Glu121-Lys319 salt bridge between catalytic and N-terminal domains is pivotal for the activity and stability of Escherichia coli aminopeptidase N

Abstract

Escherichia coli aminopeptidase N (ePepN) belongs to the gluzincin family of M1 class metalloproteases that share a common primary structure with consensus zinc binding motif (HEXXH-(X18)-E) and an exopeptidase motif (GXMEN) in the active site. There is one amino acid, E121 in Domain I that blocks the extended active site grove of the thermolysin like catalytic domain (Domain II) limiting the substrate to S1 pocket. E121 forms a part of the S1 pocket, while making critical contact with the amino-terminus of the substrate. In addition, the carboxylate of E121 forms a salt bridge with K319 in Domain II. Both these residues are absolutely conserved in ePepN homologs. Analogous Glu-Asn pair in tricon interacting factor F3 (F3) and Gln-Asn pair in human leukotriene A₄ hydrolase (LTA₄H) are also conserved in respective homologs. Mutation of either of these residues individually or together substantially reduced or entirely eliminated enzymatic activity. In addition, thermal denaturation studies suggest that the mutation at K319 destabilizes the protein as much as by 3.7°C, while E121 mutants were insensitive. Crystal structure of E121Q mutant reveals that the enzyme is inactive due to the reduced S1 subsite volume. Together, data presented here suggests that ePepN, F3, and LTA₄H homologs adopted a divergent evolution that includes E121-K319 or its analogous pairs, and these cannot be interchanged.

Introduction

Aminopeptidase N (EC 3.4.11.2) from Escherichia coli (ePepN) is a member of the M1 class peptidases.¹ The other members of this class include human aminopeptidase N (hAPN), human endoplasmic reticulum aminopeptidase N (hErPepN), human leukotriene A₄ hydrolase (LTA₄H), aminopeptidase A, aminopeptidase B, puromycin sensitive aminopeptidase (PSA), oxytocinase, tricon interacting factor F3 (F3) among others.^{2, 3} M1 class aminopeptidases comprise large group of metalloproteases classified as gluzincins with a consensus zinc-binding motif (HEXXH-(X18)-E) and an exopeptidase motif (GXMEN) in their active site. Crystal structures of ePepN, hErPepN [not published, deposited in protein data bank (PDB) with ID: 2XDT], F3, LTA₄H, PepN from Neisseria meningitidis (NmPepN) and from Plasmodium falciparum (PfPepN) have been reported.^4–10 Within the M1 class peptidases, all enzymes have four domains except, LTA₄H which has only three domains. First two domains (Domains I and II) share highest sequence and structural homology among all these structures including the LTA₄H. The LTA₄H has a unique C-terminal domain instead of Domains III and IV, probably required to perform the leukotriene hydrolase activity. All the M1 family proteins have thermolysin (bacterial endopeptidase) like catalytic domain (Domain II)¹¹ [Fig. 1(a)]. The N-terminal β-domain (Domain I) and the catalytic domain share a large surface area. Although proteins from each of the M1 class members perform distinct physiological function, they all act as aminopeptidases.^12–15

Figure 1.

(a) Domains I and II of ePepN are shown in surface diagram, while the other two domains are in line form. Ball and stick diagram represents the inhibitor bestatin in the S1 and S1′ subsites indicated by arrows within the active site grove (PDB ID: 2HPT). Domain I blocks the active site grove at one end. (b) A close-up view of the active site grove. Bestatin is shown in the green sticks. E121 from Domain I (blue surface) is shown in yellow, blocks the active site, while making intimate contact with the P1 residue and amino-terminus of the substrate. Purple sphere indicates the zinc ion. (c) Crystal structure of the tryptophan (brown) bound ePepN structure (PDB ID: 3B3B) showing the active site. K319 makes three-centered interaction with three carboxylates including one from E121 residue. M260 and M263 sandwiches the indole ring. Arrow indicates the direction of the peptide. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Thermolysin is a bilobal neutral endopeptidase with a pronounced active-site cleft, and it has been already shown that the specificity is derived by the optimized interactions of the substrate in the S₂, S₁, S₁′, and S₂′ subsites.¹¹ As in thermolysin, the active site cleft in ePepN is comparably extended. However, E121 from Domain I in ePepN blocks the cleft at P2 binding site thereby converting the enzyme to an exopeptidase. A similar pattern is observed in all the other M1 class aminopeptidases [Fig. 1(b,c)]. For example, Q136 in LTA₄H, E183 in hErPepN, E101 in F3, E117 in NmPepN, and E319 in PfPepN play analogous role. Another interesting aspect of this residue in all the structures is that its main chain peptide adopts a cis-configuration.

