Exquisite Modulation of the Active Site of Methanocaldococcus

Apr 6, 2016 - ... Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore 560064, India. Telephone: +91-80-22082812. E-mail: [email protected]...
0 downloads 16 Views 2MB Size
Download PDF
Article pubs.acs.org/biochemistry

Exquisite Modulation of the Active Site of Methanocaldococcus jannaschii Adenylosuccinate Synthetase in Forward Reaction Complexes Vishakha Karnawat,† Sonali Mehrotra,‡ Hemalatha Balaram,*,‡ and Mrinalini Puranik*,† †

Indian Institute of Science Education and Research, Pune 411008, India Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore 560064, India

‡

S Supporting Information *

ABSTRACT: In enzymes that conduct complex reactions involving several substrates and chemical transformations, the active site must reorganize at each step to complement the transition state of that chemical step. Adenylosuccinate synthetase (ADSS) utilizes a molecule each of guanosine 5′-monophosphate (GTP) and aspartate to convert inosine 5′-monophosphate (IMP) into succinyl adenosine 5′-monophosphate (sAMP) through several kinetic intermediates. Here we followed catalysis by ADSS through high-resolution vibrational spectral fingerprints of each substrate and intermediate involved in the forward reaction. Vibrational spectra show differential ligand distortion at each step of catalysis, and band positions of substrates are influenced by binding of cosubstrates. We found that the bound IMP is distorted toward its N1-deprotonated form even in the absence of any other ligands. Several specific interactions between GTP and active-site amino acid residues result in large Raman shifts and contribute substantially to intrinsic binding energy. When both IMP and GTP are simultaneously bound to ADSS, IMP is converted into an intermediate 6-phosphoryl inosine 5′-monophosphate (6-pIMP). The 6-pIMP·ADSS complex was found to be stable upon binding of the third ligand, hadacidin (HDA), an analogue of L-aspartate. We find that in the absence of HDA, 6pIMP is quickly released from ADSS, is unstable in solution, and converts back into IMP. HDA allosterically stabilizes ADSS through local conformational rearrangements. We captured this complex and determined the spectra and structure of 6-pIMP in its enzyme-bound state. These results provide important insights into the exquisite tuning of active-site interactions with changing substrate at each kinetic step of catalysis. auling first proposed that enzyme catalysis takes place by distortion of substrates in sequential steps, which leads to the formation of a transition state.1 Understanding these substrate distortions is fundamentally important to deducing catalytic mechanisms and also to designing effective substrate analogues for the enzyme. Efficacy of substrate analogues as inhibitors could be enhanced if the allosteric interactions of the enzyme are understood well and can be exploited to further stabilize nonproductive complexes of the enzyme. X-ray crystal structures have been very enlightening in providing information about various enzyme·substrate (ES) complexes and the interactions between them. While X-ray crystal structures are indispensible for initial high-resolution direct information about the ES complex, these are static snapshots. To obtain information about the different steps of catalysis, a laborious

P

© 2016 American Chemical Society

task of crystallizing the same enzyme with various analogues must be performed, yet information about the dynamical equilibrium in conformational states of ES complexes in solution remains inaccessible. Thus, there is a need to apply alternative high-resolution techniques that can provide rapid information about very specific molecular interactions between substrates and enzymes in solution. Knowledge of the strength of these interactions and the resulting structural changes in the substrate is fundamental to understanding the origin and mechanism of catalysis. A powerful but infrequently employed Received: December 31, 2015 Revised: March 24, 2016 Published: April 6, 2016 2491

DOI: 10.1021/acs.biochem.5b01386 Biochemistry 2016, 55, 2491−2499

Article

Biochemistry

to its expulsion. At the same time, Asp43 reverts a proton to the N1 position of the nucleotide, which results in the resonance stabilization of C6 carbonium ion. This leads to the nucleophilic attack of L-aspartate at the C6 position, leading to the formation of sAMP.4 In this catalytic mechanism, Mg2+ coordinates with the α- and β-phosphoryl group of GDP, the oxygen atom of the 6-phosphoryl group, L-aspartate, and the carbonyl of Gly70. Thus, Mg2+ ion helps in the assembly of the reactants to facilitate the final catalytic step.7 Extensive structural studies at various steps of the catalytic cycles in Escherichia coli (Ec) and mouse muscle (mm) ADSS are available.6,8−13 These studies suggest that unligated ADSS is disordered while IMP alone is capable of organizing the active site.12−14 By the effective use of hadacidin (HDA), an analogue of L-aspartate, along with IMP and GTP, a fully ligated conformation of ADSS was achieved. ADSSs from many organisms, viz., E. coli, Plasmodium falciparum, mouse, and plant, have been found to involve movement of dynamic loops upon ligand binding. These loops in mmADSS have been divided into the preswitch loop (65−69), the switch loop (70− 83), the IMP loop (152−165), the Val loop (304−310), the GTP loop (448−452), and the Asp loop (330−336).8 The largest displacement of 9 Å in mmADSS and EcADSS is observed in the switch loop.8,15 Conformational change in loop 299−304 is driven by the binding of L-aspartate or HDA.15 A study with analogues of GTP shows that the enzyme does not undergo any conformational change in response to the binding of guanine nucleotide: IMP and/or Mg2+ is required for the recognition of the pyrophosphate group of GTP.13 We have employed UVRR to study enzyme−nucleobase interactions. A Raman spectrum can probe a subtle change in the molecular geometry of the substrate that occurs upon binding to the enzyme.16−23 It reports on the conformation, protonation state, and structure of nucleotide molecules at the active site of enzymes. Here, we have captured distortions in complexes of Methanocaldococcus jannaschii ADSS (MjADSS) with substrate, substrate analogues, and intermediates. We also report on the extent of distortion in these ligands upon binding to MjADSS. Understanding distortion in these molecules provides an insight into the driving force of ADSS-catalyzed reaction. In view of the absence of the MjADSS structure, we have used structural information from Pyrococcus horikoshii (Ph) ADSS, mmADSS, and EcADSS to interpret UVRR spectral data. These interpretations are extrapolated to the structure and function of MjADSS, and distortion of ligands is inferred. Multiple-sequence alignment of PhADSS and MjADSS shows a sequence identity of 68% between the two.24 Active-site residues that play a key role in catalysis and in formation of noncovalent interactions with the ligand are conserved in PhADSS, EcADSS, mmADSS, and MjADSS.24

