The role of evolutionarily conserved hydrophobic contacts in the quaternary structure stability of Escherichia coli serine hydroxymethyltransferase Rita Florio 1 , Roberta Chiaraluce 1 , Valerio Consalvi 1 , Alessandro Paiardini 1 , Bruno Catacchio 1,2 , Francesco Bossa 1,3 and Roberto Contestabile 1 1 Dipartimento di Scienze Biochimiche ‘A. Rossi Fanelli’, ‘Sapienza’ Universita ` di Roma, Italy 2 CNR, Istituto di Biologia e Patologia Molecolari, ‘Sapienza’ Universita ` di Roma, Italy 3 Centro di Eccellenza di Biologia e Medicina Molecolare (BEMM), ‘Sapienza’ Universita ` di Roma, Italy Pyridoxal 5¢-phosphate (PLP)-dependent enzymes are a large ensemble of biocatalysts that make use of the same cofactor but have distinct evolutionary origins and protein architectures [1–3]. According to their 3D structure, PLP-dependent enzymes are grouped into five evolutionarily unrelated superfamilies, correspond- ing to as many different folds (fold types) [4]. The fold type I group, also referred to as the aspartate amino- transferase family [5], is the largest, functionally most diverse and best characterized. Its members are catal- ytically active as homodimers, although they may assemble into higher order complexes. A single subunit folds into two domains [6]. The central feature of the N-terminal, larger domain is a seven-stranded b-sheet. In some instances, the N-terminal tail does not partici- pate as a part of the large domain but comprises a separate structural element. The small, C-terminal domain, comprises a three- or four-stranded b-sheet, covered with helices on one side. The active site is located at the interface of the domains and is delimited by amino acid residues that are contributed by both subunits of the catalytic dimer. Remarkably, the Keywords conserved hydrophobic contacts; fold type I enzymes; pyridoxal phosphate; quaternary structure; serine hydroxymethyltransferase Correspondence R. Contestabile, Dipartimento di Scienze Biochimiche, ‘Sapienza’ Universita ` di Roma, Piazzale Aldo Moro 5, 00185 Rome, Italy Fax: +39 0649 917566 Tel: +39 0649 917569 E-mail: roberto.contestabile@uniroma1.it Website: http://w3.uniroma1.it/bio_chem/ sito_biochimica/EN/index.html (Received 18 September 2008, revised 23 October 2008, accepted 27 October 2008) doi:10.1111/j.1742-4658.2008.06761.x Pyridoxal 5¢-phosphate-dependent enzymes may be grouped into five struc- tural superfamilies of proteins, corresponding to as many fold types. The fold type I is by far the largest and most investigated group. An important feature of this fold, which is characterized by the presence of two domains, appears to be the existence of three clusters of evolutionarily conserved hydrophobic contacts. Although two of these clusters are located in the central cores of the domains and presumably stabilize their scaffold, allow- ing the correct alignment of the residues involved in cofactor and substrate binding, the role of the third cluster is much less evident. A site-directed mutagenesis approach was used to carry out a model study on the impor- tance of the third cluster in the structure of a well characterized member of the fold type I group, serine hydroxymethyltransferase from Escherichia coli. The experimental results obtained indicated that the clus- ter plays a crucial role in the stabilization of the quaternary, native assem- bly of the enzyme, although it is not located at the subunit interface. The analysis of the crystal structure of serine hydromethyltransferase suggested that this stabilizing effect may be due to the strict structural relation between the cluster and two polypeptide loops, which, in fold type I enzymes, mediate the interactions between the subunits and are involved in cofactor binding, substrate binding and catalysis. Abbreviations CHC, conserved hydrophobic contact; eSHMT, Escherichia coli serine hydroxymethyltransferase; H 4 PteGlu, tetrahydropteroylglutamate; PLP, pyridoxal 5¢-phosphate; SCR, structurally conserved region. 132 FEBS Journal 276 (2009) 132–143 ª 2008 The Authors Journal compilation ª 2008 FEBS superimposition of fold type I enzymes reveals that the location of the cofactor in the active site is virtually identical in all members of the group [7]. Despite the high similarities of their 3D structures, many fold type I enzymes show very little sequence identity, highlighting the need to identify the structural features that determine the common fold. Accordingly, a computational analysis that made use of 23 nonre- dundant crystal structures and 921 sequences of fold type I enzymes identified 17 structurally conserved regions (SCRs), which form the common cores of the large and small domains. Within these SCRs, there are three clusters of evolutionarily conserved hydrophobic contacts (CHCs) [8]. The first and second cluster are located in the cores of the large and small domains, respectively, and appear to stabilize their protein scaf- folds, allowing the proper positioning of the residues involved in PLP binding, substrate binding and modu- lation of the cofactor’s catalytic properties. The third cluster forms a hinge between two conserved a-helices (which correspond to two SCRs), located at the begin- ning and at the end, respectively, of the large domain (Fig. 1). Examination of the contact network shows that the CHCs lie along one side of each helix, forming a buried spine at positions i, i + 4, and i +7. By apparent contrast to the two previously described clus- ters, the third cluster does not appear to be directly involved in the proper positioning of any active site residue, suggesting that its high degree of evolutionary conservation could be due to a merely structural, rather than functional role. In the present study, the importance of the third hydrophobic cluster as a structural determinant of the Escherichia coli serine hydroxymethyltransferase (eSHMT) overall native fold was investigated by decreasing the hydrophobic contact area of the cluster, using a site-directed