Based on the ePepN structure in complex with various amino acids, it was concluded that the S1 pocket assumes a cylindrical hydrophobic pocket, while top and bottom are polar.¹³ E121 in the Domain I along with several other residues in the catalytic domain contribute to S1 pocket. E121 plays a crucial role in conversion of endopeptidase activity of a thermolysin to an exopeptidase function of the ePepN. In addition to interacting with the amino-terminus of the substrate, the side chain of E121 also forms a salt bridge with the K319 in Domain II [Fig. 2(a)]. K319 precedes one of the zinc coordinating glutamates in the gluzincin motif, HEXXH-X(17)-K³¹⁹E³²⁰. Apart from interacting with E121, K319 also forms hydrogen bonds with E264 and E320 side chains displaying a three-centered interaction. A similar Glu-Lys salt-bridge pairing can also be found in NmPepN and PfPepN structures. However, Glu-Lys pairing in ePepN is replaced by a Glu-Asn (E183-N375) in hErPepN, (E101-N287) in F3, and Gln-Asn (Q136-N317) pair in LTA₄H [Figs. 1(a) and 2(b,c)].

Figure 2.

(a) Sequence alignment of six M1 family aminopeptidases whose structures have been deposited in the PDB. Active site zinc binding motif and the exopeptidase motif are shown with underline. K319 and E121 residues positions in the ePepN structure are marked as asterisk. Phylogenetic tree for E. coli PepN and its structural homologs deposited in the PDB was generated by neighbor-joining method using MEGA4 software. The numbers shown next to the nodes indicate percent bootstrap values of the 100 replicates. The scale bar represents branch length (number of amino acid substitutions/100 residues). (b) Full-length sequence was used to generate this tree. The observed phylogenetic relationships showed that there is a close sequence similarity between human LTA₄H and microbial enzymes. (c) Amino acid sequence from Domains I and II only are used to draw phylogenetic tree. Interestingly, the observed pattern shows that the LTA₄H shifts into the mammalian enzymes sharing sequence homology. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

It seems that there are three possible combinations for these pairs: Glu-Lys, Glu-Asn, and Gln-Asn. It is not clear if these pairs occur in random or are part of a larger evolution process. It would not only be interesting to understand the role of each of these residues on the structure and activity relationship, but also possible interchangeability of these pairs between proteins is worth pursuing to gain more insights for better understanding on the catalytic activity. In this study, we discuss the bioinformatics approach to understand the evolutionary routes followed by M1 class proteins, in addition to observations we made on the mutation of E121 and K319 in ePepN, individually and also in combination with other amino acids that represent situations like in LTA₄H and F3.

Results

Three distinct evolution patterns of M1 class aminopeptidases

We used crystal structures of six M1 class aminopeptidases deposited in the PDB to generate sequence alignment and phylogenetic tree. The sequence similarity between ePepN and other peptidases varied between 18% and 49% (F3 35.7%; hErPepN 17.6%; LTA₄H 22.6%; PfPepN 35.7%; and NmPepN 47.6%) [Fig. 2(a)]. The root mean square deviations (r.m.s.d) for the C-α atom based structural alignments of ePepN with others varied between 1.3 and 2.4 Å (F3 1.6 Å; hErPepN 2.4 Å; LTA₄H 2.3 Å; PfPepN 1.6 Å; and NmPepN 1.3 Å). These six sequences fall into three different branches of phylogenetic tree [Fig. 2(b)]. Although ePepN, PfPepN, and NmPepN form one branch of the tree, F3 and hErPepN form second branch, whereas LTA₄H alone forms a separate branch. We observed absolutely similar pattern by using only Domains I and II [Fig. 2(c)]. It is interesting to note that the first branch contains the Glu¹²¹-Lys³¹⁹ signature as in ePepN, while the second and third branches have Glu-Asn and Gln-Asn pairing, respectively.