technique for obtaining such information is vibrational spectroscopy. We have demonstrated the utility of ultraviolet resonance Raman spectroscopy (UVRR) on the enzymes of the nucleotide synthesis pathway in informing on substrate distortion during enzyme catalysis.2,3 In this account, we utilize this technique to understand substrate binding and dynamics in adenylosuccinate synthetase (ADSS) as it proceeds through a complex kinetic cycle involving several cosubstrates. ADSS is an essential enzyme of the de novo nucleotide synthesis pathway that performs the conversion of inosine 5′monophosphate (IMP) (Figure 1) to succinyl adenosine 5′-

Figure 1. Ligands of MjADSS at physiological pH. AMP and HDA act as competitive inhibitors of IMP and L-aspartate, respectively.

monophosphate (sAMP). The catalytic mechanism of ADSS is a two-step process4−6 (Figure 2); the first step is a phosphoryl

Figure 2. Reaction catalyzed by ADSS. Asp43 abstracts a proton from N1H of the purine ring, resulting in the formation of 6-oxy anionic IMP. Electrophilic attack by γ-phosphate of GTP at O6 of IMP results in the formation of an intermediate, 6-phosphoryl IMP. In the subsequent step, L-aspartate replaces the phosphoryl group to form the product sAMP. RP stands for the ribose sugar and phosphate group.

■

transfer reaction in which Asp43 (mouse muscle ADSS numbering is used unless otherwise specified) acts as a catalytic base and abstracts a proton from N1-H of IMP leading to the formation of 6-oxy anionic IMP. His71, on the other hand, provides a proton from its ND1 position to the leaving βphosphate group of GDP. The γ-phosphate group of GTP is transferred to the 6-oxo position of IMP, resulting in the formation of an intermediate 6-phosphoryl IMP (6-pIMP). This involves the first transition state of the reaction. The second step is the phosphoryl displacement reaction in which the nucleophilic displacement of the 6-phosphoryl group by Laspartate takes place to form sAMP. In this step, Asp43 and His71 both act as a catalytic acid. His71, from the NE2 position, donates a proton to the 6-phosphoryl group, leading

METHODS Enzyme Preparation. MjADSS was overexpressed in E. coli BL21(DE3) cells that had been transformed with plasmid pET23d containing the gene of interest. The enzyme was purified as reported previously.24 The concentration of the enzyme was estimated by Bradford’s method using bovine serum albumin as a standard.25 The activity of the enzyme was determined after each batch of enzyme preparation. The assay was conducted at 70 °C using a Peltier-controlled cell holder fitted to a UV−visible spectrophotometer (Varian Cary 100, Agilent Technologies, Santa Clara, CA). A reaction mixture containing 30 mM MES (pH 6.5), 15 mM magnesium acetate, 2492

DOI: 10.1021/acs.biochem.5b01386 Biochemistry 2016, 55, 2491−2499

Article

Biochemistry 10 mM L-aspartate, and 500 μM IMP along with 0.65−0.85 μg of MjADSS was preincubated at 70 °C for 30 s before the kinetic assay was begin. The reaction was initiated by adding 250 μM GTP. Conversion of IMP to sAMP was detected by monitoring an increase in absorbance at 290 nm for 5 min. The Δε value used for this calculation is 3390 M−1 cm−1.24 ADSS Sample Preparation. IMP, GTP, L-aspartate, magnesium acetate, and sodium nitrate were purchased from Sigma-Aldrich. HDA was procured from the Develomental Therapeutic Program of the National Institutes of Health. UVRR samples were prepared in 30 mM MES (pH 6.5). UVRR spectra of the enzyme and nucleobase were acquired in buffer containing 5 mM magnesium acetate, 30 mM NaNO3, and 10 mM HDA; 500 μM nucleobase and 150 μM enzyme were used for the Raman experiments. Deuterium Isotopic Labeling. The enzyme ADSS was exchanged with D2O buffer as described previously.2 The stocks of nucleotides, magnesium acetate, and NaNO3 were prepared in D2O. The full Raman spectra recorded in D2O MES buffer are not shown, but the shifts obtained are discussed to support the spectra of H2O MES buffer. UVRR. The UVRR spectra were acquired using a 260 nm excitation wavelength generated by a tunable Ti-Sapphire laser (Indigo, Coherent Inc., Santa Clara, CA). A detailed description has been provided previously.26 The typical, average power at the sample was ∼0.6 mW. Calibration was done with the known band positions of solvents dimethylformamide, cyclohexane, indene, acetonitrile, trichloroethylene, and 2propanol. Data Analysis. Each spectrum is the average of three samples each of which was acquired over ∼15 min. The absence of photodamage in the sample upon exposure to the laser was ensured by comparing the first and last spectra acquired. SynerJY version 7.0 (Horiba Jobin-Yvon, Kyoto, Japan) was used for data acquisition and analysis. A Lorentzian line shape function was used to fit all UVRR spectra and obtain the wavenumber in each spectrum. The area under the curve was used as a measure of intensity analysis. Th wavenumber mentioned in the text and tables corresponds to the average wavenumber obtained from three data sets, whereas the wavenumbers in figures correspond to the positions of bands obtained on a particular data set. Subtraction Protocol. Contributions of apoenzyme and unbound ligand were removed from the spectrum of each enzyme−ligand complex.3 NaNO3 was added as an internal standard in all the samples. Its band at 1048 cm−1 was used as a reference band to remove the spectral contribution from the enzyme and unbound nucleobase. The spectrum of MjADSS (A) was subtracted from the spectrum of the nucleotide· MjADSS complex (B) to obtain the spectrum (C) using the NaNO3 band as an internal reference. Spectrum C contains the contribution from both bound and unbound nucleotide. In the second step, the contribution from unbound nucleotide was estimated using NaNO3 to normalize the spectrum (B) and the spectrum of the nucleotide (free) in solution. In the final step, the contribution of the unbound nucleotide was subtracted from spectrum C using the numerical factor obtained in the second step.

