DOI: 10.1002/cbic.201300173

Facile Fluorescence-Based Detection of PAD4-MediatedCitrullinationErin Wildeman and Marcos M. Pires*[a]

Introduction

Nucleosomes, the basic unit of the chromatin, are composedof both DNA and proteins. For each chromatin, 147 base pairsof the genomic DNA are wrapped around eight histone pro-teins.[1] Two copies of each of the histone proteins (H2A, H2B,H3, and H4) form the structural scaffold that associates withthe genomic DNA. It is increasingly recognized that the pro-truding N-terminal tails of the histone proteins can be heavilypost-translationally modified. The number of post-translationalmodifications and the diversity of the covalent tags (acetyla-tion, methylation, phosphorylation, ubiquitylation, etc.) ob-served within these short protein segments suggest that thisprocess is highly dynamic and potentially vital to the properregulation of gene expression.[2, 3] Indeed, the breadth of possi-ble tags on histone tails has led to the idea that large combi-natorial arrangements of modifications could potentially gener-ate distinct cellular consequences. This compilation of possiblemodifications, which can be inheritable and can alter gene ex-pression profiles, has been referred to as the “histone code”.[4, 5]

Studies related to histone modifications fall under the class ofepigenetics. Due to a recent surge of interest, we are now be-ginning to understand how various tags covalently added tohistones can mediate alterations in gene expression. In thepast decade, it has become evident that epigenetic-based cel-lular signatures can have a direct influence on both cellular ho-meostasis and disease states.[6] A large group of enzymes cova-lently tags molecules to histones; another group subsequentlyremoves them. As a whole, we have an incomplete under-standing of how these enzymes function and how they coordi-nate with their binding partners.

The protein–arginine deiminase enzyme family is a group ofpost-translational modification enzymes that convert peptidylarginine to peptidyl citrulline (Cit) (Figure 1 A).[7–9] The isoformsof this family of enzymes (PAD1, PAD2, PAD3, PAD4, and PAD6)have been detected in numerous tissues. PAD1 is known to beexpressed primarily in the epidermis and the uterus, whereasPAD3 expression has been detected in keratinocytes and hairfollicles. Embryos/oocytes have elevated expression levels ofPAD6, whereas PAD2 has been detected in several cancertypes and in the brain, secretory glands, and skeletal muscles.Recent studies have revealed information about the functionsand localizations of each individual isozyme, but a more com-

The post-translational modifications of histone proteins arehighly diverse and dynamic processes. It is becoming increas-ingly evident that modifying histone proteins can have a directinfluence on both cellular homeostasis and disease states. Pro-tein arginine deiminase 4 (PAD4) is an enzyme that convertspeptidyl-arginine to citrulline. The overexpression of PAD4 hasbeen found in numerous types of human cancer and auto-immune diseases. We report a new, facile, fluorescence-based

assay for the detection of PAD4 activity that exploits the sub-strate specificity of trypsin to monitor the citrullination reac-tion carried out by PAD4 based on the fact that, upon citrulli-nation, the positively charged arginine side chain is convertedto the neutral citrulline. We show that the assay can be per-formed rapidly with readily available reagents and that itresponds accordingly to a known PAD4 inhibitor.

Figure 1. A) A schematic representation of the octameric histones found ina basic unit of a nucleosome. PAD4 is able to citrullinate a number of argi-nine residues on histone proteins. B) Z-Arg-AMC was overlaid with the natu-ral substrate (hidden for clarity) of PAD4. The figure was generated in Pymolby using the crystal structure of PAD4 (PDB ID: 2DEY).

[a] E. Wildeman, Prof. M. M. PiresDepartment of Chemistry, Lehigh University6 E. Packer Ave., Bethlehem, PA 18015 (USA)E-mail : map311@lehigh.edu

Supporting information for this article is available on the WWW underhttp ://dx.doi.org/10.1002/cbic.201300173.

� 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2013, 14, 963 – 967 963

CHEMBIOCHEMFULL PAPERS

plete picture of the interplay of these proteins in response tocellular cues could give us greater insight into their role incells.[10] The last isomer of this family, PAD4, is the best studiedof the isomers and has been found to localize within the nu-cleus.[11] Consistent with its localization, PAD4 has been shownto deiminate a number of nuclear targets, including arginineresidues on the N-terminal tails of histones H3 and H4.