Conservation of E121 and K319 pair in ePepN related proteins

We analyzed Homology derived Secondary Structures of Proteins (HSSP files) for ePepN, F3, and LTA₄H structures deposited in the PDB to understand the conservation of amino acids E121 and K319 across M1 class aminopeptidases. HSSP alignment includes all proteins that have sequence identity as low as 30% with the query protein. Of the 869 sequences aligned with ePepN, residues 297 to 326 comprise the metal binding and the catalytic residues include HEXXH-X(17)-K³¹⁹E³²⁰ signature. Of these 30 residues, only three were conserved less than 90%, while 14 residues were conserved 100% and 11 more at the 96% plus conservation. Both, E121 and K319 in ePepN were absolutely conserved (Fig. 2). A similar pattern was observed in LTA₄H homologs. Corresponding Q136 and N317 residues in LTA₄H were conserved at 86% (rest 14% are Glu) and 97%, respectively. HSSP file derived for the F3 structure indicates that E101 and N287 are conserved at 95% each. Together these data suggest that whenever there is a lysine at position 319 like in ePepN, always a glutamate is present at 121 to form a salt bridge. However, if it is an asparagine like in LTA₄H and F3, a glutamine or a glutamate can be present at the 121 position (Fig. 3, Table I).

Figure 3.

Chemical diagrams of the active site of some M1 class aminopeptidases. The number in the brackets represents the percentage of conservation. (a) Schematic representation of hydrogen bond pattern that links the E121, K319, and the inhibitor bestatin which in turn interacts with the zinc ion in the active site. Note that E121 interacts with the ε-amino group of K319 and α-amino group of bestatin simultaneously. (b) Although a complex crystal structure for either F3 or hErPepN with a substrate or inhibitor is not determined so far, we can assume that E101 interacts with N287 via a water molecule and directly with bestatin as shown here. (c) Although the amino group of Q183 interacts with N375 via a water molecule, the carbonyl forms hydrogen bond with the bestatin. Note that Q183 is conserved only 86%, while the other 14% is conserved by a Glu like in the F3 family.

Table I.

Conservation Pattern in Three Different Subclasses of M1 Family Aminopeptidases

	Domain I	Metal binding motif	XXX“GXMEN” motif
ePepN (869)	121E100(X4)R100	297HEY99F100H(X17)KE	258F99 N99M91GA100MEN
F3 (334)	101E95 (X4)R100	265HEL81A95 H(X17)NE	227F96 a/s A69 GA98 MEN
LTA4H (279)	136Q86 (X4)R100	295HEL60A75 H(X17)NE	265F94 P98 y/f GG98 MEN

The number in parenthesis indicates the total number of proteins in each subclass. The subscript in front of the amino acid shows the position of the amino acid in the representative protein sequence, and the superscript shows the percentage conservation of that particular amino acid.

It seems that there might be a strong correlation in conservation between the Glu-Lys pair and zinc binding HEXXH motif as well as exopeptidase GXMEN motif (Table I). In ePepN and its homologs, the metal binding motif is ₂₉₇HEY⁹⁹F¹⁰⁰H and the exopeptidase motif and its flanking region is ₂₅₈F⁹⁹N⁹⁹M⁹¹GA¹⁰⁰MEN, where superscript indicates the percentage of conservation (identity) and the subscript indicates the amino acid position in the ePepN sequence. In F3 and LTA₄H homologs, although they seem to share a similar sequence within the zinc binding motif, a very different signature was observed in the exopeptidase motif (Table I). Therefore, from our results, it is clear that the M1 class aminopeptidases containing the Glu-Lys salt bridge have a HEYFH zinc binding motif and the FNMGAMEN exopeptidase motif. Similarly, F3 homologs having a Glu-Asn pair within the Domains I and II interface, contain the HELAH and Fa/sAGAMEN, while the LTA₄H with Gln-Asn pair has HELAH and FPy/fGGMEN motifs.

Activity of the wild type and mutant enzymes

To understand the role of a specific amino acid at positions 121 and 319 and their combinations, several mutations were designed and constructed on ePepN. Initially, E121Q, K319N, and the corresponding double mutant mimicking F3 and LTA₄H situations were prepared. In parallel, E121A, K319A, and K319Q single mutants and E121Q-K319A, E121Q-K319Q, and E121Q-K319N double mutants were also designed and constructed to understand if other combinations can preserve the activity of the enzyme. All mutants and the wild type ePepN proteins were expressed using the BL21 (DE3) E. coli expression strain. Using talon resin, all proteins were purified to homogeneity in a single step judged by the sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel. The overall yield of the protein varied between 100 to 150 mg per 1.5 L culture.

