Báo cáo Y học: Synthesis, characterization and application of two nucleoside triphosphate analogues, GTPcNH2 and GTPcF

doi:10.1046/j.1432-1033.2002.03003.x

Eur. J. Biochem. 269, 3270–3278 (2002) (cid:3) FEBS 2002

Synthesis, characterization and application of two nucleoside triphosphate analogues, GTPcNH2 and GTPcF

Michael Stumber1, Christian Herrmann2, Sabine Wohlgemuth2, Hans Robert Kalbitzer1, Werner Jahn1 and Matthias Geyer1,* 1Max-Planck-Institut fu¨ r medizinische Forschung, Department of Biophysics, 69120 Heidelberg, Germany; 2Max-Planck-Institut fu¨ r molekulare Physiologie, Department of Structural Biology, 44227 Dortmund, Germany

to that for the natural substrate GTP. For GTPcF we ob- tained a similar enthalpy of DH(cid:2) ¼ 3.9 kcalÆmol)1 while the )1. The magnesium association constant is only Ka ¼ 0.2 mM application of both guanine nucleotide analogues to the GTP-binding protein Ras was investigated. The rate of hydrolysis of GTPcNH2 bound to Ras protein lay between the rates found for Ras-bound GTPcS and GppNHp, while Ras-catalysed hydrolysis of GTPcF was almost as fast as for GTP. The two compounds extend the variety of nucleotide analogues and may prove useful in structural, kinetic and cellular studies.

Keywords: nucleotides; nucleotide analogues; NMR spectro- scopy; GTP hydrolysis; Ras.

Guanosine triphosphate nucleotide analogues such as GppNHp (also named GMPPNP) or GTPcS are widely used to stabilize rapidly hydrolyzing protein-nucleotide complexes and to investigate biochemical reaction path- ways. Here we describe the chemical synthesis of guanosine 5¢-O-(c-amidotriphosphate) (GTPcNH2) and a new synthe- sis of guanosine 5¢-O-(c-fluorotriphosphate) (GTPcF). The two nucleotides were characterized using NMR spectrosco- py and isothermal titration calorimetry. Chemical shift data on 31P, 19F and 1H NMR resonances are tabulated. For GTPcNH2 the enthalpy of magnesium coordination is DH(cid:2) ¼ 3.9 kcalÆmol)1 and the association constant Ka is )1. The activation energy for GTPcNH2ÆMg2+ 0.82 mM complex formation is DH(cid:2) ¼ 7.8 ± 0.15 kcalÆmol)1, similar

processes such as signal transduction (Ras and Rho families) or protein and vesicle trafficking (Ran and Rab families, respectively), which are combined with GTP- hydrolysis (reviewed in [1–3]). Another aspect of nucleotide mediated transformation is the transfer of the leaving phosphoryl group (mostly the c-phosphate group) to acceptors like water, amino-acid residues, or other nucleo- tides. Often the association of a metal ion, usually magnes- ium, with the phosphate groups of the nucleotide is crucial for these events.

Nucleotides are fundamental components in cellular meta- bolism. Acting as substrates for nucleotide binding proteins, they are the protagonists of a large variety of cellular processes. Nucleotides can regulate enzymatic activity by transitions between their mono-, di- and triphosphate bound forms. These transitions often induce conformational changes in the proteins, referred to as the (cid:3)active(cid:4) and (cid:3)inactive(cid:4) conformations. Perhaps the best known example is the energy metabolism of adenosine nucleotides: hydrolysis of ATP to ADP leads to functional molecular rearrange- ments in the actomyosin mediated muscle contraction. Guanine nucleotide-binding proteins on the other hand are specialized in the control of intracellular communication

The study of nucleotide-binding proteins, their function, structure and mechanism, often demands use of nonhy- drolyzable or slowly hydrolyzable nucleotide analogues. These modifications become necessary when stabilization of a specific isoform of the protein is required. In cellular assays the triphosphate analogues GTPcS and ATPcS are most commonly used, usually in order to generate the constitutively active form of a protein. In structural biology, long-term stability of the protein-nucleotide complex is required in order to grow homogeneous crystals or to obtain a single state of the protein. Here, the most commonly used triphosphate analogues are GppNHp (also named GMPPNP or GDPNP) and to a minor extent GppCH2p (also named GMPPCP) and their respective adenosine counterparts AppNHp and AppCH2p. Another application of substrate analogues is the use of caged nucleotides to characterize unstable protein intermediates by X-ray crys- tallography [4]. Nucleotide modifications can also serve as an approach to designing dominant negative forms of a protein [5] or to solve the phase problem in crystallography [6]. Even more specific is the application of aluminium fluoride, beryllium fluoride or orthovanadate in the presence

Correspondence to M. Geyer, Max-Planck-Institut fu¨ r medizinische Forschung, Abteilung Biophysik, Jahnstraße 29, D-69120 Heidelberg, Germany. Fax: + 49 6221 486 437, Tel.: + 49 6221 486 396, E-mail: geyer@mpimf-heidelberg.mpg.de Abbreviations: GTPcNH2, guanosine 5¢-O-(c-amidotriphosphate); GTPcF, guanosine 5¢-O-(c-fluorotriphosphate); GppNHp, guanosine 5¢-O-(b,c-imidotriphosphate); GppCH2p, guanosine 5¢-O- (b,c-methylenetriphosphate); GTPcS, guanosine 5¢-O-(c-thiotriphos- phate); ITC, isothermal titration calorimetry; DCC, dicyclohexylcar- bodiimide; DSS, sodium 2,2-dimethyl-2-silapentane-5-sulfonate; THC, triethylammonium hydrogencarbonate. *Present address: Max-Planck-Institut fu¨ r molekulare Physiologie, Department of Physical Biochemistry, 44227 Dortmund, Germany. (Received 23 January 2002, revised 8 May 2002, accepted 17 May 2002)

(cid:3) FEBS 2002

Nucleotide analogues GTPcNH2 and GTPcF (Eur. J. Biochem. 269) 3271

of a nucleoside diphosphate. These compounds can form stable analogues that mimic the transition state of the terminal leaving group of the nucleotide within a protein- nucleotide complex [7,8].