mutagenesis approach. The effects of L85A, L276A and L85A⁄ L276A mutations on the native structure of the enzyme were analyzed (Fig. 1). Results The consequences of the mutations on the native struc- ture of eSHMT were evaluated by analyzing and comparing the ultracentrifuge sedimentation, cofactor binding, catalytic and spectral properties of wild-type and mutant apo- and holoenzymes. Quaternary structure analysis The subunit assembly of apo- and holo-forms of wild- type and mutant eSHMTs was characterized by analyt- ical ultracentrifugation. Table 1 shows the values of sedimentation coefficient and dissociation constant (K d ) calculated from combined sedimentation velocity and equilibrium approaches. As established in the available literature [9,10], wild-type eSHMT exists as a dimer in both apo- and holo-forms, with a molecular mass of approximately 91 kDa [9]. The ultracentrifuga- tion experiments confirmed that the depletion of the cofactor does not have any effect on the dimeric assembly of the enzyme. The sedimentation velocity Fig. 1. Schematic representation of the monomeric structure of eSHMT. Cartoon representation of a single subunit of the eSHMT ternary complex with glycine and 5-formyl H 4 PteGlu (Protein Data Bank: 1dfo) [14], showing the N-terminal tail (residues 1–61) colored in orange, the large domain (residues 62–211) in salmon, the interdomain segment (residues 212–279) in green and the small domain in blue. The PLP- Gly complex is shown as yellow sticks, with the phosphorus atom in orange, the oxygen atoms in red and the nitrogen atoms in blue. The two a-helices involved in the formation of the third cluster of CHCs are enclosed in a circle. A magnified view of these helices shows the residues that form the CHCs represented both as sticks and as transparent spheres. L85 and L276 are indicated by arrows. R. Florio et al. Role of hydrophobic contacts in serine hydroxymethyltransferase FEBS Journal 276 (2009) 132–143 ª 2008 The Authors Journal compilation ª 2008 FEBS 133 patterns of apo- and holo-forms are indeed almost superimposable, with a single symmetrical peak char- acterized by a sedimentation coefficient, S 20,w , of 5.5S (Fig. 2), which is the value expected for a hydrated eSHMT dimer endowed with an approximately spheri- cal shape. The sedimentation equilibrium experiments showed that, in the 2.5–25 lm subunit concentration range, the wild-type eSHMT is a dimer, either in the presence or absence of cofactor. In the same concen- tration range, the L85A and L276A mutant holoen- zymes are also dimers, with sedimentation coefficients of 5.5S. Interestingly, the frictional ratio (f ⁄ f 0 , the ratio between the experimentally calculated friction coeffi- cient and the minimum friction coefficient of an anhy- drous sphere) of dimeric wild-type, L85A and L276A eSHMT holoenzymes is close to that of a spherical protein, namely 1.2–1.3, suggesting that the single mutations did not result in significant changes in the shape of the dimeric protein. Depletion of the cofactor affected the quaternary structure of both single mutants, which showed an extra sedimentation peak, at approximately 2.7S in the case of L85A and at 3.1S with the L276A mutant (Fig. 2). The smaller sedimentation coefficient corre- sponds to that of monomeric eSHMT. Therefore, the single mutant apoenzymes exist as an equilibrium mix- ture of dimers and monomers, which interconvert slowly with respect to the period of elapsed time in the sedimentation velocity experiments. A comparison of the dissociation constants of the single mutant apo- forms, obtained by a fitting of the sedimentation equi- librium curves to a monomer-dimer model, indicates that the destabilizing effect of PLP depletion is greater in L276A (K d = 2.7 · 10 )6 m )1 ) than in L85A (K d = 4.0 · 10 )9 m )1 ). When both mutations are pres- ent, as in the L85A ⁄ L276A double mutant, the apoen- zyme exists as a monomer in the range of concentrations tested (2.5–25 lm). The frictional ratio of this monomeric species was calculated to be approx- imately 1.2, suggesting that the dissociation determined by the double mutation was not accompanied by large structural changes. Compared to that observed with the single L85A and L276A mutants, cofactor binding to the L85A ⁄ L276A double mutant apoenzyme did not shift the equilibrium completely in favor of the dimer. In the double mutant holoenzyme, obtained by adding PLP to 98% of saturation (as calculated from the dis- sociation constant of the related cofactor binding equi- librium; see below), a residual 35% fraction of monomer is in equilibrium with the dimer (Table 1 and Fig. 2). A dissociation constant of 1.7 · 10 )6 m )1 was calculated for this equilibrium. Because it is known that PLP bound to eSHMT through a Schiff base linkage to the active site lysine residue absorbs light maximally at 420 nm [9], a sedimentation velocity experiment was performed on a double mutant holoen- zyme sample (33 lm), measuring absorbance at this wavelength. The presence of a 3.1S peak in the sedi- mentation pattern indicated that the cofactor was covalently bound to the monomeric form of the enzyme (Fig. 2). The lower percentage of monomer present in this sedimentation profile (12% instead of 35%; Fig. 2 and Table 1) is accounted for by the higher concentration of enzyme employed in the exper- iment, and as calculated by using the equation describ- Table 1. Sedimentation and dissociation constants calculated from ultracentrifuge experiments on apo- and holo-forms of wild-type and mutant eSHMTs. Values are shown of the S 20,w sedimentation coefficient calculated in sedimentation velocity experiments on enzyme samples at 2.5 l M subunit concentration, in 50 mM NaHepes buffer (pH 7.2), containing 200 lM dithiothreitol and 100 l M EDTA, at 20 °C. Percentages in parenthesis correspond to the fraction of enzyme subunits that sediment with the related coefficient and were calculated from an integration of the sedimen- tation