Conserved residue His-257 of Vibrio cholerae flavin transferase ApbE plays a critical role in substrate binding and catalysis

The flavin transferase ApbE plays essential roles in bacterial physiology, covalently incorporating FMN cofactors into numerous respiratory enzymes that use the integrated cofactors as electron carriers. In this work we performed a detailed kinetic and structural characterization of Vibrio cholerae WT ApbE and mutants of the conserved residue His-257, to understand its role in substrate binding and in the catalytic mechanism of this family. Bi-substrate kinetic experiments revealed that ApbE follows a random Bi Bi sequential kinetic mechanism, in which a ternary complex is formed, indicating that both substrates must be bound to the enzyme for the reaction to proceed. Steady-state kinetic analyses show that the turnover rates of His-257 mutants are significantly smaller than those of WT ApbE, and have increased Km values for both substrates, indicating that the His-257 residue plays important roles in catalysis and in enzyme-substrate complex formation. Analyses of the pH dependence of ApbE activity indicate that the pKa of the catalytic residue (pKES1) increases by 2 pH units in the His-257 mutants, suggesting that this residue plays a role in substrate deprotonation. The crystal structures of WT ApbE and an H257G mutant were determined at 1.61 and 1.92 Å resolutions, revealing that His-257 is located in the catalytic site and that the substitution does not produce major conformational changes. We propose a reaction mechanism in which His-257 acts as a general base that deprotonates the acceptor residue, which subsequently performs a nucleophilic attack on FAD for flavin transfer.

The flavin transferase ApbE plays essential roles in bacterial physiology, covalently incorporating FMN cofactors into numerous respiratory enzymes that use the integrated cofactors as electron carriers. In this work we performed a detailed kinetic and structural characterization of Vibrio cholerae WT ApbE and mutants of the conserved residue His-257, to understand its role in substrate binding and in the catalytic mechanism of this family. Bi-substrate kinetic experiments revealed that ApbE follows a random Bi Bi sequential kinetic mechanism, in which a ternary complex is formed, indicating that both substrates must be bound to the enzyme for the reaction to proceed. Steady-state kinetic analyses show that the turnover rates of His-257 mutants are significantly smaller than those of WT ApbE, and have increased K m values for both substrates, indicating that the His-257 residue plays important roles in catalysis and in enzymesubstrate complex formation. Analyses of the pH dependence of ApbE activity indicate that the pK a of the catalytic residue (pK ES1 ) increases by 2 pH units in the His-257 mutants, suggesting that this residue plays a role in substrate deprotonation. The crystal structures of WT ApbE and an H257G mutant were determined at 1.61 and 1.92 Å resolutions, revealing that His-257 is located in the catalytic site and that the substitution does not produce major conformational changes. We propose a reaction mechanism in which His-257 acts as a general base that deprotonates the acceptor residue, which subsequently performs a nucleophilic attack on FAD for flavin transfer.
Remarkably, the incorporation of the FMN cofactor through this phosphoester bond is not autocatalytic, as in all other reported cases, but is carried out by ApbE (alternative pyrimidine biosynthesis protein, subunit E) (20,22,23), the only flavin