We characterized the stability and metal ion binding properties of the two nucleotide analogues by NMR spectroscopy and isothermal titration calorimetry. Both nucleotides were bound to the small GTP-binding protein Ras and the rates of hydrolysis were determined in comparision to other nucleotide triphosphate derivatives. Finally, the suitability for spectroscopic and structural studies was tested by formation of the complex between RasGTPcNH2 and the Ras-binding domain of the Ras effector protein c-Raf-1.

M A T E R I A L S A N D M E T H O D S

General description of synthesis

Mechanistic studies to analyse the enzymatic activity of a nucleotide binding protein usually benefit from the avail- ability of a broad range of different nucleotide phosphate analogues. Here, advantage can be taken of the individual characteristics of the nucleotide when applied to a protein. Differences in metal ion binding properties as well as charge distribution and hydrophobicity determine the specific features of a nucleotide that provide insights into the biological system. Also, nucleotide modifications such as spin labeling make the protein-nucleotide complex access- ible to spectroscopic techniques. Most prominent is the use for kinetic of fluorescent analogues (e.g. mant-GTP) measurements by fluorescence spectroscopy and 17O-labe- ling for EPR or NMR techniques.

(GTPcNH2),

High pressure liquid chromatography (HPLC) was done on a Beckman (cid:3)System Gold(cid:4)(cid:4). Nucleotides were analysed by ion-pair chromatography on a reversed phase Super ODS column, 50 · 4.6 mm (TOYOPEARL(cid:4)) at a flow rate of 1.2 mLÆmin)1, using a linear gradient from 100% 10 mM tetrabutyl-ammonium bromide/10 mM sodium phosphate buffer (pH 6.8) to 100% acetonitrile within 10 min. Detec- tion was at 260 and 340 nm. The retention times given are for orientation only.

Here we investigate two modified nucleoside triphos- phates which are stable and show distinct characteristics: guanosine 5¢-O-(c-amidotriphosphate) for which we describe the first synthesis, and guanosine 5¢-O- (c-fluorotriphosphate) (GTPcF) [9], which we synthesized by the method of Wittmann [10]. Both are shown in Fig. 1.

GTP-triethylammonium salt was prepared by applying GTP sodium salt to a Super Q column (TOYOPEARL(cid:4)) and elution with a gradient from 0 to 1 M triethylammo- nium hydrogencarbonate (THC). The eluate containing the nucleotide (retention time in HPLC 4.65 min) was evapor- ated under reduced pressure, redissolved in methanol, again evaporated and dried over P4O10. Monoamido-phosphoric acid, H2PO3NH2, was prepared as described [11].

Synthesis of GTPcNH2 and GTPcF

To the solution of 0.8 g GTP triethylammonium salt in 5 mL dimethylsulfoxide were added 0.8 g DCC and 80 mg pyridinium hydrochloride. After 20–24 h at room tempera- ture the mixture was treated with about 5 mL concentrated ammonia in water for 30 min. The solution was diluted with 60–70 mL water and, after filtration, applied to a Super Q column (2.5 · 20 cm). The column was eluted at a rate of 5 mLÆmin)1 with a gradient from 0 to 1 M THC within 120 min. Fractions containing the GTPcNH2 (as checked by UV absorption and HPLC, retention time 4.20 min) were collected and evaporated. Any remaining THC was removed by dissolving in methanol and repeated evapor- ation under reduced pressure, yield of the pure GTPcNH2 was 50–60%.

One gram of GTP triethylammonium salt was added to a stirred solution of 2.5 mL tributylamine and 1.2 g 2,4-dinitrofluorobenzene in about 10 mL dimethylforma- mide. After 6–8 h a clear solution was obtained. The mixture was kept for 20–24 h at room temperature. The crude product was precipitated with 100 mL acetone and 300 mL diethyl ether. The pellet was dissolved in water (about 20 mL) and applied to a Super Q column (18 · 2.6 cm). The column was eluted with a gradient of 0–1 M THC within 2 h at a flow rate of 5 mLÆmin)1. Fractions containing the reaction product (retention time of 4.47 min, no absorption at 340 nm) were collected and evaporated as described for the GTP triethylammonium salt. The product was dissolved in 20 mL methanol and

Fig. 1. Chemical structure of the nucleoside triphosphate analogues synthesized. (A) Guanosine 5¢-O-(c-amidotriphosphate) (GTPcNH2) and (B) guanosine 5¢-O-(c-fluorotriphosphate) (GTPcF). Displayed is the D-riboside form of the respective nucleoside.

3272 M. Stumber et al. (Eur. J. Biochem. 269)

(cid:3) FEBS 2002

precipitated by addition of a solution of 250 mg NaClO4 in a few ml methanol to remove part of the colored by-products. The pellet was dissolved in water (20 mL) and purified on a Super Q column as described above, giving an almost colorless substance (yield in the range 5–10%).

31P chemical shift d (p.p.m.)

19F chem. shift d

Table 1. NMR chemical shifts of GTPcNH2, GTPcF, and GTP in aqueous solution. Spectra were recorded in 90%/10% H2O/D2O at pH 7.4 and 25 (cid:2)C. 31P and 19F chemical shifts were referenced to 85% phosphoric acid and trifluoroacetic acid, respectively, using the indirect reference method with parameters adopted from IUPAC [12].

Nucleotide a b c

GTPcNH2 GTPcNH2ÆMg2+ GTPcF GTPcÆFMg2+ GTP GTPÆMg2+ )11.46 )11.33 )11.66 )11.89 )10.74 )10.41 )22.76 )21.50 )23.49 )23.19 )21.22 )19.01 )1.12 )0.34 )18.16 )18.63 )5.51 )5.30 – – 0.91 0.97 – –

Table 2. J-Coupling constants of GTPcNH2 and GTPcF in aqueous solution.

Preparation of NMR samples and NMR spectroscopy 31P and 19F NMR spectra of free nucleotides were recorded in aqueous solution of 90%/10% H2O/D2O. Typically the lyophilized nucleotide was redissolved to a final concentra- tion of 2–10 mM and 2500 lL of the sample volume was placed in 10 mm NMR tubes (Wilmad). For titration experiments various amounts of MgCl2 were added from a 100-mM stock solution. Proton NMR experiments were performed using 500 lL sample volume in 5 mm NMR tubes (Wilmad). 31P NMR spectra of C-terminal truncated wildtype Ras protein (residues 1–167) complexed with GTPcNH2ÆMg2+ were recorded in 40 mM Tris/HCl, 5 mM MgCl2 and 2 mM DTE at pH 7.4. Here, sample volumes of 2500 lL of 1.0 mM concentrated protein were measured containing 10% D2O.

