Dimerization of Bacterial Diaminopimelate Decarboxylase Is Essential for Catalysis*

Diaminopimelate decarboxylase (DAPDC) catalyzes the final step in the diaminopimelate biosynthesis pathway of bacteria. The product of the reaction is the essential amino acid l-lysine, which is an important precursor for the synthesis of the peptidoglycan cell wall, housekeeping proteins, and virulence factors of bacteria. Accordingly, the enzyme is a promising antibacterial target. Previous structural studies demonstrate that DAPDC exists as monomers, dimers, and tetramers in the crystal state. However, the active oligomeric form has not yet been determined. We show using analytical ultracentrifugation, small angle x-ray scattering, and enzyme kinetic analyses in solution that the active form of DAPDC from Bacillus anthracis, Escherichia coli, Mycobacterium tuberculosis, and Vibrio cholerae is a dimer. The importance of dimerization was probed further by generating dimerization interface mutants (N381A and R385A) of V. cholerae DAPDC. Our studies indicate that N381A and R385A are significantly attenuated in catalytic activity, thus confirming that dimerization of DAPDC is essential for function. These findings provide scope for the development of new antibacterial agents that prevent DAPDC dimerization.

Diaminopimelate decarboxylase (DAPDC) 3 (E.C. 4.1.1.20) is a member of the pyridoxal 5Ј-phosphate (PLP)-dependent decarboxylases (1). It catalyzes the irreversible and stereospecific decarboxylation of meso-diaminopimelate (meso-DAP) in the final step of the diaminopimelate (DAP) biosynthesis pathway of bacteria and plants (2,3). The product of the reaction, L-lysine, is an important building block for the biosynthesis of the peptidoglycan cell wall, housekeeping proteins, and virulence factors of bacteria. Consequently, DAPDC represents a promising target for the development of novel antibacterial agents (4).
A DAPDC reaction mechanism has been previously proposed (5,6). The reaction mechanism is thought to be initiated by the formation of a Schiff base between PLP and a conserved active site lysine from the (Y/F)AXKA motif (7). The substrate, meso-DAP, then binds and forms a subsequent Schiff base with PLP. This is suggested to be coordinated by the highly conserved CE(S/T)XD motif provided by an adjacent subunit (7). Decarboxylation of meso-DAP is then mediated by the sulfhydryl nucleophile provided by the conserved cysteine of the CE(S/T)XD motif thereby leaving the substrate cradled between two monomers (8). This proposed mechanism suggests DAPDC functions as a dimer.
To support this assertion, previous structural studies of DAPDC from Methanocaldococcus jannaschii (PDB codes 1TWI and 1TUF) (9), Aquifex aeolicus (PDB code 2P3E), and Brucella melitensis (PDB code 3VAB) show that the enzyme crystallizes as a homodimer, with each monomer comprised of two domains (9). Domain I forms an ␣/␤ barrel, whereas domain II is comprised of a mixed ␤-sheet flanked by ␣-helices (9). The overall fold is similar to the functionally related enzyme ornithine decarboxylase, which also uses PLP as a cofactor in a decarboxylation of a secondary amine (10). Although the crystal structures of DAPDC from all species are shown to adopt the same tertiary fold, significant differences are observed at the quaternary level in the crystal state. For example, the asymmetric unit of DAPDC from Escherichia coli (PDB codes 1KNW and 1KO0) (Fig. 1A) (11), Thermotoga maritima (PDB code 2YXX), and Helicobacter pylori (PDB code 2QGH) (8) is a monomer; whereas DAPDC from Vibrio cholerae (PDB code 3N2B) adopts a tetramer (Fig. 1B). Interestingly, DAPDC from Mycobacterium tuberculosis has been crystallized as both a dimer (PDB codes 1HKW and 1HKV) (7) and a tetramer (PDB code 2O0T) (Fig. 1, C and D) (12). However, the tetrameric form observed in the asymmetric unit of DAPDC from M. tuberculosis (Fig. 1D) differs from that observed for V. cholerae (Fig. 1B).
