submitted papers

Label-Free Screening of Drug–Protein Interactions by Time-Resolved Fourier Transform Infrared Spectroscopic AssaysExemplified by Ras Interactions

CARSTEN KOTTING,* YAN SUVEYZDIS, RAVI S. BOJJA, NILS METZLER-NOLTE, andKLAUS GERWERT*Lehrstuhl fur Biophysik, Ruhr-Universitat Bochum, D-44780 Bochum, Germany (C.K., Y.S., R.S.B., K.G.);

and Lehrstuhl fur Anorganische Chemie I, Ruhr-Universitat Bochum, D-44780 Bochum, Germany (N.M.-N.)

Time-resolved Fourier transform infrared (FT-IR) spectroscopy can

reveal molecular details of protein interactions. Analysis of difference

spectra selects the absorptions of respective protein groups involved in an

interaction against the background of the whole sample. By comparison of

the same difference spectrum with and without a small molecule, one can

determine whether the small molecule interferes with the protein or not.

Usually a marker band of a specific residue of the protein is monitored.

Here, we show three different time-resolved FT-IR assays detecting

interactions of potential small molecules for molecular therapy with the

GTPase Ras as an example for small GTPase binding proteins. Ras

regulates signal transduction processes through a switching mechanism,

cycling between an active ‘‘on’’ GTP-bound form and an inactive ‘‘off’’

GDP-bound state. Molecular defects in Ras can impair the ability of Ras

and the Ras–RasGAP complex to hydrolyze GTP, contributing to

uncontrolled cell growth and cancer. Oncogenic mutated Ras is found in

about 30% of all cancer cells. We show in vitro assays, indicating (I) the

shift of Ras into its ‘‘off’’ conformation, which inhibits the Ras pathway;

(II) down-regulation of Ras signaling by changes in the Ras–Raf effector

interaction; and (III) down-regulation of Ras signaling pathway by

catalyzing GTP hydrolysis. Since almost all molecules have characteristic

marker bands in the infrared, time-resolved FT-IR spectroscopy can be

used label-free. No artificial nucleotides that could influence the

interaction are needed. Both, sample preparation and evaluation can be

automated in order to allow for high-throughput screening.

Index Headings: Time-resolved Fourier transform infrared spectroscopy;

FT-IR spectroscopy; Label-free; Screening; GTPase; Protein interactions;

Metal macrocycles; Anticancer activity; High throughput.

INTRODUCTION

The screening for the interaction of small molecules withproteins for molecular therapy is usually carried out by indirectmethods by means of artificial labels or reporter systems.Often, the protein and/or the small molecule are labeled, e.g.,by a fluorescent molecule. In these cases one can never rule outan influence of the usually bulky label on the result. On theother hand, time-resolved Fourier transform infrared (FT-IR)spectroscopy can monitor interactions of small molecules withproteins label-free.1 Further, most reporter systems provide

only very limited information on the interaction, whereas time-resolved FT-IR can often specify the type of interaction.Absorption bands can be assigned by isotopic labeling or site-directed mutagenesis and thus the interaction can be assignedto specific groups.2 This information can be of aid indeveloping drugs that interact specifically with a protein in amolecular therapy.

GTPases are molecular switches that maintain cellulargrowth control by cycling between the active GTP-boundstate (‘‘on’’ state) and the inactive GDP-bound state (‘‘off’’state).3 In the activated GTP-bound form, Ras interacts withmultiple effector proteins involved in cell proliferation, cellmigration, and other important processes.4 Numerous studiesof Ras interaction with regulatory proteins such as GTPaseactivating proteins (GAP)5 and guanine nucleotide exchangefactors (GEF)5 or downstream effectors6 were conducted withdifferent biophysical methods, including X-ray,7,8 nuclearmagnetic resonance (NMR),9 and time-resolved FT-IR10–12 inorder to understand the molecular mechanism, which is of highimportance for medicinal chemistry because Ras is the mostfrequently mutated oncogene in both solid tumors andhaematological neoplasias.13 Remarkable progress in under-standing the mechanisms of Ras regulation and signaltransduction has provided targets for rational drug designapproaches to block or possibly revert the aberrant function ofRas signaling in human cancer. As a result, a large number ofdifferent inhibitors have been developed to target oncogenicRas itself or the components upstream and downstream of Rasin tumors, i.e., (1) interfering RNAs, (2) membrane associationof Ras, (3) Ras activation by GEFs, (4) Ras–effectorinteraction, (5) GTPase reaction of Ras, and (6) multiplesignaling pathways.14–18 Phosphate chelating agents such aszinc cyclen can down-regulate signaling by shifting theequilibrium of Ras–GppNHp towards the ‘‘off’’ conforma-tion.19 Further, molecular modeling had predicted polyaminesas potential candidates to interfere with the effector bindingdomain of Ras.17