Figure 3. Resonance Raman spectra (λexc = 260 nm) of guanine nucleotides bound to MjADSS: (a) free GTP and (b) GTP·MjADSS, obtained in H2O (MES buffer, pH 6.5). (c) Chemical structure of GTP (RP3, ribose triphosphate). (d) Free GTP and (e) GTP· MjADSS, obtained in D2O (MES buffer, pD 6.5).

this excitation wavelength, while spectra of the protein, sugar, and phosphate do not contribute significantly to the Raman signal because of their small cross section at this wavelength.27−29 In previous work, we have demonstrated that resonance Raman spectra of enzyme-bound nucleotides and nucleobases report on the protonation state and noncovalent interactions of the ligand.2,29 In this work, we report on perturbations in the structure of bound ligands at the active site of MjADSS through observed spectral shifts and intensity modulation in the UVRR spectra. We obtained UVRR spectra of bound ligands by subtracting the spectra of the apoenzyme and unbound ligands from those of the enzyme−ligand complexes (see Data Analysis). Raman band assignments of IMP and GTP were taken from previous studies of IMP and guanine26,30,31 (Table S1). GTP·MjADSS Complex: GTP Forms Strong Interactions with MjADSS. The GTP·MjADSS complex shows several large shifts in the vibrational modes of the guanine moiety (Figure 3a,b). This implies that the guanine nucleobase is held tightly by several noncovalent interactions with the protein. We note, however, that the role of GTP is to donate the γ-phosphate and the interactions of the guanine base with the protein are primarily for recognition. The band at 1684 cm−1 is assigned to the carbonyl stretch (str) coupled to N1-H bending vibrations (bend). The downshift of 28 cm−1 observed in this band is attributed to hydrogen bonding interactions formed by C6O of the purine ring with a backbone amide of Lys363 and Gly447.10 In D2O, the corresponding carbonyl str in GTP at 1664 cm−1 shows a downshift of 25 cm−1 in the GTP·MjADSS complex (Figure 3d,e). The band at 1604 cm−1 in the GTP spectrum (N1-H bend and C2-N2 str) upshifts by 11 cm−1 in the GTP·MjADSS complex. Hydrogen bonding between Asp365 with N1-H and exocyclic -NH2 of the guanine ring results in the decoupling of the N1-H bend, causing the resultant vibration to shift to a higher wavenumber. An upshift of N-H bending modes upon hydrogen bonding has been observed before, e.g., in adenosine deaminases by Callender and co-workers.22

■

RESULTS UVRR spectra of enzyme·nucleotide complexes were acquired at a Raman excitation wavelength of 260 nm (Figure 3). Vibrational modes of nucleobases are resonance-enhanced at 2493

DOI: 10.1021/acs.biochem.5b01386 Biochemistry 2016, 55, 2491−2499

Article

Biochemistry

Figure 4. Resonance Raman spectra (λexc = 260 nm) of forward reaction complexes of MjADSS in MES buffer in H2O at pH 6.5: (a) IMP free in solution, (b) IMP bound to MjADSS, (c) IMP and GTP in solution, (d) IMP and GTP bound to MjADSS, (e) IMP, GTP, and HDA in solution, and (f) IMP, GTP, and HDA bound to MjADSS. Their corresponding spectra in MES buffer in D2O at pD 6.5 are shown as spectra i−n. The IMP band at 1506 cm−1 that loses intensity in the IMP·GTP·MjADSS complex is marked with a dashed line. (g and h) Chemical structures of IMP and 6pIMP, respectively.

(Figure 3b). A similar upshift is observed when IMP is deprotonated at N1-H at pH 12.30 Thus, an upshift in the band at 1468 cm−1 indicates a distortion toward deprotonation. The triene str mode appears as an intense band at 1555 cm−1. Binding of IMP to ADSS not only results in the downshift of 5 cm−1 but also produces a marked loss of intensity. This band is known to lose intensity upon deprotonation and appears at 1512 cm−1 in deprotonated IMP.30 Thus, the loss of intensity at 1550 cm−1 is also consistent with distortion toward deprotonated IMP in the IMP·MjADSS complex (Figure 4b). A pyrimidine ring str mode is observed at 1595 cm−1 in IMP and downshifts by 3 cm−1 in the IMP·MjADSS complex. This downshift is the consequence of the formation of hydrogen bonding interactions between the N1-H moiety of the pyrimidine ring and the active-site residue Asp43 and consequent weakening of the N-H bond (Figure S1a).10 IMP·GTP·MjADSS: 6-pIMP Formation Occurs in the Absence of L-Aspartate. Extensive kinetic studies of ADSSs from various organisms such as E. coli,32 mouse,8 Zea mays,33 and Saccharomyces cerevisiae34 suggest that IMP and GTP bind in random order to the enzyme at distinct binding sites. Once both ligands are bound, structural studies show that the first catalytic intermediate, 6-pIMP, is formed even in the absence of 4,5 L-aspartate in these enzymes. By contrast, in MjADSS, it was reported that the presence of HDA was required for complete utilization of GTP, which was implicated in the formation of 6pIMP.35 To examine whether HDA plays an allosteric or direct role in the catalysis, we decided to observe changes in the structure of IMP in the presence and absence of HDA. Spectra of IMP·GTP·MjADSS and IMP·GTP·HDA·MjADSS com-