Adding to this confusion are the results of earlier studies aimed at characterizing the quaternary structure of DAPDC in solution. Analytical ultracentrifugation analysis of E. coli DAPDC suggested that the enzyme exists as a 5.4 S tetramer with a molecular mass of 200 kDa (13). Although small-angle x-ray scattering (SAXS) of M. tuberculosis DAPDC at 9 and 18 mg ml Ϫ1 found it to be mostly dimeric, with the scattering data best represented by a mixture of 93% dimer and 7% tetramer (12).
Accordingly, we set out in this study to comprehensively characterize the quaternary structure of DAPDC in aqueous solution from different bacterial species. Given that the aforementioned kinetic mechanism proposes that active site residues are contributed by adjacent subunits, we hypothesized that DAPDC functions as a dimer in solution.

Experimental Procedures
Protein Expression and Purification of DAPDC Enzymes-DAPDC from Bacillus anthracis (Ba-DAPDC), E. coli (Ec-DAPDC), and M. tuberculosis (Mt-DAPDC) were expressed and purified as described in Peverelli and Perugini (14). Expression and purification of V. cholerae DAPDC (Vc-DAPDC) was achieved as follows. The nucleotide sequence of the V. cholerae DAPDC gene (lysA) was first identified from the Kyoto Encyclopedia of Genes and Genomes database as entry VC0125. The amino acid sequence was subsequently altered to match the amino acid sequence from PDB code 3N2B. Next, the codonoptimized nucleotide sequence was commercially synthesized by GeneArt (Life Technologies) and subcloned into the expression vector pET11a (Novagen). The vector was then transformed into E. coli BL21(DE3) cells and expression and purification of the recombinant enzyme was performed as previously described (14,15).
Generation of Vc-DAPDC N381A and R385A Mutants-The QuikChange II XL site-directed mutagenesis kit (Stratagene) was employed to introduce the R385A point mutation into the V. cholerae lysA gene ligated into pET11a using the primer set of 5Ј-CAATTACAATACCCGTCCGGCTGTTGCCGAAGT-TATGGTT-3Ј and 5Ј-AACCATAACTTCGGCAACAGCCG-GACGGGTATTGTAATTG-3Ј (mismatched nucleotides are underlined). Mutagenesis was performed according to the manufacturer's instructions with successful mutations confirmed by dideoxynucleotide sequencing. By contrast, the N381A mutant was commercially synthesized by Bioneer and subcloned into the expression vector pET11a (Novagen). The recombinant N381A and R385A mutants were expressed and purified using the same protocol adopted for the wild-type enzyme.
Mass Spectrometry-To determine the molar mass of purified recombinant DAPDC enzymes, electrospray ionization time-of-flight mass spectrometry was performed as previously described by Hor et al. (16). Recombinant proteins were also subjected to trypsin digestion and the resulting tryptic fragments were sequenced by tandem mass spectrometry as reported previously (17).
Circular Dichroism (CD) Spectroscopy-CD spectroscopy was performed using an Aviv Biomedical Model 410-SF CD spectrophotometer as reported previously (16, 18 -20). Samples typically consisted of 0.15 mg ml Ϫ1 enzyme in a buffer of 20 mM Tris, 150 mM NaCl, 1 mM Tris(2-carboxyethyl)phosphine (TCEP), pH 8.0. Spectra were measured in a 1-mm quartz cuvette between 190 and 250 nm employing a step size of 0.5 nm and 4-s averaging time. Data were analyzed using the CONTINLL algorithm and SP43 database from the CDPro software package (21).
Absorbance Detected Analytical Ultracentrifugation-Sedimentation velocity and sedimentation equilibrium experiments were performed in a Beckman XL-I analytical ultracentrifuge with a four-hole An-60 Ti or eight-hole An-50 Ti rotor at 20°C as described previously (16, 18 -20, 22, 23). The partial specific volume (v ) of the DAPDC samples together with the density and viscosity of the buffers were computed using the program SEDNTERP (24). For sedimentation velocity experiments, double sector quartz cells were loaded with 400 l of reference (20 mM Tris, 150 mM NaCl, 1 mM TCEP, pH 8.0) and 380 l of sample (3 M), and data were collected at 40,000 rpm from 5.8 to 7.3 cm using a step size of 0.003 cm without averaging. Sedimentation velocity data were analyzed using the continuous size-distribution model (25,26) employing the program SEDFIT. For sedimentation equilibrium experiments of R385A and N381A, double sector quartz cells were loaded with 140 l of reference and 100 l of sample (0.4 -1.0 M) with the sample sector also containing 20 l of heavy (FC43) oil. Experiments were performed at 9,600 rpm and 14,300 rpm. When sedimentation equilibrium was attained (ϳ24 h at each speed), data were collected from 6.7 to 7.3 cm using a 0.001-cm step size with 10 averages. The resulting dual speed sedimentation equilibrium data at multiple protein concentrations were fitted globally to various self-association models, including a monomer-dimer model (16,23), using SEDPHAT software (27).