Flavin-transfer activity measurement
The flavin transferase activity of ApbE was studied using NqrC as the protein substrate. NqrC is one of the two subunits of NQR that carry the covalently-attached FMN (17)(18)(19)(20). For these experiments the transmembrane segment of NqrC was eliminated for optimal protein expression, as reported previously (23). In addition, ApbE activity was investigated in the presence of potassium, which was recently shown to specifically activate the enzyme (23), to mimic the physiological conditions that V. cholerae encounters during infection.
ApbE flavin-transfer activity can be measured by following the fluorescence of covalently-bound FMN to NqrC in SDS-PAGE gels exposed to UV light (23). However, this method has many limitations, such as moderate accuracy and a high amount of sample required. Here, we report a new spectrophotometric method to follow the flavin-transfer activity. This method exploits the absorbance difference between the free FAD and the covalently-bound FMN. As shown in Fig. 1A, the fully flavinylated NqrC (solid line) displays a distinctive absorption spectrum, which is different from that of free FAD (gray line). The difference spectrum shows a peak at 395 nm with an isosbestic point at 366 nm (Fig. 1A, inset). By following the absorbance change at 395 nm, using 366 nm as reference (Fig.  1B), the activity of ApbE can be measured accurately and inexpensively. It should be pointed out that the change in absorption spectra of the flavin is probably due to the interactions of FMN with the folded NqrC protein. Indeed, upon denaturation with 0.1% SDS, the absorption spectrum of flavinylated NqrC resembles that of free FMN (Fig. 1A, dashed line). As can be seen in Fig. 1B, the in-gel activity data points (squares) match very well with the activity trace obtained with the spectrophotometric method developed in this work. Furthermore, the spectrophotometric method allows the collection of substantially more data points, compared with the SDS-PAGE gel method, and thus the activity rates can be calculated more accurately.

Bi-substrate kinetics
To investigate the kinetic mechanism of ApbE, bi-substrate kinetic experiments were carried out measuring ApbE activity under different fixed concentrations of NqrC, while varying the concentration of FAD. Data were globally fitted to the three bi-substrate kinetic models: Random, Ordered, and Ping-Pong (Equations 1-3). Statistical analyses show that the data were best fitted to either Random or Ordered models, whereas the fitting to the Ping-Pong model yielded a higher 2 (Fig. 2, A and  B). Moreover, the double-reciprocal plots show intercepting patterns, also consistent with sequential mechanisms, in which a ternary complex (ApbE-FAD-NqrC) is formed (Fig. 2E).

Kinetic and reaction mechanisms of V. cholerae ApbE
However, this analysis cannot distinguish between Random and Ordered models. To distinguish between these mechanisms, product inhibition kinetic experiments were performed. In particular, the inhibition mechanism of AMP versus NqrC was studied. It has been previously reported that AMP behaves as an inhibitor of ApbE (23), but its mechanism of action has not been investigated. The data obtained indicate that AMP acts as a mixed inhibitor with respect to NqrC, with K ic and K iu of 17 and 5 M, respectively (Fig. 2, C and D). This type of inhibition is expected in Random kinetic mechanisms (Fig. 2E). Interestingly, a kinetic component resistant (30%) to saturating concentrations of AMP was also discovered. The resistant component is consistent with an allosteric site that is regulated by adenine nucleotides. In a previous report (23) we studied ApbE regulatory mechanisms. According to our data, ApbE activity is strongly stimulated by 1 mM ADP (6 -7-fold), whereas AMP and ATP produced a 10 -20% inhibition. It is possible that at high concentrations (Ͼ10 mM) AMP could be bound to the regulatory site, which might de-inhibit the enzyme.

Steady-state kinetic characterization of His-257 mutants
Previously, the conserved residue His-257, which directly faces the pyrophosphate moiety of FAD, was proposed as part of the catalytic site of ApbE (23). The mutation to glycine completely abolished ApbE activity, suggesting that this residue plays an essential role in enzymatic catalysis. We proposed a covalent catalysis mechanism in which His-257 could react with FAD, producing a covalent intermediate that mediates the incorporation of FMN to the acceptor residue NqrC-Thr-225 (23). To understand the role of His-257, different mutants were obtained and characterized in this work. In particular, His-257 was mutated to Gly, Thr, Asp, and Lys, to characterize the effects of the charge and size of the residue on the flavin-transfer activity. The results show that all mutants are inactive at pH 7.0 ( Fig. 3C, inset), but when the activity is measured at pH 9.0, the mutants H257G and H257T have 4 -5 times lower turnover rates compared with WT ApbE (Table 1). It should be pointed out that we previously reported that the mutant H257G was completely inactive (23). However, in the previous experiments the results were obtained in the absence of the physiologic regulator potassium, which greatly activates the enzyme. On the other hand, mutants H257K and H257E showed negligible activities at all pH values assayed, probably due to the net charges interfering with substrate binding (see below). This result indicates that His-257 is not essential for enzyme function, and that it does not form a covalent intermediate with FMN during the reaction, and thus His-257 may participate in a different mechanism. This is consistent with the Random Bi Bi sequential mechanism of the WT enzyme, because the formation of a covalent flavin-imidazole intermediate would be required for a Ping-Pong type mechanism. Mutants H257G and H257T were further characterized. Both mutants show a significant increase in the K m values for FAD and NqrC compared with the WT enzyme, respectively (Fig. 3, A and B, and Table 1), indicating that His-257 plays a major role in the formation of the enzyme-substrate complex and participates directly in catalysis.

