR 12 PIK3CG 16 PI R103 O N PIK3CD 17 N OH PIK3C2G 41 N HIPK2 290 HIPK3 310 PIP5K2C 620 PKN1 9.3 TBK1 9.3 FLT3 11 O N JAK3 (JH1domain Rcatal ytic ) 12 O MLK1 15 PKC R412 O H PKN2 15 N N YSK4 15 MLK3 17 CAMK2A 20 NH O MARK3 21 MEK5 0.16 F PFCDPK1(P.falciparum) 1.7 SRMS 21 O ZAK 41 O O BRK 48 PLX R4720 Cl S NH FGR 62 F RAF1 170

N NH KIT 180 MEK4 190 NEK11 190 OH BMPR1B 2.1 ACVRL1 2.9 MTOR 3 ACVR1 4 RET 4.8 PP R242 NH 2 NH YSK4 5.1 N N MEK5 7.3 N N ACVR2B 7.6 PRKCE 9 JAK2 (JH1domain Rcatal ytic ) 11

Nature Biotechnology: doi:10.1038/nbt.1990 8 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM Cl KIT 5.1 FLT1 9.6 PDGFRB 25 HN CSF1R 45 PTK R787 N VEGFR2 62 N PDGFRA 96 DDR1 270 FLT4 330 N CDK11 1500 FRK 1800 FLT3 0.71 STK16 1.7 O GCN2 (Kin.Dom.2,S808G ) 3.3 O F O PDGFRB 3.3 N R406 JAK2 (JH1domain Rcatal ytic ) 3.5 MLK2 3.8 O NH N NH N NH O RET 4.1 PLK4 4.2 MLK1 4.3 PLK3 5.1 CDK2 0.53 NH N NH 2 PCTK1 0.54 CDK7 0.58 O N N O CDK4 RcyclinD1 0.61 S CDK4 RcyclinD3 0.81 R547 O O F PCTK2 0.86 ICK 2.2 CDK3 3.2 F PFTAIRE2 7.2 CDK5 7.4 p38 Ralpha 12 GAK 19 N RIPK2 24 O NH NLK 25 SB R203580 S JNK3 35 N CSNK1D 37 p38 Rbeta 70 F CSNK1A1 75 CSNK1E 100 JNK2 130 MET 0.19 DYRK1A 780 N DYRK1B 1800 N JNK3 1900 S N PIP5K2C 3300 SGX R523 N N JNK1 4200 N YSK4 6400 N

ABL1 Rphos phor ylated 0.057 ABL1 Rnonphosphorylated 0.12 Cl Cl MAP4K5 0.5 LCK 0.59 HN O SKI R606 O CN ERBB3 0.77 SRC 1 N O N GAK 1.3 N FRK 1.4 ABL2 1.5 STK35 2 DDR1 1.5 HIPK4 3.3 ZAK 6.3 CF 3 O DDR2 6.6 Cl O FLT3 13 Sorafenib O NH RET 13 N NH NH CSF1R 28 KIT 28 FLT1 31 PDGFRB 37

Nature Biotechnology: doi:10.1038/nbt.1990 9 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM SLK 0.024 HN LOK 0.037 O CAMKK1 0.039 O H SNARK 0.086 Staurosporine N N PHKG2 0.14 CAMK2A 0.16 CAMKK2 0.16 MST2 0.18 O MST1 0.19 NH TAOK3 0.22 PDGFRB 0.29 FLT3 0.54 O N H C O KIT 0.68 3 NH OH PDGFRA 1.1 VEGFR2 2.3 SU R14813 NH CH F 3 CSF1R 3.6 O HUNK 3.7 NH FLT1 4.7 STK35 8.2 YSK4 12 PDGFRB 0.075 KIT 0.37 F O FLT3 0.41 PDGFRA 0.79 N Sunitinib NH DRAK1 1 NH VEGFR2 1.5

NH O FLT1 1.8 CSF1R 2.5 BIKE 5.5 PHKG1 5.5 ROS1 0.49 N ULK1 0.83 N BMPR1B 0.85

Cl O N PLK4 0.93 TAE R684 N LTK 0.95 ALK 1.1 HN N NH O O FAK 1.1 S PYK2 1.1 SNARK 1.2 FER 1.4 PIK3C2G 3.2 OH PIK3CG 5.3

NH 2 PIK3C2B 7.3 N TRPM6 7.9 N TG R100 R115 CLK2 43 PIK3CA 59 H2N N N PIK3CB 80 ADCK3 94 RIPK4 97 OH CLK4 130 GAK 1.1 JAK2(JH1domain R catalytic) 1.1 DAPK3 1.2 NH STK16 6.6 S NH TG R101348 DCAMKL3 13 OO O FLT3 13 N N DAPK1 16 N NH JAK1 (JH1domain Rcatal ytic ) 18 YSK4 19 TYK2 (JH1domain Rcatal ytic ) 21 Br RIPK2 4.6 EGFR 9.5 DDR1 11 HN ABL1 Rphos phor ylated 16 Vandetanib O F LCK 17 N RET 34 ABL1 Rnon phos phor ylated 48 O N MEK5 49 N EPHA6 50 STK35 56

Nature Biotechnology: doi:10.1038/nbt.1990 10 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM AURKA 3.9 ABL2 4 N AURKC 6.3 HN NH FLT3 6.5 AURKB 7.4 VX R680/MK R0457 NH N ABL1 Rphos phor ylated 7.5 O PLK4 9.2 N N S ABL1 Rnon phos phor ylated 13 N MLCK 15 RIPK1 20 p38 Ral pha 2.8 F p38 Rbeta 74 S N DDR1 1100 N N FGR 1300 VX R745 YES 1600 F O LYN 1700 Cl Cl ABL2 1900 FYN 2100 CSF1R 2600 BLK 3100

Nature Biotechnology: doi:10.1038/nbt.1990 11 Supplementary !Table !2. !! Compound !structures !and !highest !affinity !target !kinases !(excluding !mutants)

Compound Structure Kinase Kd, !nM CLK2 0.51 CLK4 0.53

NH CLK1 1.4 NH 2 N DYRK1A 2.1 O DYRK1B 3.9 AR674563 PRKCH 9.5 N PRKCE 11 PRKC Q 12 CIT 13 PRKG2 19 ABL1 Rnonphosphorylated 2.1 CSF1R 7.6 KIT 8.1 O S N PDGFRB 8.4 DDR1 8.7 AB R1010 N NH NH N PDGFRA 25 N DDR2 26 LCK 31 ABL1 Rphosphorylated 55 LYN 61 FLT3 0.63 PDGFRB 1.9 F KIT 2 NH 2 NH NH CSF1R 3.4 N PDGFRA 4.2 ABT R869 HN O FLT1 7.5 VEGFR2 8.1 MUSK 10 FLT4 16 EPHB6 33 FLT3 1.3 KIT 4.8 O N O PDGFRB 7.7 RET 8 N CSF1R 10 AC220 S O N PDGFRA 11 O FLT1 41 N NH NH FLT4 41 DDR1 81 VEGFR2 87 O NH PDGFRA 0.51 PDGFRB 0.57 S NH AURKC 1.3 N KIT 3.2 FLT1 5.8 AG R013736 VEGFR2 5.9 AURKB 11

N PLK4 16 CSF1R 21 ABL1 Rphosphorylated 36 KIT 3.7 O CSF1R 5.6 ZAK 8 NH NH PDGFRB 9.1 FLT4 9.7 AMG R706 N NH PDGFRA 10 FLT1 12 RET 14 N VEGFR2 26 FLT3 71 TIE1 0.29 N N CDKL2 0.52 F F ABL1 Rnonphosphorylated 0.68 F DDR1 0.69 FLT3 0.79 AST R487 O NH LOK 0.92 NH CDC2L5 0.94 NH N CDK8 1.4 N O CDK11 1.5 TAK1 1.5

Nature Biotechnology: doi:10.1038/nbt.1990 1 Supplementary !Table !2. !! Compound !structures !and !highest !affinity !target !kinases !(excluding !mutants)

Compound Structure Kinase Kd, !nM PCTK2 0.95 PCTK1 1.1 Cl O CDKL5 2.6 NH CDK7 2.8 NH CDC2L5 3.2 AT R7519 N NH CDC2L2 5.2 Cl O NH CDK9 5.8 ICK 8.3 CDC2L1 8.4 GSK3B 11 MEK5 2.8 AURKC 4.4 N NH AURKB 4.6

HN NH FLT3 8.2 O KIT 17 AZD R1152HQPA N HO F N O N PDGFRA 38 PDGFRB 41 EPHB6 50 RET 80 HIPK4 97 NH PDGFRB 0.32 KIT 0.38 O PDGFRA 0.41 O F N FLT1 0.74 VEGFR2 1.1 AZD R2171 O N DDR1 1.7 FLT4 4.3 STK35 5.4 N RET 6.1 CSF1R 13 MEK1 99 MEK2 530 HO NH O O Cl EGFR 7000 NH AZD R6244/ARRY R886

N F Br N

PLK1 0.19

N O PLK2 0.81 PLK3 4 N O NH N RPS6KA4(Kin.Dom.2 RCRterminal) 12 CAMKK1 22 BI R2536 NH N N CAMKK2 23 O MYLK 97 PIP5K2C 110 DAPK3 130 FAK 150 O MEK5 1.8

N N BIKE 2.2 N VEGFR2 2.9 PKNB(M.tuberculosis) 3.6 FLT3 3.8 BIBF R1120 !(derivative) NH TAK1 4.1 TRKA 4.5 JAK1(JH2domain Rpseudokinase) 4.8 O NH MELK 4.9 YSK4 5.2 EGFR 0.25 F ERBB2 5 ERBB4 6.3 Cl NH GAK 79 NH BLK 220 BIBW R2992 N N IRAK1 240 O N O EPHA6 340 HIPK4 360 PHKG2 470 O ABL1 Rphosphorylated 570

Nature Biotechnology: doi:10.1038/nbt.1990 2 Supplementary !Table !2. !! Compound !structures !and !highest !affinity !target !kinases !(excluding !mutants)

Compound Structure Kinase Kd, !nM p38 Ralpha 0.45 DDR1 1.9 O O N p38 Rgamma 2.9 p38 Rbeta 7.2 N O JNK2 7.3 BIRB R796 N NH NH TIE1 8.3 LOK 12 TIE2 20 DDR2 33 p38 Rdelta 78 IKK Rbeta 130 YSK4 260 CDC2L2 390 CDC2L1 420 N N ERK5 620 BMS R345541 CDK7 680 NH 2 N NH MYLK4 700 CDC2L5 800 PCTK1 890 ERN1 1000 CDKL5 1.7 PCTK1 7.1 PCTK2 13 N O O S CDC2L5 23 GSK3A 28 BMS R387032/SNS R032 S NH CDK7 31 N NH GSK3B 37 CDKL2 41 PCTK3 44 CDC2L2 48 VEGFR2 5 NH FLT1 10 PDGFRA 11 STK35 26 O KIT 36 BMS R540215 F PDGFRB 50 N FLT4 56 HO O N FGFR1 99 N FGFR2 110 DDR1 160

NH PHKG1 0.39 O YSK4 0.52 LATS2 1 SNARK 1 MKNK2 1.4 CEP R701 N N PLK4 1.5 O IRAK4 1.7 PHKG2 1.7 HO OH PKN2 1.8 JAK3(JH1domain Rcatalytic) 2.3 FLT3 0.64 MLCK 2 PDGFRB 3.8 N N F NH 2 N KIT 7.5 MINK 9.1 CHIR R258/TKI R258 NH YSK4 12 NH O TNIK 24 MEK5 32 MAST1 40 HPK1 44 DDR1 13 EPHB6 43 N YSK4 54 O CF 3 GCN2(Kin.Dom.2,S808G) 55 N NH NH LOK 60 CHIR R265/RAF R265 N N ZAK 63 CIT 87 TAOK2 140 CF 3 RET 150 TIE1 150

Nature Biotechnology: doi:10.1038/nbt.1990 3 Supplementary !Table !2. !! Compound !structures !and !highest !affinity !target !kinases !(excluding !mutants)

Compound Structure Kinase Kd, !nM EGFR 0.19 F ERBB4 6.5

O ABL1 Rphosphorylated 30 HN Cl BLK 45 HN ERBB2 56 CI R1033 N MEK5 60

N O N GAK 100 MKK7 110 O ABL1 Rnon phos p hor y lated 210 ERBB3 210

O MEK1 120 MEK2 370 O NH AURKC 1800 PDGFRB 3100 F NH CAMK2A 3400 CI R1040 F Cl

I JAK3(JH1domain Rcatalytic) 0.16 JAK2(JH1domain Rcatalytic) 0.58 JAK1(JH1domain Rcatalytic) 1.6 N TYK2(JH1domain Rcatalytic) 4.8 N N DCAMKL3 12 CP R690550 O N TNK1 120 PKN1 170 N NH SNARK 240 ROCK2 420 LCK 460

NH MET 2.1 ALK 3.3 MERTK 3.6 NN ROS1 4.1 EPHB6 6 Crizotinib Cl AXL 7.8 N LTK 12 O SLK 18 NH 2 Cl MST1R 25 F LCK 30 ABL1 Rnonphosphorylated 0.029 EPHB6 0.039 N OH ABL1 Rphosphorylated 0.046 N N N EPHA3 0.093 N ABL2 0.17 Dasatinib NH NH LCK 0.2 S BLK 0.21 O Cl SRC 0.21 EPHA5 0.24 EPHA8 0.24 EGFR 0.67 GAK 3.1 LOK 19

HN YSK4 25 SLK 26 Erlotinib O O N ABL1 Rphosphorylated 76 O MEK5 96 O N BLK 190 ABL2 200 ERBB4 230 AXL 0.093 DDR1 0.2 F NH NH MERTK 0.27

O O HIPK4 0.51 O F LOK 0.53 EXEL R2880/GSK R1363089 O RET 0.74

N O N TIE1 0.79 O FLT3 0.9 PDGFRB 0.96 EPHA3 1

Nature Biotechnology: doi:10.1038/nbt.1990 4 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM ICK 0.69 CDK4 RcyclinD3 3.3 N CDK9 6.4 CDKL5 7.1 Flavopiridol OH CDK4 RcyclinD1 9 HO O CDK7 23 MAK 28 Cl TYK2 (JH2domain Rpseudokinase ) 35 TNNI3K 55 OH O CDK11 57 OH BRAF 0.19 RAF1 6.6 NN CSNK1E 130 YSK4 910 MINK 1000 GDC R0879 N RIOK2 1200 LOK 1300 SLK 1300 N CSNK1D 1400 OH BMPR1B 1800 O PIK3CA 1.1 PIK3CD 5 N PIK3CB 16 PIK3CG 48 S N N PIK3C2B 130 GDC R0941 NH N N MTOR 200 PIK3C2G 300 O N JAK1 (JH2domain Rpseudokinase ) 430 S HIPK2 520 O JNK3 560 EGFR 1 F GAK 13 IRAK1 69 YSK4 240 O HN Cl MKNK1 290 Gefitinib N O N HIPK4 310 ERBB4 410

O N CSNK1E 430 LOK 470 ABL1 Rphos phor ylated 480 O F ALK 0.55 LTK 1.1 NH INSR 1.7 NH IGF1R 7 GSK R1838705A N INSRR 8.6 NH N NH FER 9.3 O MYLK 11 O ROS1 15 N N CLK2 16 CLK1 21 O PLK1 0.094 SNARK 23 S NH 2 N N LOK 69 O RSK2 (Kin.Dom.1 RNRterminal ) 190 GSK R461364A PIM1 250 F NEK2 260 RSK4 (Kin.Dom.1 RNRterminal ) 350 N FF BIKE 470 N PLK2 500 CAMK2D 620 AKT2 2.1

HO AKT1 2.2 PRKCH 2.4 AKT3 3 GSK R690693 H2N PRKG2 3.1 N PKNB (M.tuberculosis ) 3.2 N N PRKCE 5.3 O N N PRKX 7.2

HN O PAK7 8.9 PKAC Rbeta 13

Nature Biotechnology: doi:10.1038/nbt.1990 5 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM CSF1R 2.2

NH TRKB 36 2 TRKC 120 N TRKA 630 GW R2580 H2N N O O O

MAP4K5 0.65 EGFR 1.1 O ERBB4 2.4 N ERBB2 6 O MST3 6.5 HKI R272 N HN Cl MST4 7.4 HN N ERBB3 7.7 MAP4K3 7.7 O N YSK1 12 LOK 13 DDR1 0.7

N ABL1 Rnon phos phor ylated 1.1 ABL2 10 N NH NH N CSF1R 11 KIT 13 Imatinib N O PDGFRB 14 DDR2 15 ABL1 Rphos phor ylated 21 N PDGFRA 31 LCK 40 JAK2(JH1domain Rcatalytic) 0. 036 TYK2 (JH1domain Rcatal ytic ) 0.9 JAK3 (JH1domain Rcatal ytic ) 2 N JAK1 (JH1domain Rcatal ytic ) 3.4 N INCB018424 HN MAP3K2 41 CAMK2A 46 ROCK2 52 N N ROCK1 60 N DCAMKL1 68 DAPK1 72

NH CSF1R 3.2 KIT 3.6 O N N AMPK Ral pha2 4.1 NH AXL 5.3 JNJ R28312141 TYK2 (JH1domain Rcatal ytic ) 5.8 TAK1 7.2 RET 9.2 N N DDR1 9.4 CHEK1 9.9 O JAK2 (JH1domain Rcatal ytic ) 9.9 PDGFRB 0.29 O N PDGFRA 0.49 NH NH KIT 0.69 S CSF1R 0.83 O EPHB6 5 Ki R20227 O MEK5 7.4 O DDR1 12 VEGFR2 18 O N LOK 22 SLK 60 NH DRAK1 2.9 MLCK 4.3 N O DRAK2 4.8 YSK4 5.2 KW R2449 BIKE 9.6 MAP4K2 11 LOK 13 SLK 13 N FLT3 15 NH SRPK2 15

Nature Biotechnology: doi:10.1038/nbt.1990 6 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM EGFR 2.4 ERBB2 7 ERBB4 54 O F PIK3C2B 670 O PIK4CB 940 Lapatinib S O NH HN Cl MEK5 1100 O N SLK 3300 RIPK2 3600 N LOK 4400 MKK7 4400 GSK3B 8.3 O NH O PRKCE 8.9 PRKCD 25

N RSK4 (Kin.Dom.1 RNRterminal ) 25 LY R317615 N PRKCQ 36 PRKCH 46 N ERK8 76 N RSK2 (Kin.Dom.1 RNRterminal ) 87 DYRK1A 160 PRKG2 170 PRKCQ 2.5 NH O O PRKCD 3.6 PRKCE 11 PIM3 12 LY R333531 YSK4 48 N N HIPK3 77 ERK8 88 TAOK3 97 O N HIPK2 140 MAP3K3 170 IKK Rbeta 19 Cl N PIK4CB 270 TYK2 (JH2domain Rpseudokinase ) 500 CLK1 640 O NH MLN R120B CLK4 1000 O NH IKK Ral pha 1000 DYRK1A 2100 CLK2 2300 N PDGFRA 2.4 KIT 2.7 O NH FLT3 3 N O PDGFRB 4.5 CSF1R 4.9 MLN R518 N O DDR2 120 N EGFR 410 N O N TRKA 450 CLK1 630 IRAK3 730 AURKA 6.5 DRAK2 8.1 O F AURKC 26 N HO N AURKB 43 MLN R8054 BLK 68 NH N DRAK1 190 F FGR 220 YES 260 Cl TIE2 300 EPHB1 330 DDR1 1.1 ABL1 Rnon phos phor ylated 10 F N F ZAK 11 NH NH N ABL1 Rphos phor ylated 13 F ABL2 26 Nilotinib O N KIT 29 N DDR2 33 p38 Rbeta 36 N EPHA8 37 CSF1R 45

Nature Biotechnology: doi:10.1038/nbt.1990 7 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM PDGFRB 2 KIT 2.8 NHN N N PDGFRA 4.9 N CSF1R 7.9 N Pazopanib FLT1 14 O NH 2 VEGFR2 14 S FLT4 27 O TAOK3 45 DDR1 57 EPHB6 81 ABL1 Rphos phor ylated 0.58 Cl CSF1R 0.67 ABL2 0.69 SRC 0.71 PD R173955 N LCK 1.1 PDGFRB 1.4 Cl ABL1 Rnon phos phor ylated 1.5 S NH N N O BLK 1.5 KIT 1.8 EPHB6 2 MET 0.27 SRPK1 41

Cl TAOK3 43 NH N BIKE 47 N PHA R665752 GRK7 61 Cl O HIPK3 63 O S CAMKK1 64 O O DDR1 70 NH CAMKK2 73 YSK4 78 PIK3CA 1.5 O PIK3CB 1.7 PIK3C2B 10 N MTOR 12 PIK3CG 16 PI R103 O N PIK3CD 17 N OH PIK3C2G 41 N HIPK2 290 HIPK3 310 PIP5K2C 620 PKN1 9.3 TBK1 9.3 FLT3 11 O N JAK3 (JH1domain Rcatal ytic ) 12 O MLK1 15 PKC R412 O H PKN2 15 N N YSK4 15 MLK3 17 CAMK2A 20 NH O MARK3 21 MEK5 0.16 F PFCDPK1(P.falciparum) 1.7 SRMS 21 O ZAK 41 O O BRK 48 PLX R4720 Cl S NH FGR 62 F RAF1 170

N NH KIT 180 MEK4 190 NEK11 190 OH BMPR1B 2.1 ACVRL1 2.9 MTOR 3 ACVR1 4 RET 4.8 PP R242 NH 2 NH YSK4 5.1 N N MEK5 7.3 N N ACVR2B 7.6 PRKCE 9 JAK2 (JH1domain Rcatal ytic ) 11