J-coupling constants (Hz)

2JPaPb

2JPbPc

1JPcF

Nucleotide

the states A and B and the exchange rates k1 for A fi B and k)1 for B fi A. As B ¼ 100%–A and k)1 ¼ k1*A/B only two free parameters had to be fitted to the experimental data. The simulations were performed on the complete 31P spectra (a-, b- and c-phosphorus nuclei), using chemical shift values and J-coupling constants as listed in Tables 1 and 2.

Isothermal titration calorimetry

1H, 19F and 31P NMR experiments were performed on a Bruker AMX-500 NMR spectrometer working at reson- ance frequencies of 500 MHz, 470 MHz and 202 MHz, respectively. 31P spectra were referenced to 85% phosphoric acid enclosed in a glass sphere which was immersed in the sample and calibrated for various temperatures. 19F spectra were referenced to trifluoroacetic acid, based on the IUPAC conventions for indirect referencing relative to internal DSS [12]. Unless noted otherwise, phosphorus spectra were recorded at 20 (cid:2)C with a total spectral width of 60 p.p.m. For one dimensional 31P NMR spectra of free nucleotides, 64–512 free induction decays were summed after excitation with a 65 degree pulse using a repetition time of 3–5 s. A total of 32 K time domain data points were recorded and transformed to 16 K real data points corresponding to a digital resolution of 0.74 Hz point)1. The 31P spin-spin coupling constants of the nucleotide-Mg2+ complexes were determined from a nonfiltered 1D spectrum with a digital resolution of 0.25 Hz per point after Fourier transforma- tion.

All spectra were processed on a Silicon Graphics Indigo2 workstation using the software package UXNMR (Bruker, Karlsruhe) for data processing and data evaluation. Phos- phorus spectra used for exchange rate determination were filtered by an exponential window function causing no significant line broadening.

calculated by the

fundamental

is

Determination of exchange rates The Mg2+ exchange rates of GTPcNH2 were extracted from a series of 31P NMR exchange spectra. The spectra were analyzed and compared to simulations based on the mathematical treatment of the exchanging spin system following Nageswara Rao [13]. The simulation of the 31P spectra was built on C++ NMR library (cid:3)GAMMA(cid:4) [14], modeling a three spin system with an ABC fi A¢B¢C¢ exchange. Chemical shift and J-couplings were determined for NMR spectra of both (cid:3)pure(cid:4) states: without magnesium complexation (state A) and with saturated magnesium complexation (state B) (see Tables 1 and 2). Thus, the only parameters to be adjusted were the relative populations of

The interaction between a nucleotide and the magnesium ion was investigated by means of ITC (ITC-MCS, Micro- Cal, Inc.). Briefly, in such an apparatus the solutions are thermostatted to the desired temperature, the nucleotide at 5.0 mM placed in a cell which is accurately temperature controlled and the MgCl2 solution at 50 mM in a syringe dipping into the cell. The two solutions are mixed by computer controlled stepwise injections (typically in inter- vals of 4 min) from the syringe which serves at the same time as a stirrer. The heat consumed due the endothermic association process is measured by the detection of the heating power which is necessary to keep the cell at constant temperature [15]. All ITC experiments were performed at 25 (cid:2)C. The data were analyzed using the manufacturer’s software yielding the stoichiometry N, the binary equili- constant Ka ¼ [nucleotideÆMg2+]/ brium association [nucleotide]/[Mg2+] and the enthalpy of association DH(cid:2), the latter with the approximation that this parameter is independent of the concentration. The change of entropy DS(cid:2) relationship –RT lnKa ¼ DH(cid:2) – TDS(cid:2). The experimental error on DH(cid:2)

GTPcNH2 GTPcNH2ÆMg2+ GTPcF GTPcFÆMg2+ GTP GTPÆMg2+ 20.3 15.8 19.9 15.4 19.8 14.0 19.2 16.8 18.2 12.7 19.8 12.4 – – 936.2 934.0 – –

(cid:3) FEBS 2002

Nucleotide analogues GTPcNH2 and GTPcF (Eur. J. Biochem. 269) 3273

is 5% whereas the experimental error on Ka is about 10– 20%. In addition, the stoichiometry factor N is obtained from the fit to the data, where a value of 1 corresponds to 1 : 1 complex formation.

Protein preparation and guanine nucleotide exchange

In order to test the applicability of the two synthesized triphosphate nucleotide analogues to nucleotide binding proteins, the small GTP-binding protein Ras (residues 1–167) was synthesized in Escherichia coli and purified as described [16]. Purified GDP, GTP, GTPcS and GppNHp reagents were purchased from Sigma and GppCH2p was ordered from JenaBioScience. GDP, which binds very tightly to Ras, was replaced with the respective GTP analogue by the following procedures. For nucleotide exchange GTPcNH2, GppNHp and GppCH2p were each incubated at threefold molar excess with Ras in the presence of 200 lM ammonium sulfate, 0.1 lM zinc chloride and 1 U alkaline phosphatase per mg Ras overnight at 4 (cid:2)C. In order to load Ras with GTP, GTPcF, or GTPcS nucleotide-free Ras was produced by incubation overnight at 4 (cid:2)C in the presence of 200 lM ammonium sulfate, 0.1 lM zinc chloride and 0.2 U alkaline phosphatase per mg Ras. After size exclusion chromatography, one of the nucleotides was then added to the Ras protein. Excess nucleotide after either procedure was removed (which is important in order to obtain accurate single turnover hydrolysis rate constants). The pooled Ras fractions were concentrated to 20 mgÆmL)1 by centrifugal concentrators (Vivaspin 10 kDa cut-off, VivaScience). The buffer used in all these procedures contained 25 mM Tris/HCl at pH 7.4, 2.5 mM MgCl2, and 1 mM DTE. The Ras catalysed nucleotide hydrolysis was determined with HPLC by measuring the concentration of protein-bound GTP or its triphosphate analogues and GDP as described [17]. Intrinsic reaction rates were obtained from the decay of the (triphosphate nucleotide)/(tri- and diphosphate nucleotide) ratio with time, fitted to single- exponential curves. The Ras-binding domain of human c-Raf-1 (Raf-RBD, 81 residues) was expressed in E. coli and purified as described recently [18].