In the present study, we show that changes in time-resolvedFT-IR spectra can be used as assays in screening for smallmolecules that possibly prevent movement to the ‘‘on’’ state ofRas and/or accelerate GTP hydrolysis. Both would down-

Received 24 March 2010; accepted 10 June 2010.* Authors to whom correspondence should be sent. E-mail: carsten.koetting@rub.de, gerwert@bph.rub.de.

Volume 64, Number 9, 2010 APPLIED SPECTROSCOPY 9670003-7028/10/6409-0967$2.00/0

� 2010 Society for Applied Spectroscopy

regulate the Ras signaling pathway. Thus, time-resolved FT-IRspectroscopy is an effective tool for studying interactionsbetween small molecules and proteins with clear application fordrug design and screening. Besides the description of the assay,we give examples (Scheme 1) of molecules that could down-regulate Ras. The results from the present study show theimportance of label-free screening with the native nucleotide,elucidating the normal behavior of GTPases in the presence ofthe potential interacting therapeutic molecules.

EXPERIMENTAL

Protein Expression and Chemical Synthesis. Wild-typeand mutants of H-Ras,20 NF1,21 and RapGAP22 were preparedwith an Escherichia coli expression system. Raf RBD was akind gift from Prof. Dr. Christian Herrmann (Ruhr-UniversitatBochum).23 The P3-1-(2-nitrophenyl)ethyl (NPE) ester of GTPwas synthesized in two steps from 2-nitrophenylacetylhydra-zone and GTP with subsequent purification via anion-exchangechromatography according to a procedure from Walker andTrentham.24 The P3-para-hydroxyphenacyl (pHP) ester ofGTP was synthesized by coupling GDP and pHP-caged-Pi. Thelatter was obtained in five steps from para-hydroxyacetophe-none and dibenzylphosphate according to a procedure fromPark and Givens.25

Zinc cyclen complex, RuC, and NF1 peptides of thesequences GGGTLFR789ATTLSR, GGTLFR789ATTLSR,and GTLFR789ATTLSR with catalytic arginine finger 789were synthesized. The latter were coupled to RuC as describedpreviously.26–29 All the compounds in this study were

minimum 95% pure (as determined by high-performanceliquid chromatography (HPLC)), with structure confirmed bymass spectrometry.

Fourier Transform Infrared Measurements. For the FT-IR measurements Ras protein was loaded with the desiredcaged nucleotide according to John et al.30 The samplecomposition was a 1:1 complex of Ras with the cagednucleotide (10 mM) in MgCl2 (25 mM), DTT (10 mM), andHEPES pH 7.5 (200 mM) buffer. Measurements for assays Iand II and the GAP catalyzed (1:1 complex) experiments ofassay III were performed at 260 K using 12% ethylene glycolto prevent freezing. Intrinsic and 1% NF1 catalyzed experi-ments with Ru peptide bioconjugates were conducted in waterbuffer without ethylene glycol at 303 K and 283 K,respectively. The samples were prepared between two CaF2

windows as detailed by Cepus.31

Photolysis of the caged compounds was performed using anLPX 240 XeCl excimer laser (308 nm, Lambda Physics,Gottingen, Germany). Twelve (12) flashes within 24 ms wereused for the pHP-cage group and 40 flashes within 120 ms forthe NPE-cage group. A modified Bruker IFS 66v/s spectrom-eter in the fast scan mode32 with 2000 cm�1 cut-off filter wasused for the measurement.