The pyrimidine ring str coupled to -NH2 scissoring (sci) and the N1-H bend appears at 1578 cm−1 and downshifts to 1571 cm−1 in the GTP·MjADSS complex. Asp333 interacts with -NH2 and with N1-H of GTP, which is likely the cause of the observed downshift at the active site. The imidazole ring mode appears as the most intense band at 1487 cm−1 and downshifts by 6 cm−1 in the GTP·MjADSS complex. The corresponding band in D2O also downshifts by 3 cm−1 in the GTP·MjADSS complex. The X-ray structure of EcADSS reports on the interactions between the hydroxyl group on the side chain of Ser414 and N7 (D−A, 2.8 Å).12 The large magnitude of the shifts in the spectrum of the GTP·MjADSS complex with respect to GTP in solution indicates that the guanine nucleobase of GTP is held tightly by several noncovalent interactions at the active site (Figure S1). These shifts in the Raman wavenumber are implicated in the formation of noncovalent interactions in the GTP·MjADSS complex and thereby provide a significant amount of intrinsic binding energy to the enzyme−substrate complex. Furthermore, although shifts in wavenumber are observed, the spectral pattern of the GTP at the active site of MjADSS is preserved. The absence of any radical changes in the GTP·MjADSS complex spectrum corroborates the current catalytic mechanism of ADSS in which no chemical transformation is observed at the nucleobase moiety.4 IMP·MjADSS Complex: MjADSS Distorts IMP Structure toward Its Deprotonated Form. In the following, we discuss the shifts in three most intense bands of IMP upon binding to the active site of MjADSS: 1468, 1555, and 1595 cm−1. The band at 1468 cm−1 is primarily a N1-H bend (Figure 4a). This band upshifts to 1473 cm−1 in the IMP·MjADSS complex 2494

DOI: 10.1021/acs.biochem.5b01386 Biochemistry 2016, 55, 2491−2499

Article

Biochemistry

Figure 5. Active-site interactions of 6-pIMP (a) in the absence of HDA [Protein Data Bank (PDB) entry 1LNY, 2.20 Å] and (b) in the presence of HDA (PDB entry 1LON, 2.10 Å). The orientation of amino acid residues is similar in the two cases. (c) Cartoon depicting the effect of binding of HDA to the 6-pIMP·ADSS complex resulting in local conformational reorganization of the active site.

complex, even in the absence of the ligation of the L-aspartate pocket. However, the identity of the intermediate is only established by the following experiment in which the active site is consolidated by the binding of HDA. GTP in IMP·GTP·MjADSS. Because guanine does not participate in the catalysis, the bands corresponding to GTP retain their identity in IMP·GTP·MjADSS as expected. The bands at 1487 and 1684 cm−1 show downshifts of 6 and 27 cm−1, respectively. Similarly, the band at 1604 cm−1 shows an upshift of 12 cm−1. IMP·GTP·HDA·MjADSS: The Presence of HDA Stabilizes the Reaction Intermediate, 6-pIMP, at the Active Site of MjADSS. HDA does not absorb light at 260 nm. Consequently, its spectrum is not resonance-enhanced at this excitation wavelength and does not confound the spectrum of the ES complex. In the absence of the enzyme, the spectra of (IMP+GTP) and (IMP+GTP+HDA) are identical (Figure 4c,e). The presence of the enzyme alters the spectrum considerably, and differences between the spectra of (IMP +GTP+HDA) and the corresponding enzyme complex (Figure 4f) in MES buffer (pH 6.5) are conspicuous. In Figure 3f, while the vibrations of GTP are not influenced significantly, bands of IMP at 1555 and 1596 cm−1 disappear and its ring vibration at 1468 cm−1 downshifts to 1455 cm−1. An intense new band at 1499 cm−1 is observed in the presence of HDA. A band centered at 1363 cm−1 that contains contributions from both IMP (1350 cm−1) and GTP (1366 cm−1) is replaced by two new bands at 1347 and 1363 cm−1. Such radical changes in the IMP·GTP·HDA·MjADSS complex spectrum indicate transformation of IMP to a new species that is most likely 6-pIMP (see the next section). The band at 1318 cm−1 in Figure 4e has a contribution from IMP as well as GTP. This band downshifts by 4 cm−1 in IMP· MjADSS and GTP·MjADSS complexes, but in the IMP·GTP· HDA·MjADSS complex, this band upshifts by 4 cm−1. The band also gains 2.5 times the intensity of free IMP and GTP.

plexes were obtained by subtracting the spectra of the apoenzyme and unbound ligands from the ES spectrum and were compared. In the UVRR spectrum of (free IMP+free GTP), a few bands belonging to IMP and GTP overlap because of similar normal modes of vibrations while a few other bands are clear and can be used as specific markers of the ligands. For example, the bands at 1468, 1555, and 1595 cm−1 are markers of IMP, whereas the bands at 1487, 1578, and 1604 cm−1 are markers of GTP (Figure 4c). The spectrum of the IMP·GTP·MjADSS complex (Figure 4d) evidently shows that the bands of GTP are preserved whereas the bands of IMP either show downshifts or disappear. For instance, the IMP bands at 1555 and 1595 cm−1 in Figure 4c show downshifts of 7 cm−1 (the area under the curve is also reduced) and 3 cm−1, respectively, in Figure 3d. In the IMP·GTP·MjADSS complex, the IMP band at 1468 cm−1 disappears, and instead, a new band in the form of a shoulder at 1497 cm−1 can be observed. This new band arises from the conversion of IMP into a new species that is the next intermediate in the catalysis, 6-pIMP. An equivalent experiment in D2O confirms the presence of a new species because the corresponding IMP band at 1506 cm−1 in (IMP in D2O) shows a dramatic loss of intensity in the complex (Figures 4i and 5j). In the catalytic mechanism of ADSSs, in preparation for the formation of 6-pIMP, Asp43 abstracts a proton from N1 of the purine ring.4,5,10 IMP anion has a band at 1512 cm−1 that is close to the band present at 1497 cm−1.30 A similar observation of change in protonation state was also reported by Callender and co-workers in adenosine-5′-diphosphate ribose bound to malate dehydrogenase.19 In this study, the appearance of a new band in the region from 1300 to 1400 cm−1 was interpreted as the protonation of the N3 position of the adenine moiety by a proximal aspartate residue. On the basis of previously proposed reaction mechanisms, the data presented here suggest that the formation of 6-pIMP is initiated in the IMP·GTP·MjADSS 2495