Characterization of catalytic residues pK a and pK a predictions
To understand the role of His-257 in the reaction mechanism of ApbE, a characterization of the effects of pH on the activity was carried out, which allows the study of the pK a of residues involved in the reaction process. In a previous work we showed that the catalytic residue of WT ApbE has a pK a of 8.4 (23), corresponding to pK ES1 in Equation 5, which was assigned His-257. To study the pH dependence of ApbE activity, NqrC titrations were performed under different pH values (Fig. 3C). In these experiments 50 M FAD was used, which is near-saturating at all pH values tested (Fig. S1). Global analyses of the data using Equation 5 show that ApbE displays nearly identical pK E1 and pK ES1 of 7.2 and 6.8, close to the pK a of a histidine residue. These pK a values indicate the dissociation constants of residues whose protonation determine substrate binding (pK E1 ) and catalysis (pK ES1 ), i.e. pK E1 and pK ES1 are the pK a values at which the deprotonated enzyme is able to bind the substrate and become catalytically active, respectively (37). Thus, the data

Kinetic and reaction mechanisms of V. cholerae ApbE
suggests that this residue has important roles in catalysis and in the formation of the enzyme-substrate complex.
To characterize the effects of the mutants on these parameters, the activities were measured at near-saturating concentrations of NqrC (500 M) and FAD (50 M) at different pH values. In these conditions the activities measured are close to the V max and can help in the estimation of apparent pK ES1 values for the mutants. The results showed that the pK ES1,app values of H257G and H257T mutants were significantly shifted up, by more than 2 pH units compared with the WT enzyme (Fig. 3C, inset, and Table 2). The shift in pK ES1 strongly suggests that His-257 acts as a general base that deprotonates the FMN acceptor residue Thr-225, which could allow a nucleophilic attack on FAD and the subsequent incorporation of the FMN moiety (Fig. 4D). It should be pointed out that experiments could not be carried out at pH values higher than 10.5, due to the formation of precipitates in the reaction buffer (probably MgOH 2 ).
Deka et al. (38) demonstrated that conserved Lys-207 in NqrC is important for flavin transfer, and that this residue points directly to the acceptor residue Thr-225. It is likely that Lys-207 may lower the pK a of Thr-225, as threonine generally does not undergo deprotonation under physiologic conditions, which could allow flavin transfer assisted by His-257. To gain insight into this hypothesis, theoretical pK a predictions of NqrC residues 225 and 207 were carried out using the program PROPKA3.1 (39,40). Because PROPKA3.1 only predicts pK a values of conventional ionizable residues, Thr-225 was mutated in silico to cysteine, a structurally similar but ionizable residue, using UCSF Chimera (41). Furthermore, to examine the effects of Lys-207 on residue 225, Lys-207 was also mutated to Ala, Gln, Glu, and Arg using the same method. PROPKA 3.1 also calculates how much nearby residues contribute to a pK a change. As listed in Table 3, an Ala or Gln residue at position 207 is predicted to have no effect on the pK a of Cys-225. On the other hand, Arg and Glu shift the pK a of Cys-225 by about 0.3-0.4 pH units, consistent with the previously suggested hypothesis (38) that a positive charge can help activate Thr-225 but it is not sufficient to carry out catalysis.