There is still a great deal to discover about PAD4, includingthe full range of its cellular substrates, its activation pattern, itscoregulators, and its precise links to various human diseases.What we currently know is that the PAD4 isoform is highlyoverexpressed in a number of malignant tumors, such as lung,breast, and bone cancers,[12, 13] but its expression in severalnontumor tissues was undetectable.[12] Strikingly, the findingthat PAD4 expression levels decreased following surgical re-moval of the tumor gives strong evidence for the close associ-ation between PAD4 and tumorigenesis.[12] Additionally, it hasbeen demonstrated that PAD4 is a co-repressor for p53, amajor-tumor suppressor protein that is highly mutated in mostcancers.[14] The knock-down of PAD4 in cancer cell lines bysiRNA alone is sufficient to induce cell-death.[15] Consistentwith these findings, a recent report showed that an irreversiblePAD4 inhibitor led to a 70 % shrinkage of tumors in model or-ganisms.[16] Combined, these studies indicate that the discov-ery of PAD4 inhibitors with clinical efficacy could spearhead anentirely new class of targeted anticancer therapies. It has previ-ously been shown that aberrant activity of PAD4 is also associ-ated with several autoimmune diseases including rheumatoidarthritis, Crohn’s disease and multiple sclerosis.[17, 18] In fact,rheumatoid arthritis patients display elevated levels of anti-cit-rullinated peptide antibodies, thus making these antibodiesa reliable and specific biomarker for this disease. The link be-tween PAD4 and disease states will rely greatly on our funda-mental understanding of PAD4 itself. Unfortunately, few potentPAD4 inhibitors have been discovered. With better small-mole-cule-inhibitor probes that possess increased potency and spe-

cificity, we might gain greater insight into the function ofPAD4 and the roles of other proteins in controlling its activity.

Assays that readily and reliably measure the activity of PAD4would also enhance our understanding of PADs and would aidin the search for potent and specific inhibitors. Previous PAD4assays have relied on a calorimetric readout of ammonia re-lease,[19] an antibody-based detection of citrulline,[20] an affinityprobe with fluorescence polarization,[21] a chemoselective reac-tion between glyoxal and citrulline under acidic conditions,[22]

and (recently reported) noncovalent fluorescence dequench-ing.[23] Herein, we describe a new mechanism-based fluores-cence assay that quickly and reliably measures the activity ofPAD4. The assay has several advantages over existing technolo-gies, including a strong signal-to-noise ratio, speed of analysis,and robustness of measurement. Most importantly, it is techni-cally facile and can be performed with readily available re-agents. In addition, our assay format is compatible with high-throughput screening platforms.

Results and Discussion

Substrate-analogue design

PAD4 (along with the isozymes PAD1, PAD2, PAD3, and PAD6)is responsible for the catalytic conversion of an arginine resi-due to the non-tRNA-encoded residue citrulline. When actingupon substrate histones 3 and 4, PAD4 converts a positivelycharged to a neutral side chain through the citrullination reac-tion, which is currently thought to be irreversible (Scheme 1).The neutralization of this positive charge has important impli-cations for the association of histones with the negativelycharged genomic DNA, potentially modulating transcriptionalregulation and promoting the formation of neutrophil extracel-lular traps (NETs).[24] We envisioned that the citrullination ofarginine would also interfere with the ability of trypsin to hy-drolyze the carboxyl side of the amide bond. Therefore, PAD4

Scheme 1. Schematic representation of the PAD4 assay. The citrullination reaction catalyzed by PAD4 renders the substrate (Z-Arg-AMC) resistant to trypsin-mediated amide hydrolysis, thus leading to a change in fluorescence levels.

� 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2013, 14, 963 – 967 964

CHEMBIOCHEMFULL PAPERS www.chembiochem.org

citrullination could be linked to amide hydrolysis. The incorpo-ration of a masked fluorophore (7-amino-4-methylcoumarin,AMC) on the carboxyl side of arginine could be used to gener-ate a signal upon hydrolytic release by trypsin. Acylation ofAMC is known to greatly reduce its fluorescence intensity, a fea-ture that has been utilized in a number of protease assays.[25]

Likewise, acylation of AMC onto the arginine is expected togreatly reduce the fluorescence of the potential PAD4 sub-strate Z-Arg-AMC (Scheme 1). In addition, we included a car-boxybenzyl (Cbz, Z) group on the amino side of the arginineto mimic the typical protein-based substrate of PAD4.