Nature Biotechnology: doi:10.1038/nbt.1990 8 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM Cl KIT 5.1 FLT1 9.6 PDGFRB 25 HN CSF1R 45 PTK R787 N VEGFR2 62 N PDGFRA 96 DDR1 270 FLT4 330 N CDK11 1500 FRK 1800 FLT3 0.71 STK16 1.7 O GCN2 (Kin.Dom.2,S808G ) 3.3 O F O PDGFRB 3.3 N R406 JAK2 (JH1domain Rcatal ytic ) 3.5 MLK2 3.8 O NH N NH N NH O RET 4.1 PLK4 4.2 MLK1 4.3 PLK3 5.1 CDK2 0.53 NH N NH 2 PCTK1 0.54 CDK7 0.58 O N N O CDK4 RcyclinD1 0.61 S CDK4 RcyclinD3 0.81 R547 O O F PCTK2 0.86 ICK 2.2 CDK3 3.2 F PFTAIRE2 7.2 CDK5 7.4 p38 Ralpha 12 GAK 19 N RIPK2 24 O NH NLK 25 SB R203580 S JNK3 35 N CSNK1D 37 p38 Rbeta 70 F CSNK1A1 75 CSNK1E 100 JNK2 130 MET 0.19 DYRK1A 780 N DYRK1B 1800 N JNK3 1900 S N PIP5K2C 3300 SGX R523 N N JNK1 4200 N YSK4 6400 N

ABL1 Rphos phor ylated 0.057 ABL1 Rnonphosphorylated 0.12 Cl Cl MAP4K5 0.5 LCK 0.59 HN O SKI R606 O CN ERBB3 0.77 SRC 1 N O N GAK 1.3 N FRK 1.4 ABL2 1.5 STK35 2 DDR1 1.5 HIPK4 3.3 ZAK 6.3 CF 3 O DDR2 6.6 Cl O FLT3 13 Sorafenib O NH RET 13 N NH NH CSF1R 28 KIT 28 FLT1 31 PDGFRB 37

Nature Biotechnology: doi:10.1038/nbt.1990 9 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM SLK 0.024 HN LOK 0.037 O CAMKK1 0.039 O H SNARK 0.086 Staurosporine N N PHKG2 0.14 CAMK2A 0.16 CAMKK2 0.16 MST2 0.18 O MST1 0.19 NH TAOK3 0.22 PDGFRB 0.29 FLT3 0.54 O N H C O KIT 0.68 3 NH OH PDGFRA 1.1 VEGFR2 2.3 SU R14813 NH CH F 3 CSF1R 3.6 O HUNK 3.7 NH FLT1 4.7 STK35 8.2 YSK4 12 PDGFRB 0.075 KIT 0.37 F O FLT3 0.41 PDGFRA 0.79 N Sunitinib NH DRAK1 1 NH VEGFR2 1.5

NH O FLT1 1.8 CSF1R 2.5 BIKE 5.5 PHKG1 5.5 ROS1 0.49 N ULK1 0.83 N BMPR1B 0.85

Cl O N PLK4 0.93 TAE R684 N LTK 0.95 ALK 1.1 HN N NH O O FAK 1.1 S PYK2 1.1 SNARK 1.2 FER 1.4 PIK3C2G 3.2 OH PIK3CG 5.3

NH 2 PIK3C2B 7.3 N TRPM6 7.9 N TG R100 R115 CLK2 43 PIK3CA 59 H2N N N PIK3CB 80 ADCK3 94 RIPK4 97 OH CLK4 130 GAK 1.1 JAK2(JH1domain R catalytic) 1.1 DAPK3 1.2 NH STK16 6.6 S NH TG R101348 DCAMKL3 13 OO O FLT3 13 N N DAPK1 16 N NH JAK1 (JH1domain Rcatal ytic ) 18 YSK4 19 TYK2 (JH1domain Rcatal ytic ) 21 Br RIPK2 4.6 EGFR 9.5 DDR1 11 HN ABL1 Rphos phor ylated 16 Vandetanib O F LCK 17 N RET 34 ABL1 Rnon phos phor ylated 48 O N MEK5 49 N EPHA6 50 STK35 56

Nature Biotechnology: doi:10.1038/nbt.1990 10 Supplementary Table 2. Compound structures and highest affinity target kinases (excluding mutants)

Com pound Structure Kinase Kd, !nM AURKA 3.9 ABL2 4 N AURKC 6.3 HN NH FLT3 6.5 AURKB 7.4 VX R680/MK R0457 NH N ABL1 Rphos phor ylated 7.5 O PLK4 9.2 N N S ABL1 Rnon phos phor ylated 13 N MLCK 15 RIPK1 20 p38 Ral pha 2.8 F p38 Rbeta 74 S N DDR1 1100 N N FGR 1300 VX R745 YES 1600 F O LYN 1700 Cl Cl ABL2 1900 FYN 2100 CSF1R 2600 BLK 3100

Nature Biotechnology: doi:10.1038/nbt.1990 11 © 2011 Nature America, Inc. All rights reserved. of 300 protein kinases in duplicate. Our goals were goals Our duplicate. in kinases protein 300 of panel a against inhibitors kinase known 178 of screen parallel scale NATURE BIOTECHNOLOGY NATURE 1 is inhibition a tive kinase major challenge However, achieving compounds. chemical by synthetic ATP-binding that of ence pocket conserved a highly targets because of their central roles in cellular classes important most the are among kinases Protein Theonie Anastassiadis reveals features of kinase inhibitor selectivity assay Comprehensive of kinase catalytic activity to identify by conventional target-centric screens target-centric by conventional identify to target primary tha rather targets kinase specific of number small a is focus activity inhibitory whose inhibitors, multitargeted intere of kinases for inhibitors new unexpected i identify approach parallel This optimized. chemically then id are patterns selectivity desired showing Compounds Received Received 1 September; accepted 23 September; publis should be addressed to J.R.P. (jeffrey.peterson@fcc inhibit multiple kinases. of Interpretation experim agents. Owing to evolutionary conservation of the A protein Small-molecule kinase inhibitors are widely binant protein kinases to reveal the selectivity of selectivity the to reveal binant kinases protein pa comprehensive a against manner target-blind a in libraries screening involves approach, alternative An are then tested for against selectivity a panel of rep and a of of kinase molecules The interest. res small centric manner involving high-throughput screening unknown. important an inhibi functional predict to assays these of ability of catalyt inhibition functional than rather to kinases, cases, however, these methods measure the binding of small molecules against sizable fractions against of sizable fractions the 518 human protein k tar kinase to profile of methods to development led the technologi of Recent part kinome. the a substantial However, of inhibitors the kinase is selectivity sel a in research of both inhibitors the effects preting of kinase inhibitors for their targets is critical fo elucidate elucidate kinase function and for interpreting the specific, diverse kinases. The results have implica interactions occur with seemingly unrelated kinases unexpected interactions between protein kinases and available kinase inhibitors against a panel of 300 kinase comprehensive selectivity of most inhibitors Cancer Cancer Biology Program, Fox Chase Cancer Center, Ph e sd hg-hogpt nyai asy o conduc to assay enzymatic high-throughput a used We Traditionally, kinase inhibitors have been discover been have inhibitors kinase Traditionally, 11,12 . Indeed, multitargeted inhibitors are challenging challenging are inhibitors multitargeted Indeed, .

1 ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE , Sean , W Sean Deacon 1–6 c.edu). . Knowing the selectivity . Knowing the selectivity signaling and signaling the pres- iladelphia, Pennsylvania, iladelphia, USA. Pennsylvania, r r predicting and inter- hed hed online 30 October 2011; dom assessed dom across assessed each compoundeach nd clinical settings. settings. nd clinical resentative kinases. resentative kinases. 13 2 ulting compounds ulting tion thus remains remains thus tion cal advances have advances cal inases , , Karthik Devarajan to identify novel identify to . n toward a single single a toward n of large numbers can be exploited can be exploited of compounds compounds of results results of experiments involving kinase inhibitors. ents ents that use these compounds is confounded by a la recombinant recombinant protein kinases. Quantitative analysis ic activity. The activity. ic tions tions for drug development and provide a resource f ed in a target- a in ed s predicted to to predicted s of therapeutic of therapeutic nel of recom- of nel TP-binding TP-binding site, most kinase inhibitors that target get selectivity get selectivity . . Here we used functional assays to profile the act used to elucidate cellular signaling pathways and , , revealing how large-scale profiling can identify highly selec- highly st and reveal reveal and st entified and and entified 7,8 kinase inhibitors, with a wide spectrum of promisc ed toward ed a large- a t . . In many 9,10 . constructs constructs and substrates used is provided in listin complete A tested. compounds the of 87.6% of protein kinase families ( families kinase protein Table 1Supplementary inhibit kinases from all inhibit major kinases protein subfa kinase compoun 178 comprised library The tools. research as and testing, compounds in compounds drugs, clinical Administration Drug and Food US included inhibitors assays for kinase inhibition are compared are inhibition kinase for assays which against standard the is method well-validated specifi a toward activity catalytic kinase measures w assays, filter-binding conventional on based assay ric a WeHotSpot, used kinases. protein human recombinant using assays kinase low-volume conducted we itors, To directly test the kinase selectivity of a large map interaction A kinase-inhibitor RESULTS points for the development of multitargeted kinase i for orphan kinases for which few inhibitors currentl commonly used kinase inhibitors. In addition, we re small-molecule inhibition, and unexpected off-target commonly inhibited by many compounds, kinases that a Systematic, quantitative analysis of the results re measuring pairwise inhibition of a single enzyme by resultsprisesgenerated from>100,000independent knowledgelargestavailableputhe type itsofthe in result The families. kinase major the of all geting clinic and compounds research used widely represent specificities of a large panel of kinase inhibitors. inhibitor chemotypes for specific kinase targets and doi:10.1038/nbt.2017 2 Reaction Reaction Biology Corporation, Malvern, Pennsylvania h kns pnl etd nlds ebr o al maj all of members includes tested panel kinase The 1 , , Haiching Ma 2 & Jeffrey R Peterson ). Fig. 1b Fig. ) and includes the intended targets targets intended the includes and ) multitargeted inhibitors multitargeted of this site promiscuously ivity ivity of 178 commercially revealed revealed complex and often or or selecting compounds to are are promising therapeutic ck ck of data on the uity. Many off-target 7 . Our collection of kinase kinase of collection Our . Supplementary Supplementary Table 2 number of kinase inhib- The compounds tested , , USA. Correspondence vealed kinases that are E C R U O S E R ing data set, to our to set, data ing a single compound. port potential leads, 1 milies (milies to reveal the target y exist and starting blic domain,blic com- nhibitors. activities of many c substrate. This This substrate. c functionalassays g of the kinase kinase the of g more indirect indirect more primarily used primarily a panel of 300 300 of panel a l gns tar- agents al re resistant to ds known to to known ds hich directly directly hich Fig. Fig. 1a –approved –approved or human human or radiomet- and 1 . © 2011 Nature America, Inc. All rights reserved. (0.18% of the data set) from the analysis analysis the from (Online Methods set) and data the of (0.18% and identified eliminated disparate replicates We activity). kinase remaining as to referred (henceforth reactions control solvent of age percent- a as results phosphorylation strate sub- average the expressed and duplicate in inhibitor combination (kinase-inhibitor pair) activity. inhibitory off-target weaker capture targets of 66 nM, we chose to use 0.5 0.5 use to chose we nM, 66 of targets primary their toward compounds these for loaded or analyzed within a browser window ( window set browser a within data analyzed or the of loaded queries specific kinase or compound webs internet an database, (KIR) Resource Inhibitor in map heat a as presented is pair tor observed. was inhibition kinase 20% in at pairs which least for kinase-inhibitor all replicate, second the versus replicate one in activity kinase remaining the as plotted pair kinase-inhibitor one represents point each which in plot scatter a as set data resulting Figure 1c in in of the inhibitor library, by kinase family. ( library, kinase by inhibitor the of aiu ihbtr cnetain (IC concentration inhibitory maximum presented in in presented is map heat this of version high-resolution labeled, activity. A fully kinase of map a heat as presented map interaction kinase-inhibitor entire the of analysis clustering hierarchical 2 2 0.5 of concentration a at tested were compounds all simplicity, For (analysis. 1Figure E C R U O S E R basis for isoform-specific inhibition isoform-specific for basis revealed have studies Aurora structural of and reported inhibitors been isoform-specific finding, this molecules. small by targeted differentially be can 10 10 lution form in in form lution of kinase activity was observed. ( observed. was activity kinase of inhibition >20% which for pair kinase-inhibitor each 2 for replicate 1 versus replicate in Fig. 2 Fig. ( kinases family and FGFR PDGFR Aurora, relevant by of t members included inhibited Exceptions were of compounds. terns and clustered often were identity r closely kinases Similarly, t together. grouped ally of compoun related similarity the structurally on expected, As patterns. based cl to inhibitors and performed kinases was clustering hierarchical Two-way ref. 33). ( 33). ref. on Technology based Signaling Cell of courtesy reproduced is and adapted was illustration (kinome kinome human the representing a on dendrogram dots blue by represented is panel screening the in kinases of distribution data table in in table data ( tool online (KIR) Resource Inhibitor Kinase the http://kir.fccc.edu/ We tested each protein kinase and kinase kinase and kinase protein each Wetested h ma rmiig iae ciiy o ec kinase- each for activity kinase remaining mean The M upeetr Tbe 3 Table Supplementary M M ATP. Despite an average reported half- ), suggesting the possibility that members of these of members that possibility the suggesting ), Large-scale kinase-inhibitor interaction interaction kinase-inhibitor Large-scale c a illustrates the reproducibility of the ) Scatter plot of the kinase activity activity kinase the of plot ) Scatter ) Distribution of the intended targets targets intended the of ) Distribution Supplementary Figure 2 Figure Supplementary Supplementary Table3 Supplementary upeetr Fgr 2 Figure Supplementary ). Ctrl, control. Ctrl, ). Supplementary Fig.1 M M in the presence of presence the in M . In addition, we created the Kinase Kinase the created we addition, In . d ) ) Two-way and via via and , as , a as b 14 M ) The ) The M to to M . 50 Figure 1d Figure ). ) and as a spreadsheet spreadsheet a as and http://kir.fccc.edu/ d a inhibited inhibited elated by sequence by sequence elated Most Least Consistent with with Consistent Supplementary Supplementary , in high-reso- in , Distribution ofintended ds were gener- were ds the structural structural the ite that allows allows that ite Kinases kinases have have kinases to be down- be to similar pat- similar heir activity activity heir TK inhibitor targets he clinically clinically he Kinase activity uster both both uster CK1 No data 0–25 26–50 51–75 76–100 (% ctrl) families families CMGC inhibi- TKL Inactive STE AGC ). ). Other Atypical Lipid CAMK pound bindingpound denatura thermal from kinases protecting involves ing >50% inhibition, as expected. kinase-inhibitor pairs with low affinity (weaker tha Conversely,on (>50%). inhibition functional showed affinities for a large number of kinase-inhibitor i to kinases. Two recent studies used a competitive b data with previous large-scale studies of the bindi ity remains uncertain. To assess this, we compared o extent to which these binding assays predict inhibi tion of kinase catalytic activity. Although convenie detect kinase-compound interactions without directl A variety of high-throughput screening approaches platforms assay h multiple across data of Comparison kinase-inhibitor interactions with high affinity (s activitymeasured oursingle-dosein study ( the withagreement good generallyshowed affinities dred fifty-four kinase-inhibitor pairs overlapped w An alternative approach to monitoring kinase-compou monitoring to approach alternative An CMGC CAMK ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE c b TK Kinases screened 3 . . Toinhibiti kinase approach to this predict assess AGC CK1 Compounds STE TKL

Replicate no. 2 Kinase activity (% ctrl) 60 80 20 40 0 NATURE BIOTECHNOLOGY NATURE 04 080 60 40 20 0 Fig.2a tronger than 100 nM nteractions ith our study and their ng of small molecules Kinase activity(%ctrl) tion of catalytic activ- inding assay to derive nt for screening, the n 1 ur kinase inhibition y measuring inhibi- ). Indeed,). 90.2%of ave been devised to Replicate no. 1 Replicate no. expected kinase expected M ly 13.1% of the of 13.1% ly M tion by com- by tion 1,2 K d . Six hun- nd bind- nd ) showed K on, on, d ) © 2011 Nature America, Inc. All rights reserved. showed showed numbe significant a However, predicted. as activity, inhibiti showed also temperature melting kinase the (pair kinase-inhibitor NATURE BIOTECHNOLOGY NATURE temperature melting reported in change of the function func our in activity kinase remaining the plotted we 2Figure kinases with respect to a selectivity score (S score selectivity a to respect with kinases thi do To inhibition. small-molecule to sensitivity be inhibited by a given compound or whether kinases dif was panel the in kinase each We whether asked next druggability kinase of Analysis cor significantly are inhibition and binding ity, although inhi that functionally compounds to ability predict appreciable false-positive and false-negative rates w bindi kinase-inhibitor that suggest together, taken these from findings The quadrant). dashed left lower including FLT3, TRKC and HGK/MAP4K4 were broadly in su broadly were a HGK/MAP4K4 and TRKC contrast, FLT3, By including successful. less be may scaffolds traditi using screens for list a which target suggest panel were not inhibited by any of the compounds tested by compounds any of the inhibited not were panel by (>50% activity catalytic the inhibited that tested compounds tor set. The untargeted kinases, including COT1, NEK6 COT1, including kinases, untargeted The set. tor b kinome the of coverage good demonstrating inset), tion of kinase activity without exhibiting exhibiting without activity kinase of tion 117 out s pairs of inhibitor 3,926 rant). Likewise, by activity kinase >50% (inhibiting In In g the of inhibitor an for activity kinase remaining (dotte curve a theoretical with compared be can and dose-respons a sigmoidal to fit were data resulting ( pr interaction kinase-inhibitor previous with study compound binding affinity ( affinity binding compound inhibitor binding assay binding inhibitor pairs kinase-inhibitor overlapping of interactions denaturation by kinases in the presence of individu of presence the in kinases by denaturation affect thermal stability yet show >50% inhibition o inhibition >50% show yet stability thermal affect in without stabilization thermal significant showed thres 50% the highlights line horizontal dashed The In In to solvent control, caused by compound binding for for binding compound by caused control, solvent to a , b b a , remaining kinase activity is plotted as a functio as plotted is activity kinase , remaining ) Scatter plots compare our results with studies th studies with results our compare plots ) Scatter remaining kinase activity is plotted against the c the against plotted is activity kinase remaining Selectivity score, S(50%) 0.1 0.2 0.3 0.4 Comparison of functional inhibition data generated data inhibition functional of Comparison T 0 m Fig. Fig. 3 changes >4 °C, the hit threshold used previously and COT1/MAP3K8 P38 MAPKAPK3 CTK MATK DYRK4 Supplementary Table 4Supplementary
HIPK1 WNK3 GRK2 GRK3 VRK1 NEK7 NEK6 1,2 JNK3 P38 /MAPK13 Fig. Fig. 2b (  a K ), or an assay measuring resistance to thermal thermal to resistance measuring assay an or ),

d ) for 654 kinase-inhibitor pairs. The The pairs. kinase-inhibitor 654 ) for ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE ). Generally, compounds that increased ). that increased Generally, compounds Individual kinases Fig. Fig. 2b T , , upper right quad- dashed m (50%) changes >4 °C ( °C >4 changes ). Only 14 kinases in ). the 14 Only kinases iven affinity. affinity. iven by a quantitative kinase- a quantitative by HGK MAP4K4 ofiling studies. studies. ofiling KHS MAP4K5 ARK5/NUAK1 e curve (solid line) line) (solid e curve hibiting kinase activity whereas the lower left qua left lower the whereas activity kinase hibiting f catalytic activity. Ctrl, control. Ctrl, activity. f catalytic al inhibitorsal YES/YES1 3,926 kinase-inhibitor pairs. The dashed vertical l vertical dashed The pairs. kinase-inhibitor 3,926 hold for inhibition of catalytic activity. The resu The activity. catalytic of inhibition for hold d line) for expected expected for d line) howed >50% howed inhibi- >50% n of kinase-n of TRKC FLT3 hange in in hange ), the fraction of all all of fraction the ), RET at examined examined at onal ATP-mimetic onal ith ith respect to their bit catalytic activ- bit catalytic ng assays exhibit exhibit assays ng s, we ranked the the ranked we s, r of compounds compounds of r tional assay as a as assay tional equally likely to likely equally bset of kinases kinases of bset on of catalytic catalytic of on comparisons, comparisons, of each kinase kinase each of y this inhibi- this y in this this in fered fered in their /7 and p38 and /7 ( related. ( T T 3 Fig. 3 Fig. m m 3 ( , , without , relative , relative ) of each ) of each b Fig. 2b Fig. hibited hibited ). ). , left , left D , , in in comp The (right). kinases inhibited frequently most compound any by inhibited not were that kinases 14 k individual an of score selectivity the represents activi catalytic inhibit that inhibitors tested all of the intended kinase target ( target kinase intended t the than of subfamily kinase different a from were pound the analysis. On average, 42% of the inhibit kinases targets, we removed the top ten most promiscuous compo out of frequency apparent the increase would compounds versus outside the family of the primary target. As same the within are that targets kinase of fraction quantitatively this Totest sequence. acid amino by most kinases with found be to likely more are actions that assumption on the based kinases related closely a against testing by assessed frequently is inhibitors selectivity The target(s). inhibitor the knowing on d experiments such of results the of Interpretation the of inactivation acute of consequence biological t tools research as used commonly are inhibitors Kinase selectivity inhibitor Kinase c our in biases hidden reflect could results the that possibi the exclude We completely cannot kinases. tive in should panels that screening and panels suggests selectivity inhibitor kinase of assessment the for import has molecules small to sensitivity kinase of inhibition. chemical to susceptible highly kinases by large numbers of compounds (right inset), potenti (S 3Figure assess kinase inhibitor selectivity. A selectivity selectivity A selectivity. qua inhibitor to kinase assess proposed been have methods Various function. tools research as valuable most are targets kinase possible. as kinases of panel a broad as the of importance of the assessing selectivity kina re These compounds. tool research than specificity to optimized more likely therefore, are, and set compound clinica of the all almost include compounds these that par in explained, be may compounds kinase-targeting selecti within-family The kinases. serine/threonine off- of 24% kinases, tyrosine against developed tors defined, where S is the number of kinases bound by by bound kinases of number the is S where defined, b a (50%) Supplementary Table 4Supplementary Inhibitors that exhibit selectivity for a very limi very a for selectivity exhibit that Inhibitors Kinase activity (% ctrl) (this study) 100 150 50 ) for all tested kinases. This score corresponds to corresponds score This kinases. tested all ) for 0 Kinase selectivity. A ranked bar chart of selectiv of chart bar A ranked selectivity. Kinase –10 9–8 –9 log lting upper right quadrant includes compounds that that compounds includes quadrant right upper lting drant contains compounds which only marginally marginally only which compounds contains drant K ine denotes the the denotes ine d (M) . 7– –5 –6 –7 Supplementary Fig. 3 Fig. Supplementary T