Fluorophore Labeling of Vc-DAPDC-Prior to labeling, 1.0 ml of Vc-DAPDC (7 mg ml Ϫ1 ) was dialyzed overnight at 4°C against 5 liters of 100 mM sodium bicarbonate, pH 8.3, 1 mM TCEP and containing 0.1 mM PLP to protect the active site lysine (Lys 72 ) from modification. 160 nmol of Alexa Fluor 488 5-tetrafluorophenyl ester (dissolved in 60 l of dimethyl sulfoxide) was then added to the dialyzed Vc-DAPDC sample and the mixture was incubated at room temperature for 80 min with continuous mixing on a rocking platform. Unincorporated fluorophore was removed by Superose 12 size exclusion liquid chromatography on a 1.0 cm ϫ 30-cm column previously equilibrated with 20 mM Tris, 150 mM NaCl, 1 mM TCEP, 0.1 mM PLP, pH 8.0. The optimized labeling procedure yielded a ratio of 1.2:1 fluorophore to Vc-DAPDC monomer (determined using manufacturer's instructions).
Fluorescence Detected Analytical Ultracentrifugation (AU-FDS)-Sedimentation velocity and sedimentation equilibrium experiments were performed in a Beckman XL-A analytical ultracentrifuge equipped with a fluorescence detection system (Aviv Biomedical) using Alexa Fluor 488-labeled Vc-DAPDC at nanomolar concentrations. For sedimentation velocity experiments, double sector cells containing sapphire windows were loaded with 400 l of sample. Data were collected at 40,000 rpm from 5.8 to 7.3 cm at multiple photomultiplier voltages to yield ϳ3,000 fluorescence counts for the highest concentration in each cell. The resulting fluorescence versus radial data were analyzed using a continuous size distribution model (25,26) employing SEDFIT software. For sedimentation equilibrium experiments, double sector sapphire cells were loaded with 120 l of sample and 40 l of heavy oil. Samples were centrifuged at 9,000 and 14,000 rpm until sedimentation equilibrium was attained (ϳ24 h for each speed). At equilibrium, data were collected from 6.7 to 7.3 cm at multiple photomultiplier voltages to optimize the signal-to-noise ratio and initial fluorescence counts to ϳ1,000. Resulting data were corrected for photomultiplier voltage differences and normalized to pseudo-absorbance values. The pseudo-absorbance profiles were then fitted to various self-association models, including a monomer-dimer model (16,23), using SEDPHAT software (27).
Small-angle X-ray Scattering-SAXS experiments were performed on the SAXS/WAXS beamline at the Australian Synchrotron (Melbourne, Australia). Buffer and enzyme samples at multiple concentrations were loaded into 96-well plates and incubated on a temperature-controlled sample platform at 20°C for the duration of the experiment. Scattering data of the x-ray beam were collected using a Pilatus-1 M pixel-array detector (Dectris, Switzerland). The sample-to-detector distance was 1.6 m, covering a range of momentum transfer 0.01 Å Ϫ1 Ͻ q Ͻ 0.6 Å Ϫ1 (q ϭ 4sin/, where 2 is the scattering angle and ϭ 1.03 Å is the x-ray wavelength). Data sets of ϳ20 frames covering a 2-s exposure time were collected from a sample exposed to an x-ray beam through a 1.5-mm quartz capillary. Frames were converged and buffer scattering subtracted using ScatterBrain software (Australian Synchrotron). Guinier plots and P(r) distributions were produced using PRIMUS (28 -31), whereas scattering fits to crystal structures were performed using CRYSOL (32). Data generated for Ec-DAPDC and Vc-DAPDC were fitted against their respective crystal structures (PDB codes 1KNW and 3N2B), whereas the data for Mt-DAPDC were fitted against its dimeric crystal structure (PDB code 1HKV) or the tetramer generated from the T. maritima DAPDC (PDB code 2YXX). Data collected on Ba-DAPDC were fitted against PDB code 1HKV, as M. tuberculosis DAPDC is the closest homolog available in the PDB. Ab initio models of each DAPDC enzyme were generated using DAMMIN software (available from ATSAS online) (33).