Fig. 2. 31P NMR spectra of GTPcNH2 (A) and GTPcF (B) (top) and their respective magnesium ion complexes (bottom). Spectra were recorded at pH 7.4 and 20 (cid:2)C in aqueous solution.

R E S U L T S

NMR spectra, chemical shift data and J-coupling constants of the two nucleotides

solution. The resonance lines could be assigned by their J-coupling constants and by comparison to unmodified GTP. While the chemical shift of the a-phosphate group changed only little upfield compared to GTP, the b-phosphate was shifted upfield by about )1.5 p.p.m. and the terminal c-phosphate shifted by almost 5 p.p.m. down- field by the replacement of the hydroxy OH– with an amide NH2. Complexation of GTPcNH2 with Mg2+ led to an additional downfield shift of all phosphate groups, with the b-phosphate changing most. This observation was similar to the change in GTP when coordinated with magnesium, but the absolute shift change was almost 1 p.p.m. smaller (from 1.26 p.p.m. to 2.21 p.p.m.) than in the natural substrate. The 2JPP-coupling constants of GTPcNH2 and GTPcNH2Mg2+ analogues showed smaller alterations when compared to GTP. In both cases the b-phosphate groups appeared as triplets as the coupling constants between Pa–Pb and Pb–Pc were almost identical, while coordination with magnesium again decreased the coupling constants.

Proton, phosphorus and fluorine NMR measurements confirmed the chemical structure and the high degree of purification of the two synthesized triphosphate nucleotides. As expected, 1H NMR measurements of both GTPcNH2 and GTPcF in aqueous solution at 20 (cid:2)C, pH 7.4 showed no difference to the natural substrate GTP [19] as the guanine base is not affected by the modifications and as the c-phosphate amide hydrogens are in fast exchange with the solvent. In Fig. 2 31P NMR spectra are shown for GTPcNH2 and GTPcF, and their respective metal ion complexes with Mg2+.

In GTPcF four phosphorus lines appeared as the coupling between the natural spin ½ nuclei 31P and 19F led to a splitting of the terminal phosphate resonance. This direct coupling constant 1JPcF was about 936 Hz and hardly changed upon magnesium coordination (934 Hz), indica- ting a strong interaction between the two nuclei. Chemical shift changes of GTPcF compared to GTP were much more distinct than for GTPcNH2. All three phosphate groups shifted upfield; in the case of the c-phosphate the shift was )12.6 p.p.m. By contrast, coordination to magnesium caused only slight chemical shift changes, of which the

For GTPcNH2 the appearance of three discrete reson- ance lines with similar intensity confirms the uniformity and the conformational identity of the substrate. The observed mean half width of, e.g. 4.7 Hz for the c-resonance line is typical for a molecule of 523 Da mass at 20 (cid:2)C in aqueous

3274 M. Stumber et al. (Eur. J. Biochem. 269)

(cid:3) FEBS 2002

largest was )0.5 p.p.m. for the c-phosphate. This might be an effect of the low magnesium binding affinity, as will be discussed later. A similar observation was made for the 19F NMR resonance line at position 0.91 p.p.m. which changed only to 0.97 p.p.m. upon magnesium saturation. Finally, the J-coupling values between the three phosphates again tended to be very insensitive to modifications, and fell by around 25% on complexation with magnesium. All chem- ical shift data and J-coupling constants reported are summarized in Tables 1 and 2.

Nucleotide stability

– + NH4

reciprocal

temperature shows,

of the b-resonance line and the corresponding simulations are shown (Fig. 4). The fitted exchange rates in aqueous solutions ranged from 900 to 9000 Hz with relative margins from ± 22% at 5 (cid:2)C to ± 8% at 30 (cid:2)C. As the plot against the simulated exchange rates k nicely fit to the Arrhenius equation k ¼ k0exp(–DH(cid:2)/RT) with R the gas constant and T the absolute temperature (Fig. 5). Based on these values the activation energy DH(cid:2) for the GTPcNH2Mg2+ complex formation was determined to be 7.8 ± 0.15 kcalÆmol)1. This result is similar to the activation energy for magnesium binding of the natural substrate ATP which has been determined to be 8.1 kcalÆmol)1 [20].

We next tested the stability of the GTPcNH2 nucleotide derivative. In 0.1 M triethanolamine/HCl buffer at pH 7.6 the spontaneous hydrolysis of GTPcNH2 at room tem- perature was less than 1% in five days. In contrast, at pH 4.5 in 0.1 M potassium phosphate buffer the nucleotide was hydrolysed to GDP with a half time of about 48 h. Titration of GTPcNH2 with HCl/NaOH monitored by 31P NMR spectroscopy showed no variation of the chemical shifts of the three-fold negatively charged phosphate groups from pH 3 to pH 11. At pH 2.8 the intrinsic hydrolysis 3– was increased (so called acidic hydrolysis) and GTPcNH2 + by two rapidly transformed to GDP3– + H2PO4 water molecules. The intermediate compound phosphor- acid-amidate H2PO3NH2 was not observed by NMR. As a control, we titrated H2PO3NH2 in the range from pH 11 to pH 1.8. The 31P chemical shifts for the three different protonation states were found to H2PO3NH2 at )6.90 p.p.m., [HPO3NH2]– at )2.65 p.p.m., and [PO3NH2]2– at +7.97 p.p.m. The pKa values between these three states were determined to pK(0/1–) ¼ 3.02 ± 0.05 and pK(1–/2–) ¼ 8.46 ± 0.02 using a least square fit to 15 individual measured chemical shift values (data not shown). Since the resonance lines for the a- and b-phosphate groups of GDP at pH 2.8 were located at )10.73 and )10.20 p.p.m., respectively, a possible signal overlap between GTPcNH2, GDP, HPO3NH2 and H3PO4 (Pi) could be excluded. We therefore assume that at low pH (pH < 3) GTPcNH2 is first transformed to ammonia and GTP, the latter being subsequently hydrolysed to GDP and Pi.