DOI: 10.1021/acs.biochem.5b01386 Biochemistry 2016, 55, 2491−2499

Article

Biochemistry

observed in these control experiments. All IMP bands are intact in the spectra of IMP·GDP·MjADSS, IMP·GDP·L-aspartate· MjADSS, and IMP·GMP-PNP·HDA·MjADSS complexes, and no new bands are observed in these ES complexes. Previous studies show that the presence of GDP, IMP, and Mg2+ is sufficient for complete organization of dynamic loops into the active conformation.13 In the study presented here, the IMP bands, in the presence of GDP/GMP-PNP, show shifts that are similar to those observed in the (IMP+GTP+MjADSS) complex. Our results show that the guanine moiety in GTP is most likely to be the allosteric effector of the IMP binding site. In general, although the bands in the Raman spectrum of the ES complex show a positional shift with respect to the free ligand spectrum, the spectral pattern is preserved. Only if a covalent bond is formed or broken does the spectral pattern change dramatically. In the IMP·GTP·HDA·MjADSS complex, where 6-pIMP is formed, the aromaticity of the purine ring is expected to change considerably, and this leads to the large change in the vibrational modes. The difference between (IMP +GTP+HDA) and IMP·GTP·HDA·MjADSS spectra is so radical that the mere environmental effect of the active-site residues on highly polarizable groups like CO, CN, O−H, and N−H (enzyme-induced electronic perturbations) cannot account for it. These changes point to a covalent modification of IMP. The new molecule is identified as 6-pIMP on the basis of the Raman band positions and comparison with the computed spectrum.

These results suggest that although the formation of new species, 6-pIMP, also occurs in the IMP·GTP·MjADSS complex, 6-pIMP is stabilized and remains bound to the protein only when the L-aspartate pocket is ligated with HDA. Establishing the Formation, Identity, and Structure of 6-pIMP. To determine the identity of the species formed upon simultaneous binding of IMP and GTP and stabilized by the presence of HDA, we used isotope labeling of the substrate and the enzyme. The enzyme−substrate complexes were prepared in D2O, which leads to replacement of the labile hydrogens with deuterium atoms. The effect of isotope substitution on individual bands of the starting substrates, IMP and GTP, is known.30 Spectra in Figures 4k and 5l show that the new species formed in the presence of HDA has novel isotopeinduced shifts. A new, intense band appears at 1503 cm−1 in the IMP·GTP·HDA·MjADSS complex spectrum in D2O that corresponds to the characteristic band of the intermediate at 1499 cm−1 in H2O. To obtain the spectral identity of the species that is formed in the fully ligated ADSS complex, the spectrum of the GTP· MjADSS complex was subtracted from the spectrum of the IMP·GTP·HDA·MjADSS complex. The resultant spectrum, containing bands only from the new species, was compared with the Raman spectrum of 6-pIMP computed using density functional theory. The computed Raman wavenumbers are in excellent agreement with experimental values.29 We further compared the UVRR spectra of AMP with that of 6-pIMP (Figure S2). AMP serves as a stable model for 6-pIMP as a purine with the sp3 carbon at C6 because of the similarity in their structures that arises from, first, the presence of a single bond between C6 and the exocyclic group attached to it and, second, the absence of any hydrogen at the N1 position that results in similar purine ring properties in 6-pIMP, AMP, and deprotonated IMP. The similarity in the spectral pattern of AMP and the new species confirms that the enzyme-bound intermediate is indeed 6-pIMP. To obtain further proof that the changes observed in this ES complex spectrum are due to the formation of 6-pIMP and not mere a distortion of IMP or an alternative covalent modification, we conducted three control experiments of MjADSS complexes with inactive analogues of GTP. In the first two control experiments, we replaced GTP with GMPPNP. GMP-PNP, a nonhydrolyzable analogue of GTP, cannot transfer its γ-phosphate group to the C6 carbonyl of IMP. Spectra of enzyme complexes with IMP and GMP-PNP in the presence and absence of HDA are shown in Figure S3. Both these complexes should be able to achieve the catalytically competent conformation of ADSS and distort IMP but should not be able to form 6-pIMP in the absence of transferrable γphosphate group. The positions and intensities of IMP bands in the ES complex indeed do not change with respect to that of the free ligand as seen in Figure S3. This confirms that the altered bands in the (IMP+GTP+ADSS) complex do not arise from a distorted IMP structure but are indeed due to a covalent modification of IMP. In the third control experiment, we examined the spectrum of the enzyme complex with IMP, GDP, and L-aspartate. GDP serves as an analogue of GTP and binds to the enzyme in the same mode as GTP. GDP is incapable of donating a phosphate group, and it is expected to leave the IMP spectrum unperturbed. This expectation is indeed validated by the spectrum observed in Figure S3. The band at 1499 cm−1 that was regarded as a marker band of 6-pIMP in the IMP·GTP·HDA·MjADSS complex is not