Activity profile was determined for the ePepN at different pH and salt conditions with Leu-pNA as a substrate. Maximum activity was found to be between 8.0 and 8.5 pH units in the 10 mM Tris-HCl buffer with 100 mM NaCl. For the same enzyme with Ala-pNA as substrate, maximal activity was reported to be between 7.5 and 8.0 pH units.¹⁵K_m and k_cat for the ePepN with Leu-pNA as substrate were determined to be 161.1 μM and 13.5 s⁻¹, respectively, and for Ala-pNA are 907.1 μM and 185.7 s⁻¹.

K319N single mutant that represents the situation in the F3 has shown only 1% of the activity with Leu-pNA (K_m = 43.5 μM, k_cat = 8.7 × 10⁻⁶ s⁻¹) and Ala-pNA (K_m = 412.6 μM, kc_at = 0.56 × 10⁻³ s⁻¹) compared with the wild type ePepN. This mutant has no activity on other substrates like Arg-pNA, Lys-pNA, and Glu-pNA. Interestingly, we found that the other mutants are completely inactive against any of the substrates. Either altering the reaction pH between 4 and 11, supplementation of 100 μM extra zinc chloride, increasing the enzyme concentration up to 26-fold and/or substrate by 50-fold did not improve the activity profile of the mutants. From structure-based characterization, it was expected that E121A mutation may create a large cavity near the substrate amino-terminus binding region and also, a negative charge would be missing. In addition to other substrates, we incubated the E121A mutant with N-acetyl Leu-pNA. However, no activity was observed.

Thermal stability

To understand the role of created mutations within the scope of this study on thermal stability of the enzyme, we performed thermal denaturation studies with purified enzymes. E121 mutation to either alanine or glutamine did not have any effect on the thermal stability. However, mutation of K319 residue resulted in the decrease of stability by a minimum of 3.1°C.

E121Q crystal structure analysis

The overall structure of E121Q was similar to that of the wild type except for few changes in the active site. Zinc ion in the active site was found to be present at full occupancy. We observed noticeable geometric changes both in the main chain and side chains in Q119-E121Q and M260-M263 regions due to the change in the negatively charged Glu to neutral but polar Gln [Fig. 4(a)]. Within the main chain, a maximum of 0.7 Å displacement was observed at Cα of E121Q in the direction toward the center of the S1 pocket, while the cis-peptide conformation was still preserved. About 1.1 Å shift was noticed in the side chain of the new glutamine compared with the glutamate in the wild type [Fig. 4(a)]. In this new position, the carbonyl of E121Q still interacts with K319. However, the amide NH₂, apart from interacting with E264, recruits a new water molecule. Due to this new water position, highly conserved geometry of the M263 was disturbed that was involuntarily pushed into the middle of the S1 pocket. Also, due to this new water molecule, the side chain geometry of the Q119 has also changed, again resulting in the shrinkage of the S1 pocket volume. From previous structures, M260 was found to be inherently mobile adjusting according to the substrate/product in the S1 pocket playing the role of a gate-keeper. In this structure, this side chain moves into the S1 pocket in close proximity to M263 sharing necessary hydrophobic interactions (closest distance 3.7 Å) [Fig. 4(b)]. Modeling of a tryptophan into the active site suggests that new position of Q119, E121Q, M260, and M263 results in the shrinkage of the total volume of the S1 pocket that occludes the substrate from binding and provides a clue to the complete inactivity [Fig. 4(b)].

Figure 4.

(a) Overlay of the crystal structure of E121Q mutant of ePepN shown in green and tryptophan (not shown) complex (in grey). Water molecules are displayed in red and grey spheres, while zinc in blue and grey in respective structures. Note that the amide group of the E121Q recruits a new water molecule, which displaces the side chain of the M263 which moves into the middle of the S1 pocket. This also affects in the position of Q119 side chain. The amide NH₂ in addition to interacting with the water molecule also interacts with the E264, while the carbonyl group interacts with K319. (b) Stereo view of the overlay of the crystal structure of E121Q mutant of modeled with tryptophan (brown). Note that in the new position of M263, a substrate cannot bind in the active site. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Eight sulfate anions have been identified in the entire structure. One of the sulfates occupies the active site probably, where the carboxylate of a P₃′ or P₄′ residue in a tetra- or pentapeptide substrate would bind. This sulfate connects Domains II and IV by making strong interactions with K274, Y275, R783, and R825. Note that these two arginines are conserved more than 90% suggesting a conserved functional role in ePepN homologs. These residues are much better ordered in this structure than in previously published structures. The position of this sulfate ion can form a structural basis for why ePepN has higher affinity for tri- and pentapeptides than dipeptides [15]. All the other sulfate ions are found on the surface of the protein.