Magnesium binding and magnesium exchange rates

Association of magnesium ions with different nucleotides

To analyse the metal ion binding properties of GTPcNH2 we first performed a magnesium titration series and a temperature series by NMR spectroscopy. Complete line- shape analysis simulations of the complex formation of GTPcNH2 with Mg2+ were performed on the entire 31P NMR spectra (a-, b- and c-phosphorus nuclei) and showed a reasonably good agreement for all three resonance lines. This is demonstrated in Fig. 3 where the part of the NMR spectra and simulations that show the b-phosphate is displayed. The b-resonance line underwent the biggest resonance shift and was therefore most sensitive to changes in the exchange rate, as the titration with magnesium from null to complete saturation indicates (Fig. 3).

Next, we determined the binding energy of GTPcNH2 to magnesium by a complete lineshape analysis of a series of NMR spectra. We adjusted the saturation of GTPcNH2 with Mg2+ to 45% and varied the temperature from 5 (cid:2)C to 65 (cid:2)C in 13 steps of 5(cid:2). Five representative 31P NMR spectra

In biological systems it is the complex between the nucleotide and the magnesium ion which is bound to an ATP or GTP binding enzyme rather than the nucleotide only. Therefore, ITC was employed to quantify the interaction between the nucleotides and the magnesium ion (Fig. 6 and Table 3). As expected all nucleotides bound one magnesium ion as indicated by the stoichiometry factor N ¼ 1 (Table 3). Basically, for all complex formation reactions an unfavorable enthalpy change was observed, which was counteracted by a TDS(cid:2) value two to three times as large. In comparison to GTP the affinity for the magnesium ion was lower for GTPcNH2 and GTPcF. For GTPcS the association constant was only two-fold smaller whereas for GTPcNH2 and GTPcF this constant was significantly smaller, namely 34-fold and 140-fold, respectively. Most probably this is due to the decreased negative charge at the c-position in GTPcNH2 and GTPcF

Fig. 3. 31P NMR spectra of a MgCl2 titration series added in increasing amount to GTPcNH2. The resonance line of the b-phosphate in the experimental measurements (left) and its corresponding simulation (right) are shown. Note the shift and the intermediate broadening of the resonance line. The amount of GTPcNH2Mg2+ complexes relative to free GTPcNH2 nucleotide is indicated on the left. The determined exchange rates based on the exchanging spin system simulation are shown right. The spectra were measured at 20 (cid:2)C and pH 7.4 in aqueous solution.

(cid:3) FEBS 2002

Nucleotide analogues GTPcNH2 and GTPcF (Eur. J. Biochem. 269) 3275

Fig. 6. Isothermal titration calorimetry of GTPcNH2 with MgCl2. To a solution of 5.0 mM GTPcNH2 placed in the cell of the calorimeter a solution of 50 mM MgCl2 was injected in steps of 6 lL each (the first step was 2 lL only). The increase in heating power was detected (upper panel). The power pulses were integrated and plotted vs. the molar ratio of injected MgCl2 and nucleotide (lower panel). A fit to the experimental data yields the stoichiometry factor N ¼ 0.96, the )1 and the enthalpy of association association constant Ka ¼ 0.82 mM DH(cid:2) ¼ 3.9 kcalÆmol)1.

Fig. 4. 31P NMR spectra of a temperature series of GTPcNH2Mg2+. The b-phosphate resonance line at )22.19 p.p.m. is shown in an intermediate exchange state at 45% Mg2+ saturation. Experimental measurements (left) and simulated spectra (right) are displayed showing temperature values and the simulated exchange rates, respectively. Lyophilized GTPcNH2 was dissolved to 2.1 mM con- centration in aqueous solution and adjusted to pH 7.4 with HCl/ NaOH. MgCl2 was added to 1 mM concentration. The precise saturation was determined from the chemical shift position at 20 (cid:2)C (see Fig. 3 and Table 1).

Table 3. Thermodynamic parameters for the association of magnesium ions with different nucleotides obtained by isothermal titration calori- metry. DS(cid:2) is calculated according to the Gibbs-Helmholtz equation.

)1)

Nucleotide N(Nucl./Mg) (mol/mol) Ka (mM DH(cid:2) (kcalÆmol)1) DS(cid:2) (calÆmol)1ÆK)1)

are predominantly due to lower DS(cid:2) values, possibly reflecting the release of less water into bulk upon complex formation.

Application of the nucleotides to the GTPase Ras

GDP GTP GTPcS GTPcNH2 GTPcF 1.0 0.99 1.0 0.96 0.84 3.3 28 14 0.82 0.20 2.3 3.0 4.1 3.9 3.9 24 30 33 26 24

the reciprocal absolute temperature (1/T )

The suitability of the two nucleotide analogues for biolo- gical macromolecules was finally tested using the small GTP-binding protein Ras (reviewed in [21,22]). We success- fully loaded the nucleotide analogues onto the 21 kDa GTPase Ras using the alkaline phosphatase method, which yielded a tightly bound protein-nucleotide complexes, as found for RasÆGTP [23]. First the intrinsic GTPase rate of wild-type H-Ras (1–167) complexed with magnesium ions and various guanosine triphosphate nucleotide analogues was determined by HPLC measurement (Table 4). At 37 (cid:2)C the intrinsic hydrolysis rate of Ras-bound fluorotriphos- phate GTPcF was only twofold lower than for the natural

where the protic hydroxy group is replaced by the amino and fluoride groups, respectively. In contrast, the sulfur in GTPcS may take on the role of the oxo-group. It should be noted that the smaller affinities of GTPcNH2 and GTPcF

Fig. 5. Arrhenius plot of the simulated magnesium ion exchange rate constants (k) vs. for GTPcNH2. The activation energy DH(cid:2) for GTPcNH2ÆMg2+ complex formation is determined to 7.8 ± 0.15 kcalÆmol)1.