Data acquisition was performed in the ‘‘double-sidedforward backward’’ mode; the four spectra calculated fromeach complete cycle of the scanner (4 cm�1 resolution) wereaveraged. The data were analyzed between 1900 and 900 cm�1

using a global fit method.33 In this analysis, the absorbancechanges DA are analyzed with sums of nr exponentials with

SCHEME 1. Structures of investigated compounds.

968 Volume 64, Number 9, 2010

apparent rate constants kl and amplitudes al:

DAðm; tÞ ¼Xnr

l¼1

alðmÞe�klt þ a‘ðmÞ ð1Þ

For all assays shown in this paper a single exponentialfunction is used (nr ¼ 1). This rate describes either aconformational change (assays I and II) or GTP hydrolysis(assay III). In the figures, negative amplitude spectra areshown; the disappearing bands face downwards and theappearing bands face upwards. Rates for assay III arecalculated by plotting the difference between the intensitiesof a-GDP and a-GTP with respect to time and fitting to a singleexponential equation.34

Analytic Methods. The HPLC system was a ProStar, Model210 from Varian (Walnut Creek, CA) with a column of AlltimaC18LL, 5 lm, 250 3 4.6 mm i.d. (Alltech, Deerfield, IL). Themobile phases were (A) water–TFA (100:0.025, v/v) and (B)acetonitrile–TFA (100:0.025, v/v) with a gradient of 0 to 20%B in A in the first 5 min, 20 to 40% B in A from 5 to 10 min, 40to 80% B in A from 10 to 15 min, 80 to 100% B in A from 15to 20 min, and 100 to 0% B in A from 20 to 30 min. The flowrate was 1.0 mL/min, and 10 lL was injected in all analytic

experiments. The UV conditions set for detection are twochannel modes at 220 nm and 254 nm. For preparative HPLC,the respective fractions were collected directly from the columnaccording to their retention times. All positive electrosprayspectra were acquired using a Bruker Daltonics Esquire 6000(Bremen, Germany).

RESULTS AND DISCUSSION

Experimental Approach. Figure 1 shows our experimentalsetup for time-resolved FT-IR spectroscopy.1 The spectrometer(Fig. 1A) is equipped with a globar as the light source. Theinfrared light passes an aperture, an interferometer, the sample,and finally reaches a liquid-nitrogen-cooled mercury cadmiumtelluride (MCT) detector. The sample can further be irradiatedby an ultraviolet (UV) laser. The sample cell consists of twoCaF2 windows separated by a spacer (Fig. 1B). Due to thestrong infrared absorbance of water, the path length is limitedto 10 lm and a high concentration of the protein is necessary.In the case of GTPases concentrations of ;200 lg/lL are used.In order to obtain difference spectra, the reaction has to betriggered within the spectrometer. In the difference spectra,only the absorptions of groups participating in the reaction areseen, but all other bands are canceled out. For GTPasestriggering is done by means of caged GTP, which is alreadyplaced within the binding pocket but does not hydrolyze asshown in Fig. 1C.35 By a laser flash GTP is formed andsubsequent reactions can be followed.

The GTPase Cycle. Figure 2 shows the GTPase cycle ofRas. In our experiments we enter this cycle by photolysis ofRas–caged-GTP. In this process RasoffGTP is formed.

FIG. 1. Experimental approach: The setup of (A) the spectrometer with (B) thesample cuvette is shown. The sample contains (C) Ras in buffer with boundcaged GTP. For the assays small molecules and the effector RafRBD are addedas detailed in the text. The switch-region of Ras is shown in blue. Formeasurements at higher temperatures the more stable NPE-caged GTP (left)should be used, while at lower temperature the faster pHP-caged GTP isrecommended.

FIG. 2. Possible anticancer strategies involving down-regulation of Rasprotein: (Assay I) Inhibition of active ‘‘ON’’ conformation of the Ras protein byshifting the equilibrium towards ‘‘OFF’’ conformation, which has leastinteraction with effectors; (Assay II) interruption of the effector interactionwith activated RasonGTP-state with suitable ligands; (Assay III) stimulation oracceleration of the Ras-bound GTP hydrolysis to GDP with GAP mimics.