profiles shown in Fig. 2. The dissociation constants of dimer– monomer equilibria (K d ) were determined from sedimentation equilibrium experiments carried out on enzyme samples in the 2.5–25 l M subunit concentration range. S 20, w (S) K d (M )1 ) a Holoenzyme forms WT 5.5 ND L85A 5.5 ND L276A 5.5 ND L85A ⁄ L276A 5.5 (66%) 3.3 (34%) 5.5 (88%) b 3.1 (12%) b 1.7 · 10 )6 Apoenzyme forms WT 5.5 ND L85A 5.5 (90%) c 2.7 (8%) c 4.0 · 10 )9 L276A 5.5 (65%) 3.1 (35%) 2.7 · 10 )6 L85A ⁄ L276A 3.1 ND a Dissociation constants could not be calculated for wild-type and single mutant holoenzymes and for the double mutant apoenzyme because these were completely either in the dimeric or monomeric state in the range of protein concentration used (ND, not deter- mined). However, the detection limit of the instrumentation employed, which may be estimated to be approximately 1% (per- centage of detectable monomer in a dimeric sample or vice versa), restricts the K d for the dimeric holo-forms to values £ 8 · 10 )9 M )1 and the K d for the monomeric double mutant apoenzyme to values ‡ 5 · 10 )4 M )1 (calculated on the basis of Eqn (1), assuming that 1% of undetected dimer or monomer were present in the sedimen- tation velocity experiments carried out at a subunit concentration of 2.5 l M). b Calculated on data collected at 420 nm with an enzyme sample at a subunit concentration of 33 l M. Data of all other exper- iments were collected at 277 nm. c In the sedimentation velocity experiments on the L85A apoenzyme, approximately 2% of sub- units sedimented very slowly in the form of aggregates. Role of hydrophobic contacts in serine hydroxymethyltransferase R. Florio et al. 134 FEBS Journal 276 (2009) 132–143 ª 2008 The Authors Journal compilation ª 2008 FEBS ing the dissociation equilibrium (Eqn 1). A complete shift of the equilibrium in favor of the dimeric form was obtained when l-serine (1 mm) was added to the double mutant holoenzyme (Fig. 2). PLP binding properties The affinity of wild-type and mutant forms for the cofactor was measured to evaluate the impact of the mutations on the structure of the PLP binding site. Because PLP binding to apo-eSHMT is known to quench the intrinsic fluorescence emission of the enzyme [10], without changing the wavelength of maxi- mum emission, the dissociation constant of the binding equilibrium was calculated from saturation curves obtained by measuring the fluorescence emission of apoenzyme samples (26 nm subunit concentration) at increasing PLP concentrations (Fig. 3). Table 2 shows that the apparent K d values calculated from least square fitting of experimental data points to Eqn (2) are essentially the same for all enzyme forms. The cal- culated relative fluorescence intensities in the absence of PLP (F 0 ) or in the presence of saturating concentra- tions of cofactor (F inf ) were: F 0 = 125 ± 1 and F inf =65 ± 1 for the wild-type enzyme; F 0 = 139 ± 1 and F inf = 66 ± 1 for L85A; F 0 = 122 ± 1 and F inf = 63 ± 1 for L276; and F 0 = 103 ± 1 and F inf = 73 ± 1 for L85A ⁄ L276A. The higher fluores- cence with respect to wild-type observed with the L85A apoenzyme may be explained by the presence of a small percentage of subunits present as aggregates (Table 1). A remarkable difference is noted with respect to the intrinsic fluorescence emission intensities of apo- and holo-forms of the L85A ⁄ L276A double mutant: the relative fluorescence emission of the double mutant apoenzyme is considerably lower than that of the other apo-forms and PLP binding does not quench fluores- cence to the same extent it does with the other holo- forms, although the wavelength of maximum emission is the same for all enzyme forms (Fig. 3, insets). In the light of the results obtained from the ultra- centrifuge experiments, it should be noted that, at a concentration of 26 nm, the association state of subun- its is expected to vary among wild-type and mutant enzymes. Indeed, it may be calculated, using the disso- ciation constants showed in Table 1 and according to Eqn (1), that, at this concentration, the apoenzyme subunits of wild-type eSHMT are mostly in the dimeric state (for a fraction ‡ 90%), the apo-L85A is approximately 75% dimeric, whereas the apo-L276A and the apo-L85A ⁄ L85A mutants are in the mono- meric state. The association state of the holoenzymes can be estimated to be ‡ 90% dimeric for all enzyme forms, except for the L85A ⁄ L276A double mutant, Fig. 2. Sedimentation velocity distributions obtained with apo- and holo-forms of wild-type and mutant eSHMTs. Sedimentation velocity measurements were performed at 116 480 g on 2.5 l M (subunit concentration) holoenzyme (—) and apoenzyme (- - -) samples kept at 20 °Cin50m M NaHepes buffer (pH 7.2), containing 200 lM dithiothreitol and 100 lM EDTA. L-serine (1 mM) was added to a sample of the L85A ⁄ L276A double mutant holoenzyme (ÆÆÆÆÆ). Absorbance data were all collected at 277 nm, except in the case of a sedimentation experi- ment carried out on the double mutant holoenzyme at a subunit concentration of 33 l M, when the absorbance of protein-bound cofactor was measured at 420 nm (Æ-Æ-Æ). R. Florio et al. Role of hydrophobic contacts in serine hydroxymethyltransferase FEBS Journal 276 (2009) 132–143 ª 2008 The Authors Journal compilation ª 2008 FEBS 135 which is expected to be mostly in the monomeric state. Therefore, the similar values of dissociation constant obtained with wild-type and mutant enzymes suggest that the cofactor binds to monomeric and dimeric forms with similar affinities. Catalytic properties SHMT catalyses the reversible transfer of the C b of l-serine to tetrahydropteroylglutamate (H 4 PteGlu), with formation of glycine and 5,10-methylene-H 4 Pte- Glu. However, in the absence of H 4 PteGlu, it also accelerates the cleavage of several different l-3-hydr- oxyamino acids to glycine and the corresponding aldehyde [11]. Both erithro and threo forms of l-3- phenylserine are rapidly cleaved to glycine and benzal- dehyde [12,13]. The serine hydroxymethyltransferase and l-threo-phenylserine aldolase activities of all eSHMT forms were assayed using enzyme samples (at 0.05 and 3 lm subunit concentrations, respectively) saturated with PLP. The calculated kinetic parameters of both reactions are shown in Table 2. The L85A mutation had virtually no effect on the catalytic prop- erties of the enzyme. Minor differences with respect to Table 2. Dissociation constants of PLP binding equilibrium and kinetic parameters determined with wild-type and mutant eSHMTs. Para- meters are expressed as the mean ± SD determined by nonlinear least squares fitting of data to the related equation (see Experimental procedures). K d PLP a (nM) aK m b Ser (m M) aK m b H 4 PteGlu (l M) k cat SHMT c (min )1 ) K m /-Ser d (mM) k cat /-Ser d (min )1 ) Wild-type 5.11 ± 1.14 0.14 ± 0.01 7.03 ± 0.88 686.6 ± 21.7 36.3 ± 1.4 202.1 ± 4.2 L85A 5.88 ± 0.73 0.15 ± 0.01 7.16 ± 1.57 647.4 ± 36.3 37.4 ± 2.9 257.8 ± 13.7 L276A 5.01 ± 0.87 0.20 ± 0.01 4.35 ± 0.52 400.5 ± 10.8 42.6 ± 1.8 132.3 ± 3.2 L85A ⁄ L276A 6.50 ± 1.69 0.20 ± 0.01 11.20 ± 1.34 400.0 ± 15.3 42.1 ± 4.5 173.9 ± 10.7 a Dissociation constant of PLP binding equilibrium. b Apparent K m of either L-serine or H 4 PteGlu in serine hydroxymethyltransferase reaction when the other substrate is at saturating concentration. c Catalitic constant of the serine hydroxymethyltransferase reaction. d Kinetic param- eters of the L-threo-3-phenylserine cleavage reaction. Fig. 3. Comparison of PLP-binding saturation curves obtained with wild-type and mutant enzymes. Apoenzyme samples (26 nM) were mixed with different concentrations of PLP (1–400 n M)in50mM NaHepes (pH 7.2), containing 200 lM dithiothreitol and 100 lM EDTA, at 20 °C. Fluorescence emission spectra were measured in a 1-cm quartz cuvette with excitation at 280 nm. The graphs report the relative fluores- cence intensity at 336 nm (F r ) as a function of the total PLP concentration (i.e. the concentration of free and enzyme-bound PLP). The contin- uous lines are those calculated from the least square fitting of experimental data to Eqn (2). Insets show comparisons between the intrinsic fluorescence emission spectra of holoenzyme (—) and apoenzyme (- - -) forms of wild-type (thick lines) and mutant (thin lines) eSHMT. Role of hydrophobic contacts in serine hydroxymethyltransferase R. Florio et al. 136 FEBS Journal 276 (2009) 132–143 ª 2008 The Authors Journal compilation ª 2008 FEBS wild-type were observed with the L276A and the L85A ⁄ L276A mutants, which yielded similar kinetic parameters overall: with both reactions tested, k cat values were decreased by up to two-thirds and the K m for amino acid substrates were slightly increased. Given the dimerization effect observed in the ultracen- trifuge experiments upon addition of l-serine to the double mutant holoenzyme, it may be assumed that the slightly different kinetic parameters of L276A and L85A ⁄ L276A mutants do not depend on the oligomer- ization states of the enzymes. It may be inferred that the minor changes of catalytic properties observed with the double mutant enzyme are determined by the L276A mutation. Far and near-UV circular dichroism spectra The far-UV CD spectra of wild-type and mutant apo- and holoenzymes at a subunit concentration of 2.5 lm were virtually identical (data not shown), indicating that the mutations did not alter the secondary structure of the enzyme. The tertiary structure of all apo- and holo-forms was analyzed measuring and comparing their near-UV CD spectra at a subunit concentration of 35 lm. A substantial similarity was observed among the aromatic CD spectra of wild-type, L85A and L276A apo-eSHMTs, with respect to fine structure and representative bands (Fig. 4). It may be estimated that the L276A mutant at this concentration is approxi- mately 80% dimeric, whereas the other apo-forms are completely dimeric. In the case of the L85A ⁄ L276A double mutant apoenzyme, the fine structure of the spectrum is markedly less resolved than it is with all other apo-forms, despite the similarity in overall ellipt- icty (Fig. 4). The loss of fine structure may be accounted for by the fact that, at a concentration of 35 lm, the enzyme exists mostly as a monomer (for a fraction ‡ 90%). Because PLP binds to the monomeric and dimeric apoenzyme with similar affinities, it may be deduced that the structure of the monomer is analo- gous to that of the subunits in the dimeric enzyme. The protein-bound cofactor contributes to the CD spectra of all the holoenzymes with a broad positive band centered at 420 nm and a negative band at approximately 340 nm, which are similar in all enzyme forms (data not shown). The presence of the cofactor negative band determines a significant increase of the overall negative ellipticity below 320 nm (Fig. 4). PLP binding to the wild-type apoenzyme also determines an increase of the aromatic 285 nm band contribution. A very similar spectral change is observed with the single mutants that, at a concentration of 35 lm and in the holo-form, are fully dimeric. PLP binding to the dou- ble mutant apoenzyme definitely improves the resolu- Fig. 4. Comparison of the near-UV CD spec- tra of apo- and holo-forms of mutant (contin- uous lines) and wild-type (dashed lines) eSHMTs. Enzymes samples (35 l M) were dissolved in 50 m M NaHepes (pH 7.2), con- taining 200 l M dithiothreitol and 100 lM EDTA, at 20 °C. CD spectra were measured using a 1 cm quartz cuvette. R. Florio et al. Role of hydrophobic contacts in serine hydroxymethyltransferase FEBS Journal 276 (2009) 132–143 ª 2008 The Authors Journal compilation ª 2008 FEBS 137 tion of the fine structure of the CD spectrum. How- ever, the CD spectrum of the double mutant holoenzyme (85% dimeric at a concentration of 35 lm) shows lower negative ellipticity in the aromatic region and a less pronounced 285 nm band (Fig. 4), indicating that the formation of the dimer induced by PLP binding is not accompanied by a complete