m Kinase activity (% ctrl)

shift threshold used in ref. 3. ref. in used threshold shift (this study) 1 020 10 4 0 –10 ty by >50%. Each bar bar Each >50%. by ty inase. Insets identify the the identify Insets inase. 150 lete table is presented presented is table lete 50 E C R U O S E R 0 highly promiscuous score S( score se se inhibitors against (left) and the seven seven the and (left) T for probing kinase kinase probing for ed ed by a given com- m with small kinase kinase small with ompound library.ompound clude these sensi- clude these This broad range range broad This shift ( shift the fraction of fraction the kinase subfamily subfamily kinase , we assessed the the assessed we , ir kinase targets. targets. kinase ir epends critically critically epends limited panel of panel limited ant implications implications ant ally ally representing of novel kinase kinase novel of target hits were were hits target ity scores scores ity off-target inter- off-target vity of tyrosine tyrosine of vity ted number of of number ted sults highlight highlight sults closely related related closely l agents in our in l agents ). For inhibi- For ). lity,however, he subfamily subfamily he t, by the fact fact the by t, ° an inhibitor inhibitor an with regard regard with C) x o reveal the the reveal o unds unds before ) has been been has ) ntitatively ntitatively -of-family -of-family 3 © 2011 Nature America, Inc. All rights reserved. Table 5 analogs exhibited the lowest Gini scores ( scores Gini lowest the exhibited analogs but did not observe a consistent linear correlation linear consistent a observe not did but with to respect properties either the Gini score or promiscuity correlations fied p compound between physicochemical hav studies profiling kinase-inhibitor previous promiscuity, spots). (darker targets kinas residual lower producing by score Gini higher masitin dendrograms), right and (middle scores Gini were kinases targeted by the compounds with the medi comparab a Although panels. bottom the in shown are highest and median lowest, the with compounds three tar The kinases. family ErbB of inhibitors distinct ( compounds Amo selective spectrum. target broad known their with sistent tested cuous to the most selective ( selective to most the cuous of results to analysis this rank the from compounds t scor (a Gini kinases tested all across equally uted score Gini (a target single a only toward directed of sum as inhibition the (calculated of a compound aggregate the which to degree the 1, to 0 of re scale score Gini The concentration. screening compound strongly is it although threshold, hit arbitrary an CHK2. Each bar corresponds to the percent remaining percent the to corresponds bar Each CHK2. Erb for selective highly inhibitor, a multitargeted benzyloxyanilino)-6,7-dimethoxyquinazoline, Technology. ( Signaling Cell of courtesy reproduced is and adapted was dendrogram kinome The activity. 25–50% lightest, activity; 10–25% lighter, activity; remaining 0–10% darkest, potency: coefficientb selectivity inhibitor kinase for metric calculated described therefore We scored. be not thres could hit kinase the meet not did that compounds addition, similarity sequence acid amino on based kinases human all of a dendrogram on presented are compounds representative three of selectivity in presented is table complete The compounds. scoring highest five the inset, right the and scores Gini lowest the with compounds five the highlights inset Left inhibition). (selective kinase one only of inhibition 1 indicates of a score whereas inhibition) (promiscuous kinases all of inhibition equal indicates 0 of score A Gini selectivity. inhibitor 4 4 ( 4Figure (with an greater than affinity E C R U O S E R sorted by Gini score Gini by sorted hit thresholds can produce different rank orders of orders rank different produce can thresholds hit scores generated from the selectivity same data set (no threshold this below just kinases other of tion gre a despite kinases, of number limited a with old met they favorably scored compounds because t several as inhibition percent arbitrary an using data our on an arbitrary hit threshold on (hit an threshold arbitrary a ) A ranked list of kinase inhibitors inhibitors kinase of list ) A ranked To understand the molecular features that contribute contribute that features molecular the Tounderstand 2 . A critical limitation of the selectivity score is score of selectivity the limitation . A critical ). Not surprisingly, staurosporine and several of its s and several ). staurosporine Not surprisingly, Kinase inhibitor selectivity. selectivity. inhibitor Kinase 33 . Spot color represents inhibitory inhibitory represents color . Spot 15 13,16 b Supplementary Table 5Supplementary . Importantly, this method does not depend on defin not depend does method . this Importantly, ) Target spectrum of 4-(4-of ) Target spectrum . We analyzed a variety of compound physicochemical 15 as a measure of a measure as Fig. 4a Fig. Fig. 4a Fig. x x , right inset) were several structurally structurally several were inset) right ,

M

M M) divided M) by divided the number of kinases M). M). For example, we when analyzed . Below, the . the Below, ; complete list in list ; complete Fig. 4a Fig. B family members, a limited number of other tyrosin other of number a limited members, B family the selectivity score the selectivity e of 0). e Weof 0). the used b a activity for an individual kinase. individual an for activity inhibitory activity activity inhibitory influenced by the the by influenced but using different of 1) or is distrib- is or 1) of t shown). Indeed, Indeed, shown). t get spectra of the the of spectra get

Gini score con- inset), left , ased on the Gini Gini the on ased 0.25 0.50 0.75 1.00 at deal of inhibi- of deal at compounds for all kinases) is kinases) for all tuoprn GTP-14564 Staurosporine with any single single any with he he most promis- he hit criterion, criterion, hit he Supplementary Supplementary its dependence its dependence e activity in its its in activity e 0 the hit thresh- an an and highest a previously previously a roperties and roperties ib achieves a achieves ib le number of of number le hold for any any for hold ng the most most the ng to inhibitor inhibitor to Gini scores scores Gini flects, on a on flects, Cdk1/2 inhibitorIII tructural tructural Staurosporine PKR inhibitor e identi- e SB 218078 K-252a 2 . In . ing ing EGFR/ErbB-2/ErbB-4 inhibitor EGFR/ErbB-2 inhibitor 4-(4-benzyloxyanilino)-6,7-dimethoxyquinazoline family The ErbB kinases. of number cancer-relevant ag activities off-target with inhibitors of examples for CHK2, CHK2, a critical component of the DNA damage checkpserine/ the of inhibition potent surprisingly, most inhibition of a few tyrosine kinases beyond ErbB fa most still targeted multiple kinases with similar p similar with kinases multiple targeted still most identified inhibitors selective most the among Even inhibitors kinase uni-specific of targets Novel interest.of targets multiple toward activity with identification the can facilitate data sets similar min Data interest. therapeutic of kinases divergent that show acti novel scaffolds unanticipated reveal CHK2 CHK2 are trialsin clinical in and therapies cancer of DNA-damaging impact apeutic CHK2 inhibition has been proposed as a to strategy in ity against specific kinases of interest of kinases specific against ity that show promi scaffolds compound to approach identify be has libraries inhibitor large of profiling kinase Importantly, this stringent criterion excludes compounds excludes criterion stringent this Importantly, termed we characteristic a panel, our in other any any whether compounds a inhibited mor kinase single We kinase. single a of function the elucidate to tools us their confounding therefore dendrogram), rightmost develop single compounds with a desired target spectrum challe technical a difficult it remains pharmacology, selectivity selectivity a manner. to linear simple, a in parameter related strongly be to unlikely is promiscuity suggest studies previous the of findings discrepant compound property ( property compound tic effect tic than effect targeting a single kinase relevant targets with the goal of clinically achieving li a toward selectivity significant show should they inhibitors promiscuous from differ compounds such Ideally, to to increasing interest multitargeted kina in so-called Tandutinib Sorafenib Masitinib The clinical success of some kinase inhibitors that show poor kinase ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE in vitro Masitinib (e.g., dasatinib (Sprycel), sunitinib (Sutent)) has l e kinase targets and the serine/threonine kinase kinase serine/threonine the and targets e kinase Supplementary Fig. 4 Fig. Supplementary Kinase activity (% ctrl) 100 120 20 40 60 80 21 0 . This illustrates . how can profiling This illustrates kinase BRK MER CHK2 EGFR 18 9,19 ErbB4 CAS no.179248-61-4 EGFR/ErbB-2 inhibitor ErbB2 . . Despite the promise of poly- . We interrogated our data data our Weinterrogated . NATURE BIOTECHNOLOGY NATURE ). This finding and the the and finding This ). LCK of inhibitor scaffolds of scaffolds inhibitor mily mily members and, en suggested as an an as suggested en vity against highly highly against vity greater greater therapeu- threonine kinase kinase threonine 20 mited number of of number mited se se inhibitors that compound compound that otency ( otency nge to rationally nge to rationally ny one physical physical one ny ‘uni-specificity’. ‘uni-specificity’. therefore asked asked therefore kinase inhibitor inhibitor kinase ing of this and and this of ing showed potent potent showed crease crease the ther- by the screen, screen, the by e e potently than ainst a limited limited a ainst oint ( e as research e as research 18,19 that target target that sing activ- sing hibitors of hibitors . Parallel Fig. 4b Fig. 4a Fig. in that that in 12,17 YES ed ed ). ). , . © 2011 Nature America, Inc. All rights reserved. its most sensitive target and the right boundary re boundary right the and target sensitive most its t kinase other any than potently more kinase single inhibited targets. Only compounds with a differenti with compounds Only targets. inhibited o length horizontal the Thus, 5%). of bins in shown NATURE BIOTECHNOLOGY NATURE ( target inhibited potently o activity kinase remaining the indicates edge most kin inhibited potently most the for activity kinase denotes bar the of edge leftmost the which in graph as results the Weplotted least. to score) numerical kinase. Compounds were then ranked from most uni-specific (highest the of most next the from activity the of kinase inhibited activity kinase remaining the subtracting by for calculated was score uni-specificity A panel. ing homolog close without targets kinase for bias a has subfamily. same the from if isoforms related even closely are potency, similar with kinase one than more 5Figure their primary and secondary targets (short bars in in bars (short targets secondary and primary their differ potency slight only showed these of most and targ these against potency absolute the indicates bar two most sensitive kinase targets, and the left-right i of potency differential the denotes bar, therefore, other kinase in ( panel the kinase other po more 20% least at target Ninet primary their kinases. inhibited related closely of ac of inhibition challenge ential the highlights finding This panel). with a reported IC reported a with tor, a EGFR 4,6-dianilinopyrimidine inhibitor (CAS n uni- most the In fact, (EGFR). receptor factor growth kinases specific are several inhibitors intended to is a negative control compound for the p38 MAP kina MAP p38 the for compound control a is negative e the panel, gray.in right shown the In are target a targets, intended potency, their differential 20% 01 03 544 06 08075 70 60 4540 50 35 30 20 15 10 5 Few compounds in the panel showed any degree of uni-s any showed degree panel in the Few compounds Uni-specific kinase inhibitors. The left panel pre panel left The inhibitors. kinase Uni-specific 25 Kinase activityin5%bins 50 of 21 nM for EGFR for nM 21 of 565 55 Fig. 5 Fig. Fig. 5 Fig.

ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE , middle). Among these 19 uni-most Among these , middle). , leftmost panel). The length of each each of length The panel). leftmost , 22 , inhibited EGFR catalytic catalytic EGFR inhibited , ffect of the individual compounds on each kinase in kinase each on compounds individual the of ffect GSK-3b inhibitor p38 MAPkinase N niio llJNK JNK inhibitorVlll JNK inhibitorlX Chk2 inhibitorll flects the potency with which the next most sensiti most next the which with potency the flects EGFR inhibitor Akt inhibitor X Akt inhibitor nd their most sensitive targets. Six compounds for for compounds Six targets. sensitive most their nd ested. The left boundary of each horizontal bar dep bar horizontal each of boundary left The ested. SB 202474* ATM kinase GTP-14564 al potency of at least 5% are shown. The central ta central The shown. are 5% least at of potency al PD 158780 PD 174265 f each bar reflects the differential activity of th of activity differential the reflects bar f each adtnb387867-13-2 Tandutinib SC-68376 inhibitor lll AG 1478 AG 1024 Gefitinib inhibitor BPIQ-l target the epidermal Name se inhibitor SB 202190. **, ATM incl **, not was 202190. SB kinase inhibitor se DMBI Vlll potently inhibited inhibited potently positioning positioning of this ase and the right- the and ase nhibition of these these of nhibition f the second most f second the een compounds compounds een a horizontal bar bar horizontal a each compound compound each ets. s in the screen- the in s Fig. 5 Fig. o. 879127-07-8) tently than any than tently sents a graphical table of compounds ranked based o based ranked compounds of table a graphical sents specific inhibi- specific ivn differ- hieving the remaining remaining the ences between between ences most potently potently most In addition, it it addition, In those kinases kinases those , leftmost leftmost , 8812- ATM** 587871-26-9 184475-35-2 172747-50-1 894804-07-0 175178-82-2 581098-48-8 925681-41-0 318480-82-9 487021-52-3 171179-06-9 174709-30-9 216163-53-0 312917-14-9 516480-79-8 879127-07-8 pecificity pecificity 34823-86-4 65678-07-1 5812-07-7 CAS no. FLT3/ c-KIT p38 MAPK NCIEHaspin INACTIVE KIT/ FLT3 c-FMS/ c- Intended PDGFR IGF-1R PDGFR S- GSK-3 GSK-3 MRCK by activity >94% but its inhibited next most poten inhibition among the closely related ErbB family kinase family ErbB related closely the among inhibition this compound the also ability highlights to achieve ( int were they kinases the than potently more kinases target. therapeutic important inh of development the to devoted attention unequal or, EGFR of features unique reflect could here tified EG of and this uni-specific other selectivity matic gis te oe tres ee ofre ( confirmed were targets novel the against intended andnovel cases their targets.In both all the dose-response relationship for five uni-specific sensitivity of different to kinases the same compou identified. compounds specific one of the AG1024, inhibitor kinase tyrosine of IGF1R the sens more much a as RIPK2 kinase serine/threonine the exampl For target. intended the of subfamily kinase falls hit off-target potent most the DMBI, compound cas one but all in Remarkably, compounds. these of gets unknow hitherto represent hits off-target potent more I panel. ATM of screening not the a because was part wa inhibitor all kinase ATM The against plot. sorted a compounds as 6 panel these of 5 of activity the GREGFR EGFR GREGFR EGFR EGFR EGFR EGFREGFR EGFR H2CHK2 CHK2 target p38 AKT JNK Fig. 5 Fig. Strikingly, 6 of the top 17 uni-specific compounds in compounds uni-specific 17 top 6 of the Strikingly, To validate the use of our single-dose screening da screening single-dose our of use the validate To
, center, gray rows). The rightmost panel of of panel rightmost The rows). gray center, , A PDGFR Primary RIPK2 EGFR EFGR EGFR JNK3 JNK3 CK1 PIM3 FLT3 FLT3 PIM3 p38 , by only 22%. In contrast to other EGFR inhibitors EGFR other to contrast In 22%. only by , hit

Kinase activity(%ctrl) e corresponding inhibitor against its two most pote most two its against inhibitor e corresponding 100 100 100 100 100 10 10 10 10 10 ve kinase is inhibited (% remaining kinase activity kinase remaining (% inhibited is kinase ve the panel is shown in a ranked plot. *, SB 202474 202474 SB *, plot. a in ranked shown is panel the which the most sensitive target is not the intended intended the not is target sensitive most the which ble identifies the compounds that showed at least least at showed that compounds the identifies ble icts the potency with which the compound inhibits inhibits compound the which with potency the icts 1 1 1 1 1 uded in our test panel. Ctrl, control. Ctrl, panel. test our in uded n the compound’s ability to inhibit a inhibit to ability compound’s n the Individual kinases 0100 300 100 10 300 100 10 300 100 10 300 100 10 10 upeetr Fg 5 Fig. Supplementary E C R U O S E R the greater potency greater potency the compounds against FR inhibitors iden- FR inhibitors tly inhibited target, tly inhibited nd, we determined isoform-selective isoform-selective more likely, the likely,the more Figure 5 Figure n all cases these these cases n all e, we identified identified we e, ended to target target to ended kinases in the the in kinases s not included included not s outside of the the of outside ta to rank the the rank to ta ibitors of this this of ibitors hibited other other hibited n kinase tar- kinase n s e, that of the the of that e, 22 itive target target itive . The dra- . The most uni-most tested, tested, shows shows ntly ntly 300 is ). ). 5 © 2011 Nature America, Inc. All rights reserved. of new inhibitors selective for particular kinase s kinase for particular selective inhibitors of new data set in guiding both tool compound selection an uti the illustrate findings These subfamilies. other su this of divergence A sequence greater to inhibit due perhaps selectively to employed be can chemotypes variet a that demonstrating subfamily, kinase AGC the othe all almost are compounds these by shared targets no clear structural similarity in the compounds. In warranted if attempting to extrapolate these these extrapolate to attempting if warranted ro and rapid economical, is activity kinase measure compounds. these for unk previously represent compounds these cases, In all in their than potently more kinases other inhibit that comp tion of Six new uni-specific targets. inhibitor s a We only is to target inhibited. priorit metric this used which within window dosing widest the vide e are uni-specificity of degree greatest the exhibiting and tar most its sensitive kinase next most sensitive i an of activity differential the assessing titatively Inhibitor IV and Y-27632 together ( together Y-27632 and IV Inhibitor Rockou clustered spectrum target on based h clustering Strikingly, ROCK. than KHS and PRKX of inhibition other kinase (data not shown). potent By more contrast, fasudil significantly II and I ROCK both indeed, inhibited and, (0.738) glycyl-H-1152 for selectivity agent fasudil (HA-1077) fasudil agent t and Y-27632 IV), Inhibitor Kinase (Rho glycyl-H-1152 kinase (Rho-associated ROCK kinase AGC the subfamily four well-establishe contains collection the compound functi kinase investigating for inhibitor cho selective for basis powerful a provide should interest of targeting inhibitors multiple across scores Gini of prediction of cellular efficacy. First, our screen w our First, screen efficacy. of cellular prediction for arbitrary hit thresholds used by previous methods by previous used thresholds hit arbitrary for a a measure of inhibitor kinase selectivity to elucidate kinase biology. We therefore applied t agents of therapeutic efficacy may be not selectivity kinase strong Although reported. profil kinase ever-larger as important increasingly is supported by our data. Quantitative assessment of i of the importance of broad profiling kinase co these anti-cancer target, though few inhibitors have been contributes to chromosomal organization and has bee characterized kinase inhibitors kinase characterized number pected of with interactions kinases off-target re have studies profiling inhibitor kinase Previous DISCUSSION multitargeted kinase inhibition. inhibitors for their novel targets and in some case compounds the here uni-specific described provide n 6 6 ( protei nuclear cell–specific germ haploid the kinase, inhibitor 23), SB202190 showed (ref. significant inhibition of onlyp one the of analog inactive an SB202474, Additionally, FLT of inhibitor potent highly a be to DMBI inhibitor, growth platelet-derived weak the revealed results the compounds. these for targets new inhibited potently These findings confirm the accuracy of our single-d E C R U O S E R Fig. 5 Fig. Although the high-throughput assay used here to sys to here used assay high-throughput the Although a w as of uni-specificity concept the We introduce also ). This atypical family kinase phosphorylates histone phosphorylates kinase family atypical This ). 28,29 27 . Gini score analysis revealed greatest greatest revealed analysis score Gini . , it is critical , for it used compounds tool is critical 1,2 . These findings have emphasized emphasized have findings These . Supplementary Fig. 2 Fig. Supplementary 15 , , thus avoiding the necessity in vitro in he Gini coefficient as ize the characteriza- ize the as carried out in the as carried showed more potent s, starting s, starting points for ubfamilies. reported nhibitor toward its its toward nhibitor nhibitor selectivity fact, fact, the secondary ose ose data and reveal d the development ounds ounds were found lity of the present present the of lity on. For example, example, For on. gets. Compounds Compounds gets. a specific kinase kinase specific a mpounds mpounds and are n suggested as an vealed an unex- an vealed n kinase Haspin kinase n ing data sets are sets data ing , , even for highly ew ew and selective 38 MAP kinase kinase MAP 38 this compound compound this bust, caution is is caution bust, xpected to pro- to xpected osing the most most the osing tended targets. targets. tended factor receptor factor 2 results to the the to results d inhibitors of d inhibitors t, Rho Kinase Kinase Rho t, . Comparison . Comparison r members of members r bfamily from from bfamily For example, example, For nown targets targets nown essential for essential GC kinases, kinases, GC ingle kinase kinase ingle y of distinct distinct of y 3 and TrkC. and 3 ly than any than ly ay ay of quan- tematically tematically ): Rockout, ): Rockout, 24–26 ierarchical ierarchical he clinical clinical he ), despite despite ), H3 and H3 . Thus, obtain diffracting crystals of unliganded kinases of unliganded crystals diffracting obtain plasticity, can which make conformational considerable studies may also benefit from the present study. Pr new tool tify compounds to stimulate new research. profil kinase large-scale how of example one provides their function. Our identification of a uni-specific facilitat greatly would kinases understood poorly ing Note: Supplementary information is available on the version of the paper at http://www.nature.com/naturavailable are references associated any and Methods METHODS models. cell in functions kinase elucidate to tool will kinases, protein of human majority the against ity characterize that we the collection inhibitor expect such maps interaction kinase-inhibitor robust using developed tion ( tion multitargeted for opportunities new reveal to mined present the how illustrate we addition, In studies. kinase, per nine average on candidates, of library tutr determination structure for c kinases to stabilize can be used inhibitors kinase subset of the kinome inhibitor interactions generally correlates with cellular cellular with correlates generally interactions inhibitor anal biochemical and interactions, kinase-inhibitor informati of resource rich a provide here presented Neverth expected. reasonably be can kinases untested not included in the panel. Thus, additional off-target a minorit a activity, catalytic kinase of measurements kinase panel tested here is among the largest available and only one such state was assayed for each kinase. conformation multiple adopt can kinases many addition, In or kinase in the context of differ the full-length compo a with interactions whose constructs truncated by panel the in kinases many Second, context. cellular may here determined kinases inhibited of order rank presence presence of 10 binding and the cellular concentration of ATP of concentration cellular the and binding co Michaelis-Menten the by also but kinase, the for affin intrinsic the by only not dictated is context for ATP. Potency of ATP-competitive inhibito kinase written by J.R.P. with input from the other authors by T.A. and J.R.P. with input from S.W.D. and H.M., bycollected S.W.D., statistical analysis was perfo The study was conceived by J.R.P., S.W.D. and H.M., R44 CA114995 to H.M.). supported partially by the US National Institutes o CA006927 to Fox Chase Cancer Center. HotSpot techno by US National Institutes of Health awards RO1 GM08 Keystone Program in Head and Neck Cancer of Fox Cha This work was supported by a W.W. Smith Foundation Corp. for developing the Kinase Inhibitor Resource laboratory for comments on the manuscript and R. Ha We gratefully acknowledge B. Turk, A. Andrews and m HTML HTML version of the paper at http://www.nature.com/ The authors declare competing interests:financial AUTHOR CONTRIBUTIONS ACKNOWLEDGMENTS COMPETING FINANCIAL INTERESTSFINANCIAL COMPETING Protein kinase research has been predominantly focu Fig. 4b Fig. ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE 13 to of analysis facilitate polypharmacolo drug potential ). Indeed, new statistical methods have been recent been have methods statistical new Indeed, ). M M M ATP regardless of the of affinity kina individual 31 . The identification of selective inhibitors target 3 Te aa e peetd ee rvds a provides here presented set data The . rmed rmed by K.D., data were analyzed . f f Health (RO1 HG003818 and details details accompany the full-text NATURE BIOTECHNOLOGY NATURE