APPLIED SPECTROSCOPY 969

Afterwards the formation of RasonGTP can be observed bytime-resolved FT-IR.12 A molecule shifting the equilibriumtowards the ‘‘off’’ state could prevent oncogenic Ras fromsignaling. Such a molecule can be screened for in our firstassay, which detects changes in this equilibrium. Subsequently,RasonGTP can bind to an effector. This process is measured inassay II. The next process in the GTPase cycle is thetermination of the signaling by GTP hydrolysis, which canalso be monitored by time-resolved FT-IR (assay III). In thefollowing we elaborate the three assays and give examples ofmolecules that can interfere with the Ras protein. All threeassays can find molecules that down-regulate Ras signaling; bystabilizing RasoffGTP (I), by preventing Raf binding to Ras(II), or by forming RasoffGDP (III).

Assay I: Switching of Conformation States. Ras wt showsafter photolysis of the caged-group the conformational changefrom RasoffGTP towards RasonGTP.12 The correspondingdifference spectrum is shown in Fig. 3a (red). On the otherhand, Ras T35S (green) remains in the ‘‘off’’ state, showing noconformational change within this period. A screening can bedone by measuring Ras wt together with a small molecule (black

spectrum) and simple comparison of the resulting spectrum withthe two spectra, the red spectrum standing for ‘‘no interaction’’and the green spectrum standing for ‘‘interaction’’. Noassignment of any band to specific residues is necessary toprove an influence. However, several of the bands are alreadyassigned by isotopic labeling of Ras or GTP.36,37 Thus,interactions can even be assigned to a specific site.

Instead of using the fingerprint of the whole spectrum, theabsorption of the asymmetric c-GTP vibration at 1143 cm�1,which is the most intense marker band for the RasonGTPstate,12 can be used for the assay (Fig. 3b). While for Ras wt(red) an absorption increase is observed with a rate of about 6s�1, the absorbance is not changed for Ras T35S (green).

Previous results from NMR studies19 showed the modulationof the conformation states by interaction of zinc cyclen withRas–GppNHp. In our time-resolved FT-IR experiments weused instead of GppNHp the native nucleotide GTP (black).Interestingly, in this case there is no shift towards the ‘‘off’’state by zinc cyclen, as clearly shown by the similarity of thespectra to Ras wt without zinc cyclen (Fig. 3a). The same resultcan be obtained by investigation of the marker band (Fig. 3b).The observed rate for Ras wt þ zinc cyclen is about 6 s�1,unchanged compared to Ras wt alone. This indicates that thephosphate chelating agent zinc cyclen has no effect on shiftingan active ‘‘on’’ conformation state of the Ras–GTP proteintowards an ‘‘off’’ conformation state. We have further screenedmembers of the polyamine family, because molecular modelingstudies predict an interaction with Ras.17 Our results indicatethat in the presence of the magnesium complexes of spermine,spermidine, and putrescine, there were no changes associatedwith the final equilibrium ‘‘on’’ conformation state of the Ras–GTP protein in photolysis difference spectra and representativeswitching kinetics (data not shown, the potential drugs areshown in Scheme 1).

Assay II: Ras–Effector Interaction. This assay is a

FIG. 3. Assay I: Interaction of Ras–GTP protein with zinc cyclen. (a)Amplitude spectra represent the conformational changes associated withinteraction of zinc cyclen with Ras–GTP. Data for RasT35S (green), Ras wildtype (red), and Ras wt in the presence of 203 zinc cyclen (black) are shown;scaling is done at 1689 cm�1 representing the rate k1. (b) Rate of switching ofRas protein in the presence of zinc cyclen. The plot shows the generation of thec-phosphate vibration at 1143 cm�1 of the RasonGTP in the presence of the zinccyclen at 203 concentration compared to the concentration of the Ras protein(black, 6.1 s�1). Control Ras protein without zinc cyclen is shown in red (6.1s�1). The second control is RasT35S mutant alone (green, kinetics withoutchanges).