Although the modification of arginine with the unnaturalmoieties AMC and Cbz has the potential to disrupt the recog-nition of Z-Arg-AMC by the active site of PAD4, it is evidentfrom the crystal structure available for PAD4 bound to its natu-ral substrate (Figure 1 B) that the incorporation of both Cbzand AMC would be well tolerated. The active site of PAD4 isnot imbedded in the core of the protein, and the residuesflanking the central arginine are also solvent exposed, a factthat gives greater flexibility in substrate binding; however, thisfeature can also hinder the discovery of potent reversibleactive-site inhibitors. We expected that if Z-Arg-AMC werea substrate for PAD4, the guanidinium group would be con-verted efficiently to urea upon co-incubation (Scheme 1,bottom). Furthermore, we hypothesized that the product, Z-Cit-AMC, would not be hydrolyzed by subsequent incubationwith excess trypsin. If the protein were inhibited, then the Z-Arg-AMC would remain intact, and fluorescent AMC would bereleased in the trypsin hydrolysis step.

Initial assay evaluation

Initially, we investigated the fluorescence propertiesof the acylated ACM in the context of its trypsin hy-drolysis. Upon incubation of Z-Arg-AMC with trypsin,we observed an approximately 6.5-fold increase influorescence signal (Figure 2 A). This expected in-crease in fluorescence was consistent with the factthat arginine is the preferred substrate of trypsin andthat AMC does not interfere with substrate recogni-tion. On the other hand, when we incubated Z-Cit-AMC with trypsin under similar conditions, we ob-served a slight decrease in the fluorescence intensity(Figure 2 B). These critical results confirmed that neu-tralizing Arg to Cit causes a dramatic change in sus-ceptibility to trypsin hydrolysis, and that differencecan be linked to the unmasking of the AMC fluoro-phore. Next, we determined the ability of PAD4 torecognize and modify the unnatural substrate Z-Arg-AMC. We utilized RP-HPLC to track the potentialchemical modification of Z-Arg-AMC by PAD4through changes in polarity. We observed that thepeak corresponding to Z-Arg-AMC disappears afterincubation with PAD4 (Figure 2 C) and is replaced bya peak with a retention time that is consistent witha less-polar species, as expected for the conversionof arginine to citrulline. Authentic Z-Cit-AMC was

shown to have the same retention time as this product (Fig-ure 2 D). Together, these initial results confirmed that a signifi-cant fluorescence increase accompanies the trypsin digestionof Z-Arg-AMC, whereas no fluorescence increase is seen for thecitrulline form. Moreover, HPLC data are consistent with Z-Arg-AMC being a substrate for PAD4, thus highlighting the tolera-bility of our N- and C-terminal modifications to arginine.

Next, we set out to determine the suitability of our assay forreporting on the enzymatic activity of PAD4. In particular, wewanted to perform the assay in 96-well plates to demonstrateits utility for high-throughput screening. Similar to what weobserved in Figure 2, the fluorescence level of Z-Arg-AMCalone was relatively low (Figure 3), but the introduction oftrypsin yields a fluorescence level that is about 6.5 timeshigher than with the addition of buffer only. Next, we incubat-ed Z-Arg-AMC with PAD4 and observed low fluorescence levelsin the absence of trypsin. The introduction of trypsin to this so-lution resulted in almost no statistically significant increase influorescence signal. Clearly, Z-Arg-AMC has the potential to bea fast and facile reporter of PAD4 activity. The assay proved tobe extremely robust, displaying excellent interassay consisten-cy and strong signal differences between the presence and ab-sence of PAD4. Additionally, we verified that no difference wasobserved between the fluorescence signals from Z-Cit-AMCbefore and after the introduction of trypsin. To further explorethe possibility of using our assay for high-throughput screen-ing, we performed similar activity assays on PAD4 in a 384-wellplate. We found that the fluorescence signals were nearly iden-tical to those of the 96-well plates, and thus demonstratedthat miniaturization of the assay should be feasible (see Fig-ure S1 in the Supporting Information). Overall, we have shown

Figure 2. Top: After excitation at 340 nm, fluorescence emission was monitored from theincubation of A) Z-Arg-AMC or B) Z-Cit-AMC with trypsin (c) or buffer (a). Condi-tions: 50 mm Z-Arg-AMC, 10 mm PAD4. Bottom: RP-HPLC analysis (monitored at 322 nm)of C) Z-Arg-AMC in the presence (c) and absence (c) of PAD4 and D) Z-Arg-AMC inthe presence of PAD4 (c) and of Z-Cit-AMC (c). Conditions: 25 mm Z-Arg-AMC, 5 mm

PAD4.