Nature Biotechnology and the manuscript was (KIR) (KIR) web application tool. experimental data experimental was rtman rtman of Biology Reaction 3025 3025 to J.R.P. and P30 nbt/index.html. embers embers of the Peterson Award, funding from the logy developmentlogy was se se Cancer Center and in the cellular milieu. milieu. in the cellular 30 32 otein kinases exhibit . Thus, the relative relative the Thus, . ity of the inhibitor inhibitor the of ity inhibitor of Haspin . . ATP-competitive d d here, with activ- Third, though the Crystallographic Crystallographic ebiotechnology/. data set can be be can set data to support such such support to are represented represented are rs in the cellular rs in the cellular y of kinases are are kinases of y for biochemical e elucidation of e elucidation nstant for ATP for nstant rystallographic rystallographic ctivities ctivities against ysis of kinase- of ysis as this. Finally, Finally, this. as on concerning concerning on kinase inhibi- kinase in the online online the in eless, the data data the eless, sed sed on a small be a powerful a powerful be ing can iden- can ing differ in the the in differ it difficult to it difficult efficacy und could could und website. al states states al 30 . ses ses gy gy ly ly - © 2011 Nature America, Inc. All rights reserved. 3. Fedorov, O. Fedorov, 3. NATURE BIOTECHNOLOGY NATURE M.A. Fabian, 1. reprints/index.html. o available is information permissions and Reprints Published online at http://www.nature.com/nbt/index . Davies,S.P., Reddy, H.,Caivano, 5.Cohen, &M. P. . aaa, M.W. Karaman, 2. . Bain, J. 4. 12. Morphy, R., Kay, C. & Rankovic, Z. From magic bullets to designed multiple ligands. 14. Dodson, C.A. 14. Dodson, 13. Metz, J.T.13. Metz, 1 ab H, pct K & lrc, . taeis o vroe eitne o targeted to resistance overcome to Strategies A. Ullrich, & K. Specht, H., 11. Daub, 0 iuuu C.V.10. Miduturu, 9. Goldstein, D.M., Gray, N.S. & Zarrinkar, P.P. High-throughput kinase profiling as a as P.P.profiling Zarrinkar, kinase & High-throughput N.S. Gray, D.M., Goldstein, 9. . Smyth, L.A. & Collins, I. Measuring and interpreting the selectivity of protein kinase 8. 6. Bain, J., McLauchlan, H., Elliott, M. & Cohen, P.Cohen, kinase & protein M. of Elliott, specificities H., The McLauchlan, J., Bain, 6. 7. Ma, H., Deacon, S. & Horiuchi, K. The challenge of selecting protein kinase assays kinase protein selecting of challenge The K. Horiuchi, & S. Deacon, H., Ma, 7. 16. Bamborough, P., Drewry, D., Harper, G., Smith, G.K. & Schneider, K. Assessment K. Schneider, & G.K. Smith, G., Harper, P.,D., 16. Drewry,Bamborough, 15. Graczyk, P.P. Gini coefficient: a new way to express selectivity of kinase inhibitors kinase of selectivity express to way new 15. P.P.a Graczyk, coefficient: Gini Biotechnol. inhibitors. Ser/Thr kinases. Ser/Thr J. of somecommonlyof usedprotein kinaseinhibitors. Drug Discov. TodayDiscov. Drug on t MN04 isgt it slciiy n du design. drug and selectivity into insights MLN8054: to bound protein kinase inhibitors. kinase protein platform for drug discovery. drug for platform inhibitors. inhibitors: an update. an inhibitors: toward the discovery of new kinase inhibitors. kinase new of discovery the toward for lead discovery optimization. discovery lead for J. Med. Chem.Med. J. of chemical coverage of kinome space and its implications for kinase drug discovery. against a family of kinases. of family a against (2010).19–28

408 , 297–315 (2007).297–315 , et al. et al.et Nat. Biotechnol.Nat. J. Chem. Biol.Chem. J.

26 et al.et The selectivity of protein kinase inhibitors: a further update. et al.et t al.et

, 127–132 (2008).127–132 , t al.et 51 Navigating the kinome. the Navigating t al.et Proc. Natl. Acad. Sci. USASci. Acad. Natl. Proc. A systematic interaction map of validated kinase inhibitors with inhibitors kinase validated of map interaction systematic A , 7898–7914 (2008).7898–7914 ,

Crystal structure of an Aurora-A mutant that mimics Aurora-B mimics that mutant Aurora-A an of structure Crystal sal oeuekns itrcin a fr lncl kinase clinical for map interaction molecule-kinase small A 9 ihtruhu kns poln: mr efiin approach efficient more a profiling: kinase High-throughput , 641–651 (2004).641–651 , qatttv aayi o kns ihbtr selectivity. inhibitor kinase of analysis quantitative A Biochem. J.Biochem.

Nat. Rev. Drug Discov.Drug Rev. Nat. 2

, 131–151 (2009).131–151 , 23 Nat. Rev. Drug Discov.Drug Rev. Nat. J. Med. Chem.Med. J.

, 329–336 (2005).329–336 , Expert Opin. Drug Discov.Drug Opin. Expert ADVANCE ONLINE PUBLICATION PUBLICATION ONLINE ADVANCE

371 , 199–204 (2003).199–204 , Nat. Chem. Biol.Chem. Nat.

50

104 Specificity andmechanism actionof , 5773–5779 (2007).5773–5779 , Chem. Biol.Chem.

Biochem.J. , 20523–20528 (2007).20523–20528 , 3 nline at http://www.nature.com/ at nline hm. .html. , 1001–1010 (2004).1001–1010 ,

7 , 391–397 (2008).391–397 ,

7

3 18 , 200–202 (2011).200–202 ,

, 607–621 (2008).607–621 , 351 , 868–879 (2011).868–879 , ice. J.Biochem. , 95–105,(2000). Biochem.

427 Nat. , 1 aen, . Godn, . el yl knss s hrpui tres o cancer. for targets therapeutic as kinases cycle Cell A. Giordano, & S. 21. Lapenna, 2 hn, Q.22. Zhang, 8 io JK, eo M & oa K Ro iae RC) inhibitors. (ROCK) kinase Rho K. Noma, & M. Seto, J.K., 28. Liao, 23. 23. J.C. Lee, 7 ave S, jlol S & amn, . e prdgs n niacr therapy: anticancer in paradigms New E. Raymond, & S. Djelloul, S., 17. Faivre, 31. Fedorov, O., Muller, S. & Knapp, S. The (un)targeted cancer kinome. cancer (un)targeted The S. Knapp, & S. Muller, 31. O., Fedorov, 5 ura, D. 25. Huertas, 24. 24. Cuny,G.D. 20. 20. S. Cockerill, 6 ani, D. 26. Patnaik, 7 ece, .. Wn, .. rmsuu dus oprd o eetv drugs selective to compared drugs Promiscuous L.G. Wang, & S.K. 27. Mencher, 8 opy R Slciey oslcie iae niiin srkn te ih balance. right the striking inhibition: kinase nonselective Selectively R. 18. Morphy, 9 arw JC, oskw, . Cars GT & icio, .. cenn fr cell for Screening T.J. Mitchison, & G.T. Charras, G., Totsukawa, J.C., 29. Yarrow, 9 ngt ZA, i, . Soa, .. agtn te acr ioe through kinome cancer the Targeting K.M. Shokat, & H. Lin, Z.A., 19. Knight, 30. Shokat,Knight,FeaturesK.M.selecti&Z.A. of 32. Huse, M. & Kuriyan, J. The conformational plasticity of protein kinases. protein of plasticity conformational The J. Kuriyan, & 32. M. Huse, 33. Manning, G., Whyte, D.B., Martinez, R., Hunter, T. & Sudarsanam, S. The protein The S. Sudarsanam, T.& Hunter, R., Martinez, D.B., Whyte, 33. G., Manning, Nat. Rev. Drug Discov.Drug Rev. Nat. combinatorial kinase-directed heterocycle library. biosynthesis. (2006). agtn mlil sgaig ahas ih iae inhibitors. kinase with pathways signaling multiple targeting 621–637 (2005).621–637 kinase Haspin. kinase and DYRK2 kinase inhibitors. kinase DYRK2 and of the EGFR and c-erbB-2. and EGFR the of 2011). (promiscuity can be a virtue). a be can (promiscuity J. Med. Chem.Med. J. (2006).407–420 Biol. migration inhibitors via automated microscopy reveals a Rho-kinase inhibitor. polypharmacology. Pharmacol. Biol. resonance energy transfer assay.transfer energy resonance haspin by high-throughput screening using a homogen kinase complement of the human genome. human the of complement kinase (2002).275–282

12 6 , 166–169 (2010).166–169 , , 385–395 (2005).385–395 , et al.et t al.et

et al.et 50 t al.et t al.et et al.et Nature A protein kinase involved in the regulation of inflammatory cytokine inflammatory of regulation the in involved kinase protein A , 17–24 (2007).17–24 ,

Oncogene 53 icvr o EF slcie 4,6-disubstituted pyri selective EGFR of Discovery Structure-activity relationship study of acridine analogs as haspin as analogs acridine of study relationship Structure-activity Identification of small molecule inhibitors of the of inhibitors molecule small of Identification niuo atvt o a ml-oeue niio o te histone the of inhibitor small-molecule a of activity Antitumor Indazolylamino quinazolines and pyridopyrimidines as inhibitors as pyridopyrimidines and quinazolines Indazolylamino , 1413–1437 (2010).1413–1437 , Nat. Rev. CancerRev. Nat.

372

8 , 547–566 (2009).547–566 , published online, doi:10.1038/onc.2011.335 (1 August (1 doi:10.1038/onc.2011.335 online, published , 739–746 (1994).739–746 , Bioorg. Med. Chem. Lett.Chem. Med. Bioorg. Bioorg. Med. Chem. Lett.Chem. Med. Bioorg. BMC Clin. Pharmacol.Clin. BMC J. Biomol. Screen.Biomol. J.

10 , 130–137 (2010).130–137 , Science ve kinaseinhibitors.ve J. Am. Chem. Soc.

eous time-resolved fluorescence 13 298

5 11 , 3 (2005).3 , , 1025–1034 (2008).1025–1034 ,

, 1912–1934 (2002).1912–1934 , E C R U O S E R 20 , 1401–1405 (2001).1401–1405 , , 3491–3494 (2010).3491–3494 , ei. Oncol.Semin.

128 Chem.Biol. iie fo a from midines . Cardiovasc. J. ioi kinase mitotic , 2182–2183 Nat. Chem. Nat. Cell Chem.

109

33 12 7 , , , © 2011 Nature America, Inc. All rights reserved. reaction buffer; 20 mM Hepes pH 7.5, 10 mM MgCl mM 10 7.5, pH Hepes mM 20 buffer; reaction c required with along pairs kinase/substrate specific “HotSpot” the using Corporation Biology Reaction at tions. IC ( vehicle to compared samples test in activity kinase as wereexpressed data activity kinase enzyme, inactive acid. After subtraction of background derived from cont filter of washing extensive by removed was phosphate ting of the reactions onto P81 ion exchange filter NATURE BIOTECHNOLOGY NATURE Kinase assays. Table 2Supplementary kinases recombinant of description complete A >98%. with Laboratories LC or Biosciences EMD from either Materials. METHODSONLINE were estimated using maximum likelihood methods likelihood maximum using estimated were of a mixture of ATP (Sigma) and Compounds were delivered into the reaction, followed Fig. 1a Fig. distrib exponentially double be to determined was D quantile-quanti and estimation density kernel Using each for computed were observations duplicate from ( variation of coefficient the and analysis duplica further across values identical or observations ing kinase-inhibitor and zero to truncated were values in dupli pair kinase-inhibitor for each activity of compound versus D is displayed in D is versus displayed methods.Statistical reactionbiology.com/apps/kinome/mapper/LaunchKinome 10 of details of individual kinase reaction components se Brij35, 0.02 mg/ml BSA, 0.1 mM Na for for D, double the in lines vertical gray the by determined distributi of the mean of the 1 s.d. within observations meas assay the in noise inherent the for account To on the estimated mean and s.d. of this distribution this of s.d. and mean estimated the on region enclosed by these vertical lines contains 75.6% of the observations based Kinome tree representations were prepared using Kin M Supplementary Fig. 1aSupplementary M. Reactions were carried out at 25 °C for 120 min, 120 for °C 25 at out carried were Reactions M. , b 50 ). Its location and scale parameters (and hence the hence (and parameters scale and location Its ). values and curve fits were obtained using Prism (G iae niios ( inhibitors Kinase

In vitro

Outlier Outlier detection profiling of the 300 member kinase panel was perfo Supplementary Figure 1e Figure Supplementary . ) for further analyses of ) analyses compound activity. for The further 33 upeetr Tbe 1 Table Supplementary P ATP (PerkinElmer) to a final concentration 3 VO . . Raw data were as measured percentage 4 , , 2 mM DTT, 1% DMSO (for specific paper paper (Whatman). Unbound CV) and the difference (D) (D) difference the and CV) e e for all pairs of data points. of points. data pairs for all . The red vertical lines in in lines vertical red The . ome Mapper ( ofactors were prepared in in preparedwere ofactors exponential density plot plot density exponential dimethyl sulfoxide) reac- sulfoxide) dimethyl Supplementary Supplementary Table 2 ~20 min later by addition tes were removed from from removed were tes 34 pairs with either miss- either with pairs le plots, the difference difference the plots, le urements, we retained retained we urements, 2 rol reactions containing assay platform. Briefly, platform. assay the percent remainingpercent the on of differences D (as on of differences s in 0.75% phosphoric phosphoric 0.75% in s uted ( uted , 1 mM EGTA, 0.02% 0.02% EGTA, mM 1 , . A scatter plot of CV CV of plot scatter A . kinase-inhibitor pair. kinase-inhibitor an average purity of of purity average an used is provided in in provided is used raphPad Software). .htm followed by spot- by followed ) were obtained obtained were ) cate. All negative negative All cate. Supplementary Supplementary mean and s.d.) s.d.) and mean ). http://www. rmed ). ). calculated according to the equation: activity = (1 = activity equation: the to according calculated Kinase activity Kinase analysis.activity extremevalues. and VGAM libraries using ment data the to applied was No scaling th linkage. average as correlation rank Spearman – 1 on based tering heat map of compound was activity obtained using two t at truncated were >100 values and zero to truncated Hierarchical clustering Hierarchical in line horizontal pink The expected. are vations computing the threshold beyond which a certain presobservat extreme whether determine to performed then log-normal distribution and its parameters were robustl the top and bottom 1%) were fit to the quantile-quanti tion assumes a Hill coefficient of 1 for the binding an binding of 1 the for coefficient a Hill assumes tion M these data ( pairs. kinase-inhibitor remaining the th and data of set current the from excluded were tions obser these represent region this within circles black 1e Figure Supplementary described earlier for earlier Ddescribed distribution of CV was determined and its parameter se reduced on this based CV to the applied then was all kinases.all data set in in set data t within are shown points) data (black outliers The exclude and observations outlying as identified were band, this of and 0.5 outside CV cut-off the above observation the of remainder the approach, twofold this c CV a to corresponds and threshold this represents 4 e, .. Hsi, .. eue-ak etr eeaie lna models. linear generalized vector Reduced-rank T.J. Hastie, & T.W. 34. Yee, 5 a dr o, ... itiuinbsd ule dtcin o uiait data. (K constant inhibition univariate the between Relationship W.H. for Prusoff, & Y. 36. detection Cheng, outlier Distribution-based M.P.J. Loo, der 35. van M))) and the Cheng-Prusoff equation Cheng-Prusoff the and M))) The distribution-based outlier detection method out method detection outlier The distribution-based Computations were carried out in the R statistical l statistical R the in out carried were Computations enzymatic reaction. enzymatic h cnetain f niio wih ass 0 e cn ihbto (I inhibition cent per 50 causes which inhibitor of concentration the Modelling Discussion paper 10003 (Statistics Netherlands, The Hague, 2010).Hague, The Netherlands, (Statistics 10003 paper Discussion Supplementary Figure 1f Figure Supplementary Supplementary Fig. 1c

3 , 15–41 (2003).15–41 , Biochem. Pharmacol.Biochem. 34 . Negative values for remaining kinase activity wer activity kinase remaining for values Negative . The theoretical kinase activity curve in curve activity kinase The theoretical . The log-normal distribution . provided distribution The the best log-normal fit also represent these limits whereas the green and and green the whereas limits these represent also , d ). For outlier detection, the data (excluding . 36

22 relating relating , 3099–3108 (1973).3099–3108 , represented by blue circles, by blue circles, represented Supplementary Figure 1e Figure Supplementary he context of the complete complete of the context he . K s estimated using methods pecified pecified number of obser- 00 − (100/(1 + (IC + (100/(1 − 00 t of data points. First, the the First, t points. of data i ut-off of ~0.5. Based on Based ~0.5. of ut-off vations. These observa- These vations. and IC and d from further analysis. analysis. further from d d d a le plot positions for le the plot positions hat value. A reordered reordered A value. hat -way hierarchical -way clus- hierarchical e distance metric and and metric distance e anguage and environ- and anguage y y estimated. A test was lined by lined van der Loo e CV recomputed for for recomputed CV e doi:10.1038/nbt.2017 K ions are outliers by outliers are ions s that were located located were that s m,ATP 50 . This calcula- This . Figure 2a Figure of 10 of 10 50 M o an of ) 1 50 M M for and ) was Stat. /0.5 /0.5 for 35 e

SUPPLEMENTARY DATA

Supplementary Figure 1. Identification and elimination of statistical outliers. (a) Kernel density

plot of difference (D) from duplicate observations for each kinase-inhibitor pair. The estimated

double exponential density (y-axis) is plotted against D (x-axis). The grey vertical lines represent

one standard deviation on either side of the mean of the distribution of D. ( b) Quantile-quantile

plot of D from duplicate observations for each kinase-inhibitor pair. Quantiles of the estimated

double exponential distribution (y-axis) are plotted against D (x-axis). (c) Kernel density plot of

the coefficient of variation (CV) from duplicate observations for each kinase-inhibitor pair. The

estimated log-normal density (y-axis) is plotted against CV (x-axis). ( d) Quantile-quantile plot of

CV from duplicate observations for each kinase-inhibitor pair. Quantiles of the estimated log-

normal distribution (y-axis) are plotted against CV (x-axis). ( e) A scatter plot of CV versus D for

all kinase-inhibitor pairs. The green and black circles that lie inside the band formed by the red

vertical lines represent observations within one standard deviation of the mean of the estimated

double exponential distribution of D. These observations were considered to be within

acceptable noise levels in assay measurements and were retained for further analyses of

compound activity. The pink horizontal line represents the CV threshold for outlier detection.