FIG. 4. Assay II: Switching kinetics of RasT35S at 1143 cm�1 in the presenceof the effector Raf RBD and zinc cyclen (103): 1.3 s�1 (purple); Lupeol (0.23):4.7 s�1 (blue); and ruthenium NF1 peptide bioconjugate (23): 5.0 s�1 (black).Measurement of T35S and RafRBD as a control is shown in red: 5.0 s�1. Thesecond control is T35S alone (green, kinetics without changes).

970 Volume 64, Number 9, 2010

modification of assay I. It uses the fact that Ras T35S can beswitched ‘‘on’’ by interaction with RafRBD.12 In Fig. 4, thekinetics of the same band as used for assay I are shown. WhileT35S remains in the ‘‘off’’ state (green), T35SþRafRBD showsagain the switch towards the ‘‘on’’ state (red). A moleculeinterfering with the Ras–Raf interaction would prevent thisswitching. Thus, for drug-screening, Ras T35SþRafRBD willbe measured in the presence of a small molecule. In the casethat Ras–Raf interactions are disturbed, the kinetics of thismarker band should deviate from the kinetics without the smallmolecule. We screened zinc cyclen, lupeol,38 and a combina-tion of ruthenium bioconjugates of the NF1 catalytic sequences(see Scheme 1). For lupeol (blue) and Ru-peptides (black) nochanges are observed, indicating no strong interaction withRas. On the other hand, the reaction is slowed down for thezinc cyclen (purple). Thus, in this case, an interaction isproved. Even though the interaction is weak and it only slowsdown the binding but does not prevent it, monitoring by time-resolved FT-IR spectroscopy is possible. Note that the use ofthe natural nucleotide GTP is important for these assays,because GppNHp already stabilizes the ‘‘off’’ conformation.

Assay III: GTP Hydrolysis. In this assay we measure thekinetics of GTP hydrolysis. This can be done best by theabsorption difference of the asymmetric a-GDP (1236 cm�1)and a-GTP (1263 cm�1) stretching vibrations (see also Fig.3a).34 The advantage of taking two reference points close toeach other instead of monitoring the trace of a singlewavenumber is an annihilation of baseline drifts. This artificewas not necessary for the measurement of the fast processes inthe previous assays, but it is very helpful for the much slowerintrinsic hydrolysis of Ras. In this assay (Fig. 5), the smallmolecule should shift the kinetics of Ras (red) towards thekinetics of Ras–NF1 (green).

As seen in Fig. 5a, the intrinsic GTPase reaction of Ras has arate of 0.012 min�1. Instead of a small molecule, wedemonstrate the assay by RapGAP,22 a GAP for the GTPaseRap without the crucial arginine finger (black). Surprisingly,this RapGAP can also accelerate Ras, obviously by a differentmechanism than RasGAP. The observed rate was 25-foldfaster, despite the 33 K lower temperature. The temperature-corrected acceleration is about 500 fold. This shows thatRapGAP can bind to Ras and also close the binding niche ofRas, shifting the electron distribution towards the productstate.11,39 While this is probably only possible due to the highconcentrations in time-resolved FT-IR samples,22 the GAP1family members are known to accelerate both Ras and Rapeven at low concentrations,40 showing the related mechanismsof Ras and Rap acceleration.

We further tested the Ru-bioconjugates compounds (Scheme1). Ruthenium macrocycle (RuC) was conjugated to the N-terminus of the NF1 peptide. The designed bioconjugate wasexpected to bind to Ras by coordination of the arginine with theGTP and additional interaction of the ruthenium macrocycle tothe carboxylates above the bound nucleotide in the Switch IIregion. While the results with assay III were not conclusive, weadditionally tried a competitive assay. We added 1% NF1 toour sample. If the Ru compounds bind to Ras, the binding toNF1 should be attenuated and the reaction rate should bedelayed. Indeed, we observe this deceleration of the GTPhydrolysis by a factor of about 8 (Fig. 5b). Thus, the Ru-bioconjugate binds to Ras. We are currently modifying thecompound in order to achieve acceleration of the hydrolysis.