Crystal structures of ApbE
To fully understand of role of His-257, the crystal structures of WT ApbE and H257G were investigated. Crystals grown under conditions supplemented with FAD and MgCl 2 (as described under "Experimental procedures") were used for data collection. Resolutions for WT and H257G were determined to be 1.61 and 1.92 Å, respectively. Overall, ApbE WT structure shows a significant similarity to previously published ApbE structures from Salmonella enterica (42), Pseudomonas stutzeri (22), Treponema pallidum (43), Escherichia coli (38), and Ther-

Kinetic and reaction mechanisms of V. cholerae ApbE
motoga maritima (44) (Fig. S2). Superposition of the two structures showed that no significant changes occurred due to the mutation (r.m.s. deviation ϭ 0.418 Å, calculated by UCSF Chimera, Fig. 4A (Fig. S3). In addition, His-257 in the WT also forms a hydrogen bond through the N␦1 of its imidazole side chain with the 3Ј-OH of the FAD ribose moiety (Fig. 4B) and locks FAD inside the pocket (Fig. 4C). Thus, the crystallographic data reveal that mutation of His-257 did not result in significant changes in the enzyme structure, but it could have an effect on FAD binding, further demonstrating that His-257 does not play a structural role. Instead, it is an important residue in catalysis.

Discussion
The newly discovered family ApbE is the only known flavin transferase that covalently integrates FMN cofactors into different respiratory enzymes of various pathogenic bacteria. In this work we characterized the kinetic and reaction mechanisms of flavin transfer carried out by ApbE, providing insight to understand its catalytic mechanism.

Structural insight into ApbE mechanism
High-resolution X-ray crystallographic structures were obtained for ApbE WT and H257G. Superposition of the WT and mutant structures show that the mutation does not cause significant conformational changes. Examination of the active site in both crystal structures demonstrate that the FAD-binding pocket remains undisturbed upon mutation of His-257, in terms of both the protein-FAD interactions and the active site conformation (Fig. 4A and Fig. S2). However, kinetic studies show that the mutation substantially changed the K m values for both substrates, which in the case of FAD can be explained by the fact that His-257 forms a hydrogen bond with this molecule. The data indicate that His-257 does not contribute to the overall structural integrity of the enzyme. Rather, it is directly

Kinetic and reaction mechanisms of V. cholerae ApbE
involved in the catalytic process and in the formation of a ternary complex with the substrates.

Random Bi Bi kinetic mechanism of ApbE
ApbE uses FAD as a substrate and covalently attaches the FMN moiety to the protein substrate (NqrC), following a bisubstrate kinetic mechanism. There are three basic bi-substrate models: Random, Ordered, and Ping-Pong. To determine the ApbE mechanism, steady-state kinetics were investigated. Both the Random and Ordered mechanisms are sequential processes in which the substrates are bound simultaneously before they can be converted into the products (37). On the other hand, in the Ping-Pong mechanism, one of the substrates reacts with the enzyme, leading to the formation of an intermediate and the release of the first product. Subsequently, the second substrate is bound and reacts with the intermediate and the enzyme catalyzes the formation of the final product. The Ping-Pong mechanism is common in many group transfer reactions (37), in which the first substrate carries the group that is later on transferred to the second substrate. Therefore, it would be reasonable to postulate that ApbE flavin-transfer reaction also follows a Ping-Pong mechanism in which FAD provides the FMN moiety and NqrC is the acceptor substrate. However, bi-substrate kinetic analyzes show that the kinetic behavior of ApbE is best explained by a sequential mechanism in which a ternary enzyme-substrates complex is formed during the reaction, suggesting that both ApbE and NqrC participate in enzyme catalysis.
To distinguish between Random and Ordered mechanisms, product inhibition kinetic experiments were performed, using AMP. The results reveal that AMP behaves as a mixed inhibitor against NqrC. As illustrated in Fig. 2E, AMP competes with NqrC for the free form of the enzyme (E), and it is also bound "uncompetitively" to the modified form (E-NqrC-FMN), explaining the mixed behavior. On the other hand, if ApbE followed an ordered mechanism with either FAD or NqrC binding first, AMP would act as an uncompetitive inhibitor with respect to NqrC.
Interestingly, AMP inhibition kinetics have also demonstrated that the ternary complex with the bound inhibitor remains catalytically active, yet to a lesser extent. Our previous study has shown that ApbE is activated by ADP, probably through an allosteric site (23). Thus, it is possible that at high concentrations AMP can be bound to the regulatory site, de-inhibiting the enzyme and producing the partial inhibitory behavior. Further work is being conducted to clarify the location of this site and its role.