■

DISCUSSION There has been a plethora of structural information about Michaelis complexes acquired using various techniques, viz., Xray crystallography, NMR, etc. These structures provide tremendous insight into the mechanisms of enzyme action and the chemical origin of complementarity. A rational understanding of the mechanistic origin of enzymes requires further, precise complementary information about the distortion of the substrate by the enzyme in complex. We employed UVRR, which provides the dual advantages of enhancing the signal from substrate while being sensitive to subtle distortions in substrate structure. We have been able to systematically investigate the extent of enzyme-induced distortion at each step of catalysis and infer the allosteric role of various ligands. This work reveals exquisite tuning of the active site as the initial substrate morphs into various intermediates within the same binding pocket. GTP·MjADSS. GTP transfers the γ-phosphate group to the C6 oxyanion of IMP. Substantial shifts in the band positions of the GTP·MjADSS complex with respect to GTP in solution implicate hydrogen bond interactions at the active site and suggest that GTP is held tightly at the active site. The larger magnitude of shifts in the GTP·MjADSS complex with respect to GTP quantifies a stronger hydrogen bond between the nucleobase moiety of GTP and the active-site residues. IMP·MjADSS. The results discussed above show that IMP bound at the active site of MjADSS is substantially distorted to resemble the deprotonation form. The deprotonation is not complete because the overall spectrum of neutral IMP is preserved in the IMP·MjADSS complex. The catalytic mechanism of ADSS involves abstraction of a proton from N1 of IMP [donor−acceptor (D−A) distance, 2.9 Å; Protein Data Bank (PDB) entry 1IWE],10 resulting in the delocalization of the C6O bond and formation of the C6 oxyanion that in 2496

DOI: 10.1021/acs.biochem.5b01386 Biochemistry 2016, 55, 2491−2499

Article

Biochemistry turn is converted to 6-pIMP. This is the first step of the catalytic mechanism that involves phosphoryl transfer and is expected to have a transition state in which N1 is deprotonated. The data presented here clearly demonstrate the distortion of IMP toward the deprotonated form and thus capture the pretransition state of the phosphoryl transfer step. Further, the distortions in the bands comprised of the imidazole ring mode arise from the interactions between N7 of the purine ring and Asn256 (D−A, 2.8 Å; PDB entry 1IWE) and between C6O and Asn256 (D−A, 4.0 Å; PDB entry 1IWE). IMP·GTP·MjADSS. Perturbation of the IMP spectrum of the singly ligated IMP·MjADSS complex indicated distortion of IMP but no covalent bond alteration. When IMP and GTP are added together to MjADSS, the downshifts in IMP bands suggest greater distortion and the appearance of a new shoulder at 1497 cm−1 indicates the formation of 6-pIMP. However, intact IMP bands suggest that IMP·MjADSS is present along with 6-pIMP·MjADSS. This implies that in the absence of its stabilization, 6-pIMP formed in the ternary complex (IMP· GTP·MjADSS) falls off from the active site and reverts back to IMP. Thus, the spectrum of IMP·GTP·MjADSS is contributed by bound IMP as well as bound 6-pIMP. IMP·GTP·HDA·MjADSS. It is evident that the spectrum of the IMP·GTP·HDA·MjADSS complex features both the shifted and new bands. The complete disappearance of the marker bands of IMP and the appearance of new bands suggest that IMP is completely transformed to 6-pIMP. Most conspicuous new bands in the resultant spectrum are the bands at 1583, 1498, and 1324 cm−1. The position of these bands is compared with the UVRR spectrum of the model compound, AMP, which has a spectral pattern similar to that of 6-pIMP. The protonation state of the purine ring at the N1 position and aromatic character of the ring in AMP and 6-pIMP are similar. Thus, it is evident from the UVRR spectra that 6-pIMP is formed in the IMP·GTP·HDA·MjADSS complex. 6-pIMP Is Stabilized in the Presence of HDA/LAspartate. The complete transition of IMP to 6-pIMP upon ligation of the L-aspartate pocket suggests that the stabilization of 6-pIMP at the active site of MjADSS is not facile in absence of L-aspartate/HDA. Addition of HDA stabilizes 6-pIMP at the active site. This is also corroborated by a previous study in which during RP-HPLC, in the absence of L-aspartate, the GDP concentration was not formed instantaneously but was built over time. On the other hand, upon addition of HDA, complete utilization of GTP was observed within 30 s.35 Ligation of the Laspartate pocket by either L-aspartate or HDA is known to bring about the conformational changes in the binding pockets of ADSS, more specifically in the Asp loop. Though the crystal structures of mmADSS with and without HDA show similar active-site organization (Figure 5), the corresponding Raman spectra capture subtle local conformational changes around the ligands that stabilize the intermediate at the active site. The His Residue Plays a Key Role in Stabilizing 6pIMP. The crystal structure of EcADSS with IMP, HDA, and NO3− suggests a possible role of His41 (equivalent to His71 in mmADSS) that acts as a catalytic acid in the phospho-transfer step and donates a proton from its ND2 position to GDP to stabilize the charge development while the γ-phosphate leads to 6-oxyanionic IMP.12 In the second step where phosphoryl displacement occurs, His41 undergoes a rotation of 90° around the Cβ−Cγ bond and donates an ε′ proton to the 6-phosphoryl group (Figure 5). The rotation of His41 is driven by the ligation of the L-aspartate pocket. The interaction of HDA at