The blue circles represent observations whose CV exceeded this threshold and were identified as

outliers. ( f) A scatter plot of the kinase activity in replicate 1 versus replicate 2 for all kinase-

inhibitor pairs. Data points in black represent the identified outliers that were removed from the

final data.

Nature Biotechnology: doi:10.1038/nbt.2017 Nature Biotechnol og y: do i: 10 .1 03 8/ nb t. 20 17 Supplementary Figure 2. A high-resolution version of Figure 1d showing two-way hierarchical

clustering analysis of the entire kinase-inhibitor interaction network presented as a heatmap of

inhibitory activity.

Nature Biotechnology: doi:10.1038/nbt.2017 Nature Biotechnology: ! CAMK2d CAMK2a CHK1 EM#G E MARK2/P PKA <#E% p70S6Kb/RPS6KB2 PAK2 P PAK4 CAMK1a RIPK5 <"#GPSA I%O7PSE ?!?@L E'>5PE># F>>G X'I#% HIPK4 HIPK2 SRPK1 "E#%HP"EW#% MRCKa/CDC42BPA PYK2 ALK IRAK4 IRAK1 IKKe/IKBKE CAMKK CK DYRK2 EW#5 AJ#%P8384*(&> ACDK1/cyclin D1CDK4/cyclin AJ#LP8384*(&J% S"EL PKCa PKCd PKCg PKCtheta PKCeta PKCepsilon PKCb2 PKCb1 CK1epsilon CK1d ?!?@5 !?@ ERBB4/HER4 ?@#PIE#6 CSK EPHB4 EPHB1 ZAK/ML DDR2 >IXM% >IX<% EPHA8 EPHA1 EPHA2 EPHB2 EPHA4 EPHA5 "E#%OP1J@G PKAcg PHKg2 GRK6 GRK7 PKG1a PKN2/PRK2 PDK1/PDPK1 PKG1b I@#/ ROCK1 PAK1 "E#56PY"#G EPHA7 PAK5 '!?G@ MNK1 IR DAPK1 ZIPK/D EE# NEK4 NEK9 SGK2 PASK MKK6 D MEK2 MEK1 CAMK1d PBK/ YES/YES1 ?!@ WV#P"E#GK ? ROCK2 ?>"P?I" ULK2 NEK1 E E KHS KHS MAP4K5 @>E HGK MAP4K4 ? E@#A Aurora B GRK2 PLK2 AE#&S U1#% CO DYRK4 NEK7 I%OR HIPK1 NEK6 I%O9PS#% 1'#PS##% JAK2 JAK1 ULK1 8%S>@ ?!?@% EY#GP RIPK2 BRK LIMK1 SIK2 MS/P>E# ME# E/# IJ!?@7 ABL1 LYN B EPHA6 8%#*, ?S" IJ!?@+ BLK LCK HCK ABL2/ARG LYN 8%"@A ?Y1 A !"#%+ AJ#HP8384*(&EG Kcyclin CDK9 D1CDK6/cyclin AJ#NP8384*(&J% BCDK1/cyclin ECDK2/cyclin CDK5/p25 ACDK2/cyclin AJ#6P:%6 "W#P"E#5 ?!?@G ? #J@P=>!?@5 CLK4 CLK2 \W#% MY S"E%P"E#5L PKN1/PRK1 ?>@ Aurora C E@#M SW#GPS!?@% E% EGP=>!?@G V#%PU'# R 2a2 V%+ #%7 EGPS

Bosutinib AMPK Inhibitor, Compound C Sunitinib Dovitinib Syk Inhibitor !"#%%&'()*+*,-.&'/ Staurosporine0&1%+2( zo 34% SU11652 Indirubin Derivative E804 SB 218078 Staurosporine, Streptomyces sp. #%56570&1-87.9*-:;*;&; p. Cdk1/2 Inhibitor III Gˆ 6976 PKR Inhibitor J<#%&'()*+*,-.&=' JAK Inhibitor I Isogranulatimide Met Kinase Inhibitor '##%5&'()*+*,-.&'= =>!?&@282:,-.&5&#*(7;2&'()*+*,-.&''' Aurora Kinase/Cdk Inhibitor Alsterpaullone 0&5%A yanoethyl Cdk4 Inhibitor Alsterpaullone SU9516 Aminopurvalanol A !"#%%&'()*+*,-.&/ ?4,%%&'()*+*,-.&'' Indir B+*(%%C%D-( oxime !"#%%&'()*+*,-.&/''' Cdk2 Inhibitor I =, NU6140 SU6656 !"#%+&'()*+*,-.&/''0&EF"GGH& Dasatinib =andetanib Lck Inhibitor Eozasertib IJ!?& RE#&'()*+*,-. Pazopanib @-%%5%KL%5 M*;*(9-434D742*D*92&'= PKCb Inhibitor Bisindolylmaleimide I !-&NHO% JNK Inhibitor II PKR Inhibitor, Negative Control Aloisine A, RP107 Aloisine, RP106 Purvalanol A Compound 52 Cdk1 Inhibitor, CGP74514A Nilotinib IJ!?&@282:,-.&E3.-;*(2&#*(7;2&'()*+*,-.&'= Sorafenib IJ#GP!?&@282:,-.&5&#*(7;2&'()*+*,-.&'' =>!?&@282:,-.&5&#*(7;2&'()*+*,-.&' ?4,%%&'()*+*,-.&''' =>!?&@282:,-.&5&#*(7;2&'()*+*,-.&'= IIG&<(74-R&''0&G1S%IIG J<#%&'()*+*,-.&'' Src Kinase Inhibitor I MI'T%' AG 1478 Compound 56 >!?@P>.+M%5P>.+M%L&'()*+*,-.& >!?@P>.+M%5&'()*+*,-. PD 158780 Erlotinib Gefitinib Kenpaullone Cdk/Crk Inhibitor U1#&'()*+*,-.&= SEV%NKH A7;2*(&#*(7;2&''&'()*+*,-.&'''0&EMA< A985%W* ke Kinase Inhibitor 0&E!KK% :%O&S!?&@282:,-.&E3.-;*(2&#*(7;2&'()*+*,-.&'''0&#@1N%% =>!?&@282:,-.&E yrosine Kinase Inhibitor II Masitinib Imatinib IJ!?&@282:,-.&E yrosine Kinase Inhibitor III =atalanib A!W&5KL% Eandutinib I'&%%#R&'()*+*,-.& I'&%%#R&'()*+*,-.&''& E:45&#*(7;2&'()*+*,-. PD 174265 Lapatinib JNK Inhibitor, Negative Control Rho Kinase Inhibitor III, Rockout ROCK Inhibitor0&Y%5ZN%5 @)-&#*(7;2&'()*+*,-.&'= Roscovitine Cdk2 Inhibitor III Bohemine =/%ZK5 MNK1 Inhibitor >@#&'()*+*,-.&''0&?@GOK5KL E!?%+&@'&#*(7;2&'()*+*,-. E!?%+&@'&'()*+*,-.&''' "A%NO%ZN IC261 Casein Kinase I Inhibitor, D4476 Aurora Kinase Inhibitor III Eofacitinib Fortmannin ?ascaplysin, Synthetic AG 1296 AG 1295 Chk2 Inhibitor II Cdk4 Inhibitor III Cdk4 Inhibitor II, NSC 625987 Chelerythrine Chloride Cdk1/5 Inhibitor Cdk1 Inhibitor 8?S"&@282:,-.&E yrosine Kinase Inhibitor I#A+''P>!?@&'()*+*,-. IJ!?&@282:,-.&E yrosine Kinase Inhibitor II I'%GK% II% PD 98059 LY 294002 U1#&'()*+*,-.&=''' U1#&'()*+*,-.&'/ 1?%QM&<8,* vation Inhibitor #1%N5 #1%H% LY&%K%6GG%&12R7,* ve control MK2a Inhibitor MEK1/2 Inhibitor J<#%&'()*+*,-.&'= MEK Inhibitor I MEK Inhibitor II AG 1024 !"#%%+&'()*+*,-.&=''' !"#%%+&'()*+*,-.&'' AG 112 Syk Inhibitor III AG 9 AG 490 Mubritinib SB 202474, Negative control f-.&:%O&S!?@&'()*+*,-. J1<%I#&'()*+*,-.&'' '!?%G@&'()*+*,-.&'' Herbimycin A, Streptomyces sp. !"#%%+&'()*+*,-.&' ERK Inhibitor III J1<%I#&'()*+*,-.&= J1<%I#&'()*+*,-.&''' Diacylglycerol Kinase Inhibitor II AES&#*(7;2&'()*+*,-.

containing their intended targets. All compounds intended to target a kinase in the indicated

specific subfamily of kinases were analyzed as to whether the kinases they inhibit fall within the

subfamily of their intended target (blue) or outside of that family (red, percent of total shown).

The final pie chart presents aggregate data for all of the subfamilies presented. “n” reports the

number of compounds analyzed for each target subfamily. The following compound types were

not included in this analysis: compounds intended to target lipid kinases, inactive control

compounds, compounds intended to target kinases from multiple subfamilies, compounds that

did not inhibit any kinases. For example, none of the four compounds intended to target STE

subfamily kinases inhibited any kinases significantly.

Nature Biotechnology: doi:10.1038/nbt.2017 Na ture Biotechnol og y: do i: 10 .1 03 8/ nb t. 20 17 Supplementary Figure 4. No single physicochemical property correlates with inhibitor

selectivity. All 178 test compounds were analyzed with regard to molecular weight, predicted

LogP, polar surface area (PSA), number of hydrogen bond acceptors, and number of hydrogen

bond donors. These features are plotted for each compound as a function of inhibitor selectivity

from the screening data reported either as Gini score (from Supplementary Table 5) or as

Selectivity score (S (50%) ). The Selectivity score of a compound corresponds to the number of

kinases that it inhibited by at least 50% divided by the number of kinases against which the

compound was tested. More selective inhibitors are associated with lower Selectivity scores and

higher Gini scores.

Nature Biotechnology: doi:10.1038/nbt.2017 Na ture Bi otec hn ol og y: do i: 10 .1 03 8/ nb t. 20 17 Supplementary Figure 5. Validation of novel uni-specific kinase inhibitors. Complete in vitro

kinase assay dose-response results are shown for the five indicated uni-specific compounds ( a-e)

from Figure 5 against both the intended targets of each inhibitor and their novel, more potently

inhibited target(s). Data for intended kinase target(s) are shown with open symbols and novel

target(s) with closed symbols. Data points represent averages of two independent replicates.

Error bars denote standard error of the mean. Data exhibiting significant inhibition was fitted

with a sigmoidal dose-response curve to derive IC 50 values.

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Figure 5 a 125 Akt Inhibitor X CAS #: 925681-41-0 100

Kinase IC 50 (nM) 75 AKT1 > 30,000 C l AKT2 > 30,000 50 O AKT3 > 30,000 N N PIM1 2263 25 PIM2 > 30,000 Kinase activity (% ctrl) ctrl) (% activity Kinase PIM3 514 0 -9 -8 -7 -6 -5 -4 log concentration (M) b DMBI 125 CAS #: 5812-07-7

100 Kinase IC 50 (nM) !"#$%r > 30,000 75 !"#$%s > 30,000 c-Kit 84 50 O TRKA 1829 HN TRKB 487 25 N TRKC 82

Kinase activity (% ctrl) ctrl) (% activity Kinase FLT3 70 0 -9 -8 -7 -6 -5 -4 log concentration (M) c SC-68376 100 CAS #: 318480-82-9

75 Kinase IC 50 (nM) !()r*+,!-./ 5156 !()s*+,!-.. > 30,000 50 N > 30,000 CH 3 !()u*+,!-.( O > 30,000 25 !()t N 2-.u 864

Kinase activity (% ctrl) ctrl) (% activity Kinase 2-.v 3360 0 -9 -8 -7 -6 -5 -4 log concentration (M) d 125 SB202474 100 CAS #: 172747-50-1

75 Kinase IC 50 (nM) !()r*+,!-./ 8197 50 N H !()s*+,!-.. > 30,000 N !()u*+,!-.( > 30,000 25 O N !()t > 30,000 Kinase activity (% ctrl) ctrl) (% activity Kinase Haspin 731 0 -9 -8 -7 -6 -5 -4 log concentration (M) e 125 AG 1024 100 CAS #: 65678-07-1

75 Kinase IC 50 (nM) IGF1R > 30,000 50 N Br RIPK2 522

25 HO

N Kinase activity (% ctrl) ctrl) (% activity Kinase 0 -9 -8 -7 -6 -5 -4 log concentration (M)

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1. Compounds used in this study.

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

Target Name CAS # Structure family

N Br

AG 1024 65678-07-1 TK HO N

NH2 N AG 112 144978-82-5 TK

HO N N

AG 1295 71897-07-9 N TK

N

O

AG 1296 146535-11-7 N O TK

N

HN Cl AG 1478 175178-82-2 O TK N

O N

O HO N AG 490 133550-30-8 H TK

HO N

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

N AG 9 2826-26-8 TK O N

N N AGL 2043 226717-28-8 S TK

N N

N N Akt Inhibitor IV 681281-88-9 AGC S N+ I- N

N N H2N N Akt Inhibitor V, 35943-35-2 HO N AGC Triciribine N O

OH OH

N O

N N Akt Inhibitor VIII, N Isozyme-Selective, 612847-09-3 NH AGC Akti-1/2 N N H

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O

Akt Inhibitor X 925681-41-0 N Cl HCl AGC

N

N

Aloisine A, RP107 496864-16-5 N N CMGC H

OH

N

Aloisine, RP106 496864-15-4 N N CMGC H

O

H O N

Alsterpaullone 237430-03-4 CMGC HN O N+ O- H O N

Alsterpaullone, 852527-97-0 CMGC 2-Cyanoethyl N HN O N+

O-

Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

NH2

HN Cl Aminopurvalanol A 220792-57-4 CMGC N N

HO N N N H

O

N N N AMPK Inhibitor, 866405-64-3 CAMK Compound C N

N

O O N S ATM Kinase Inhibitor 587871-26-9 ATYPICAL S

O

Cl F O Cl Cl S ATM/ATR Kinase O- 905973-89-9 N N N N+ ATYPICAL Inhibitor H H H O

N

HNN NH Aurora Kinase 879127-16-9 OTHER Inhibitor III O F N F H F

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

N F

O Aurora Kinase/ O F OTHER, 443797-96-4 H2N Cdk Inhibitor S N CMGC NH N 2 O N N H

OO S BAY 11-7082 19542-67-7 OTHER N

O

NH2

Bcr-abl Inhibitor 778270-11-4 TK F O N F F N N H

NH

Bisindolylmaleimide I 133052-90-1 N AGC N O

NH O

NH

Bisindolylmaleimide IV 119139-23-0 HN AGC O

NH O

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

NH

N Bohemine 189232-42-6 N CMGC

HO N N N H

Bosutinib 380843-75-4 TK

HN Br

BPIQ-I 174709-30-9 N TK N

N N

O NH2

Casein Kinase I CK1, CMGC, 301836-43-1 Inhibitor, D4476 TKL N NH O N

O Br O Br Casein Kinase II OH 934358-00-6 CK1 Inhibitor III, TBCA Br Br

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O S Cdc2-Like Kinase 300801-52-9 CMGC Inhibitor, TG003 O N

O O HO HN N Cdk/Crk Inhibitor 784211-09-2 CMGC N N Cl Cl

H N Cl

Cdk1 Inhibitor 220749-41-7 CMGC

O N H

HN Cl Cdk1 Inhibitor, 190654-01-4 N CMGC CGP74514A N

N N N H

NH2

F

HN F S Cdk1/2 Inhibitor III 443798-55-8 O O CMGC S N H2N N NH2 N N H

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

N NH2

Cdk1/5 Inhibitor 40254-90-8 CMGC N N N H

NH

N O N

Cdk2 Inhibitor III 199986-75-9 HO CMGC N N N

OH

Cdk2 Inhibitor IV, 444723-13-1 OO CMGC NU6140 N N N

N N N H H

H N O O

Cdk4 Inhibitor 546102-60-7 CMGC

Br N N H H

S O

Cdk4 Inhibitor II, 141992-47-4 CMGC NSC 625987

N H

O

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O H N N Cdk4 Inhibitor III 265312-55-8 CMGC S

O

NH2

cFMS Receptor N Tyrosine Kinase 870483-87-7 TK

Inhibitor H2N N O O O

O

O Chelerythrine Chloride 3895-92-9 AGC N+ O

O Cl-

O

HN NH Cl 2 Chk2 Inhibitor II 516480-79-8 N CAMK

O

Cl NH

Compound 52 212779-48-1 N CMGC N HO N N N H

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O NH Compound 56 171745-13-4 O TK Br N N

Dasatinib 302962-49-8 TK

F

O Diacylglycerol Kinase 120166-69-0 N LIPID Inhibitor II N NH S F

O

HN DMBI 5812-07-7 TK N

O O N DNA-PK Inhibitor II 154447-35-5 ATYPICAL

O

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O N DNA-PK Inhibitor III 404009-40-1 ATYPICAL

OH O

O N DNA-PK Inhibitor V 404009-46-7 ATYPICAL

OH O

Dovitinib 405169-16-6 TK

NN

HN NH EGFR Inhibitor 879127-07-8 TK O F N F H F

O

EGFR/ErbB-2 Inhibitor 179248-61-4 TK HN O N

O N

Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

F

EGFR/ErbB-2/ErbB-4 HN Cl 881001-19-0 H TK Inhibitor N N O N N

HN N

N NH2 ERK Inhibitor II, N 865362-74-9 CMGC FR180204

N N

O N

HN NH2 ERK Inhibitor III 345616-52-6 CMGC -O N+ N O- O N+ O

Erlotinib 183319-69-9 TK

Cl-

N+ Fascaplysin, Synthetic 114719-57-2 CMGC N H O

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O O S O Flt-3 Inhibitor 301305-73-7 NH TK

O

H2N

HO OH

Flt-3 Inhibitor II 896138-40-2 TK N N H H O

O N N Flt-3 Inhibitor III 852045-46-6 TK S N H

Gefitinib 184475-35-2 TK

H N O

Gö 6976 136194-77-9 AGC N N

N

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O

O HN

Gö 6983 133053-19-7 O N AGC N

HN

Br OH N

GSK-3 Inhibitor IX 667463-62-9 CMGC

N NH H O

O

Br O N GSK-3 Inhibitor X 740841-15-0 CMGC

N NH H O

N NH

HN

GSK-3 Inhibitor XIII 404828-08-6 N CMGC

N

CMGC, GSK-3b Inhibitor I 327036-89-5 O N O AGC, TK

N S

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

I N

GSK-3b Inhibitor II 478482-75-6 O CMGC S N N

O N O GSK-3b Inhibitor VIII 487021-52-3 N CMGC N+ N H O S H -O

H N O O

GSK-3b Inhibitor XI 626604-39-5 NN CMGC N N

OH OH

GSK3b Inhibitor XII, O 601514-19-6 NH CMGC TWS119 2 N

N N H

GTP-14564 34823-86-4 O TK N N H

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

N

HCl OOS HCl NH AGC, CK1, H-89, Dihydrochloride 127243-85-0 HN CAMK

Br

HN

N HCl HA 1077, HCl Dihydrochloride 103745-39-7 O AGC S Fasudil O

N

O

O O N Herbimycin A, H 70563-58-5 TK Streptomyces sp. O O

O O O

O NH2

O O

IC261 186611-52-9 O CK1

O N H

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

Cl

O N IGF-1R Inhibitor II 196868-63-0 TK N N H H O

O

NH2

IKK-2 Inhibitor IV 507475-17-4 OTHER NH S NH O 2 F

Imatinib 220127-57-1 TK

OH

OH

Indirubin Derivative O CMGC, TK E804 854171-35-0 N

NH N H O

OH

N Indirubin-3 ′-monoxime 160807-49-8 CMGC

N NH H O

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

-O N N+ O NH N IRAK-1/4 Inhibitor 509093-47-4 O TKL

N

O

H O N O CAMK, Isogranulatimide 244148-46-7 CMGC, ATYPICAL N N N H

F

N JAK Inhibitor I 457081-03-7 TK O N H HN

OH

HN Br JAK3 Inhibitor II 211555-04-3 O TK N

O N

O

N JAK3 Inhibitor IV 58753-54-1 TK

HCl

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O S O OH

H JAK3 Inhibitor VI 856436-16-3 N N TK

O N H

N NH

JNK Inhibitor II 129-56-6 CMGC

O

N

JNK Inhibitor IX 312917-14-9 NH CMGC S O

OH2

N

N JNK Inhibitor V 345987-15-7 NH CMGC H N N

S N

NH2 N O JNK Inhibitor VIII 894804-07-0 O CMGC

O N N O H

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

N N JNK Inhibitor, INACTIVE Negative Control 54642-23-8

O

O OH

H O O K-252a, N 97161-97-2 N AGC Nocardiopsis sp.