Role of His-257 in catalysis
In our previous report His-257 was found to be essential for ApbE function (23). To investigate the role of this residue in catalysis, several mutants of His-257 were generated and characterized in this study. Two of the mutants, H257E and H257K, showed minimal activities even in the presence of the activator potassium in all conditions tested, probably due to their charged side chains. The negatively charged side chain of Glu is likely to repel the pyrophosphate group of FAD, pushing it away from the active site. On the other hand, the long side chain of Lys may introduce local rearrangements of the active site that might prevent catalysis. In contrast, the other two mutants, H257G and H257T, showed substantially decreased yet observable turnover rates at pH 9.0 compared with the WT enzyme. A significant increase on the K m for both substrates was also observed, which together with the similar pK E1 and pK ES1 values of around 7 of WT ApbE indicate that His-257 directly participates in catalysis and in the formation or stability of the enzyme-substrate complex. Although the later process is difficult to understand with the data currently available, we can propose that His-257 could participate directly in substrate binding, as seen in the crystal by the formation of a hydrogen bond with FAD, or that it might "kinetically" stabilize the enzyme-substrate complex pushing the equilibrium in the forward direction or decreasing the dissociation rate of the complex (i.e. k off ). Further studies are required to clarify this role.
Previously, we demonstrated that ApbE has a pK ES1 of 8.4, which helped in the identification of His-257 as a possible catalytic residue (23). In this work, we carried a systematic characterization of the effect of pH on the activity of WT ApbE. Global analysis of the NqrC titrations at saturating concentrations of FAD measured in different buffers show that the pK E1 and pK ES1 kinetic parameters are very close to that of histidine. This result suggests that His-257 (the only conserved His residue in the sequence of ApbE) plays an important role in the catalytic reaction. It should be mentioned that the pK ES1 value calculated in this report (6.9) is significantly lower compared with our previous report. This difference is due to the addition of potassium ions, which appears to activate ApbE by lowering the optimal pH of the enzyme.
To clarify the role of His-257 in catalysis, the pH profiles of the activity of H257G and H257T under near-saturating concentrations of both substrates were examined, which can provide insight into the apparent pK ES1 of the mutants. The mutants displayed pH dependence explained by Equation 5 that results in a sigmoidal curve (37), from which the critical kinetic parameter pK ES1 was obtained. Both mutants display increased pK ES1 values by more than 2 units, demonstrating that His-257 helps in the deprotonation of the catalytic residue, and that it likely acts as a general base. Consequently, it is possible that His-257 deprotonates the FMN acceptor residue Thr-225 in NqrC, which activates this residue. However, the canonical pK a of threonine found in aqueous solution is above 13 (45), significantly higher than the pK ES1 values found in the mutants. Interestingly, a conserved Lys-207 in NqrC was discovered to be essential for flavin transfer (38). It was proposed that this residue serves as a general base that deprotonates Thr-225 due to its proximity to the threonine residue. Our data suggest that Lys-207 may have a different role. pK a predictions by PROPKA3.1 show that the pK a of residue 225 is clearly influenced by the charge of residue 207, i.e. a positive charge will shift the pK a down. In addition, the predictions showed that the pK a shift caused by residue 207 is within 1 pH unit. In general, such change does not support the role of Lys-207 as the catalytic residue. We propose that Lys-207 lowers the pK a of Thr-225, which allows its deprotonation by His-257 and/or stabilizes the alkoxide form through its positive charge (46).