Enzyme Kinetics Assay-DAPDC enzyme kinetic assays were performed as described in Peverelli and Perugini (14). Briefly, pre-initiation mixtures comprised of 200 mM Tris buffer, pH 8.0, 0.25-40 mM DAP, 2.5 M saccharopine dehydrogenase, 0.16 mM NADH, 1 mM TCEP, 25 mM ␣-ketoglutarate, and 0.1 mM PLP were incubated for 10 min at 37°C, DAPDC (final concentration of 5-500 nM) was then added to initiate the reaction, which was monitored by measuring the decrease in absorbance at 340 nm as a function of time using a temperaturecontrolled Varian Cary 4000 UV-visible spectrophotometer at 37°C. Assays were performed in triplicate and analyzed using a Michaelis-Menten kinetic model to determine the K m for meso-DAP and k cat as reported previously (14).

PISA Analyses of DAPDC Crystal
Structures-To support our hypothesis that DAPDC functions as a dimer, we first performed computational analyses of DAPDC crystal structures from E. coli (PDB code 1KNW), M. tuberculosis (PDB code 1HKV), and V. cholerae (PDB code 3N2B). As reported earlier, these orthologs crystallize as monomers, dimers/tetramers, and tetramers, respectively. The in silico analyses were performed using the Protein Interfaces, Surfaces and Assemblies (PISA) web-based server (34), which calculates the nature of interactions stabilizing protein-protein interfaces in crystal structures. The PISA algorithm can therefore aid in discriminating between interfaces resulting from crystal artifacts versus those that are structurally relevant (34,35). The PISA results conducted on 1KNW, 1HKW, and 3N2B suggest the enzymes exist as dimers (Fig. 1, E-G). We were therefore interested in confirming these predictions by studying the quaternary structure of the enzymes in aqueous solution.
Recombinant DAPDC Enzymes Are Folded and Enzymatically Active-Before assessing and comparing the quaternary structure of Ec-DAPDC, Mt-DAPDC, and Vc-DAPDC, we first needed to express and purify milligram quantities of the recombinant enzymes. To expand the diversity of orthologs to also include a DAPDC from a Gram-positive bacterium, we also expressed and purified Ba-DAPDC. All four enzymes were produced in milligram quantities (1.69 -132 mg) with Ͼ95% purity as assessed by SDS-PAGE. Mass spectrometric analyses showed that the recombinant products were comprised of the correct primary structure. CD spectroscopy studies were also performed to confirm that Ba-DAPDC, Ec-DAPDC, Mt-DAPDC, and Vc-DAPDC adopt ␣/␤ structures in solution (Table 1), consistent with x-ray crystal structure analyses (7-9, 11, 12). The enzymatic activity of the folded recombinant enzymes were next determined and compared using a recently optimized coupled assay employing saccharopine dehydrogenase (14,36). The enzyme kinetic data for Vc-DAPDC is shown in Fig. 2. The resultant data fitted well to the Michaelis-Menten equation (R 2 ϭ 0.99) thereby imparting a K m meso-DAP of 1.9 mM, a V max of 27 mol min Ϫ1 mg Ϫ1 , and a catalytic rate constant (k cat ) of 22 s Ϫ1 (Table 2). This is similar to the kinetic parameters determined for Ba-DAPDC, Ec-DAPDC, and Mt-DAPDC (Table 2). We were next interested in determining the quaternary structure of the active recombinant enzymes in solution.