O N H

H O N

CMGC, CK1, Kenpaullone 142273-20-9 TK HN

Br

N

O S O O KN-62 127191-97-3 CAMKII

N N N O S O O

N

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O OH O S KN-93 139298-40-1 O N CAMK

N

Cl

Lapatinib 231277-92-2 TK

O

NH2 Lck Inhibitor 213743-31-8 TK N

N N

O

LY 294002 154447-36-6 O N LIPID O

O

LY 303511- Negative 154447-38-8 INACTIVE control O N NH

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

Masitinib 790299-79-5 TK

N

NH2 OH

MEK Inhibitor I 297744-42-4 S STE

NH N 2

O O N MEK Inhibitor II 623163-52-0 STE O Cl O

F N F F MEK1/2 Inhibitor 305350-87-2 S STE

NH2 NH 2

O

N N

Met Kinase Inhibitor 658084-23-2 O O N TK S H N O N H

Cl

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

OH

O NH MK2a Inhibitor 41179-33-3 CAMK

F

NH2 H N N MNK1 Inhibitor 522629-08-9 CAMK N F N N H

Mubritinib 366017-09-6 TK

O

NF-kB Activation HN 545380-34-5 OTHER Inhibitor H2N N

N

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

Nilotinib 641571-10-0 TK

F O p38 MAP Kinase 219138-24-6 CMGC Inhibitor N Cl N H N

NH p38 MAP Kinase 581098-48-8 N CMGC Inhibitor III H N S N

F

Pazopanib 444731-52-6 TK

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

HN Br PD 158780 171179-06-9 H TK N N N N

N

H N O- PD 169316 152121-53-4 N+ CMGC N O

F

HN Br H PD 174265 216163-53-0 N TK N O N

O

H2N

PD 98059 167869-21-8 O STE

O

O O PDGF Receptor Tyrosine Kinase 249762-74-1 TK N N Inhibitor II H H O

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

H O N

N O PDGF Receptor Tyrosine Kinase 205254-94-0 N TK Inhibitor III O N

O N

H N N PDGF Receptor O Tyrosine Kinase 627518-40-5 TK N F Inhibitor IV H O

H H N N

OH2 S O PDGF RTK Inhibitor 347155-76-4 O TK O OH2

OH2

O N

O

331253-86-2 N AGC, PDK1/Akt/Flt Dual and N N ATYPICAL, Pathway Inhibitor 329710-24-9 TK N N N

O NH

S O PI 3-Kg Inhibitor 648450-29-7 LIPID N

N

Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O NH

S O PI 3-Kg Inhibitor II 648449-76-7 LIPID F O

F O

O

N

O PI-103 371935-74-9 N LIPID N OH N

H N O O

HN

PKCb Inhibitor 257879-35-9 N AGC

N N

H N O O

PKCbII/EGFR Inhibitor 145915-60-2 TK, AGC

NH HN

F F

N

S NH N PKR Inhibitor 608512-97-6 OTHER O N H

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

Cl

OH PKR Inhibitor, 852547-30-9 INACTIVE Negative Control Cl Cl O N H

NH2 PP1 Analog II, 221244-14-0 N TK 1NM-PP1 N N N

NH2

N N PP3 5334-30-5 N N TK

Cl NH

N Purvalanol A 212844-53-6 N CMGC HO N N N H

HO

O

O O OH Rapamycin 53123-88-9 N AGC H O O O O O HO O O

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

N

Rho Kinase Inhibitor III, 7272-84-6 AGC Rockout

N H

O

H2N

N O Rho Kinase Inhibitor IV 913844-45-8 N AGC S O

N

H N O O

Ro-32-0432 151342-35-7 AGC N N

HCl

N

HCl O ROCK Inhibitor, HCl 146986-50-7 H AGC Y-27632 N N OH2 H NH2

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

Roscovitine 186692-46-6 CMGC

N H N SB 202190 152121-30-7 OH CMGC N

F

N SB 202474, Negative H control for p38 MAPK 172747-50-1 N INACTIVE inhibition studies O N

N

H N SB 203580 152121-47-6 S CMGC N O

F

H O N O

SB 218078 135897-06-2 CAMK

N N O

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

NH2 NH

NN

SB220025 165806-53-1 N CMGC

N

F

N SC-68376 318480-82-9 CMGC O

N

N

SKF-86002 72873-74-6 CMGC N F S N

Sorafenib 284461-73-0 TK

H S N

N Sphingosine Kinase OH LIPID Inhibitor 312636-16-1

HCl

Cl

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O

HN Src Kinase Inhibitor I 179248-59-0 O TK N

O N

H N O

N N Staurosporine, AGC, 120685-11-2 O H N-benzoyl- CMGC, TK

O N O

HN

O H O

N Staurosporine, N AGC, CAMK, 62996-74-1 Streptomyces sp. TK

O N H

N O

STO-609 52029-86-4 N OH O CAMK

O OH

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

O

Cl N N H SU11652 326914-10-7 TK N H O N H

O O N SU6656 330161-87-0 S H TK N O N H

O NH SU9516 666837-93-0 N CMGC N O H

Sunitinib 557795-19-4 TK

O Syk Inhibitor 622387-85-3 O N TK S

H2N O N H

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

OH2 F OH2 NH O F Syk Inhibitor II 227449-73-2 F HCl TK N NH2 HCl

H N 2 N N H

O

+ O N Syk Inhibitor III 1485-00-3 O- TK

O

Tandutinib 387867-13-2 TK

O O

H N TGF-b RI Inhibitor III 356559-13-2 TKL N

N

HCl

H N N TGF-b RI Kinase 396129-53-6 TKL Inhibitor

N N

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

Tofacitinib 477600-75-2 TK

Tozasertib 639089-54-6 OTHER

N F

HN Cl Tpl2 Kinase Inhibitor 871307-18-5 N CMGC HN

N N

Vandetanib 443913-73-3 TK

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

Vatalanib 212141-51-0 TK

O

VEGF Receptor 2 O 15966-93-5 TK Kinase Inhibitor I N H O N H

VEGF Receptor 2 N 288144-20-7 Br H TK Kinase Inhibitor II O N H

VEGF Receptor 2 N 204005-46-9 H TK Kinase Inhibitor III O N H

O

N VEGF Receptor 2 N 216661-57-3 TK Kinase Inhibitor IV N

S

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 1 Anastassiadis et al.

Cl O

N VEGF Receptor H Tyrosine Kinase 269390-69-4 NH TK Inhibitor II

N

Cl H H N N

O VEGF Receptor O Tyrosine Kinase 286370-15-8 TK O Inhibitor III, KRN633 N

O N

VX-702 745833-23-2 CMGC

O O O O Wortmannin 19545-26-7 ATYPICAL O

O O O

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2. Kinase constructs and substrates used in this study.

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # baculovirus in ABL1 ABL1 Abltide NP_005148.2 P00519 full-length - C-terminal His Sf21 insect cells baculovirus in ABL2/ARG ABL2 Abltide NP_009298 P42684 full-length - C-terminal His Sf21 insect cells baculovirus in ACK1 TNK2 Abltide NP_005772.3 Q07912 aa 110-476 - N-terminal GST Sf21 insect cells baculovirus in AKT1 AKT1 Crosstide NP_005154 P31749 full-length N-terminal His Sf21 insect cells baculovirus in Sf21 insect AKT2 AKT2 Crosstide NP_001617 P31751 full-length N-terminal His cells, activated by PDK1 baculovirus in AKT3 AKT3 Crosstide NP_005456 Q9Y243 full-length N-terminal His Sf21 insect cells ALK ALK pEY NP_004295.2 Q9UM73 cytoplasmic - Insect N-terminal GST ALK1/ACVRL cytoplasmic ACVRL1 Casein NP_000011.2 P37023 - Insect N-terminal GST 1 aa139-503 cytoplasmic ALK2/ACVR1 ACVR1 Casein NP_001096.1 Q04771 - Insect N-terminal GST aa145-509 Baculovirus ALK4/ACVR1 ACVR1B Casein NP_004293 P36896 aa 150-505 - infected insect N-terminal GST B cells ALK5/TGFBR TGFBR1 Casein NM_004612 P36897 aa200-503 - Sf9 cells GST-HIS fusion 1 Baculovirus YY301- N-terminal GST- ARAF ARAF MEK1 (K97R) NM_001654 P10398 aa 282-end infected Sf9 302DD tag cells ARK5/NUAK1 NUAK1 CHKtide NP_055655 O60285 full-length - Insect N-terminal His ASK1/MAP3K MAP3K5 MBP NP_005914 O99683 full-length - Insect N-terminal GST 5 baculovirus in N-terminal His6- Aurora A AURKA Kemptide NP_940839 O14965 full length - Sf21 insect cells tag Aurora B AURKB Kemptide NP_004208.2 Q96GD4 full-length - Insect N-terminal His AAH75064, Aurora C AURKC Kemptide Q9UQB9 full-length - Insect N-terminal His NP_003151

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # Baculovirus AXL AXL Abltide + Mn NP_068713 P30530 aa 473-894 - infected Sf9 C-terminal His cells BLK BLK pEY NP_001706 P51451 full-length - Insect N-terminal His baculovirus in BMX/ETK BMX pEY NP_001712 P51813 full length - C-terminal His Sf21 insect cells BRAF BRAF MEK1 (K97R) NP_004324.2 P15056 full-length - Insect N-terminal GST BRK PTK6 pEY + Mn NP_005966 Q13882 full-length - Insect C-terminal His N-terminal GST- BRSK1 BRSK1 CHKtide NP_115806 Q8TDC3 full-length - Insect tag GenBank baculovirus in N-terminal His6- BRSK2 BRSK2 ZIPtide Q8IWQ3 full-length - NM_003957 Sf21 insect cells tag N-terminal His6- BTK BTK pEY NP_000052 Q06187 full-length - Insect tagged Autocamtide baculovirus CAMK1a CAMK1 NP_003647.1 Q14012 full length - N-terminal His-tag 2 + Ca-CaM insect cell Autocamtide GenBank baculovirus N-terminal GST- CAMK1b PNCK Q6P2M8 full length - 2 + Ca-CaM NM_012040 insect cell tag Autocamtide baculovirus CAMK1d CAMK1D NP_705718.1 Q8IU85 full length - N-terminal His-tag 2 + Ca-CaM insect cell Autocamtide GenBank C-terminal baculovirus N-terminal GST- CAMK1g CAMK1G Q96NX5 - 2 + Ca-CaM NM_020439 truncation insect cell tag Autocamtide baculovirus CAMK2a CAMK2A NP_741960 Q9UQM7 full length - C-terminal His-tag 2 + Ca-CaM insect cell Autocamtide baculovirus N-terminal His6- CAMK2b CAMK2B NP_742078.1 Q13554 full length - 2 + Ca-CaM insect cell tag ZIPtide + Ca- baculovirus C-terminal His6- CAMK2d CAMK2D NP_742113 Q13557 full length - CaM* insect cell tag ZIPtide + Ca- GenBank C-terminal baculovirus in N-terminal GST- CAMK2g CAMK2G Q13555 - CaM* NM_172169 truncation Sf9 insect cells tag ZIPtide + Ca- N-terminal GST- CAMK4 CAMK4 NP_001735 Q9UQM7 full length - E. coli CaM* tag MBP + Ca- GenBank baculovirus in CAMKK1 CAMKK1 Q8N5S9 full-length - N-terminal GST CaM NM_032294 Sf9 insect cells

Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # Baculovirus MBP + Ca- CAMKK2 CAMKK2 NP_757380.1 Q96RR4 full-length - infected insect N-terminal GST CaM cells CDK1: N-terminal GST- CDK1/CCNA NM_001786/ P06493; full length / full baculovirus in CDK1/cyclin A Histone H1 - tag / N-terminal 2 NM_001237 cyclin A: length Sf9 insect cells GST-tag P20248 CDK1: C-terminal His6- CDK1/CCNB NP_001777/B P06493; full length / full baculovirus CDK1/cyclin B Histone H1 - tag / N-terminal 1 P_114172 cyclin B: length insect cell His6-tag P14635 CDK2: N-terminal His6- CDK2/CCNA NP_001789, P24941; CDK2/cyclin A Histone H1 full-length - Insect tag / N-terminal 2 NP_001228 cyclin A: His6-tag P20248 EMBL CDK2: C-terminal His6- CDK2/CCNE M68520, P24941; full length / full baculovirus in CDK2/cyclin E Histone H1 - tag / N-terminal 1 GenBank cyclin E: length Sf21 insect cells GST-tag NM_001238 P24864 CDK3: CDK3/CCNE NM_001258, Q00526; full length / full baculovirus in N-terminal GST- CDK3/cyclin E Histone H1 - 1 NM_001238 cyclin E: length Sf9 insect cells tag P24864 CDK4: CDK4/cyclin CDK4/CCND NP_000066, P11802; baculovirus N-terminal GST- RB-CTF full length - D1 1 NP_444284 Cyclin D1: insect cell tag P24385 CDK4: CDK4/cyclin CDK4/CCND NM_000075, P11802; baculovirus in N-terminal GST- RB-CTF full length - D3 3 NM_001760 Cyclin D3: Sf9 insect cells tag P30281 CDK5: N-terminal His6- CDK5/CDK5 NP_004926.1, full length / full baculovirus CDK5/p25 Histone H1 Q00535; p25: - tag / N-terminal R1 NP_003876 length insect cell Q15078 GST-tag

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # CDK5: N-terminal His6- CDK5/CDK5 NP_004926.1, full length / full baculovirus CDK5/p35 Histone H1 Q00535; p35: - tag / N-terminal R1 NP_003876 length insect cell Q15078 His6-tag CDK6: Baculovirus CDK6/cyclin CDK6/CCND NP_001250, Q00534; RB-CTF full-length - infected insect N-terminal GST D1 1 NP_444284 Cyclin D1: cells P24385 CDK6: N-terminal His6- CDK6/cyclin CDK6/CCND X66365, Q00534; full length / full baculovirus in RB-CTF - tag / N-terminal D3 3 M90814 Cyclin D3: length Sf9 insect cells GST-tag P30281 CDK7: P50613; NP_001790, CDK7/CCNH/ Cyclin H: CDK7/cyclin H Histone H1 NP_001230, full-length - Insect N-terminal His MNAT1 P51946; NP_002422.1 MNAT1: Q6ICQ7 CDK9: N-terminal His6- NP_001252, P50750; full length / full CDK9/cyclin K CDK9/CCNK PDKtide - Insect tag / N-terminal NP_003849 Cyclin K: length His6-tag O75909 CDK9: CDK9/cyclin CDK9/CCNT NP_001252, P50750; PDKtide full-length - Insect N-terminal His T1 1 NP_001231 Cyclin T1: O60563 GenBank baculovirus in CHK1 CHEK1 CHKtide O14757 full length - N-terminal His NM_001274 Sf9 insect cells baculovirus in CHK2 CHEK2 CHKtide NP_009125 O96017 full-length - C-terminal His Sf21 insect cells baculovirus CK1a1 CSNK1A1 CK1tide NP_001883.4 P48729 full length - GST-tag insect cell baculovirus N-terminal GST- CK1d CSNK1D CK1tide NP_620693 P48730 full length - insect cell tag baculovirus CK1epsilon CSNK1E CK1tide NP_001885 P49674 full length - C-terminal His-tag insect cell CK1g1 CSNK1G1 CK1tide NP_071331 Q9HCP0 full-length - Insect N-terminal GST

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # baculovirus CK1g2 CSNK1G2 CK1tide NP_001310 P78368 full length - C-terminal His-tag insect cell NP_004375.1, CK1g3 CSNK1G3 CK1tide Q9Y6M4 full-length - Insect N-terminal GST NP_004375.2 baculovirus C-terminal GST- CK2a CSNK2A1 CK2 sub NP_001886 P68400 full length - insect cell tag baculovirus N-terminal GST- CK2a2 CSNK2A2 CK2 sub NP_001887 P19784 full length - insect cell tag c-Kit KIT pEY + Mn NP_000213 P10721 aa 544-976 - Insect His6-tag full catalytic CLK1 CLK1 MBP NP_004062 P49759 - E. coli N-terminal GST domain catalytic baculovirus CLK2 CLK2 MBP NP_003984 P49760 domain aa137- - N-terminal GST insect cell 498 CLK3 CLK3 MBP NP_003983 P49761 full-length - Insect N-terminal GST CLK4 CLK4 MBP NP_065717 Q9HAZ1 full-length - Insect N-terminal GST baculovirus c-MER MERTK pEY NP_006334.2 Q12866 aa 578-872 - N-terminal GST insect cell baculovirus c-MET MET MBP NP_000236.2 P10721 aa 956-1390 - N-terminal His insect cell Baculovirus COT1/MAP3K MAP3K8 MEK1 (K97R) NP_005195 P41279 aa 30-397 - infected insect N-terminal GST 8 cells CSK CSK pEY NP_004374 P41240 full-length - E. coli C-terminal His c-Src SRC pEY NP_005408 P12931 full-length - Insect C-terminal His CTK/MATK MATK pEY NP_647611 P42679 full-length - Insect C-terminal His DAPK1 DAPK1 ZIPtide NP_004929 P53355 full catalytic - Insect N-terminal GST ZIPtide + Ca- DAPK2 DAPK2 NP_055141 Q9UIK4 full catalytic - Insect N-terminal GST CaM Baculovirus Autocamtide DCAMKL2 DCLK2 NP_689832 Q8N568 full-length - infected insect N-terminal GST 2 + Ca-CaM cells baculovirus in DDR2 DDR2 AXLtide + Mn NP_006173.2 Q16832 aa 424-855 - N-terminal GST insect cells

Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # DMPK DMPK AXLtide NP_004400 Q09013 full-length - Insect N-terminal GST DRAK1/STK1 STK17A ZIPtide NP_004751 Q9UEE5 full-length - Insect N-terminal GST 7A DYRK1/DYRK DYRK1A Casein NP_001387 Q13627 full-length - Insect N-terminal GST 1A DYRK1B DYRK1B Casein NP_004705 Q9Y463 full-length - Insect N-terminal GST GenBank baculovirus in DYRK2 DYRK2 Casein Q92630 full length - His6-tag NM_003583 Sf21 insect cells DYRK3 DYRK3 Casein NP_003573 O43781 full length - Insect N-terminal GST DYRK4 DYRK4 Casein NP_003836.1 Q9NR20 full length - Insect N-terminal GST EGFR EGFR pEY + Mn NP_005219.2 P00533 cytoplasmic - Insect N-terminal GST EPHA1 EPHA1 pEY + Mn NP_005223.2 P21709 cytoplasmic - Insect N-terminal GST EPHA2 EPHA2 pEY NP_004422.2 P29317 cytoplasmic - Insect N-terminal GST EPHA3 EPHA3 pEY + Mn NP_005224 P29320 cytoplasmic - Insect C-terminal His baculovirus in EPHA4 EPHA4 pEY + Mn NP_004429.1 P54764 aa 616-887 - N-terminal GST insect cells baculovirus in EPHA5 EPHA5 pEY + Mn NP_004430.1 P54756 aa 595-1037 - N-terminal GST insect cells NM_0010804 Catalytic (561- EPHA6 EPHA6 pEY Q9UF33 - Insect N-terminal GST 48 end) catalytic baculovirus N-terminal GST- EPHA7 EPHA7 pEY NP_004431.1 Q15375 domain (aa - insect cell tag 579-998) catalytic baculovirus N-terminal GST- EPHA8 EPHA8 pEY NP_065387 P29322 domain (aa - insect cell tag 565-1005) EPHB1 EPHB1 pEY NP_004432 P54762 cytoplasmic - Insect N-terminal GST EPHB2 EPHB2 pEY + Mn NP_004433 P29323 aa 616-889 - Insect GST EPHB3 EPHB3 pEY NP_004434 P54753 cytoplasmic - Insect N-terminal GST EPHB4 EPHB4 pEY NP_004435 P54760 cytoplasmic - Insect C-terminal His GenBank ERBB2/HER2 ERBB2 pEY + Mn P04626 aa679-1255 - Insect N-terminal GST X03363 baculovirus ERBB4/HER4 ERBB4 pEY + Mn NP_005226 Q15303 aa 708-993 - N-terminal GST insect cell ERK1/MAPK3 MAPK3 MBP NP_002737 P27361 full-length - Insect N-terminal GST

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # ERK2/MAPK1 MAPK1 MBP NP_620407 P28482 full-length - E. coli N-terminal GST FAK/PTK2 PTK2 pEY NP_722560 Q05397 full-length - Insect N-terminal GST baculovirus FER FER pEY NP_005237 P16591 aa540-822 - N-terminal GST insect cell FES/FPS FES pEY NP_001996 P07332 full-length - Insect C-terminal His FGFR1 FGFR1 pEY + Mn NP_000595 P11362 cytoplasmic - Insect C-terminal His FGFR2 FGFR2 pEY + Mn NP_075261 P21802 cytoplasmic - Insect C-terminal His FGFR3 FGFR3 pEY + Mn NP_000133 P22607 cytoplasmic - Insect N-terminal His baculovirus N-terminal His6- FGFR4 FGFR4 pEY + Mn NP_002002 P22455 aa 460-802 - insect cell tag baculovirus C-terminal His6- FGR FGR pEY NP_005239 P09769 full-length - insect cell tag FLT1/VEGFR baculovirus N-terminal GST FLT1 pEY + Mn NP_002010 P17948 aa 781-1338 - 1 insect cell tagged baculovirus C-terminal His6- FLT3 FLT3 Abltide NP_004110 P36888 aa 564-958 - insect cell tag FLT4/VEGFR NP_891555.1, FLT4 pEY + Mn P35916 aa 800-1297 Q890H Insect N-terminal GST 3 AAA85215 FMS CSF1R pEY + Mn NP_005202 P07333 cytoplasmic - Insect C-terminal His FRK/PTK5 FRK pEY + Mn NP_002022 P42685 full-length - Insect N-terminal GST FYN FYN pEY NP_694592 P06241 full-length - Insect C-terminal His N-terminal His6- GCK/MAP4K2 MAP4K2 MBP NP_004570.2 Q12851 full-length - Insect tagged baculovirus GRK2 ADRBK1 Casein AAH37963 P25098 full length - C-terminal His-tag insect cell GRK3 ADRBK2 Casein NP_005151 P35626 full-length - Insect N-terminal GST GRK4 GRK4 Casein NP_892027 P32298 full-length - Insect N-terminal GST GRK5 GRK5 Casein NP_005299 P34947 full-length - Insect N-terminal GST NP_00100410 GRK6 GRK6 Casein P43250 full-length - Insect N-terminal GST 6 GRK7 GRK7 Casein NP_631948 Q8WTQ7 full-length - Insect N-terminal GST Phospho- Glycogen baculovirus GSK3a GSK3A NP_063937.2 P49840 full length - C-terminal His-tag Synthase insect cell peptide