Kinetic and reaction mechanisms of V. cholerae ApbE Proposed mechanism of flavin transfer
Based on the data obtained in this study, a reaction mechanism of ApbE-catalyzed flavin transfer is proposed and illustrated in Fig. 4D. In the first step of the reaction, the two substrates, FAD and NqrC, are bound to the enzyme in a random order, forming the ternary complex. The binding of FAD seems to be mediated through the coordination with the Mg 2ϩ ions, as shown in the crystal. When both substrates are bound, Lys-207 lowers the pK a of Thr-225 to a physiologically feasible value as its positive charge shifts the ionization equilibrium of Thr-225 toward the alkoxide form. His-257 subsequently acts as a general base, removing a proton from Thr-225. The deprotonated Thr-225 then performs a nucleophilic attack on the pyrophosphate group of FAD and eventually forms the phosphoester bond between FMN and NqrC, finalizing the flavin-transfer process.
Similar cases have been reported in other protein families such as the Fic domain (47,48) and serine protease (49). The Fic domain is an adenylyltransferase that integrates AMP to either Thr or Tyr residues of the Rho-family GTPases, which then inhibits downstream signaling pathways (50,51). Two independent studies by Luong et al. (47) and Xiao et al. (48) have demonstrated that a conserved His residue in the Fic domain is essential for the adenylyltransferase activity. The authors proposed that the His residue functions as a base that deprotonates the AMP acceptor Thr/Tyr residue, leading to a subsequent nucleophilic attack on ATP by the deprotonated residues and hence AMPylation of the protein substrate. In the case of the well-studied serine proteases, the classic catalytic triad is composed of Asp, His, and Ser residues. Peptide hydrolysis is initiated by His, which as a base abstracts a proton from Ser, allowing it to attack the peptide bond, whereas Asp stabilizes the protonated His (49). These mechanisms are analogous to the one proposed in this work, describing a general base catalysis mechanism in which the His residue deprotonates the hydroxyl groups of Thr/Ser/Tyr, activating these residues for subsequent nucleophilic attacks.

Conclusions
The flavin transferase family ApbE catalyzes the incorporation of flavin cofactors into a variety of bacterial respiratory enzymes and hence plays a critical physiological role in the survival and pathogenicity development of these bacteria. In this work, we have demonstrated that ApbE follows a random sequential mechanism and that the conserved His-257 residue participates in the catalytic mechanism of this enzyme as a general base. These results allow us to gain a deeper understanding of the flavin-transfer mechanism catalyzed by ApbE and the interplay between this enzyme and a variety of essential bacterial respiratory enzymes.

Recombinant plasmid construction
Recombinant V. cholerae ApbE and NqrC, as a substrate of the flavin-transfer reaction, were heterologously expressed in E. coli (23), as previously described. Briefly, the apbe and nqrc genes were engineered to carry a His 6 tag at the C terminus for protein purification. The leader sequence of ApbE, and the N-terminal transmembrane section of NqrC, were eliminated during the cloning procedure, allowing high yields of protein expression (20,23). The engineered fragments were inserted into the pBAD/HisB vector, transformed into E. coli DH5␣ for plasmid production, and subsequently used for transformation of E. coli BL21 for protein expression.

Site-directed mutagenesis
To study the role of His-257, mutants Gly, Thr, Asp, and Lys were obtained as previously described using the QuikChange site-directed mutagenesis kit (Agilent Technologies). The primers used are listed in Table 4 (23).