� 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2013, 14, 963 – 967 965

CHEMBIOCHEMFULL PAPERS www.chembiochem.org

that our trypsin-linked assay can be performed within a concisetimeframe in small volumes, it uses readily obtainable re-agents, and it can be performed at reasonable temperatures,thus creating a distinct advantage over the classical COLDERPAD4 assay.[19]

Further assay characterization and control experiments

Having established that our assay platform is well suited toreport on the activity of PAD4, we performed additional experi-ments to further characterize our system. A time-course experi-ment was performed to confirm that the reaction of PAD4 onthe substrate was following kinetics similar to those previouslyreported in vitro for PAD4.[26, 27] After approximately 30 min ofincubation of PAD4 in the reaction buffer, we observed almostcomplete conversion of Z-Arg-AMC (Figure S2). It should benoted that, as expected, the reaction kinetics for PAD4 wereshown to be substrate dependent.[28] Nonetheless, the efficien-cy of PAD4 in utilizing Z-Arg-AMC as a substrate under ourassay conditions is sufficiently high for completion to bereached within a reasonable time.

Numerous reports have identified calcium as the metal ionof choice for the proper function of PAD4.[29] Upon binding fivecalcium ions, the protein undergoes a conformational changethat forms the active site. We also observed a large increase influorescence when there was no exogenous calcium in thereaction solution; this is consistent with the reduced catalyticefficiency of PAD4 (Figure S3).

Finally, we showed that our PAD4 assay can respond accord-ingly to the presence of a known PAD4 inhibitor. A mecha-nism-based PAD4 inhibitor, Cl-amidine, has been shown toinhibit the citrullination reaction.[30] Specifically, Cl-amidine co-valently modifies an essential cysteine residue (C645) in theactive site of PAD4, thus rendering the enzyme inactive. Weshowed that titrating increasing concentrations of Cl-amidineled to a concentration-dependent increase in fluorescence

levels (Figure 4 A). These data are consistent with the inhibitionof PAD4, which appears to be nearly fully inhibited by 50 mm

Cl-amidine.[30] Only in our assay does the presence of a PAD4inhibitor lead to an increase in fluorescence signal. The in-crease in fluorescence for successful inhibitors is less prone tofalse positives resulting from small molecules quenching thefluorophore, as in assays in which inhibitors result in a decreasein fluorescence signal.

We envision that our assay format will be useful for screen-ing large libraries of small molecules for the next generation ofPAD4 inhibitors. However, a potential point of concern for theuse of trypsin-mediated signal generation in the screening ofthousands/millions of compounds is the inadvertent inhibitionof the trypsin itself. However, we believe that any possibleinhibition of trypsin could be circumvented by simply usinghigher concentrations of trypsin for the final step of the assaywith little to no loss of signal-to-noise. Therefore, we do notthink that small molecules that inhibit trypsin would disruptour assay. When a similar assay was performed with either 100or 200 mm of phenylmethanesulfonylfluoride (PMSF, a potentcovalent inhibitor of trypsin) included in the reaction mixture,we saw virtually identical fluorescence levels to those from re-actions that lacked PMSF (Figure 4 B) Likewise, when PMSF was

Figure 3. Fluorescence assay for the activity of PAD4. Fluorescence was mea-sured (lex = 345 nm, lem = 465 nm) in a 96-well format under the designatedconditions. The addition of trypsin causes a large increase in fluorescencelevels when the final molecule is arginine and not citrulline. Conditions:25 mm Z-Arg-AMC, 5 mm PAD4.

Figure 4. Fluorescence assay for the activity of PAD4. Fluorescence was mea-sured (lex = 345 nm, lem = 465 nm) in a 96-well format under the designatedconditions. A) Upon the introduction of increasing concentrations of theknown PAD4 inhibitor Cl-amidine, a fluorescence increase was observed.B) Addition of the potent trypsin inhibitor PMSF did not interfere with thefluorescence signal in the assay. Conditions: 25 mm Z-Arg-AMC, 5 mm PAD4.