3276 M. Stumber et al. (Eur. J. Biochem. 269)

(cid:3) FEBS 2002

Table 4. Intrinsic GTPase rate of wild-type H-Ras(1)167)Mg2+ protein at 37 (cid:1)C bound to various guanosine triphosphate nucleotide analogues. Hydrolysis rates were determined with HPLC by measuring the con- centration of protein-bound tri- and diphosphate nucleotides. Buffer conditions: 25 mM Tris/HCl at pH 7.4, 2.5 mM MgCl2 and 1 mM DTE.

GTPase rate (10)5 min)1) Nucleotide

can be observed also by heteronuclear NMR [25,26] or in different crystal forms of Ras protein [27,28]. Flexibility in the active center of G-proteins has been also observed for RanGTP [29] and in different conformations of the switch regions in the crystal structures of Rap2A complexed with GTP, GDP and GTPcS [30]. As shown in Fig. 7 (bottom spectrum) this feature was preserved for Ras bound to GTPcNH2. At low temperature (5 (cid:2)C) the b-phosphate resonance was split into a less populated high field shifted peak (b1, (cid:2) 27%) and a highly populated down field shifted peak (b2, (cid:2) 73%). A temperature series from 2 (cid:2)C to 30 (cid:2)C revealed the coalescence of both lines at approximately 15 (cid:2)C which is typical for a two-site exchange with a transition from slow to fast exchange (data not shown). As described for the intrinsic hydrolysis of GTPcNH2, the Ras- catalysed hydrolysis of bound GTPcNH2 did not lead to the observable formation of the compound H2PO3NH2 in the NMR spectra (which is expected at )2.7 p.p.m.). Instead, the resonance signals for Pi and Ras-bound GDP increased during the time course of the experiment (Fig. 7, compare bottom and top spectra) suggesting the formation of ammonia and Ras-bound GTP before hydrolysis.

substrate GTP while the rate for the thiotriphosphate GTPcS was about 11-fold lower. Most stable with up to 190-fold lower hydrolysis rates were the two triphosphate analogues with b,c-substitutions GppCH2p and GppNHp. The Ras-catalysed hydrolysis rate of GTPcNH2 finally lay midway between the rates for GTPcS and GppNHp, with a 3-fold difference to both.

The more stable RasÆGTPcNH2ÆMg2+ complex was subsequently studied by 31P NMR spectroscopy. A partic- ular feature of the Ras protein is the flexibility of the effector loop which can be detected in the triphosphate bound form by a line splitting of the phosphorus resonances [24]. The exchange is due to at least two distinct conformations which

A concentration series with the effector protein Raf-RBD at 5 (cid:2)C added in increasing amount from 0.2 to 1 molar ratio showed the progressive stabilization of one particular conformation due to its high affinity for triphosphate bound Ras (Fig. 7). Most remarkably, also the c-phosphate group is perturbed by this interaction and shifted about )0.8 p.p.m. upfield (Table 5). These data indicate the ability of GTPcNH2 to function as a triphosphate nucleotide analogue with characteristic properties.

2820 1427 252 84.6 25.6 15.0 RasGTP RasGTPcF RasGTPcS RasGTPcNH2 RasGppNHp RasGppCH2p

D I S C U S S I O N

The data reported here demonstrate the synthesis of the two guanosine triphosphate nucleotide analogues GTPcNH2 and GTPcF, their biochemical characterization and appli- cation to the GTP-binding protein Ras. GTPcNH2 was prepared by the method described by Knorre et al. [31] for ATP derivatives. c-Amide derivatives of GTP were des- cribed by Babkina et al. who used, e.g. the c-(4-azido) anilide of GTP to substitute efficiently for GTP as a photoaffinity label in the elongation factor protein EF-Tu [32]. The GTPcF substrate analogue was first prepared by Eckstein et al. [9], by the method of Haley & Yount [33], and used to study its interaction with the GTP-binding site of adenylyl cyclase [34]. We synthesized this substance by

31P chemical shift (p.p.m.)

Table 5. 31P chemical shifts of Ras(1)167)ÆGTPcNH2ÆMg2+ at pH 7.4, 5 (cid:1)C. Spectra were recorded in 25 mM Tris/HCl buffer, 2.5 mM Mg2+ and 1 mM DTE. The splitting of the a- and b-phosphate resonance lines in protein bound triphosphate-nucleotides is a specific feature of the Ras protein, indicating different conformations of the active center [24].

Proteinnucleotide c a(1) a(2) b(1) b(2) into two states the b-phosphate resonance – – RasGTPcNH2Mg2+ RafRasGTPcNH2Mg2+ RasGppNHpMg2+ )11.80 )16.15 )16.85 1.90 )16.81 )11.76 – 1.07 )3.41 )0.32 )2.69 )11.15 )11.85 Fig. 7. 31P NMR spectra of protein bound RasÆGTPcNH2ÆMg2+ and concentration series with Raf-RBD. The ratio of Raf-RBD to Ras varies from 0 (bottom spectrum) to 1 (top spectrum) as indicated on the right. Spectra were recorded at pH 7.4 and 5 (cid:2)C in 25 mM Tris/ HCl buffer, 2.5 mM MgCl2 and 1 mM dithioerythritol. Note the splitting of for RasÆGTPcNH2ÆMg2+ (bottom spectrum) and the stabilization of one conformation upon complexation with Raf. Excess of free and bound phosphate groups are labelled.

(cid:3) FEBS 2002

Nucleotide analogues GTPcNH2 and GTPcF (Eur. J. Biochem. 269) 3277

A C K N O W L E D G E M E N T S

the simple method of Wittmann [10], which works very well with adenine nucleotides. The presence of a guanine base gives rise to the formation of yellow by-products, probably due to the reaction of 2,4-dinitrofluorobenzene with the amino group of GTP. Thus in this case the simplicity of the method is at the expense of yield.

We thank John Wray and Roger S. Goody for discussions and Ulrich Haeberlen, Alfred Wittinghofer and Kenneth C. Holmes for continu- ous support. M.G. acknowledges support by the Peter und Traudl Engelhorn Stiftung (Penzberg, Germany).