Protein expression and purification
Recombinant proteins were expressed and purified as described previously (23). Briefly, E. coli cells carrying the proteins of interest were grown in terrific broth media (supplemented with 0.4% glycerol). 0.05% arabinose was added to induce the expression of the proteins. The cells were then harvested by centrifugation and lysed by sonication in washing buffer, containing 300 mM NaCl, 1 mM MgCl 2 , 5 mM imidazole, 50 mM Na 2 HPO 4 , pH 8.0. Subsequently, the soluble cytosolic proteins were obtained by ultracentrifugation, after which the supernatant was collected and subject to two steps of chromatographic purification: nickel-nitrilotriacetic acid affinity chromatography and DEAE-Sepharose cation exchange chromatography. Purities Ͼ95% were obtained for both proteins, as determined by SDS-PAGE.

Protein crystallization
Prior to crystallization, protein samples were subject to size exclusion chromatography with a Superdex 75 10/300 column (GE Healthcare) in buffer containing 10 mM HEPES, 150 mM NaCl, pH 8.0, to remove possible protein aggregates. Protein samples were supplemented with 2 mM FAD and 5 mM MgCl 2 , and subsequently subject to size exclusion chromatography with a Superdex 75 10/300 column (GE Healthcare) in buffer The initial crystallization conditions were determined with a sparse crystallization matrix at 16°C temperatures using the sitting-drop vapor-diffusion technique using MCSG crystallization suite (Microlytic) and PEG/Ion HT screen (Hampton Research). The crystals grew in multiple conditions after a couple weeks of incubation. The best crystals of WT and H257G mutant ApbE protein were obtained from MCSG-4 screen, reagent formulation number 86 (0.1 M sodium acetate, 25% PEG 4000, 8% isopropyl alcohol) and MCSG-4 screen, reagent formulation #66 (0.1 M sodium acetate, 22% PEG 4000, 0.1 M HEPES buffer, pH 7.5), respectively. Crystals selected for data collection were soaked in the crystallization buffer supplemented with 15% ethylene glycol and flash-frozen in liquid nitrogen.

Data collection, structure determination, and refinement
Single-wavelength X-ray diffraction data were collected at 100 K temperature at the 19-ID beamline (52) of the Structural Biology Center at the Advanced Photon Source at Argonne National Laboratory using the program SBCcollect. The intensities were integrated and scaled with the HKL3000 suite (53). The structures were determined by molecular replacement using the HKL3000 suite (53) incorporating the following programs, MOLREP (54), SOLVE/RESOLVE (55), and ARP/wARP (56). The coordinates of S. enterica ApbE protein (PDB 3PND) (42) were used as the starting model for the structure solution. Several rounds of manual adjustments of structure models using COOT (57) and refinements with the Refmac program (58) from the CCP4 suite (59) were done. The stereochemistry of the structure was validated with PHENIX suite (60) incorporating MOLPROBITY (61) tools. A summary of data collection and refinement statistics is given in Tables 5 and 6.

Coordinates
The atomic coordinates and structure factors of WT and H257G mutant ApbE protein were deposited into the Protein Data Bank as 6NXI and 6NXJ, respectively.

Flavin-transfer activity measurement
We are reporting a completely novel method of flavin-transfer activity measurement catalyzed by ApbE, based on the change in UV-visible absorption spectrum of free FAD versus covalently-bound FMN. The flavin transferase activity of ApbE was followed at 395-minus-366 nm as the difference spectrum of holo-NqrC-minus-FAD shows that the most significant difference is observed at 395 nm, whereas 366 nm is the isosbestic point (Fig. 1A, inset). The molar extinction coefficients of free FAD and free FMN were 12.50 mM Ϫ1 cm Ϫ1 at 450 nm (62) and 12.02 mM Ϫ1 cm Ϫ1 at 445 nm, respectively (63). In addition, the molar extinction coefficient of the flavinylated NqrC at 450 nm was determined to be 13.91 mM Ϫ1 cm Ϫ1 .

Steady-state kinetics
To investigate the kinetic properties of WT ApbE and the mutants, steady-state kinetic experiments were performed in reaction buffer containing 100 mM KCl, 1 mM EDTA, 5 mM MgCl 2 , 50 mM Tris, pH 9.0. NqrC titrations were carried out at a fixed concentration of FAD (50 M), using an ApbE concentration of 0.5 M. On the other hand, FAD titrations were run at a fixed concentration of NqrC (100 M), at an enzyme concentration of 0.01 M to allow a more accurate measurements of the reaction rates.