the L-aspartate pocket stabilizes the hydrogen bond between His41 and the 6-phopshoryl group.4−6,12 Direct observation of this hydrogen bond has been possible in the UVRR spectrum where 6-pIMP is found to be stabilized upon ligation of the Laspartate pocket. Because of the low stability of this intermediate, its isolation has been difficult. It is evident that in MjADSS, formation of 6-pIMP does not require ligation of the aspartate pocket but the stabilization does. The absence of any stabilization of 6-pIMP in a PfADSS mutant (Thr307Val) that does not bind to HDA indicates that the conformational change in His41 is caused by the ligation of HDA only.36 On the basis of the analysis of the available crystal structure of EcADSS, it is observed that if no 6-pIMP forms (IMP·GDP· HDA·EcADSS, PDB entry 1GIM) then Asp13 (equivalent to Asp43 in mmADSS) does not lie in the inner coordination sphere of Mg2+ (D−A distance, 3.2 Å). In a structure (6-pIMP· HDA·EcADSS, PDB entry 1CGO), where 6-pIMP is formed and HDA is also present, a conformational change in the side chain of His41 causes Asp13 to enter the inner coordination sphere of Mg2+.15 A coordinate bond between Asp13 and Mg2+ (D−A distance, 2.1 Å) makes Asp13 a strong catalytic acid that can now return a proton to N1 of the purine ring. This distance is greater (4.1 Å) in the IMP-bound complex of mmADSS (PDB entry 1IWE).10 In another complex (IMP·HDA·PPi· EcADSS, PDB entry 1KKF), when no 6-pIMP is formed at the active site then even in the presence of HDA, Asp13 does not lie in the inner coordination sphere of Mg2+. These observed conformational changes in crystal structures along with the Raman data presented here suggest that coordination of Asp43 with Mg2+ along with a hydrogen bond between His71 and 6pIMP stabilizes the intermediate to the active site of ADSS. The conformational changes necessary for the formation of these bonds are driven by ligation of HDA.

■

CONCLUSION Complexes of MjADSS in the forward reaction have been studied, and the structure−function relationship has been established. We find that IMP alone, though distorted toward the deprotonated form, remains bound in the neutral form in the absence of GTP and L-aspartate. Formation of 6-pIMP does not require ligation of the L-aspartate pocket, but its stabilization at the active site does. The results capture the local conformational changes that are necessary to hold the intermediate at the active site. Results obtained in this study have bearing on important attributes such as the extent of substrate distortion in the mechanism of enzyme action and development of potent drug molecules against ADSS.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.5b01386.

■

Table S1 and Figures S1−S3 (PDF)

AUTHOR INFORMATION

Corresponding Authors

*Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore 560064, India. Telephone: +91-80-22082812. E-mail: [email protected]. 2497

DOI: 10.1021/acs.biochem.5b01386 Biochemistry 2016, 55, 2491−2499

Article

Biochemistry *Chemistry, Indian Institute of Science Education and Research, Pune 411008, Maharashtra, India. Phone: +91-735 069 4600. E-mail: [email protected] or puranik. [email protected].

Adenylosuccinate Synthetase from Escherichia-Coli. J. Mol. Biol. 254, 431−446. (15) Hou, Z. L., Wang, W. Y., Fromm, H. J., and Honzatko, R. B. (2002) IMP alone organizes the active site of adenylosuccinate synthetase from Escherichia coli. J. Biol. Chem. 277, 5970−5976. (16) Callender, R., and Deng, H. (1994) Nonresonance Raman Difference Spectroscopy - a General Probe of Protein-Structure, Ligand-Binding, Enzymatic Catalysis, and the Structures of Other Biomacromolecules. Annu. Rev. Biophys. Biomol. Struct. 23, 215−245. (17) Carey, P. R. (1999) Raman spectroscopy, the sleeping giant in structural biology, awakes. J. Biol. Chem. 274, 26625−26628. (18) Carey, P. R. (2006) Spectroscopic characterization of distortion in enzyme complexes. Chem. Rev. 106, 3043−3054. (19) Deng, H., Burgner, J., and Callender, R. (1991) RamanSpectroscopic Studies of Nad Coenzymes Bound to MalateDehydrogenases by Difference Techniques. Biochemistry 30, 8804− 8811. (20) Deng, H., and Callender, R. (1999) Raman spectroscopic studies of the structures, energetics, and bond distortions of substrates bound to enzymes. Methods Enzymol. 308, 176−201. (21) Deng, H., Callender, R., Schramm, V. L., and Grubmeyer, C. (2010) Pyrophosphate Activation in Hypoxanthine-Guanine Phosphoribosyltransferase with Transition State Analogue. Biochemistry 49, 2705−2714. (22) Deng, H., Kurz, L. C., Rudolph, F. B., and Callender, R. (1998) Characterization of hydrogen bonding in the complex of adenosine deaminase with a transition state analogue: A Raman spectroscopic study. Biochemistry 37, 4968−4976. (23) Deng, H., Lewandowicz, A., Schramm, V. L., and Callender, R. (2004) Activating the phosphate nucleophile at the catalytic site of purine nucleoside phosphorylase: A vibrational spectroscopic study. J. Am. Chem. Soc. 126, 9516−9517. (24) Mehrotra, S., and Balaram, H. (2007) Kinetic characterization of adenylosuccinate synthetase from the thermophilic archaea methanocaldococcus jannaschii. Biochemistry 46, 12821−12832. (25) Bradford, M. M. (1976) Rapid and Sensitive Method for Quantitation of Microgram Quantities of Protein Utilizing Principle of Protein-Dye Binding. Anal. Biochem. 72, 248−254. (26) Jayanth, N., Ramachandran, S., and Puranik, M. (2009) Solution Structure of the DNA Damage Lesion 8-Oxoguanosine from Ultraviolet Resonance Raman Spectroscopy. J. Phys. Chem. A 113, 1459−1471. (27) Fodor, S. P. A., Rava, R. P., Hays, T. R., and Spiro, T. G. (1985) Ultraviolet Resonance Raman-Spectroscopy of the Nucleotides with 266-Nm, 240-Nm, 218-Nm, and 200-Nm Pulsed Laser Excitation. J. Am. Chem. Soc. 107, 1520−1529. (28) Spiro, T. G. (1974) Resonance Raman-Spectroscopy - New Structure Probe for Biological Chromophores. Acc. Chem. Res. 7, 339− 344. (29) Karnawat, V., and Puranik, M. (2015) Solution structure of ligands involved in purine salvage pathway. Spectrochim. Acta, Part A 151, 679−686. (30) Gogia, S., Jain, A., and Puranik, M. (2009) Structures, Ionization Equilibria, and Tautomerism of 6-Oxopurines in Solution. J. Phys. Chem. B 113, 15101−15118. (31) Wen, Z. Q., and Thomas, G. J. (1998) UV resonance Raman spectroscopy of DNA and protein constituents of viruses: Assignments and cross sections for excitations at 257, 244, 238, and 229 nm. Biopolymers 45, 247−256. (32) Soans, C., and Fromm, H. J. (1991) Studies of Ligand-Binding to Escherichia-Coli Adenylosuccinate Synthetase. Arch. Biochem. Biophys. 291, 107−112. (33) Walters, E. W., Lee, S. F., Niderman, T., Bernasconi, P., Subramanian, M. V., and Siehl, D. L. (1997) Adenylosuccinate synthetase from maize - Purification, properties, and mechanism of inhibition by 5′-phosphohydantocidin. Plant Physiology 114, 549−555. (34) Lipps, G., and Krauss, G. (1999) Adenylosuccinate synthase from Saccharomyces cerevisiae: homologous overexpression, purifica-