R E F E R E N C E S

the natural substrate [20,35],

1. Bar-Sagi, D. & Hall, A. (2000) Ras and Rho GTPases: a family reunion. Cell 103, 227–238. 2. Nakielny, S. & Dreyfuss, G. (1999) Transport of proteins and RNAs in and out of the nucleus. Cell 99, 677–690. 3. Zerial, M. & McBride, H. (2001) Rab proteins as membrane organizers. Nat. Rev. Mol. Cell Biol. 2, 107–117.

4. Scheidig, A.J., Burmester, C. & Goody, R.S. (1998) Use of caged nucleotides to characterize unstable intermediates by X-ray crys- tallography. Methods Enzymol. 291, 251–264.

5. Ahmadian, M.R., Zor, T., Vogt, D., Kabsch, W., Selinger, Z., Wittinghofer, A. & Scheffzek, K. (1999) Guanosine triphos- phatase stimulation of oncogenic Ras mutants. Proc. Natl Acad. Sci. USA 96, 7065–7070.

6. Gruen, M., Becker, C., Beste, A., Reinstein, J., Scheidig, A.J. & Goody, R.S. (1999) 2¢Halo-ATP and -GTP analogues: rational phasing tools for protein crystallography. Protein Sci. 8, 2524– 2528. 7. Wittinghofer, A. (1997) Signaling mechanistics: aluminum fluoride for molecule of the year. Curr. Biol. 7, 682–685.

The kinetic parameters determined reveal the biochemi- cal properties of the two nucleotide analogues. While the activation energy of magnesium binding for GTPcNH2 is similar to that of the association constant Ka for magnesium is significantly smaller for both nucleotides analysed. Therefore, the more stable c-amido triphosphate analogue may be particularly useful for the study of the role of divalent cation binding, e.g. to analyse a proposed reaction mechanism. For example, the influence of the magnesium binding affinity on the kinetics of the Ras guanine nucleotide exchange factor Sos [36] can be studied with the nucleoside diphosphate GDPcNH2 derivative loaded onto Ras. Additionally, an intermediate magnesium-free state may be stabilized more easily with GTPcNH2 or ATPcNH2 analogues. The analogue ATPcNH2 may provide new insights into the equilibrium between different conforma- tions of myosin [37]. Finally, specific labeling of the amide group with 15N isotopes may be useful for nitrogen selective heteronuclear NOE experiments for the structural analysis of the active center in solution. In combination with specific labeling of single residues in the protein [38] this may yield detailed insights into the dynamics of the nucleotide binding site.

8. Schlichting, I. & Reinstein, J. (1999) pH influences fluoride coordination number of the AlFx phosphoryl transfer transition state analog. Nat. Struct. Biol. 6, 721–723.

9. Eckstein, F., Bruns, W. & Parmeggiani, A. (1975) Synthesis of guanosine 5¢-di- and -triphosphate derivatives with modified terminal phosphates: effect on ribosome-elongation factor G-dependent reactions. Biochemistry 14, 5225–5238. 10. Wittmann, R. (1962) Neue Synthese von Monofluor-Phos- phorsa¨ ure-Monoestern. Angew. Chem. 74, 214–215. 11. Brauer, G., ed. (1975) Handbuch der Pra¨parativen Anorganischen Chemie, pp. 558–559. Ferdinand Enke Verlag, Stuttgart.

12. IUPAC (1998) Recommendations for the presentation of NMR structures of proteins and nucleic acids. Pure Appl. Chem. 70, 117–142.

13. Nageswara Rao, B.D. (1989) Nuclear magnetic resonance line- shape analysis and determination of exchange rates. Methods Enzymol. 176, 279–311.

The GTPcF analogue may be particularly useful because of the high sensitivity of 19F NMR spectroscopy. In a previous report the GTP binding site of the 110 kDa protein tubulin was studied using the fluorotriphosphate [39]. Here, fluorine relaxation rates were determined to analyse the location of the divalent cation site relative to the exchangable nucleotide. We are analysing the possi- bility of two distinct nucleotide binding sites in the human guanylate-binding protein 1 (hGBP1) [40] by titration of GTPcF or GDPbF to the noncomplexed protein, assu- ming different chemical environments for each putative nucleotide binding site (data not shown). Finally, the suitability of both nucleotide derivatives for the use in solid state NMR spectroscopy should be tested in future experiments.

For

14. Smith, S.A., Levante, T.O., Meier, B.H. & Ernst, R.R. (1994) Computer simulations in magnetic resonance. An object-oriented programming approach. J. Magn. Reson. 106, 75–105.

15. Wiseman, T., Williston, S., Brandts, J.F. & Lin, L.N. (1989) Rapid measurement of binding constants and heats of binding using a new titration calorimeter. Anal. Biochem. 179, 131–137.

16. Franken, S.M., Scheidig, A.J., Krengel, U., Rensland, H., Lautwein, A., Geyer, M., Scheffzek, K., Goody, R.S., Kalbit- zer, H.R., Pai, E.F. & Wittinghofer, A. (1993) Three-dimen- sional structures and properties of a transforming and a nontransforming glycine-12 mutant of p21h–Ras. Biochemistry 32, 8411–8420.

17. Schweins, T., Geyer, M., Scheffzek, K., Warshel, A., Kalbitzer, H.R. & Wittinghofer, A. (1995) Substrate-assisted catalysis as a mechanism for GTP hydrolysis of p21ras and other GTP-binding proteins. Nat. Struct. Biol. 2, 36–44.

the slowly hydrolysing Ras protein the two nucleotides described here close the 10-fold gap in intrinsic hydrolysis rates between bound GTP and GTPcS, and GTPcS and GppNHp (Table 4). This broad variety allows an almost individual selection of hydrolysis stability from protein bound GTP to protein bound GppCH2p for all different purposes. The application onto the GTP-binding protein Ras confirmed that the flexibility of the effector loop of Ras is preserved (Fig. 7). For nucleotide binding proteins the flexibility of the nucleotide binding site in its active center is a vital prerequisite for the dynamic function of the protein. This is particularly important for mechanistic studies and the analysis of binding intermedi- ates, as shown for the complex formation of Ras with the effector protein Raf-RBD [24,41]. With these two deriv- atives characterized we extend the variety of nucleotide analogues available for kinetic, structural, and cellular studies.