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # Phospho- Glycogen baculovirus C-terminal His6- GSK3b GSK3B NP_002084 P49841 full-length Synthase insect cell tag peptide Haspin GSG2 Histone H3 NP_114171.1 Q8TF76 aa471-798 - Insect C-terminal His Src Substrate HCK HCK NP_002101 P08631 full-length - Insect C-terminal His peptide

HGK/MAP4K4 MAP4K4 MBP NP_004825 O95819 full catalytic - Insect N-terminal GST baculovirus HIPK1 HIPK1 MBP NP_689909 Q86Z02 aa 158-555 - N-terminal GST insect cell baculovirus HIPK2 HIPK2 MBP NP_073577.3 Q9H2X6 aa 165-564 - N-terminal GST insect cell baculovirus N-terminal His6- HIPK3 HIPK3 MBP NP_005725.2 Q9H422 aa 163-562 - insect cell tag baculovirus HIPK4 HIPK4 MBP NP_653286 Q8NE63 full-length - N-terminal GST insect cell baculovirus IGF1R IGF1R pEY + Mn NP_000866 P08069 aa 960-1367 - C-terminal His insect cell Baculovirus N-terminal 6X-His IKKa/CHUK CHUK IKKtide NP_001269 O15111 full-length - infected insect tag cells IKKb/IKBKB IKBKB IKKtide NP_001547 O14920 full-length - Insect N-terminal GST IKKe/IKBKE IKBKE Casein NP_054721.1 Q14164 full-length - Insect N-terminal GST cytoplasmic baculovirus domain of the IR INSR pEY + Mn P06213 - expression GST-tag β-subunit (aa system 941-1343) Baculovirus IRAK1 IRAK1 MBP NP_001560 P51617 aa 197-721 - infected insect N-terminal GST cells NP_057207, baculovirus N-terminal His6- IRAK4 IRAK4 MBP Q9NWZ3 full-length - AAH13316 insect cell tag IRR/INSRR INSRR AXLtide NP_055030 P14616 cytoplasmic - Insect N-terminal GST

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # ITK ITK MBP NP_005537 Q08881 full-length - Insect N-terminal GST JAK1 JAK1 pEY NP_002218.2 P23458 aa 866-1154 - Insect N-terminal GST aa 809-1132 JAK2 JAK2 pEY NP_004963 O60674 - Insect N-terminal GST +g JAK3 JAK3 JAK3tide NP_000206 P52333 aa 781-1124 - Insect N-terminal GST baculovirus insect cell, JNK1 MAPK8 ATF2 NP_002741.1 P45983-2 full length - N-terminal His-tag activated by MAP2K7 baculovirus in insect cell, JNK2 MAPK9 ATF2 NP_002743 P45984-1 full length - N-terminal His-tag activated by MAP2K7 baculovirus in JNK3 MAPK10 ATF2 NP_002744 P53779 full length - N-terminal GST insect cell baculovirus in insect cell, C-terminal His6- KDR/VEGFR2 KDR pEY + Mn NP_002244 P35968 aa 789-1356 - activated by tagged autophophorylat ion baculovirus in KHS/MAP4K5 MAP4K5 MBP NP_942089 Q9Y4K4 full length - N-terminal GST insect cell baculovirus in C-terminal His6- LCK LCK pEY + Mn NP_005347 P06239 full-length - insect cells tagged baculovirus in catalytic insect cells, N-terminal His6- LIMK1 LIMK1 Cofilin 1 NP_002305 P53667 domain (aa - activated by co- tag 285-638) expression with ROCK1 STK11:Q1583 1; NM000455/A STK11/STRA STRADA:Q7 baculovirus in LKB1 LKB1tide F308302/NM_ full length - DA/CAB39 RTN6; Sf21 cells 016289 CAB39:Q9Y3 76

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # GenBank baculovirus in N-terminal His6- LOK/STK10 STK10 Axltide O94804 aa 1-348 - NM_005990 Sf21 insect cells tag baculovirus in LRRK2 LRRK2 LRRKtide NP_940980.2 Q5S007 aa 970-2527 - N-terminal GST insect cells baculovirus in C-terminal His6- LYN LYN pEY NP_002341 P07948-1 full length - insect cells tag LYN B LYN pEY+ Mn NP_002341 P07948-2 full-length - Insect C-terminal His Glycogen Synthase- E. coli, activated MAPKAPK2 MAPKAPK2 NP_116584 P49137 ful length N-terminal His-tag derived by MAPK14 peptide Glycogen Insect cell, Synthase- MAPKAPK3 MAPKAPK3 NP_004626 Q16644 full length - activated by N-terminal His-tag derived MAPK14 peptide Glycogen MAPKAPK5/P Synthase- baculovirus in N-terminal His6- MAPKAPK5 NP_003659 Q8IW41 full length - RAK derived insect cells tag peptide baculovirus in N-terminal GST- MARK1 MARK1 CHKtide NP_061120.1 Q9P0L2 full length - insect cells tag MARK2/PAR- baculovirus in N-terminal GST- MARK2 CHKtide NP_059672.2 Q7KZI7-4 full length - 1Ba insect cells tag baculovirus in N-terminal GST- MARK3 MARK3 CHKtide NP_002367.3 P27448 full length - insect cells tag baculovirus in N-terminal GST- MARK4 MARK4 CHKtide NP_113605 Q96L34 full length - insect cells tag baculovirus in insect cell, N-terminal His6- MEK1 MAP2K1 ERK(K52R) NP_002746 Q02750 full length - activated by tag RAF1 in vivo baculovirus in insect cells, C-terminal His6- MEK2 MAP2K2 ERK(K52R) NP_109587 P36507 full length - activated by co- tag expression with RAF1

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # baculovirus in N-terminal GST- MEKK2 MAP3K2 MBP NM_006609 Q9Y2U5 full length - Sf9 insect cells tag baculovirus in N-terminal GST- MEKK3 MAP3K3 MBP NM_002401 Q99759 full length - Sf9 insect cells tag N-terminal GST- MELK MELK ZIPtide NP_055606.1 Q14680 aa 1-340 - Insect cell tag MINK/MINK1 MINK1 MBP NP_056531 Q8N4C8 full catalytic - Insect N-terminal GST baculovirus in MKK4 MAP2K4 JNK(K55M) NM_003010 P45985 aa 33-end N-terminal His-tag Sf9 insect cells S207E, MKK6 MAP2K6 MBP NP_002749 P52564 full length E. coli N-terminal His-tag T211E ZIPtide + Ca- Catalytic (aa MLCK/MYLK MYLK NP_444253 Q15746 Insect N-terminal GST CaM 1428-1771) MLCK2/MYLK ZIPtide + Ca- MYLK2 NP_149109 Q9H1R3 full length Insect N-terminal GST 2 CaM MLK1/MAP3K MAP3K9 Casein NP_149132.1 P80192 full catalytic - Insect N-terminal GST 9 MLK2/MAP3K MAP3K10 MBP NP_002437 Q02779 full catalytic - Insect N-terminal GST 10 MLK3/MAP3K MAP3K11 MBP NP_002410.1 Q16584 full catalytic - Insect N-terminal GST 11 aa2-end GenBank baculovirus in MNK1 MKNK1 MBP Q9BUB5 (deletion T385D N-terminal GST NM_003684 Sf9 insect cells aa165-205) Baculovirus MNK2 MKNK2 MBP NP_060042.2 Q9HBH9 full length - infected insect N-terminal GST cells Long S6 Baculovirus MRCKa/CDC Kinase C-terminal 6X-His CDC42BPA NP_055641 Q5VT25 aa 1-473 - infected insect 42BPA substrate tag cells peptide Long S6 Baculovirus MRCKb/CDC Kinase C-terminal 6X-His CDC42BPB NP_006026.2 Q9Y5S2 aa 1-473 - infected insect 42BPB substrate tag cells peptide

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # MSK1/RPS6K RPS6KA5 Crosstide NP_004746.2 O75582 full-length - Insect N-terminal GST A5 MSK2/RPS6K RPS6KA4 Crosstide NP_003933 O75676 full-length - Insect N-terminal GST A4 MSSK1/STK2 SRPK3 RS peptide NP_055185 Q9UPE1 full-length - Insect N-terminal GST 3 MST1/STK4 STK4 Axltide NP_006273 Q13043 full-length - Insect N-terminal GST MST2/STK3 STK3 MBP NP_006272.2 Q13188 full length - Insect N-terminal GST MST3/STK24 STK24 MBP NP_003567 Q9Y6E0 full length - Insect N-terminal GST MST4 MST4 MBP NP_057626.2 Q9P289 full-length - Insect N-terminal GST MUSK MUSK MBP NP_005583.1 O15146 cytoplasmic - Insect N-terminal GST ZIPtide + Ca- MYLK3 MYLK3 MYLK3 BC109097 Q32MK0 full-length Insect N-terminal GST CaM Catalytic MYO3B MYO3B MYO3B MBP NM_138995 Q8WXR4 Insect N-terminal GST (1-326) NEK1 NEK1 MBP NP_036356 Q96PY6 aa 1-505 - Insect N-terminal GST Baculovirus NEK11 NEK11 MBP NP_079076 Q8NG66 full-length - infected insect N-terminal GST cells C-terminal His6- NEK2 NEK2 MBP NP_002488 P51955 full-length - Insect tagged NEK3 NEK3 MBP NP_689933 Q8WUN5 full-length - Insect N-terminal GST NEK4 NEK4 MBP NP_003148 P51957 full-length - Insect N-terminal GST baculovirus NEK6 NEK6 MBP NP_055212 Q9HC98 aa 7-313 - C-terminal His-tag insect cell N-terminal GST NEK7 NEK7 MBP NP_598001.1 Q8TDX7 full length - Insect tag Catalytic (aa N-terminal GST NEK9 NEK9 MBP NP_149107.2 Q8TD19 - Insect 347-732) tag Catalytic (aa N-terminal GST NIK/MAP3K14 MAP3K14 MBP NP_003945.2 Q99558 - Insect 318-947) tag NLK NLK MBP NP_057315.1 Q9UBE8 full length - Insect N-terminal GST Baculovirus OSR1/OXSR1 OXSR1 CATCHtide NP_005100.1 O95747 full length - infected insect N-terminal GST cells

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # P38a/MAPK1 MAPK14 MBP NP_620581 Q16539 full-length - E. coli N-terminal GST 4 P38b/MAPK1 MAPK11 MBP NP_002742 Q15759 full-length - Insect N-terminal His 1 P38d/MAPK1 MAPK13 MBP NP_002745 O15264 full-length - Insect N-terminal His 3 P38g/MAPK1 MAPK12 MBP NP_002960 P53778 full-length - Insect N-terminal His 2 p70S6K/RPS S6K/Rsk2 NP_003152, Catalytic (aa 1- RPS6KB1 P23443 - Insect N-terminal GST 6KB1 peptide 2 T412E 421) p70S6Kb/RP S6K/Rsk2 RPS6KB2 NP_003943 Q9UBS0 full length - Insect N-terminal GST S6KB2 peptide 2 Long S6 Kinase PAK1 PAK1 NP_002567 Q13153 full length - Insect N-terminal GST substrate peptide Long S6 Kinase PAK2 PAK2 NP_002568.2 Q13177 full length - Insect N-terminal GST substrate peptide N-terminal His6- PAK3 PAK3 ZIPtide NP_002569 O75914 full length - Insect tag Catalytic (aa PAK4 PAK4 MBP NP_005875 O96013 - Insect N-terminal GST 295-591) Baculovirus N-terminal 6X-His PAK5 PAK7 ZIPtide NP_065074 Q9P286 aa 425-719 - infected insect tag cells PAK6 PAK6 ZIPtide NP_064553 Q9NQU5 full-length - Insect C-terminal His PASK PASK ZIPtide NP_055963 Q96RG2 full catalytic - Insect N-terminal GST PBK/TOPK PBK MBP NP_060962 Q96KB5 full catalytic - Insect C-terminal His Cytoplasmic PDGFRa PDGFRA pEY + Mn NP_006197 Q9DE49 Insect N-terminal GST (550-1089) Cytoplasmic N-terminal His6- PDGFRb PDGFRB pEY + Mn NP_002600 P09619 Insect (558-1106) tagged

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # N-terminal His6- PDK1/PDPK1 PDPK1 PDKtide NP_002604 O15530 full length - Insect tag N-terminal GST- PHKg1 PHKG1 ZIPtide NP_006204 Q16816 full length - Insect tag N-terminal GST- PHKg2 PHKG2 ZIPtide NP_000285 P15735 full length - Insect tag S6K/Rsk2 C-terminal His PIM1 PIM1 NP_002639 P11309 full length - Insect peptide 2 tagged PIM2 PIM2 Pim2tide NP_006866 Q9P1W9 full-length - Insect N-terminal GST GenBank Baculovirus in N-terminal His6- PIM3 PIM3 Pim2tide Q86V86 aa 2-end - AB114795 Sf21 insect cells tag Catalytic (1- N-terminal His6- PKA PRKACA PKA sub NP_002721.1 P17612 - E. coli 351) tag Long S6 Kinase baculovirus in PKAca KAPCA NM_002730 P17612 full-length - N-terminal GST substrate Sf9 cells peptide Long S6 Kinase baculovirus in PKAcb KAPCB NM_002730 P22694 full-length - N-terminal GST substrate Sf9 cells peptide Long S6 Kinase PKAcg KAPCG NM_002732 P22612 full-length - Insect N-terminal GST substrate peptide Histone H1 + PKCa PRKCA Lipid NP_002728 P17252 full length - Insect none Activator Histone H1 + baculovirus in PKCb1 PRKCB1 Lipid NP_991100.1 P05771 full length - none Sf21 insect cells Activator Histone H1 + PKCb2 PRKCB2 Lipid NP_002729 P05771 full length - insect none Activator

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # PKCe Pep + PKCd PRKCD Lipid NP_006245 Q05655 full length - insect none Activator S25 PKC PKCepsilon PRKCE NP_005391.1 Q02156 full-length - Insect none Peptide PKCe Pep + baculovirus in PKCeta PRKCH Lipid NM_006255 P24723 full length - N-terminal GST Sf9 insect cells Activator Histone H1 + baculovirus in PKCg PRKCG Lipid NP_002730 P05129 - - none insect cells Activator PKCepsilon baculovirus PKCiota PRKCI NP_002731 P41743 full length - N-terminal His-tag Peptide insect cell Glycogen PKCmu/PRK Synthase- PRKD1 NP_002733 Q15139 full length - Insect N-terminal GST D1 derived peptide Glycogen PKCnu/PRKD Synthase- PRKD3 NP_005804 O94806 full-length - Insect N-terminal GST 3 derived peptide Histone H1 + baculovirus in C-terminal His6- PKCtheta PRKCQ Lipid NP_006248 Q04759 full length P330L insect cells tag Activator PKCepsilon PKCzeta PRKCZ NP_002735 Q05513 full-length - Insect none Peptide Glycogen Synthase- PKD2/PRKD2 PRKD2 NP_057541 Q9BZL6 full-length - Insect N-terminal GST derived peptide NM_0010985 baculovirus in N-terminal His6- PKG1a PRKG1 Kemptide Q13976 full length - 12 Sf9 cells tag GenBank baculovirus in N-terminal His6- PKG1b PRKG1 Kemptide P14619 full length - NM_006258 Sf21 insect cells tag PKG2/PRKG2 PRKG2 Kemptide NP_006250 Q13237 full length - Insect N-terminal GST

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # Glycogen Synthase- PKN1/PRK1 PKN1 NP_998725 Q16512 full length - Insect N-terminal GST derived peptide Glycogen Synthase- PKN2/PRK2 PKN2 NP_006247 Q16513 full length - Insect N-terminal GST derived peptide C-terminal His6- PLK1 PLK1 Casein NP_005021 P53350 full length - Insect tag N-terminal GST- PLK2 PLK2 Casein NP_006613 Q9NYY3 full length - Insect His6 tag Catalytic (58- PLK3 PLK3 Casein NP_004064 Q9H4B4 - Insect N-terminal GST 340) PRKX PRKX Kemptide NP_722560 P51817 full length - Insect N-terminal GST PYK2 PTK2B pEY + Mn NP_004094 Q14289 full-length - Insect N-terminal GST MEK1 RAF1 RAF1 NP_002871 P04049 full catalytic - Insect N-terminal GST (K97R) RET RET CHKtide NP_066124 P07949 cytoplasmic - Insect N-terminal GST Baculovirus N-terminal 6X-His RIPK2 RIPK2 MBP NP_003812 O43353 aa 1-299 - infected insect tag cells RIPK5 DSTYK MBP NP_056190.1 Q6XUX3 full length - Insect N-terminal GST Long S6 Kinase ROCK1 ROCK1 NP_005397 Q13464 aa 1-535 - Insect N-terminal GST substrate peptide Long S6 Kinase baculovirus in ROCK2 ROCK2 NM_004850 O75116 aa 5-554 - N-terminal GST substrate Sf9 insect cells peptide Baculovirus RON/MST1R MST1R Axltide + Mn NP_002438 Q04912 aa 983-1400 - infected insect N-terminal GST cells ROS/ROS1 ROS1 IGF-1Rtide NP_002935 P08922 cytoplasmic - Insect N-terminal GST

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # Glycogen Synthase- N-terminal His6- RSK1 RPS6KA1 NP_002944 Q15418 full length - Insect derived tag peptide Glycogen Synthase- baculovirus RSK2 RPS6KA3 NP_004577 P51812 full length - C-terminal His-tag derived insect cell peptide Glycogen Synthase- N-terminal His6- RSK3 RPS6KA2 NP_066958 Q15349 full length - Insect derived tag peptide Long S6 Kinase RSK4 RPS6KA6 NP_055311 Q9UK32 full-length - Insect N-terminal GST substrate peptide NP_005618, Catalytic (60- SGK1 SGK1 Crosstide O00141 S422D Insect N-terminal GST S589D 431) full length baculovirus SGK2 SGK2 Crosstide Q9HBY8 (Met1- - expression GST-tag Cys367) system NP_037389, Catalytic (87- N-terminal GST- SGK3/SGKL SGK3 Crosstide Q96BR1 - Insect S487D 496) tag N-terminal GST- SIK2 SIK2 Kemptide NP_056006.1 Q9H0K1 full length - Insect tag N-terminal GST- SLK/STK2 SLK Histone H3 NP_055535.1 Q9H2G2 full length - Insect tag SNARK/NUA baculovirus in N-terminally fused NUAK2 MBP NM_030952 Q9H093 aa M1-T628 - K2 Sf9 insect cells to GST-His6 SRMS SRMS pEY + Mn NP_543013 Q9H3Y6 full length - Insect N-terminal GST SRPK1 SRPK1 RS peptide NP_003128 Q96SB4 full length - Insect N-terminal His-tag N-terminal GST- SRPK2 SRPK2 RS peptide NP_872633 P78362 full length - Insect tag STK16 STK16 MBP NP_003682 O75716 full length - Insect N-terminal His-tag

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # STK22D/TSS TSSK1B CHKtide NP_114417 Q9BXA7 full length - Insect C-terminal His-tag K1 STK25/YSK1 STK25 MBP NP_006365 O00506 full length - Insect N-terminal GST STK33 STK33 MBP NP_112168 Q9BYT3 full length - Insect N-terminal His-tag Modified PKA STK38/NDR1 STK38/NDR1 STK38 NM_007271 Q15208 full-length Insect N-terminal GST Substrate STK39/STLK3 STK39 CATCHtide NP_037365.2 Q9UEW8 full length - Insect N-terminal GST SYK SYK pEY NP_003168 P43405 full-length - Insect N-terminal GST full-length MAP3K7:O43 Baculovirus MAP3K7/MA NP_663306, MAP3K7; aa N-terminal 6X-His TAK1 Casein 318; - infected insect P3K7IP1 NP_006107 437-504 tag TAB1:Q15750 cells MAP3K7IPI baculovirus in TAOK1 TAOK1 MBP NM_020791 Q7L7X3 aa1-314 - N-terminal GST Sf9 insect cells TAOK2 TAOK2 MBP NP_004774 Q9UL54 full catalytic - Insect N-terminal GST TAOK3/JIK TAOK3 MBP NP_057365 Q9H2K8 full-length - Insect N-terminal GST TBK1 TBK1 CK1tide NP_037386 Q9UHD2 full-length - Insect N-terminal GST baculovirus TEC TEC pEY + Mn NP_003206 P42680 full length - C-terminal His-tag insect cell baculovirus in TGFBR2 TGFBR2 MBP NM_003242 P37173 190-end - N-terminal GST Sf9 insect cells Cytoplasmic N-terminal GST- TIE2/TEK TEK pEY + Mn NP_000450 Q02763 - Insect (817-1101) tag Catalytic TLK2 TLK2 TLK2 Casein NM_006852 Q86UE8 Insect N-terminal GST (388-end) Cytoplasmic TRKA NTRK1 pEY + Mn NP_002520 P04629 - insect C-terminal His-tag (441-796) Cytoplasmic TRKB NTRK2 pEY + Mn NP_006171 Q16620 - Insect C-terminal His-tag (526-838) catalytic baculovirus TRKC NTRK3 pEY NP_002521.2 Q16288 domain (aa - C-terminal His-tag insect cell 510-825) N-terminal His6- TSSK2 TSSK2 CHKtide NP_443732 Q96PF2 full length - Insect tag

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 2: Kinase constructs and substrates Anastassiadis et al.