recov- ery of the native tertiary structure. Discussion The decrease of hydrophobic contact area determined by the mutations in the third cluster of CHCs caused a shift of the equilibrium between dimeric and monomeric forms of eSHMT in favor of the latter. The extent to which this happened was maximum for the L85A ⁄ L276A double mutation and minimum for the L85A mutation, whereas the L276A had an intermedi- ate effect. The alteration of the dimer–monomer equi- librium determined by the single L85A and L276A mutations was not accompanied by any significant change of tertiary structure, as determined by analysis of the fluorescence and near-UV CD properties. On the other hand, the L85A ⁄ L276A double mutation had visi- ble effects on both the intrinsic fluorescence and near- UV CD properties of the enzyme. Both L276A and L85A ⁄ L276A apoenzymes are in the monomeric state at the concentration used in the intrinsic fluorescence measurements. However, the fluorescence emission of the double mutant apoenzyme is much less intense than that of the L276A apoenzyme, for which the emission spectrum, in turn, is very similar to that of the wild-type apoenzyme (Fig. 3). This indicates that the presence of both mutations slightly perturbed the native structure of the monomer. When PLP binds to the monomeric double mutant, the intrinsic fluorescence is quenched to a lesser extent than with the other enzyme forms. This difference may be attributed to the fact that, upon PLP binding, the double mutant holoenzyme stays in the monomeric state, whereas all the other forms become dimeric, as revealed by analytical ultracentrifugation. Nevertheless, at a higher enzyme concentration, when the subunits of the double mutant associate into a dimer, the aromatic CD spectrum of the double mutant indicates that minor structural changes with respect to the wild-type enzyme are present. None of the introduced mutations had large conse- quences on the catalytic properties of the enzyme. In this respect, it should be noted that substrate binding to the double mutant eSHMT was observed to stabi- lize the dimeric, catalytically competent form of the enzyme (Fig. 2). Therefore, it is expected that all enzyme forms were dimeric in the conditions used to assay their catalytic activity. It may be deduced that the structural alterations determined by the mutations had a modest impact on the overall tertiary structure of the enzyme and were largely confined to the stabil- ity of the native quaternary assembly. Initially, the observed effects of the mutations on the quaternary structure may appear to be rather surprising because the residues replaced by site-directed mutagenesis are far away from the subunit interface. Nevertheless, in SHMT, an important interaction between the subunits is established between the a-helices of the third cluster of CHCs and the N-terminal a-helix of the adjacent subunit (Fig. 5). This observation may be sufficient to explain the increase of the dissociation constant of the monomer–dimer equilibrium determined by the muta- tions. The results obtained in the present study also show that the monomeric form of the enzyme is able to bind PLP and that this binding event counteracts the effect of the mutations, stabilizing the dimeric form. The stabilizing effect is even more pronounced if l-serine binds to the cofactor at the active site, as is evident in the case of the double mutant enzyme. This is monomeric in the apo-form, exists as an equilibrium mixture of monomers and dimers even when all active sites are occupied by PLP and is fully converted into a dimer by the addition of l-serine. Evidently, the muta- tions caused a slight and indirect alteration of crucial interactions at the subunit interface. The stabilizing effect of PLP and l-serine on the dimeric assembly suggests that these alterations involve regions at the subunit interface that are contacted by cofactor and substrates, when these are bound to the active site of the adjacent subunit. Scrutiny of the eSHMT crystal structure [14] reveals that two polypeptide loops, at the N-terminal ends of the a-helices that form the third cluster of CHCs, are likely to be the relevant structural regions (Fig. 5). One is the polypeptide section made of residues 55¢–67¢ (where the primes indicate that the residues are contributed by the other subunit). In eSHMT, Y55¢ interacts with the phosphate moiety of PLP; E57¢ is crucial in the binding of the l-serine hydroxyl group [15] and in the catalysis of the hydrox- ymethyltransferase reaction [16]; and Y65¢ interacts with the carboxylate group of substrates and plays a key role in substrate binding [17]. An alignment among 63 amino acid sequences of the enzyme from several different sources showed that all three residues are invariant in SHMT [18]. The second polypeptide loop, comprising residues 258¢–264¢, is a very conserved region which interacts with the phosphate moiety of PLP. It is delimited by two invariant proline residues, P258¢ (which is in a cis configuration in all five known structures from mouse [19], human [20], rabbit [21], Role of hydrophobic contacts in serine hydroxymethyltransferase R. Florio et al. 138 FEBS Journal 276 (2009) 132–143 ª 2008 The Authors Journal compilation ª 2008 FEBS E. coli [14] and Bacillus stearothermophilus [15]) and P264¢. The importance of this structural region is testi- fied by site-directed mutagenesis studies on P258¢ and P264¢, which showed how the native conformation of the loop is pivotal to PLP binding and catalysis [22]. Given the above considerations, a series of possible, correlated outcomes of the mutations may be envis- aged. The decrease of the hydrophobic contact area in the third cluster of CHCs is expected to alter the asso- ciation of the a-helices that form the cluster and, consequently, to weaken their interaction with the N-terminal a-helix. This can be imagined to be the main cause of subunit dissociation in apo-eSHMT. At the same time, the mutations