DAPDC Exists as a Dimer in Solution-To characterize and compare the quaternary structure of Ba-DAPDC, Ec-DAPDC, Mt-DAPDC, and Vc-DAPDC, absorbance-detected sedimentation velocity experiments were initially conducted in the analytical ultracentrifuge at a protein concentration of 150 g ml Ϫ1 (3 M). The raw absorbance versus radial position data were fitted to a continuous size-distribution model yielding the c(s) distributions depicted in Fig. 3. The results show that Ba-DAPDC and Vc-DAPDC exist as a monodisperse 5.6 S species (Fig. 3, A and D; Table 3). The equivalent c(M) distribution results in a single species with a modal molecular mass of 83 kDa for Ba-DAPDC and 92 kDa for Vc-DAPDC (Table 3). This suggests that the 5.6 S species is a dimer. Similarly, Ec-DAPDC and Mt-DAPDC exist primarily as a ϳ5.6 S species, but the c(s) distributions also show a small proportion of a ϳ4 S species (Fig. 3, B and C; Table 3). The c(M) distribution results for Ec-DAPDC and Mt-DAPDC indicate that the ϳ4 S species has a molecular mass of ϳ50 kDa ( Table 3), suggesting that this species is a monomer. Nevertheless, the results of the sedimentation velocity experiments demonstrate that DAPDC from all four bacterial species primarily exist as dimers in aqueous solution. We therefore set out to validate these results using smallangle x-ray scattering.
SAXS Confirms DAPDC Is Dimeric-To support the initial analytical ultracentrifugation results, SAXS experiments were performed using a synchrotron x-ray source at DAPDC concentrations of 3-30 M. The resulting scattering data were subjected to Guinier analysis (28,29) to determine the radius of gyration (R g ) ( Table 4). The maximum dimension of the scattering particle (D max ) for each DAPDC enzyme (Table 4) were determined from the pairwise distance-distribution function (P(r)) calculated using the indirect Fourier transform method (30). Scattering data were also evaluated using CRYSOL (32) incorporating appropriate PDB entries of DAPDC enzymes ( Fig. 4; Table 4). These analyses showed that the experimental scattering profiles for Ba-DAPDC, Ec-DAPDC, and Vc-DAPDC best fit to the theoretical scattering profile of a dimer (Fig. 4, A, B, and D). However, the scattering data for Mt-DAPDC fit best to a dimer at 3 M, but to a tetramer at 30 M (Fig. 4C), which is consistent with previous studies (12). Furthermore, ab initio bead models were created using DAM-MIN (33) to establish shape envelopes from the scattering data of each DAPDC sample. Crystal structures were overlaid with the bead models as shown in Fig. 4 (insets). Consistent with the CRYSOL results, the ab initio models exhibit shape envelopes that correspond to dimers for Ba-DAPDC, Ec-DAPDC, and Vc-DAPDC (Fig. 4, A, B, and D). Whereas, the bead models for Mt-DAPDC generated from the lowest (3 M) and highest (30 M) concentration datasets displayed a dimeric and tetrameric shape envelope, respectively (Fig. 4C). The bead models provide further evidence that the structurally relevant quaternary structure of DAPDC is a dimer. We were thus interested in further probing the relationship between the dimeric structure and enzyme function. To do this, we decided to create mutant forms of Vc-DAPDC with attenuated dimerization propensity and compare the catalytic activity of the mutants to the wild-type dimer. Nevertheless, it was important to first determine the dimerization affinity of wild-type Vc-DAPDC to enable comparison with dimer attenuated mutant forms.
AU-FDS Studies of Vc-DAPDC-Although the initial sedimentation velocity studies outlined earlier indicate that Vc-DAPDC exists as a stable dimer at a concentration of 3 M (Fig.  3D); we set out to examine the concentration dependence of Vc-DAPDC dimerization by diluting the enzyme concentration by 5-fold to 30 g ml Ϫ1 (0.6 M) (Fig. 5A). However, no significant dissociation of the dimer was observed. This indicated     distribution. c Frictional ratio derived using the v bar method (24). d Determined from global nonlinear best fit to a monomer-dimer equilibrium model using sedimentation equilibrium data at multiple concentrations and rotor speeds. e ND, not determined. a Experimental data were fitted to theoretical scattering profiles generated from appropriate atomic coordinates using the program CRYSOL (32). b Errors are estimated automatically with the help of a polynomial smoothing procedure (32). perturb dimerization. At 150 g ml Ϫ1 , the labeled enzyme sedimented as a single species with an s 20,w value of 5.8 S (Fig. 5B) and a molecular mass of 98 kDa, consistent with the unlabeled enzyme ( Fig. 3D; Table 3). Sedimentation velocity experiments were thus performed in AU-FDS using 5 nM fluorescently labeled enzyme. The resulting continuous size distribution shows evidence of dissociation of the 5.6 S dimer to a 3.6 S monomer ( Fig. 5C; Table 3). Importantly, we also determined that the specific activity of the Alexa Fluor 488-labeled enzyme was 22.4 units mg Ϫ1 , which is similar to the unlabeled enzyme ( Fig. 5D; Table 2). This demonstrates that labeling did not significantly alter the catalytic function of Vc-DAPDC.