RBC Enzyme HUGO Genbank Protein Substrate Clone Mutation Expression Tag Name symbol Accession # Accession # TTK TTK MBP NP_003309 P33981 full-length - Insect N-terminal N-terminal GST- TXK TXK ABLtide NP_003319.2 P42681 full length - Insect tag Cytoplasmic N-terminal GST- TYK1/LTK LTK ABLtide NP_002335.2 P29376 - Insect (450-864) tag TYK2 TYK2 AXLtide NP_003322.2 P29597 aa 833-1187 - Insect N-terminal GST Cytoplasmic N-terminal GST- TYRO3/SKY TYRO3 pEY + Mn NP_006284 Q06418 - Insect (451-890) tag baculovirus in N-terminal GST- ULK1 ULK1 MBP BC111603 O75385 aa 1-649 - Sf9 insect cells tag baculovirus in N-terminal GST- ULK2 ULK2 MBP NM_014683 Q8IYT8 aa 1-631 - Sf9 insect cells tag N-terminal His6- ULK3 ULK3 ULK3 Casein BC157884 Q6PHR2 full-length Insect tag full length, VRK1 VRK1 Casein NM_003384 Q99986 - Sf9 cells GST-HIS fusion Met1-Lys396 WEE1 WEE1 MBP NP_003381.1 P30291 full-length - Insect N-terminal Catalytic (166- WNK2 WNK2 MBP NP_006639.3 Q9Y3S1 - Insect N-terminal GST 489) Catalytic (1- N-terminal His6- WNK3 WNK3 MBP NP_065973 Q9BYP7 - Insect 434) tag YES/YES1 YES1 pEY NP_005424 P07947 full length - Insect C-terminal His-tag ZAK/MLTK ZAK MBP NP_598407.1 Q9NYL2 full length - Insect N-terminal GST C-terminal His6- ZAP70 ZAP70 pEY NP_001070 P43403 full length - Insect tag ZIPK/DAPK3 DAPK3 ZIPtide NP_001339 O43293 full-length - Insect N-terminal GST

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 4. A ranked table of kinases sorted by Selectivity score, the fraction of all

tested inhibitors that inhibited the catalytic activity of the test kinase by >50%.

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 4: Kinases sorted by Selectivity score Anastassiadis et al.

Selectivity Selectivity Kinase Kinase Score Score COT1/MAP3K8 0.000 GRK4 0.017 CTK MATK 0.000 HIPK2 0.017 DYRK4 0.000 HIPK3 0.017 GRK2 0.000 IKKb/IKBKB 0.017 GRK3 0.000 PKCiota 0.017 HIPK1 0.000 PAK6 0.017 JNK3 0.000 CK1g1 0.022 MAPKAPK3 0.000 MEK1 0.022 NEK6 0.000 MKK6 0.022 NEK7 0.000 MSSK1/STK23 0.022 P38d/MAPK13 0.000 p70S6Kb/RPS6KB2 0.022 P38g 0.000 PAK4 0.022 VRK1 0.000 SRMS 0.022 WNK3 0.000 TEC 0.022 ALK2/ACVR1 0.006 AKT3 0.023 CAMK1b 0.006 CAMK1g 0.023 CAMK4 0.006 DAPK2 0.023 DMPK 0.006 PBK/TOPK 0.023 GRK5 0.006 TIE2/TEK 0.023 NEK2 0.006 SRPK2 0.023 NEK3 0.006 EPHB3 0.023 OSR1/OXSR1 0.006 ASK1/MAP3K5 0.028 PLK2 0.006 EPHA8 0.028 RON/MST1R 0.006 NEK9 0.028 SGK3/SGKL 0.006 PAK2 0.028 STK39/STLK3 0.006 PAK3 0.028 WNK2 0.006 RIPK5 0.028 MRCKa/CDC42BPA 0.006 TGFBR2 0.028 AKT2 0.011 TTK 0.028 ALK4/ACVR1B 0.011 ZIPK/DAPK3 0.028 CDK7/cyclin H 0.011 CK1g2 0.028 DCAMKL2 0.011 DYRK3 0.028 ERK1 0.011 CAMK1d 0.028 ERK2 MAPK1 0.011 PKAcg 0.031 HIPK4 0.011 EPHA3 0.034 IKKa/CHUK 0.011 EPHA5 0.034 JNK1 0.011 EPHA7 0.034 MAPKAPK2 0.011 GRK6 0.034 MRCKb/CDC42BPB 0.011 GRK7 0.034 NEK11 0.011 MEK2 0.034 PKCzeta 0.011 P38b/MAPK11 0.034 SRPK1 0.011 PAK5 0.034 TSSK2 0.011 PIM2 0.034 WEE1 0.011 PKCg 0.034 ZAP70 0.011 PKG1b 0.034 AKT1 0.011 PKN2/PRK2 0.034 ALK5/TGFBR1 0.011 PLK1 0.034 CAMK1a 0.011 STK25/YSK1 0.034 MAPKAPK5/PRAK 0.011 PKA 0.034 NIK/MAP3K14 0.011 SGK2 0.034 ALK1/ACVRL1 0.017 CK1g3 0.039 BRAF 0.017 EPHA1 0.039 c-MET 0.017 IGF1R 0.039 FGFR4 0.017 NEK4 0.039

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 4: Kinases sorted by Selectivity score Anastassiadis et al.

Selectivity Selectivity Kinase Kinase Score Score PAK1 0.039 MEKK2 0.067 PASK 0.039 PKCb2 0.067 PLK3 0.039 PKG2/PRKG2 0.067 PRKX 0.039 CAMKK1 0.068 ARAF 0.040 CDK1/cyclin A 0.068 PDK1/PDPK1 0.040 MST4 0.068 CK1a1 0.040 PKCepsilon 0.068 STK38/NDR1 0.040 PKCa 0.068 CAMK2b 0.045 IRAK1 0.073 DAPK1 0.045 MEKK3 0.073 DDR2 0.045 PYK2 0.073 EPHA4 0.045 SYK 0.073 EPHB2 0.045 ALK 0.073 PHKg2 0.045 DYRK1/DYRK1A 0.073 RAF1 0.045 ERBB4/HER4 0.073 TLK2 0.050 MARK3 0.073 FAK/PTK2 0.051 NLK 0.073 IR 0.051 CLK1 0.074 PKCtheta 0.051 AXL 0.079 SGK1 0.051 CAMK2a 0.079 SNARK/NUAK2 0.051 CAMK2d 0.079 NEK1 0.051 CDK3/cyclin E 0.079 PKG1a 0.051 c-MER 0.079 TAOK2/TAO1 0.051 ERBB2/HER2 0.079 ZAK/MLTK 0.051 FGFR3 0.079 CK2a 0.051 Haspin 0.079 EPHB4 0.052 MLK2/MAP3K10 0.079 CAMK2g 0.056 MUSK 0.079 CSK 0.056 PKCeta 0.079 ITK 0.056 ROCK2 0.079 ROCK1 0.056 STK16 0.079 TYRO3 SKY 0.056 ULK1 0.079 ULK2 0.056 MSK2/RPS6KA4 0.079 EPHB1 0.056 MYO3b 0.080 FES/FPS 0.056 ULK3 0.080 IRR/INSRR 0.056 CAMKK2 0.084 TAOK1 0.056 CDK9/cyclin T1 0.084 TAOK3/JIK 0.056 DYRK1B 0.084 CDK4/cyclin D1 0.062 JAK2 0.084 CLK3 0.062 LKB1 0.084 DYRK2 0.062 MARK1 0.084 EPHA2 0.062 PKCmu/PRKD1 0.084 FRK/PTK5 0.062 TYK1/LTK 0.084 JAK1 0.062 STK22D/TSSK1 0.085 MNK1 0.062 FER 0.090 MSK1/RPS6KA5 0.062 FGFR1 0.090 PKCb1 0.062 FLT1/VEGFR1 0.090 JNK2 0.062 LIMK1 0.090 P38a/MAPK14 0.062 MARK2/PAR-1Ba 0.090 CK2a2 0.067 MST3/STK24 0.090 CDK4/cyclin D3 0.067 STK33 0.090 FGFR2 0.067 PKD2/PRKD2 0.090 IKKe/IKBKE 0.067 TAK1 0.090 IRAK4 0.067 TBK1 0.090

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 4: Kinases sorted by Selectivity score Anastassiadis et al.

Selectivity Selectivity Kinase Kinase Score Score BMX/ETK 0.096 MST2/STK3 0.153 MLCK/MYLK 0.096 CHK2 0.157 PKCd 0.096 GSK3b 0.157 ROS/ROS1 0.096 TYK2 0.157 SIK2 0.096 CLK4 0.163 CHK1 0.096 FLT4/VEGFR3 0.163 p70S6K/RPS6KB1 0.096 PDGFRb 0.164 CDK6/cyclin D1 0.101 GSK3a 0.169 LYN B 0.101 LCK 0.169 MNK2 0.101 LYN 0.174 DRAK1/STK17A 0.107 MST1/STK4 0.174 PIM1 0.107 RSK4 0.174 SLK/STK2 0.107 Aurora B 0.180 TXK 0.107 EPHA6 0.180 BRSK2 0.112 ACK1 0.191 EGFR 0.112 FGR 0.191 MARK4 0.114 RSK3 0.191 ABL1 0.118 TRKB 0.191 Aurora C 0.118 LOK/STK10 0.198 CDK9 cyclin K 0.118 FMS 0.202 CK1d 0.118 MLK1/MAP3K9 0.202 CK1epsilon 0.118 MLK3/MAP3K11 0.202 JAK3 0.118 c-Kit 0.208 KDR/VEGFR2 0.118 YES/YES1 0.242 PIM3 0.118 ARK5/NUAK1 0.249 PKCnu/PRKD3 0.118 RET 0.253 PKN1/PRK1 0.119 KHS MAP4K5 0.254 BLK 0.124 HGK MAP4K4 0.264 BRK 0.124 TRKC 0.271 CDK1/cyclin B 0.124 FLT3 0.410 PDGFRa 0.124 TRKA 0.124 HCK 0.124 BTK 0.129 RIPK2 0.129 BRSK1 0.130 CDK6/cyclin D3 0.130 CLK2 0.135 MELK 0.135 MLCK2/MYLK2 0.135 ABL2/ARG 0.140 Aurora A 0.140 CDK2/cyclin A 0.140 CDK2/cyclin E 0.140 CDK5/p25 0.140 c-SRC 0.140 MINK/MINK1 0.140 PHKg1 0.140 CDK5/p35 0.146 GCK MAP4K2 0.146 LRRK2 0.146 RSK2 0.146 RSK1 0.147 FYN 0.152

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 5. A ranked table of compounds sorted by Gini score.

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 5: Compounds sorted by Gini score Anastassiadis et al.

CAS # Inhibitor name Gini score CAS # Inhibitor name Gini score

Staurosporine, 62996-74-1 0.20 404828-08-6 GSK-3 Inhibitor XIII 0.57 Streptomyces sp. VEGF Receptor 2 Kinase 97161-97-2 K-252a, Nocardiopsis sp. 0.29 15966-93-5 0.57 Inhibitor I

135897-06-2 SB 218078 0.36 371935-74-9 PI-103 0.57

443798-55-8 Cdk1/2 Inhibitor III 0.37 546102-60-7 Cdk4 Inhibitor 0.57

608512-97-6 PKR Inhibitor 0.44 40254-90-8 Cdk1/5 Inhibitor 0.57

854171-35-0 Indirubin Derivative E804 0.48 496864-16-5 Aloisine A, RP107 0.57

136194-77-9 Gö 6976 0.49 740841-15-0 GSK-3 Inhibitor X 0.57

856436-16-3 JAK3 Inhibitor VI 0.49 507475-17-4 IKK-2 Inhibitor IV 0.58

326914-10-7 SU11652 0.50 227449-73-2 Syk Inhibitor II 0.58

557795-19-4 Sunitinib 0.52 265312-55-8 Cdk4 Inhibitor III 0.58

120685-11-2 Staurosporine, N-benzoyl- 0.52 129-56-6 JNK Inhibitor II 0.58

244148-46-7 Isogranulatimide 0.52 667463-62-9 GSK-3 Inhibitor IX 0.58

146535-11-7 AG 1296 0.53 19545-26-7 Wortmannin 0.59

PDK1/Akt/Flt Dual PDGF Receptor Tyrosine 331253-86-2 0.54 249762-74-1 0.59 Pathway Inhibitor Kinase Inhibitor II

457081-03-7 JAK Inhibitor I 0.54 167869-21-8 PD 98059 0.59

VEGF Receptor 2 Kinase 622387-85-3 Syk Inhibitor 0.54 288144-20-7 0.59 Inhibitor II

5334-30-5 PP3 0.55 114719-57-2 Fascaplysin, Synthetic 0.59

Cdk2 Inhibitor IV, 444723-13-1 0.56 443797-96-4 Aurora Kinase/Cdk Inhibitor 0.59 NU6140 VEGF Receptor 2 Kinase 216661-57-3 0.56 133053-19-7 Gö 6983 0.59 Inhibitor IV

160807-49-8 Indirubin-3 ′-monoxime 0.56 175178-82-2 AG 1478 0.60

AMPK Inhibitor, Casein Kinase II Inhibitor III, 866405-64-3 0.56 934358-00-6 0.60 Compound C TBCA

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 5: Compounds sorted by Gini score Anastassiadis et al.

CAS # Inhibitor name Gini score CAS # Inhibitor name Gini score

VEGF Receptor Tyrosine 269390-69-4 0.60 219138-24-6 p38 MAP Kinase Inhibitor 0.62 Kinase Inhibitor II

171179-06-9 PD 158780 0.60 601514-19-6 GSK3b Inhibitor XII, TWS119 0.62

133052-90-1 Bisindolylmaleimide I 0.60 145915-60-2 PKCbII/EGFR Inhibitor 0.62

71897-07-9 AG 1295 0.60 522629-08-9 MNK1 Inhibitor 0.63

PKR Inhibitor, Negative 852547-30-9 0.60 648449-76-7 PI 3-Kg Inhibitor II 0.63 Control

216163-53-0 PD 174265 0.60 516480-79-8 Chk2 Inhibitor II 0.63

VEGF Receptor 2 Kinase 204005-46-9 0.60 626604-39-5 GSK-3b Inhibitor XI 0.63 Inhibitor III

212779-48-1 Compound 52 0.60 658084-23-2 Met Kinase Inhibitor 0.63

VEGF Receptor Tyrosine 380843-75-4 Bosutinib 0.61 286370-15-8 0.63 Kinase Inhibitor III, KRN633

496864-15-4 Aloisine, RP106 0.61 211555-04-3 JAK3 Inhibitor II 0.63

Cdk4 Inhibitor II, HA 1077, Dihydrochloride 141992-47-4 0.61 103745-39-7 0.63 NSC 625987 Fasudil Cdk1 Inhibitor, 405169-16-6 Dovitinib 0.61 190654-01-4 0.63 CGP74514A JNK Inhibitor, 3895-92-9 Chelerythrine Chloride 0.61 54642-23-8 0.63 Negative Control

220749-41-7 Cdk1 Inhibitor 0.61 581098-48-8 p38 MAP Kinase Inhibitor III 0.63

Alsterpaullone, 2- 852527-97-0 0.61 666837-93-0 SU9516 0.64 Cyanoethyl

146986-50-7 ROCK Inhibitor, Y-27632 0.61 345616-52-6 ERK Inhibitor III 0.64

144978-82-5 AG 112 0.61 778270-11-4 Bcr-abl Inhibitor 0.64

Herbimycin A, Streptomyces 871307-18-5 Tpl2 Kinase Inhibitor 0.61 70563-58-5 0.64 sp. PDGF Receptor Tyrosine 205254-94-0 0.62 34823-86-4 GTP-14564 0.64 Kinase Inhibitor III

896138-40-2 Flt-3 Inhibitor II 0.62 189232-42-6 Bohemine 0.64

220792-57-4 Aminopurvalanol A 0.62 648450-29-7 PI 3-Kg Inhibitor 0.64

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 5: Compounds sorted by Gini score Anastassiadis et al.

CAS # Inhibitor name Gini score CAS # Inhibitor name Gini score

2826-26-8 AG 9 0.64 154447-36-6 LY 294002 0.68

58753-54-1 JAK3 Inhibitor IV 0.65 171745-13-4 Compound 56 0.68

237430-03-4 Alsterpaullone 0.65 681281-88-9 Akt Inhibitor IV 0.68

Casein Kinase I Inhibitor, 199986-75-9 Cdk2 Inhibitor III 0.65 301836-43-1 0.68 D4476

330161-87-0 SU6656 0.65 639089-54-6 Tozasertib 0.68

852045-46-6 Flt-3 Inhibitor III 0.65 139298-40-1 KN-93 0.69

127243-85-0 H-89, Dihydrochloride 0.65 396129-53-6 TGF-b RI Kinase Inhibitor 0.69

119139-23-0 Bisindolylmaleimide IV 0.65 327036-89-5 GSK-3b Inhibitor I 0.69

PDGF Receptor Tyrosine 627518-40-5 0.66 212844-53-6 Purvalanol A 0.69 Kinase Inhibitor IV

152121-53-4 PD 169316 0.66 257879-35-9 PKCb Inhibitor 0.69

Cdc2-Like Kinase Inhibitor, 1485-00-3 Syk Inhibitor III 0.66 300801-52-9 0.69 TG003

186611-52-9 IC261 0.66 154447-38-8 LY 303511- Negative control 0.69

133550-30-8 AG 490 0.66 184475-35-2 Gefitinib 0.69

879127-16-9 Aurora Kinase Inhibitor III 0.66 19542-67-7 BAY 11-7082 0.69

221244-14-0 PP1 Analog II, 1NM-PP1 0.67 366017-09-6 Mubritinib 0.70

196868-63-0 IGF-1R Inhibitor II 0.67 509093-47-4 IRAK-1/4 Inhibitor 0.70

356559-13-2 TGF-b RI Inhibitor III 0.67 186692-46-6 Roscovitine 0.70

174709-30-9 BPIQ-I 0.67 865362-74-9 ERK Inhibitor II, FR180204 0.70

347155-76-4 PDGF RTK Inhibitor 0.67 127191-97-3 KN-62 0.70

301305-73-7 Flt-3 Inhibitor 0.67 183319-69-9 Erlotinib 0.70

345987-15-7 JNK Inhibitor V 0.67 142273-20-9 Kenpaullone 0.70

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 5: Compounds sorted by Gini score Anastassiadis et al.

CAS # Inhibitor name Gini score CAS # Inhibitor name Gini score

487021-52-3 GSK-3b Inhibitor VIII 0.70 913844-45-8 Rho Kinase Inhibitor IV 0.74

444731-52-6 Pazopanib 0.71 152121-47-6 SB 203580 0.74

Rho Kinase Inhibitor III, 7272-84-6 0.71 305350-87-2 MEK1/2 Inhibitor 0.74 Rockout

318480-82-9 SC-68376 0.71 623163-52-0 MEK Inhibitor II 0.74

41179-33-3 MK2a Inhibitor 0.71 587871-26-9 ATM Kinase Inhibitor 0.74

Diacylglycerol Kinase 65678-07-1 AG 1024 0.71 120166-69-0 0.74 Inhibitor II

478482-75-6 GSK-3b Inhibitor II 0.71 302962-49-8 Dasatinib 0.74

52029-86-4 STO-609 0.72 226717-28-8 AGL 2043 0.75

SB 202474, Negative control 53123-88-9 Rapamycin 0.72 172747-50-1 0.75 for p38 MAPK

5812-07-7 DMBI 0.72 905973-89-9 ATM/ATR Kinase Inhibitor 0.75

784211-09-2 Cdk/Crk Inhibitor 0.72 443913-73-3 Vandetanib 0.75

cFMS Receptor Tyrosine 404009-46-7 DNA-PK Inhibitor V 0.72 870483-87-7 0.75 Kinase Inhibitor

745833-23-2 VX-702 0.73 35943-35-2 Akt Inhibitor V, Triciribine 0.75

894804-07-0 JNK Inhibitor VIII 0.73 312917-14-9 JNK Inhibitor IX 0.75

545380-34-5 NF-kB Activation Inhibitor 0.73 213743-31-8 Lck Inhibitor 0.75

212141-51-0 Vatalanib 0.73 72873-74-6 SKF-86002 0.75

Sphingosine Kinase Akt Inhibitor VIII, Isozyme- 312636-16-1 0.73 612847-09-3 0.75 Inhibitor Selective, Akti-1/2

404009-40-1 DNA-PK Inhibitor III 0.73 151342-35-7 Ro-32-0432 0.76

477600-75-2 Tofacitinib 0.73 152121-30-7 SB 202190 0.76

154447-35-5 DNA-PK Inhibitor II 0.73 220127-57-1 Imatinib 0.77

297744-42-4 MEK Inhibitor I 0.74 925681-41-0 Akt Inhibitor X 0.77

Nature Biotechnology: doi:10.1038/nbt.2017 Supplementary Table 5: Compounds sorted by Gini score Anastassiadis et al.

CAS # Inhibitor name Gini score

165806-53-1 SB220025 0.77

231277-92-2 Lapatinib 0.77

641571-10-0 Nilotinib 0.78

879127-07-8 EGFR Inhibitor 0.78

179248-59-0 Src Kinase Inhibitor I 0.78

387867-13-2 Tandutinib 0.78

284461-73-0 Sorafenib 0.79

179248-61-4 EGFR/ErbB-2 Inhibitor 0.79

EGFR/ErbB-2/ErbB-4 881001-19-0 0.79 Inhibitor

790299-79-5 Masitinib 0.81

Biotechnology: doi:10.1038/nbt.2017