may indirectly alter the structure of the 55¢–67¢ and 258¢–264¢ loops. In holo- eSHMT, this alteration may compromise the interac- tions between the cofactor bound at the active site and the loops of the adjoining subunit, promoting subunit dissociation. In the holoenzyme, effects on both the N-terminal helix and on the 55¢–67¢ and 258¢–264¢ loops would be present. On the other hand, PLP bind- ing to the apoenzyme is expected to stabilize the struc- ture of the loops and, indirectly, the association between the helices of the cluster. This, as a conse- quence, would reinforce the interaction among the helices and the N-terminal helix of the other subunit, promoting dimerization. At this point, the observation that the monomeric eSHMT binds PLP (Fig. 2) and that all mutant forms show similar affinity for the cofactor (Table 2), although they exist as different mixtures of monomers and dimers at equilibrium, is rather puzzling. One possible explanation is that the affinity of monomeric eSHMT for PLP is lower than that of the dimer, although not drastically different. This difference may not be detectable in the experi- ments performed in the present study. The high degree of sequence and structural conserva- tion of the third cluster of CHCs, observed for the majority of fold type I enzymes, suggests that its stabi- lizing function, hypothesized for eSHMT, could be extended to the whole superfamily. The extent to which the third hydrophobic cluster is involved in the stabil- ization of the dimer assembly might be different amongst fold type I enzymes, depending on the size and orientation of the N-terminal arm. However, although the 55¢–67¢ loop of eSHMT is a highly diversi- fied region in fold type I enzymes, it invariably contains residues involved in PLP and substrate binding and catalysis [8,23]. Moreover, the second polypeptide loop (residues 258¢–264¢ in eSHMT) represents a very con- served region in this group of enzymes, interacting with the phosphate moiety of PLP. Therefore, the capability of the third cluster of CHCs to confer stability to the polypeptide loops of the active site is presumably important in all members of the fold type I family. An additional clue on the structural importance of the third cluster of CHCs in fold type I enzymes comes from folding studies on eSHMT. The folding mechanism of eSHMT may be divided into two phases [10,22,24,25]. In the first, relatively rapid phase, the two domains fold into a native-like intermediate that has virtually all of the native secondary and tertiary structure, but is unable Fig. 5. Schematic representation of the dimeric structure of eSHMT. Enzyme subun- its of the eSHMT•Gly•5-formyl H 4 PteGlu ternary complex (Protein Data Bank: 1dfo) [14] are shown in salmon and cyan, whereas the PLP-gly complex is shown as in Fig. 1. The a-helices involved in the for- mation of the third cluster of CHCs belong to the subunit shown in salmon. The first a-helix and the 55¢–67¢ loop are colored in red; the second a-helix and the 258¢–264¢ loop are shown in magenta. The N-terminal helix of the cyan subunit is shown in blue. Residues mentioned in the text are labeled. R. Florio et al. Role of hydrophobic contacts in serine hydroxymethyltransferase FEBS Journal 276 (2009) 132–143 ª 2008 The Authors Journal compilation ª 2008 FEBS 139 to bind PLP. In this intermediate, the N-terminus and an inter-domain segment remain exposed to solvent. In the following, slow phase, these structural elements assume their native conformation, completing the active site that acquires the capability to bind PLP. Notice- ably, one helix of the third cluster of CHCs is part of the inter-domain segment and the other helix lines the boundary between the major domain and the N-termi- nus (Fig. 1). Therefore, the formation of the third cluster may represent a key event of the final phase of folding, determining the assembly of eSHMT active site and promoting dimerization. Experimental procedures Materials Ingredients for bacterial growth were obtained from Sigma- Aldrich (St Louis, MO, USA). Chemicals for the purifica- tion of the enzymes were obtained from BDH ⁄ Merck (Whitehouse Station, NJ, USA); DEAE-Sepharose and Phenyl-Sepharose were obtained from GE Healthcare (Mil- waukee, WI, USA). Wild-type and mutant forms of eSHMT and methylenetetrahydrofolate dehydrogenase were purified as previously described [26,27]. PLP was added to protein samples during the purification procedure, but it was left out in the final dialysis step. (6S)H 4 PteGlu, was a gift from Eprova AG (Schaffhausen, Switzerland). PLP was obtained from Sigma-Aldrich (98% pure). All other reagents were obtained from Sigma-Aldrich. Preparation of apoenzyme and holoenzyme samples Apo-eSHMT was prepared using l-cysteine as previously described [10]. The apoenzyme, for which the subunit con- centration was calculated according to a molar absorptivity value of e 280 = 42790 cm )1 Æm )1 [28], was stored in 10% glycerol at )20 °C for no more than 3 days before use. A small, residual fraction (less than 5%) of holoenzyme, esti- mated with activity assays, was present in the apoenzyme samples. We noticed that different batches of purified eSHMT samples contained variable holoenzyme ⁄ apo- enzyme ratios. This observation was made with either wild-type or mutant forms of the enzyme. To carry out comparable experiments with the holo-forms, it was man- datory to prepare enzyme samples containing the same fraction of protein-bound PLP (possibly close to saturation) and, at the same time, devoid of any excess of cofactor. Holoenzymes were prepared from apoenzyme samples, by addition of PLP at the concentration needed to obtain a 98% saturation, calculated on the basis of the related disso- ciation constant of PLP binding equilibrium (Table 2), using Eqn (2A). The subunit concentration of the holo- enzyme was calculated according to a molar absorptivity value of e 280 = 44884 cm )1 Æm )1 [28]. Site-directed mutagenesis Site-directed