Funding

This work is funded by internal grants to M.P. from Indian Institute of Science Education and Research, Pune, and H.B. acknowledges Department of Science and Technology (SR/ SO/BB-060/2008), Department of Biotechnology (BT/ PR10410/BRB/10/598/2008) and internal grants from the Jawaharlal Nehru Centre for Advanced Scientific Research, India for funding. V.K. and S.M. acknowledge the Council of Scientific and Industrial Research (CSIR), India, for Ph.D. fellowships. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Pauling, L. (1948) Nature of Forces between Large Molecules of Biological Interest. Nature 161, 707−709. (2) Gogia, S., Balaram, H., and Puranik, M. (2011) Hypoxanthine Guanine Phosphoribosyltransferase Distorts the Purine Ring of Nucleotide Substrates and Perturbs the pK(a) of Bound Xanthosine Monophosphate. Biochemistry 50, 4184−4193. (3) Karnawat, V., Gogia, S., Balaram, H., and Puranik, M. (2015) Differential Distortion of Purine Substrates by Human and Plasmodium falciparum Hypoxanthine-Guanine Phosphoribosyltransferase to Catalyse the Formation of Mononucleotides. ChemPhysChem 16, 2172−2181. (4) Honzatko, R. B., and Fromm, H. J. (1999) Structure-function studies of adenylosuccinate synthetase from Escherichia coli. Arch. Biochem. Biophys. 370, 1−8. (5) Honzatko, R. B., Stayton, M. M., and Fromm, H. J. (1999) Adenylosuccinate synthetase: Recent developments. Adv. Enzymol. 73, 57−102. (6) Poland, B. W., Bruns, C., Fromm, H. J., and Honzatko, R. B. (1997) Entrapment of 6-thiophosphoryl-IMP in the active site of crystalline adenylosuccinate synthetase from Escherichia coli. J. Biol. Chem. 272, 15200−15205. (7) Choe, J. Y., Poland, B. W., Fromm, H. J., and Honzatko, R. B. (1999) Mechanistic implications from crystalline complexes of wildtype and mutant adenylosuccinate synthetases from Escherichia coli. Biochemistry 38, 6953−6961. (8) Iancu, C. V., Borza, T., Choe, J. Y., Fromm, H. J., and Honzatko, R. B. (2001) Recombinant mouse muscle adenylosuccinate synthetase - Overexpression, kinetics, and crystal structure. J. Biol. Chem. 276, 42146−42152. (9) Iancu, C. V., Borza, T., Fromm, H. J., and Honzatko, R. B. (2002) Feedback inhibition and product complexes of recombinant mouse muscle adenylosuccinate synthetase. J. Biol. Chem. 277, 40536−40543. (10) Iancu, C. V., Borza, T., Fromm, H. J., and Honzatko, R. B. (2002) IMP, GTP, and 6-Phosphoryl-IMP complexes of recombinant mouse muscle adenylosuccinate synthetase. J. Biol. Chem. 277, 26779− 26787. (11) Iancu, C. V., Zhou, Y., Borza, T., Fromm, H. J., and Honzatko, R. B. (2006) Cavitation as a mechanism of substrate discrimination by adenylosuccinate synthetases. Biochemistry 45, 11703−11711. (12) Poland, B. W., Fromm, H. J., and Honzatko, R. B. (1996) Crystal structures of adenylosuccinate synthetase from Escherichia coli complexed with GDP, IMP hadacidin, NO3- and Mg2+. J. Mol. Biol. 264, 1013−1027. (13) Poland, B. W., Hou, Z. L., Bruns, C., Fromm, H. J., and Honzatko, R. B. (1996) Refined crystal structures of guanine nucleotide complexes of adenylosuccinate synthetase from Escherichia coli. J. Biol. Chem. 271, 15407−15413. (14) Silva, M. M., Poland, B. W., Hoffman, C. R., Fromm, H. J., and Honzatko, R. B. (1995) Refined Crystal-Structures of Unligated 2498

DOI: 10.1021/acs.biochem.5b01386 Biochemistry 2016, 55, 2491−2499

Article

Biochemistry tion and characterization of the recombinant protein. Biochem. J. 341, 537−543. (35) Mehrotra, S., and Balaram, H. (2008) Methanocaldococcus jannaschii adenylosuccinate synthetase: Studies on temperature dependence of catalytic activity and structural stability. Biochim. Biophys. Acta, Proteins Proteomics 1784, 2019−2028. (36) Mehrotra, S., Mylarappa, B. N., Iyengar, P., and Balaram, H. (2010) Studies on active site mutants of P-falciparum adenylosuccinate synthetase: Insights into enzyme catalysis and activation. Biochim. Biophys. Acta, Proteins Proteomics 1804, 1996−2002.

2499

DOI: 10.1021/acs.biochem.5b01386 Biochemistry 2016, 55, 2491−2499