� 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2013, 14, 963 – 967 966

CHEMBIOCHEMFULL PAPERS www.chembiochem.org

added to a similar reaction in the absence of PAD4, we saw nosignificant change in the fluorescence levels (see the Support-ing Information). These data demonstrate that trypsin inhibi-tors would be overwhelmed by the concentration of trypsinadded, and, therefore, should not result in false-negative hits.Were the unintended inhibition of the trypsin by a potentialPAD4 inhibitor in the final step to lead to false negatives, theconcentration of trypsin could be increased without interfer-ence to the signal (Figure S4). Together, these data clearlyshow that our novel fluorescence-based PAD4 assay respondsto the presence of a known inhibitor and that known trypsininhibitors do not perturb fluorescence signals.

Conclusions

In conclusion, we have reported a new, rapid, and facile fluo-rescence-based assay of PAD4 that relies on the redesign of anarginine substrate. In particular, we incorporated a moietywhose fluorescence is controlled by the reaction of the sub-strate with PAD4. The citrullination reaction catalyzed by PAD4leads to a change in the level of blue fluorescence and yieldsa reliable read-out of PAD4 activity. We plan to apply our novelassay format to screening large small-molecule libraries forpotent and specific PAD4 inhibitors. Additionally, we shall useour trypsin-linked assay on a histone-mimetic peptide in anattempt to develop isozyme-specific probes. Finally, we plan todevelop a live-cell assay for PAD4 by implementing similar ele-ments to the assay described here.

Experimental Section

PAD4 expression and purification: The PAD4–GST fusion contain-ing pGEX-6P plasmid was transformed into competent Escherichiacoli BL21(DE3) cells. The cells were selected on an ampicillin-con-taining agar-LB plate overnight. A single colony was picked andgrown overnight in lysogeny broth (LB, 20 mL) with ampicillin at37 8C. The next morning, LB (6 L) with ampicillin was inoculatedwith the overnight growth medium. Cells were allow to continueto grow at 37 8C until the OD600 reached ~0.3. The temperaturewas decreased to 16 8C. Once the OD600 reached ~0.6, protein pro-duction was induced with isopropyl-b-d-thiogalactopyranoside(0.3 mm). Cells were incubated overnight with shaking at 16 8C.The next morning, cells were harvest by centrifugation for 30 minat 3333 g and 4 8C. The remaining cell pellet was resuspended in1 � phosphate-buffered saline (PBS, 20 mL with 1 mm PMSF and1 mm dithiothreitol) per liter of cell growth. Following a 15 minsonication step, the lysate was spun at 48 000 g at 4 8C for 30 min.The supernatant was applied to a Protino’s glutathione agarose 4Bcolumn. After being washed with 1 � PBS (3 � 10 mL per mL ofbead volume), protein was eluted with elution buffer (1 mL, 50 mm

Tris, 10 mm glutathione, pH 8.0) per mL of bead volume. Theeluted solution was further purified and concentrated by using a100 MWCO spin column. A sample was analyzed by SDS-PAGE toverify the expected molecular weight of the protein and its purify.Aliquots were made and stored at �80 8C.

Acknowledgements

This work was supported by the Lehigh University Start-up Pack-age. We thank Dr. Walter Fast for providing the GST-PAD4 plas-mid for protein expression.

Keywords: arginine · citrullination · fluorescence assays · post-translational modifications · trypsin

[1] K. Luger, A. W. Mader, R. K. Richmond, D. F. Sargent, T. J. Richmond,Nature 1997, 389, 251 – 260.

[2] S. L. Berger, Nature 2007, 447, 407 – 412.[3] T. Kouzarides, Cell 2007, 128, 693 – 705.[4] B. D. Strahl, C. D. Allis, Nature 2000, 403, 41 – 45.[5] T. Jenuwein, C. D. Allis, Science 2001, 293, 1074 – 1080.[6] G. Egger, G. Liang, A. Aparicio, P. A. Jones, Nature 2004, 429, 457 – 463.[7] T. Hagiwara, K. Nakashima, H. Hirano, T. Senshu, M. Yamada, Biochem.