High-Throughput Screening. Our results clearly show thattime-resolved FT-IR is a very powerful technique, enablingtime-resolved, label-free, direct screening for drugs at theatomic level. Nevertheless, due to the complex appearance ofFT-IR difference spectra and the elaborate technical setup, FT-IR is rarely used for screening. However, automatic sampleinlets are available and as shown in this paper, simple assaysusing only marker bands or comparison to known spectra arepossible. Thus, the evaluation of the measurements can also betaken over by automated programs.

CONCLUSION

In summary, in this paper we have developed three FT-IRassays to screen for potential interactions of small moleculesand peptide metal bioconjugates with Ras–GTP as the target.We demonstrated that we can screen for potential therapeuticligands to probe Ras–GTP protein properties such as switchingof the protein conformation states and protein-bound GTP

FIG. 5. Assay III: Analysis of GTPase activity. (a) Kinetic analysis of theintrinsic activity at 303 K (red), and GAP-catalyzed (NF1, green, and RapGAP,black, at 260 and 270 K, respectively) GTPase activity of wild type Ras. Shownis the difference between the intensities of the a-phosphate vibrations of GDP(1236 cm�1) and a-phosphate vibrations of GTP (1263 cm�1). Kinetics arefitted to single exponential functions. (b) Kinetic analysis of the GTPaseactivity of Ras wt at 283 K in the presence of 1% NF1 (blue) and 23 rutheniumNF1 peptide bioconjugates (pink).

APPLIED SPECTROSCOPY 971

hydrolysis comparable to other methods, which require eithernon-hydrolysable nucleotides or other fluorescent indicators.Practically all molecules are visible in the infrared. Thus, time-resolved FT-IR spectroscopy can be universally used inpharmacological screening. In contrast to most other assays,this method is label-free and does not use an artificial markerthat could influence the protein–drug interaction.

ACKNOWLEDGMENTS

We thank Alfred Wittinghofer and Christian Herrmann for providingRasT35S plasmid and RafRBD, and Angela Kallenbach and Sven Brucker forhelp with protein purification. Support by the SFB 642 (DFG) and the CVM(co-financed by the EU and the state North Rhine-Westphalia) is gratefullyacknowledged. Ravi S. Bojja is funded by the IMPRS in Chemical Biology(Max Planck Society).

1. C. Kotting and K. Gerwert, ChemPhysChem 6, 881 (2005).2. K. Gerwert, B. Hess, J. Soppa, and D. Oesterhelt, Proc. Nat. Acad. Sci.

U.S.A. 86, 4943 (1989).3. J. Downward, Nat. Rev. Cancer 3, 11 (2003).4. K. Wennerberg, K. L. Rossman, and C. J. Der, J. Cell. Sci. 118, 843

(2005).5. J. L. Bos, H. Rehmann, and A. Wittinghofer, Cell 129, 865 (2007).6. C. Herrmann, Curr. Opin. Struct. Biol. 13, 122 (2003).7. K. Scheffzek, M. R. Ahmadian, W. Kabsch, L. Wiesmuller, A. Lautwein,

F. Schmitz, and A. Wittinghofer, Science (Washington, D.C.) 277, 333(1997).

8. I. R. Vetter and A. Wittinghofer, Science (Washington, D.C.) 294, 1299(2001).

9. M. Spoerner, C. Herrmann, I. R. Vetter, H. R. Kalbitzer, and A.Wittinghofer, Proc. Nat. Acad. Sci. U.S.A. 98, 4944 (2001).

10. C. Kotting, M. Blessenohl, Y. Suveyzdis, R. Goody, A. Wittinghofer, andK. Gerwert, Proc. Nat. Acad. Sci. U.S.A. 103, 13911 (2006).

11. C. Kotting, A. Kallenbach, Y. Suveyzdis, A. Wittinghofer, and K. Gerwert,Proc. Nat. Acad. Sci. U.S.A. 105, 6260 (2008).

12. C. Kotting, A. Kallenbach, Y. Suveyzdis, C. Eichholz, and K. Gerwert,ChemBioChem 8, 781 (2007).

13. S. Schubbert, K. Shannon, and G. Bollag, Nat. Rev. Cancer 7, 295 (2007).14. M. R. Ahmadian, Brit. J. Haematol. 116, 511 (2002).15. D. Strumberg and S. Seeber, Onkologie 28, 101 (2005).16. A. Wittinghofer and H. Waldmann, Angew. Chem. Int. Ed. Engl. 39, 4192

(2000).