18. Herrmann, C., Martin, G.A. & Wittinghofer, A. (1995) Quanti- tative analysis of the complex between p21ras and the Ras-binding domain of the human Raf-1 protein kinase. J. Biol. Chem. 270, 2901–2905. 19. Wu¨ thrich, K. (1986) NMR of Proteins and Nucleic Acids. John Wiley &. Sons, Inc, New York.

3278 M. Stumber et al. (Eur. J. Biochem. 269)

(cid:3) FEBS 2002

32. Babkina, G.T., Jonak, J., Karpova, G.G., Knorre, D.G., Rychlik, I. & Vladimirov, S.N. (1988) Affinity labeling at the A-site of Escherichia coli ribosomes by a non-hydrolyzable c-amide analog of GTP. Biochimie 70, 597–603.

20. Vasavada, K.V., Ray, B.D. & Nageswara Rao, B.D. (1984) 31P NMR lineshapes of b-P (ATP) in the presence of Mg2+ and Ca2+: estimate of exchange rates. J. Inorg. Biochem. 21, 323–335. 21. Campbell, S.L., Khosravi-Far, R., Rossman, K.L., Clark, G.J. & Der, C.J. (1998) Increasing complexity of Ras signaling. Oncogene 17, 1395–1413. 22. Kuhlmann, J. & Herrmann, C. (2001) Biophysical characteriza- 33. Haley, B. & Yount, R.G. (1972) c-Fluoroadenosine triphosphate. Synthesis, properties, and interaction with myosin and heavy meromyosin. Biochemistry 11, 2863–2871. tion of the Ras protein. Top. Curr. Chem. 211, 61–116.

34. Pfeuffer, T. & Eckstein, F. (1976) Topology of the GTP-binding site of adenylyl cyclase from pigeon erythrocytes. FEBS Lett. 67, 354–358. 23. John, J., Sohmen, R., Feuerstein, J., Linke, R., Wittinghofer, A. & Goody, R.S. (1990) Kinetics of interaction of nucleotides with nucleotide-free H-ras p21. Biochemistry 29, 6058–6065.

35. Klaus, W., Schlichting, I., Goody, R.S., Wittinghofer, A., Rosch, P. & Holmes, K.C. (1986) 31P-NMR studies on ATP X Mg2+, p21 X nucleotide, and adenylate kinase X nucleotide complexes. Chem- ical shifts, rate and equilibrium constants. Biol. Chem. Hoppe Seyler 367, 781–786.

36. Cherfils, J. & Chardin, P. (1999) GEFs: structural basis for their activation of small GTP-binding proteins. Trends Biochem. Sci. 24, 306–311.

24. Geyer,M.,Schweins,T.,Herrmann,C.,Prisner,T.,Wittinghofer,A. & Kalbitzer, H.R. (1996) Conformational transitions in p21ras and in its complexes with the effector protein Raf-RBD and the GTPase activating protein GAP. Biochemistry 35, 10308–10320. 25. Ito, Y., Yamasaki, K., Iwahara, J., Terada, T., Kamiya, A., Shirouzu, M., Muto, Y., Kawai, G., Yokoyama, S., Laue, E.D., Wa¨ lchli, M., Shibata, T., Nishimura, S. & Miyazawa, T. (1997) Regional polysterism in the GTP-bound form of the human c-Ha- Ras protein. Biochemistry 36, 9109–9119. 37. Wray, J. & Jahn, W. (2002) c-Amido-ATP stabilizes a high- fluorescence state of myosin subfragment 1. FEBS Lett. 518, 97–100.

26. Hu, J.S. & Redfield, A.G. (1997) Conformational and dynamic differences between N-ras P21 bound to GTPcS and to GMPPNP as studied by NMR. Biochemistry 36, 5045–5052.

38. Yabuki, T., Kigawa, T., Dohmae, N., Takio, K., Terada, T., Ito, Y., Laue, E.D., Cooper, J.A., Kainosho, M. & Yokoyama, S. (1998) Dual amino acid-selective and site-directed stable-isotope labeling of the human c-Ha-Ras protein by cell-free synthesis. J. Biomol. NMR 11, 295–306. 27. Geyer, M. & Wittinghofer, A. (1997) GEFs, GAPs, GDIs and effectors: taking a closer (3D) look at the regulation of Ras-related GTP-binding proteins. Curr. Opin. Struct. Biol. 7, 786–792. 28. Wittinghofer, A. & Nassar, N. (1996) How Ras-related proteins talk to their effectors. Trends Biochem. Sci. 21, 488–491. 39. Monasterio, O. (1987) 19F nuclear magnetic resonance measure- ment of the distance between the E-site GTP and the high-affinity Mg2+ in tubulin. Biochemistry 26, 6099–6106.

29. Geyer, M., Assheuer, R., Klebe, C., Kuhlmann, J., Becker, J., Wittinghofer, A. & Kalbitzer, H.R. (1999) Conformational states of the nuclear GTP-binding protein Ran and its complexes with the exchange factor RCC1 and the effector protein RanBP1. Biochemistry 38, 11250–11260. 40. Praefcke, G.J.K., Geyer, M., Schwemmle, M., Kalbitzer, H.R. & Herrmann, C. (1999) Nucleotide-binding characteristics of human guanylate-binding protein 1 (hGBP1) and identification of the third GTP-binding motif. J. Mol. Biol. 292, 321–332.

41. Terada, T., Ito, Y., Shirouzu, M., Tateno, M., Hashimoto, K., Kigawa, T., Ebisuzaki, T., Takio, K., Shibata, T., Yokoyama, S., Smith, B.O., Laue, E.D. & Cooper, J.A. (1999) Nuclear magnetic resonance and molecular dynamics studies on the interactions of the Ras-binding domain of Raf-1 with wild-type and mutant Ras proteins. J. Mol. Biol. 286, 219–232. 30. Cherfils, J., Menetrey, J., Le Bras, G., Janoueix-Lerosey, I., de Gunzburg, J., Garel, J.R. & Auzat, I. (1997) Crystal structures of the small G protein Rap2A in complex with its substrate GTP, with GDP and with GTPgammaS. EMBO J. 16, 5582–5591. 31. Knorre, D.G., Kurbatov, V.A. & Samukow, V.V. (1976) General method for the synthesis of ATP c-derivatives. FEBS Lett. 70, 105–108.