Biophys. Res. Commun. 2002, 290, 979 – 983.[8] E. R. Vossenaar, A. J. Zendman, W. J. van Venrooij, G. J. Pruijn, Bioessays

2003, 25, 1106 – 1118.[9] K. Nakashima, T. Hagiwara, M. Yamada, J. Biol. Chem. 2002, 277, 49562 –

49568.[10] K. L. Bicker, P. R. Thompson, Biopolymers 2013, 99, 155 – 163.[11] J. E. Jones, C. P. Causey, B. Knuckley, J. L. Slack-Noyes, P. R. Thompson,

Curr. Opin. Drug Discovery Dev. 2009, 12, 616 – 627.[12] X. Chang, J. Han, L. Pang, Y. Zhao, Y. Yang, Z. Shen, BMC Cancer 2009, 9,

40.[13] S. Mohanan, B. D. Cherrington, S. Horibata, J. L. McElwee, P. R. Thomp-

son, S. A. Coonrod, Biochem. Res. Int. 2012, 2012, 895343.[14] Q. Guo, W. Fast, J. Biol. Chem. 2011, 286, 17069 – 17078.[15] P. Li, H. Yao, Z. Zhang, M. Li, Y. Luo, P. R. Thompson, D. S. Gilmour, Y.

Wang, Mol. Cell. Biol. 2008, 28, 4745 – 4758.[16] Y. Wang, P. Li, S. Wang, J. Hu, X. A. Chen, J. Wu, M. Fisher, K. Oshaben, N.

Zhao, Y. Gu, D. Wang, G. Chen, J. Biol. Chem. 2012, 287, 25941 – 25953.[17] N. K. Acharya, E. P. Nagele, M. Han, N. J. Coretti, C. DeMarshall, M. C.

Kosciuk, P. A. Boulos, R. G. Nagele, J. Autoimmun. 2012, 38, 369 – 380.[18] I. Auger, C. Charpin, N. Balandraud, M. Martin, J. Roudier, Autoimmun.

Rev. 2012, 11, 801 – 803.[19] M. Knipp, M. Vasak, Anal. Biochem. 2000, 286, 257 – 264.[20] E. A. Moelants, J. Van Damme, P. Proost, PLoS One 2011, 6, e28976.[21] B. Knuckley, J. E. Jones, D. A. Bachovchin, J. Slack, C. P. Causey, S. J.

Brown, H. Rosen, B. F. Cravatt, P. R. Thompson, Chem. Commun. 2010,46, 7175 – 7177.

[22] K. L. Bicker, V. Subramanian, A. A. Chumanevich, L. J. Hofseth, P. R.Thompson, J. Am. Chem. Soc. 2012, 134, 17015 – 17018.

[23] Q. Wang, M. A. Priestman, D. S. Lawrence, Angew. Chem. 2013, 125,2379 – 2381; Angew. Chem. Int. Ed. 2013, 52, 2324 – 2325.

[24] P. Li, M. Li, M. R. Lindeberg, M. J. Kennett, N. Xiong, Y. Wang, J. Exp. Med.2010, 207, 1853 – 1862.

[25] M. Zimmerman, B. Ashe, E. C. Yurewicz, G. Patel, Anal. Biochem. 1977,78, 47 – 51.

[26] E. M. Stone, T. H. Schaller, H. Bianchi, M. D. Person, W. Fast, Biochemistry2005, 44, 13744 – 13752.

[27] P. L. Kearney, M. Bhatia, N. G. Jones, L. Yuan, M. C. Glascock, K. L. Catch-ings, M. Yamada, P. R. Thompson, Biochemistry 2005, 44, 10570 – 10582.

[28] B. Knuckley, C. P. Causey, J. E. Jones, M. Bhatia, C. J. Dreyton, T. C. Os-borne, H. Takahara, P. R. Thompson, Biochemistry 2010, 49, 4852 – 4863.

[29] K. Arita, H. Hashimoto, T. Shimizu, K. Nakashima, M. Yamada, M. Sato,Nat. Struct. Mol. Biol. 2004, 11, 777 – 783.

[30] Y. Luo, K. Arita, M. Bhatia, B. Knuckley, Y. H. Lee, M. R. Stallcup, M. Sato,P. R. Thompson, Biochemistry 2006, 45, 11727 – 11736.

Received: March 22, 2013Published online on May 2, 2013

� 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2013, 14, 963 – 967 967

CHEMBIOCHEMFULL PAPERS www.chembiochem.org