17. H. te Heesen, A. M. Schlitter, and J. Schlitter, J. Mol. Graph. 25, 671(2007).

18. J. Kato-Stankiewicz, I. Hakimi, G. Zhi, J. Zhang, I. Serebriiskii, L. Guo, H.Edamatsu, H. Koide, S. Menon, R. Eckl, S. Sakamuri, Y. C. Lu, Q. Z.Chen, S. Agarwal, W. R. Baumbach, E. A. Golemis, F. Tamanoi, and V.Khazak, Proc. Nat. Acad. Sci. U.S.A. 99, 14398 (2002).

19. M. Spoerner, T. Graf, B. Koenig, and H. R. Kalbitzer, Biochem. Biophys.Res. Commun. 334, 709 (2005).

20. J. Tucker, G. Sczakiel, J. Feuerstein, J. John, R. S. Goody, and A.Wittinghofer, EMBO J. 5, 1351 (1986).

21. M. R. Ahmadian, U. Hoffmann, R. S. Goody, and A. Wittinghofer,Biochemistry 36, 4535 (1997).

22. P. P. Chakrabarti, O. Daumke, Y. Suveyzdis, C. Kotting, K. Gerwert, andA. Wittinghofer, J. Mol. Biol. 367, 983 (2007).

23. C. Herrmann, G. A. Martin, and A. Wittinghofer, J. Biol. Chem. 270, 2901(1995).

24. J. Walker, G. P. Reid, J. A. McCray, and D. R. Trentham, J. Am. Chem.Soc. 110, 7170 (1988).

25. C.-H. Park and R. S. Givens, J. Am. Chem. Soc. 119, 2453 (1997).26. S. I. Kirin, P. Dubon, T. Weyhermuller, E. Bill, and N. Metzler-Nolte,

Inorg. Chem. 44, 5405 (2005).27. S. I. Kirin, F. Noor, N. Metzler-Nolte, and W. Mier, J. Chem. Educ. 84,

108 (2007).28. N. Metzler-Nolte, Chimia 61, 736 (2007).29. F. Noor, R. Kinscherf, G. A. Bonaterra, S. Walczak, S. Wolfl, and N.

Metzler-Nolte, ChemBioChem 10, 493 (2009).30. J. John, R. Sohmen, J. Feuerstein, R. Linke, A. Wittinghofer, and R. S.

Goody, Biochemistry 29, 6058 (1990).31. V. Cepus, A. J. Scheidig, R. S. Goody, and K. Gerwert, Biochemistry 37,

10263 (1998).32. K. Gerwert, G. Souvignier, and B. Hess, Proc. Nat. Acad. Sci. U.S.A. 87,

9774 (1990).33. B. Hessling, G. Souvignier, and K. Gerwert, Biophys. J. 65, 1929 (1993).34. C. Kotting and K. Gerwert, Chem. Phys. 307, 227 (2004).35. A. J. Scheidig, S. M. Franken, J. E. Corrie, G. P. Reid, A. Wittinghofer, E.

F. Pai, and R. S. Goody, J. Mol. Biol. 253, 132 (1995).36. C. Allin and K. Gerwert, Biochemistry 40, 3037 (2001).37. B. Warscheid, S. Brucker, A. Kallenbach, H. E. Meyer, K. Gerwert, and C.

Kotting, Vib. Spectrosc. 48, 28 (2008).38. M. Saleem, S. Kaur, M. H. Kweon, V. M. Adhami, F. Afaq, and H.

Mukhtar, Carcinogenesis 26, 1956 (2005).39. H. te Heesen, K. Gerwert, and J. Schlitter, FEBS Lett. 581, 5677 (2007).40. B. Sot, C. Kotting, D. Deaconsescu, Y. Suveyzdis, K. Gerwert, and A.

Wittinghofer, EMBO J. 29, 1205 (2010).

972 Volume 64, Number 9, 2010