mutagenesis of E. coli glyA (SHMT encoding gene) coding region was performed with the Quick- ChangeÔ kit from Stratagene (La Jolla, CA, USA), using the pBS::glyA plasmid as template [26] and two comple- mentary oligonucleotide primers containing the mutations, which were synthesized by MWG-Biotech AG (Anzinger, Germany). The L85A and L276A mutants were produced using as primers 5¢-CGTGCAAAGAA GCGTTCGGCGC- 3¢ and 5¢-GCGGTTGCT GCGAAAGAAGCG-3¢, respec- tively, and their complementary oligonucleotides (the mutated bases are underlined). The L85 ⁄ L276A double mutant was obtained introducing the L276A mutation into a template pS::glyA plasmid that already contained the L85A mutation. E. coli DH5a cells were used to amplify the mutated plasmids . Both strands of the coding region of the mutated genes were sequenced. The only differences with respect to wild-type nucleotide sequence were those that were intended. Enzyme expression was performed using the GS1993 recA) strain of E. coli [26]. Analytical ultracentrifugation analyses Sedimentation velocity and equilibrium experiments were carried out at 20 °Cin50mm NaHepes buffer (pH 7.2), containing 200 lm dithiothreitol, 100 lm EDTA, on a Beckman XL-I analytical ultracentrifuge equipped with absorbance optics and an An60-Ti rotor (Beckman Coul- ter, Fullerton, CA, USA). In the sedimentation velocity experiments, performed at 116 480 g the protein concentra- tion was 2.5 lm for both apo- and holo-forms. Data were collected at 277 nm and at a spacing of 0.003 cm with three averages in a continuous scan mode. In the case of the double mutant L85A ⁄ L276A holoenzyme the sedimen- tation velocity experiments were also performed at 420 nm, aiming to evaluate the presence of PLP covalently bound to the monomeric species as a Schiff base with the active site lysine residue [9]. Sedimentation coefficients and integration of data were obtained using the software sed- fit (provided by P. Schuck, National Institutes of Health, Bethesda, MD, USA). The values were reduced to water (S 20,w ) using standard procedures. The buffer density and viscosity were calculated by the software Sednterp. The ratio f ⁄ f 0 was calculated from the diffusion coefficient, which, in turn, is related to the spreading of the bound- ary, using the software sedfit. Sedimentation equilibrium experiments were performed at 7280, 10 483 and 14 270 g on enzyme samples at subunit concentrations of 2.5 and 25 lm. Data were collected every 3 h at a spacing of Role of hydrophobic contacts in serine hydroxymethyltransferase R. Florio et al. 140 FEBS Journal 276 (2009) 132–143 ª 2008 The Authors Journal compilation ª 2008 FEBS 0.001 cm with ten averages in a step scan mode. Equilib- rium was checked by comparing scans for up to 24 h using the software winmatch (J. Lary, National Analyti- cal Ultracentrifugation Center, Storrs, CT, USA). Data sets were edited by reedit (J. Lary, National Analytical Ultracentrifugation Center, Storrs, CT, USA) and ana- lyzed using the software sedphat (National Institutes of Health). Data from different speeds were combined for global fitting. When a monomer–dimer association model was required for data fitting, the monomer molecular mass was fixed at the value determined from the amino acid sequence (45 320 Da). Measurement of the K d of PLP binding equilibrium PLP binding equilibria were analyzed taking advantage of the protein intrinsic fluorescence quenching observed upon the binding event [10]. Dissociation constants of binding equilibria were then calculated from saturation curves obtained measuring the protein fluorescence emission inten- sity as a function of increasing PLP concentration. The cofactor (range 1–400 nm) was added to apoenzyme sam- ples (26 nm)at20°Cin50mm NaHepes (pH 7.2), contain- ing 200 lm dithithreitol and 100 lm EDTA. Preliminary experiments demonstrated that, with all enzyme forms, the binding equilibrium was established within the mixing time. Fluorescence emission spectra (300–450 nm; 7 nm emission slit) were recorded immediately after mixing PLP and apo- enzyme with a LS50B spectrofluorimeter (Perkin-Elmer, Waltham, MA, USA), with excitation wavelength set at 280 nm (5 nm excitation slit), at the same temperature and with a 1 cm path length quartz cell. Data were analyzed according to Eqn (2). Activity assays All assays were carried out at 20 °Cin50mm NaHepes (pH 7.2), containing 200 lm dithiothreitol and 100 lm EDTA. The serine hydroxymethyltransferase activity was measured with 0.05 lm enzyme samples with l-serine and H 4 PteGlu as substrates, as previously described [9]. To determine the K m for l-serine, H 4 PteGlu was maintained at 0.23 mm and the l-serine concentration was varied in the range 0.06–6 mm. For K m determinations of H 4 PteGlu, l-serine concentrations were held constant at 30 mm and H 4 PteGlu concentrations varied in the range 3–500 lm. The rate of l-threo-phenylserine cleavage (3 lm enzyme samples) was obtained from spectroscopic measurement of benzalde- hyde production at 279 nm, employing a molar absorptivity value of e 279 = 1400 cm )1 Æm )1 [12,13]. The K m for l-threo- phenylserine was determined by varying the substrate concentration in the range 4–200 mm. Spectroscopic measurements Fluorescence emission measurements were carried out at 20 °C with a LS50B spectrofluorimeter (Perkin-Elmer) using a 1 cm path length quartz cuvette. Fluorescence emis- sion spectra were recorded from 300–450 nm (1 nm sam- pling interval), with the excitation wavelength set at 280 nm. Far- (190–250 nm), near-UV (250–310 nm) and visible (310–500 nm) CD spectra were measured using 0.2 and 1 cm path length quartz cuvettes and the results obtained were expressed as the mean residue ellipticity (Q), assuming a mean residue molecular mass of 110 per amino acid residue. UV–visible spectra were recorded with a dou- ble-beam Lambda 16 (Perkin-Elmer). Kinetic measurements in the activity assays were performed on a HP 8453 diode- array