To quantify the dimerization affinity of Alexa Fluor 488labeled Vc-DAPDC, sedimentation equilibrium experiments were conducted at two rotor speeds (9,000 and 14,000 rpm) and four enzyme concentrations (2, 4, 20, and 40 nM). The resulting data were globally fitted to various self-association models, including monomer-dimer, monomer-trimer, monomer-tetramer, and dimer-tetramer. The global nonlinear best fit resulted to a monomer-dimer model, which yielded a dimermonomer dissociation constant (K D 231 ) of 0.061 nM ( Fig. 6; Table 3). This is very tight, therefore supporting the hypothesis that dimerization is critical for DAPDC function. We next set out to design dimerization attenuation mutants to compare with the wild-type Vc-DAPDC dimer.
Dimer Interface Mutants Are Attenuated in Dimerization Propensity and Catalytic Activity-There are several residues that facilitate subunit interactions at the dimerization interface of Vc-DAPDC. This includes Arg 385 , which forms a hydrogen bond with Asn 379 on the adjacent subunit (Fig.  7A). We therefore decided to mutate Arg 385 to an Ala in an attempt to create a dimerization attenuated mutant. Suc- cessful implementation of the point mutation R385A was confirmed by dideoxynucleotide sequencing and mass spectrometry. The structure of R385A was analyzed by CD spectroscopy, which demonstrates that the point mutation did not perturb the secondary structure of the enzyme relative to wild-type Vc-DAPDC (Table 1).
Next, the quaternary structure of R385A was investigated using analytical ultracentrifugation. Sedimentation velocity experiments were performed over a concentration range of 0.4 -20 M. At 0.4 M, the continuous size distribution analysis showed that the enzyme exists in a monomer-dimer equilibrium with similar proportions of each species (Fig. 7B; Table 3). This suggests the dimerization dissociation constant of R385A is ϳ400 nM. To better quantify the dimerization affinity of R385A, sedimentation equilibrium experiments were performed at two rotor speeds (9,600 and 14,300 rpm) and multiple protein concentrations (400, 600, and 1,000 nM). The resulting data were fitted to a monomer-dimer self-association model, which yielded a K D 231 of 0.41 M consistent with the sedimentation velocity results (Fig. 7, D-F; Table 3). This indicates that dimerization of the R385A mutant is ϳ6,660-fold weaker than wild-type Vc-DAPDC.
The catalytic activity of R385A was next examined and compared with the wild-type enzyme. At a fixed concentration of 100 nM, the specific activity of R385A is 1.9 units mg Ϫ1 , which is ϳ8% of the wild-type enzyme. These results demonstrate that the wild-type Vc-DAPDC dimer is considerably more active than the dimer attenuated R385A mutant. Furthermore, the point mutant does not display a linear relationship between rate and enzyme concentrations (Fig. 7C). A distinct upward curvature is observed between 5 and 500 nM, which is indicative of a self-associating enzyme where the lower order species is significantly less active than the higher order oligomer (16,20,23). By contrast, the wild-type enzyme displays a linear rate versus concentration relationship at concentrations (5-50 nM) well above the K D 231 (Fig. 7C). To further validate that loss of catalytic function correlates with attenuation in dimerization, a second dimer interface mutant (N381A) was cloned, expressed, purified, and characterized in solution (Tables 1-3; Fig. 7, B and C). Asn 381 is posi- tioned at the dimer interface ϳ10 Å from Arg 385 , and forms an inter-chain hydrogen bond with Glu 86 . Although less pronounced than the R385A mutant, the results show that N381A is also attenuated in dimerization propensity ( Fig. 7B; Table 3) and catalytic activity (Fig. 7C). Indeed, comparison of the structure-activity properties of wild-type Vc-DAPDC, N381A, and R385A shows that specific activity (Table 2) correlates with dimerization affinity (i.e. decreasing K D 231 ) ( Table 3).