Kinetic and reaction mechanisms of V. cholerae ApbE
To understand the kinetic mechanism of ApbE, bi-substrate kinetic experiments were carried out. Initial rates were measured through FAD titrations (0, 0.1, 0.2, 0.5, 1, and 2 M) at different fixed concentrations of NqrC (5, 10, 20, and 100 M). Rates calculated were then plotted and fitted to different bisubstrate kinetic models to determine the kinetic mechanism. Three bi-substrate kinetic models used were: 1) Random, 2) Ordered, and 3) Ping-Pong, and are described below, where k cat is the turnover rate, [A] and [B] are the concentrations of FAD and NqrC, respectively, K A and K B are the Michaelis constants of the enzyme-substrate complexes, and ␣ is the factor by which the binding of one substrate changes the Michaelis constant of the other (37).
To investigate if ApbE fits in either the random or ordered kinetic model, ApbE inhibition kinetics were carried out. In particular, NqrC titrations at a fixed concentration of FAD (50 M) and different fixed concentrations of AMP (0, 0.01, 0.2, and 5 mM) were performed. Rates were calculated and globally fitted to the partial mixed inhibition equation shown below, ͓E t ͔ ϭ k cat ͓S͔ ϩ k cat R͓S͔ ͓I͔ where k cat is the turnover rate, k cat R is the turnover rate obtained at saturating concentrations of AMP, [S] is the concentration of NqrC, [I] is the concentration of AMP, K m is the Michaelis constant of the enzyme-substrate complex, K ic and K iu are the Michaelis constants of the enzyme-inhibitor complex and enzyme-substrate-inhibitor complex, respectively (37).

ApbE pH-dependence assay
To inspect the pH dependence of ApbE WT and the mutants, pH-dependence assays were carried out in buffer containing 100 mM KCl, 1 mM EDTA, 5 mM MgCl 2 , 50 mM MES, 50 mM HEPES, 50 mM Tris, 50 mM CAPS, at a series of pH values (6.0 -10.5). Initial rates at near saturating concentrations of the substrates were obtained and plotted against pH. For WT ApbE the concentrations of substrates used were: 50 M FAD (Ͼ1000 ϫ K m ) and 100 M NqrC (10 ϫ K m ). For the mutants these concentrations were: 50 M FAD (250 ϫ K m ) and 500 M NqrC (5 ϫ K m ). The data were fitted to the following equation where k cat is the turnover rate, [S] is the concentration of NqrC, K m is the Michaelis constant of the enzyme-substrate complex, pK E1 is the negative logarithm of K E1 , the equilibration constant of the enzyme deprotonation, and pK ES1 is the negative logarithm of K ES1 , the dissociation constant of the protonated enzyme-substrate complex (37).

pK a predictions
To predict the effect of the Lys-207 residue on Thr-225, the pK a values for all ionizable residues in NqrC were predicted using PROPKA3.1 (39,40). Protein models were protonated using UCSF Chimera (41). However, due to limitations of the software, pK a predictions are not possible for residues other than the standard ionizable residues. To combat the inability of PROPKA3.1 to predict the pK a of Thr-225, Thr-225 was mutated in silico to Cys-225 using UCSF Chimera (41). Although the predicted pK a of Cys-225 will not reflect the pK a of Thr-225 in an absolute manner, relative changes in calculated Cys-225 pK a can be expected to be mirrored by Thr-225. Four mutants at position 207 (K207A, K207R, K207E, K207Q) were created in silico again using UCSF Chimera (41) to elucidate the role of residue 207 on the pK a of residue 225. These mutants were chosen due to their diverse nature in charge, size, and polarity. With these four mutants and the WT, the effects of positive charge, negative charge, polarity, and hydrophobicity of residue 207 on residue 225 were determined.