spectrophotometer (Hewlett-Packard, Palo Alto, CA, USA). All spectroscopic measurements were carried out at 20 °Cin50mm NaHepes (pH 7.2), containing 200 lm dith- iothreitol and 100 lm EDTA. Data analysis Kinetic parameters were determined by nonlinear least squares fitting of initial velocity data to the Michaelis– Menten equation using the software prism (GraphPad Soft- ware Inc., San Diego, CA, USA). The concentrations of monomeric and dimeric species at equilibrium were calcu- lated from Eqn (1) [29], in which [D eq ] and [E t ] are the equilibrium concentrations of the dimer and the total con- centration of the enzyme, respectively, expressed as dimer equivalents, and K d is the dissociation constant of the related equilibrium. D eq Âà ¼ 8  E t ½þK d À ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 16  E t ½ÂK d þ K 2 d p 8 ð1Þ Fluorescence data obtained in PLP binding equilibrium experiments were analyzed according to Eqn (2), in which F rel is the measured relative fluorescence, F 0 is fluorescence in the absence of PLP, F inf is fluorescence at infinite PLP concentration, [E] is the total enzyme subunit concentration, [PLP] is the total cofactor concentration and K d is the disso- ciation constant of the equilibrium HOLO ¢ APO þ PLP: F rel ¼ F 0 À F 0 À F inf ðÞ PLP½þE½þK d À ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi À K d þ PLP½þE½ðÞ 2 þ4  PLP½þK d ðÞÀ4  E½ÂPLP½ q 2 E½ ð2Þ R. Florio et al. Role of hydrophobic contacts in serine hydroxymethyltransferase FEBS Journal 276 (2009) 132–143 ª 2008 The Authors Journal compilation ª 2008 FEBS 141 [...]... Scarsdale N, Kazanina G & Schirch V (2000) Role of tyrosine 65 in the mechanism of serine hydroxymethyltransferase Biochemistry 39, 7492–7500 18 Paiardini A, Gianese G, Bossa F & Pascarella S (2003) Structural plasticity of thermophilic serine hydroxymethyltransferases Proteins 50, 122–134 19 Szebenyi DM, Liu X, Kriksunov IA, Stover PJ & Thiel DJ (2000) Structure of a murine cytoplasmic serine hydroxymethyltransferase. .. V (1996) Structural studies on folding intermediates of serine hydroxymethyltransferase using single tryptophan mutants J Biol Chem 271, 2987–2994 25 Cai K & Schirch V (1996) Structural studies on folding intermediates of serine hydroxymethyltransferase using fluorescence resonance energy transfer J Biol Chem 271, 27311–27320 Role of hydrophobic contacts in serine hydroxymethyltransferase 26 Iurescia.. .Role of hydrophobic contacts in serine hydroxymethyltransferase The fraction present in Eqn (1) corresponds to the fraction of subunits that binds PLP at equilibrium  à ( HOLOeq =½EŠ) The concentration of holoenzyme at equilibrium ([HOLOeq]) was derived from the equation for the dissociation constant of the binding equilibrium: À  ÃÁ À  ÃÁ ½EŠ À HOLOeq... Schirch V, Hopkins S, Villar E & Angelaccio S (1985) Serine hydroxymethyltransferase from Escherichia coli: purification and properties J Bacteriol 163, 1–7 10 Cai K, Schirch D & Schirch V (1995) The affinity of pyridoxal 5¢-phosphate for folding intermediates of Escherichia coli serine hydroxymethyltransferase J Biol Chem 270, 19294–19299 11 Schirch V (1998) Mechanism of folate-requiring enzymes in one-carbon... 2008 The Authors Journal compilation ª 2008 FEBS R Florio et al 22 Fu TF, Boja ES, Safo MK & Schirch V (2003) Role of proline residues in the folding of serine hydroxymethyltransferase J Biol Chem 278, 31088–31094 23 Contestabile R, Paiardini A, Pascarella S, di Salvo ML, D’Aguanno S & Bossa F (2001) l-Threonine aldolase, serine hydroxymethyltransferase and fungal alanine racemase A subgroup of strictly... Savithri HS & Subramanya HS (2002) Crystal structure of binary and ternary complexes of serine hydroxymethyltransferase from Bacillus stearothermophilus: insights into the catalytic mechanism J Biol Chem 277, 17161–17169 16 Szebenyi DM, Musayev FN, di Salvo ML, Safo MK & Schirch V (2004) Serine hydroxymethyltransferase: role of glu75 and evidence that serine is cleaved by a retroaldol mechanism Biochemistry... mutagenesis techniques in the study of Escherichia coli serine hydroxymethyltransferase Protein Expr Purif 7, 323–328 27 Schirch V (1997) Purification of folate-dependent enzymes from rabbit liver Methods Enzymol 281, 146– 161 28 Malerba F, Bellelli A, Giorgi A, Bossa F & Contestabile R (2007) The mechanism of addition of pyridoxal 5¢-phosphate to Escherichia coli apo -serine hydroxymethyltransferase. .. Biophys Acta 1248, 81–96 4 Grishin NV, Phillips MA & Goldsmith EJ (1995) Modeling of the spatial structure of eukaryotic ornithine decarboxylases Protein Sci 4, 1291–1304 5 Jansonius JN (1998) Structure, evolution and action of vitamin B6-dependent enzymes Curr Opin Struct Biol 8, 759–769 6 Schneider G, Kack H & Lindqvist Y (2000) The manifold of vitamin B6 dependent enzymes Structure 8, R1–R6 7 Kack H,... hydroxymethyltransferase quinonoid ternary complex: evidence for asymmetric obligate dimers Biochemistry 39, 13313–13323 20 Renwick SB, Snell K & Baumann U (1998) The crystal structure of human cytosolic serine hydroxymethyltransferase: a target for cancer chemotherapy Structure 6, 1105–1116 21 Scarsdale JN, Kazanina G, Radaev S, Schirch V & Wright HT (1999) Crystal structure of rabbit cytosolic serine hydroxymethyltransferase. .. Schneider G & Lindqvist Y (1999) Crystal structure of diaminopelargonic acid synthase: evolutionary relationships between pyridoxal-5¢-phosphate-dependent enzymes J Mol Biol 291, 857–876 8 Paiardini A, Bossa F & Pascarella S (2004) Evolutionarily conserved regions and hydrophobic contacts at the superfamily level: the case of the fold-type I, pyridoxal5¢-phosphate-dependent enzymes Protein Sci 13, 2992– . The role of evolutionarily conserved hydrophobic contacts in the quaternary structure stability of Escherichia coli serine hydroxymethyltransferase Rita. impact of the mutations on the structure of the PLP binding site. Because PLP binding to apo-eSHMT is known to quench the intrinsic fluorescence emission of the enzyme