Discussion
Given that the catalytic mechanism proposed for DAPDC assumes the enzyme exists as a dimer (7)(8)(9)12), it is interesting to note that previous studies have not yet established whether the enzyme functions as a dimer in solution. Although the crystal structure of DAPDC from some species suggests the enzyme exists as a dimer, monomeric and tetrameric asymmetric units are also observed in the crystal state for several DAPDC structures. Nevertheless, we hypothesized that DAPDC functions as a dimer and therefore non-dimeric forms observed in the crystalline state are artifacts of crystallization or simply represent the asymmetric unit. Accordingly, the overall aim of this study was to characterize the quaternary structure of DAPDC from different bacterial species in aqueous solution.
Consistent with our hypothesis, we show using analytical ultracentrifugation, SAXS, and enzyme kinetics that DAPDC from B. anthracis, E. coli, M. tuberculosis, and V. cholerae are functional dimers in solution. These results offer strong support to the proposed catalytic mechanism implicating an essential role for dimerization for DAPDC function (7-9, 12). Moreover, the critical CE(S/T)XD motif required for catalysis is provided by the adjacent subunit in the dimeric complex. Dimerization has also been shown to be critical for other family members of the PLP-dependent decarboxylases, such as eukaryotic ornithine decarboxylase, where the active site is completed by the contribution of Lys 69 and Cys 360 from opposing subunits (10,37). These are equivalent to the DAPDC Lys and Cys residues in the respective (Y/F)AXKA and CE(S/T)XD motifs. Mutation of these residues in mouse ornithine decarboxylase leads to heavily attenuated activity (38,39). Clearly, these cross-interface interactions are highly important. Once sedimentation of samples were at equilibrium at 9,000 (E) and 14,000 (ƒ) rpm, fluorescence emission readings of Alexa Fluor 488 from 519 nm were collected, corrected for photomultiplier voltage differences, and finally normalized to pseudo-absorbance values (y axes) for use in the program SEDPHAT. Overlaid to the experimental data are the global nonlinear regression best fit to a monomer-dimer self-association model (global reduced 2 ϭ 0.039). Residuals are displayed above each plot for 9,000 (E) and 14,000 (ƒ) rpm data. The Interestingly, these interactions are also observed for several other enzymes, including dihydrodipicolinate synthase, which catalyzes the first committed and rate-limiting step in the DAP biosynthesis pathway (2,3). For dihydrodipicolinate synthase, dimerization has been shown to be important for completing the active site of each monomer through the interdigitation of a key catalytic triad residue (18,20,22,23,35,40). Similarly, recent studies demonstrate that dimerization is required for the catalytic activity of diaminopimelate epimerase, which also functions in the DAP pathway (16). Additionally, dimerization has been proposed to be required for reorientating the active site in the activation of apoptotic initiator caspase 9 (41). Indeed, it is predicted that oligomerization is critical for providing key active site determinants in ϳ1 in 6 enzymes (42). The importance of enzyme oligomerization is therefore not always pertinent for evolving allosteric regulation sites, because more and more examples of the role of quaternary structure for the completion of active sites are being discovered. One such example is reported in this study.
Protein-protein interfaces represent highly specific drug targets, as conservation is much higher in active sites relative to interfaces. Given the increase in drug-resistant bacteria associated with the overuse and misuse of broad spectrum antibacterials, exploiting the dimerization interface of DAPDC may provide a means to negate the incidence of broad spectrum drug resistance (2, 3, 43-45).
In conclusion, we show that DAPDC from multiple bacterial species form dimers in solution and that self-association is required for activity. Consequently, preventing dimerization of DAPDC may provide a viable approach for inhibition and subsequent discovery of new antibacterial agents.
Author Contributions-M. A. P. conceived the project; M. G. P. and M. A. P. designed the majority of the experiments and co-wrote the manuscript; M. G. P. and T. P. S. C. performed the experiments and analyzed the data; N. K. co-designed, co-performed, and assisted M. G. P. with the analysis of the SAXS experiments.