Discussion

Several of the M1 class aminopeptidases, particularly in humans have been sorted for understanding their role in physiopathology. For example, hAPN, LTA₄H, and PSA are targeted by various inhibitor molecules to find cures for cancer, inflammation, and Alzheimers disease.^{2, 16–19} Structure of ePepN can serve as a prototype example for understanding these important proteins. Crystal structures of ePepN have been reported for some product and inhibitor complexes that explain the basis for substrate recognition.^{13, 20} Several studies have been performed to understand the role of the active site residues in many M1 class aminopeptidases within the substrate binding and metal binding regions that comprise the GXMEN and HEXXH-(X18)-E motifs.²¹ However, there has been no study so far to address the interface of Domains I and II that constitute substrate binding region. In this study, we describe the role of a couple of residues on the enzyme activity and stability, one each from Domains I and II within the interface that form a buried salt bridge.

The E121Q mutant cannot interact simultaneously with K319 and substrate

Although, the S1 pocket shrinkage in E121Q mutant seems to be a plain reason for substrate not able to bind in the active site and get processed, there seems to be another angle for this. As a carboxylate was replaced by an amide group, not only the negative charge was lost but also hydrogen bonding potentials were changed. Because a carboxylic group exists as a carboxylate at physiological conditions, both the oxygen atoms act as hydrogen bond acceptors. Hence, the two oxygen atoms of E121 carboxylate could interact with two N➖H groups, one each from the K319 side chain and the substrate amino-terminus [Fig. 3(a)]. However, when the acid is changed to an amide, the carbonyl can act as an acceptor and the amino as a donor. This means that the E121Q amide can either interact with ε-N-H of K319 or the α-N-H of substrate but not both at the same time. This explains why the E121Q-K319N double mutant is completely inactive. In the E121Q crystal structure, the carbonyl interact with K319, while the amide NH₂ recruits a new acceptor in the form of water (W1), which in turn effects the position of other residues that make up the S1 pocket thereby reducing its size (Fig. 4).

Mutation of ePepN to mimic LTA₄H and F3 situations

Although K319N single mutation looks like the situation in F3, E121Q-K319N double mutant represents the LTA₄H. The K319N single mutant was active though only 1% of the wild type, and the double mutant was completely inactive. In the K319N mutation, a water molecule may probably act as bridge like in the F3 and hErPepN structures. Although one of the oxygen atoms of the carboxylate of the E121 side chain interacts with this new water, the other oxygen can facilitate the substrate binding [similar to that in Fig. 3(b)]. In the absence of a crystal structure, particularly in complex with substrate/inhibitor the above model alone does not explain low activity of this mutant. One possibility is that replacement of a positively charged lysine to a neutral and branched asparagine residue very close to the active site may result in the unexpected structural changes that either limits the substrate accessibility and/or the compromised catalytic ability. Most surprising observation is the complete loss of activity of ePepN with E121Q-K319N double mutation. In this case, one positive and one negative charge are removed that does not affect the overall charge potential of the protein. One could have expected that the new glutamine would interact with the water as in LTA₄H and could conveniently interact with the substrate amino-terminus [similar to that in Fig. 3(c)]. Again, in the absence of the structural data on this mutant and several others reported here, it is hard to imagine geometrical arrangement around the substrate binding region that results in the enzyme inactivation.

Thermal stability studies suggest that mutation of E121 to either an alanine or a glutamine, where a negative charge is removed does not affect its stability. Similarly, E121Q-K319N double mutant where both the charges were removed is also as stable as the wild type. However, in K319N single mutant, where a positive charge is replaced by a neutral amino acid, protein loses its stability by 3.7°C.

Does coevolution of Domains I and II have significance in functional difference?

So far, there has been no attempt to define a classification method for the M1 family proteins. Herein, we demonstrate that just based on the sequence, we can divide the M1 class proteases into at least three subclasses that have the following three individual motifs (based on Table I): (a) E-(X4)-R, HEYFH-(X17)-KE and FNMGAMEN; (b) E-(X4)-R, HELAH-(X17)-NE and F-X-AGAMEN; and (c) Q-(X4)-R, HELAH-(X17)-NE and FP-X-GGMEN. Exact physiological substrate for many of the M1 class proteins is not yet established although all of them display aminopeptidase activity. Within the aminopeptidase function, ePepN processes the peptides with hydrophobic and basic residues on the amino terminus.¹³ A similar pattern was observed for LTA₄H apart from LTA₄H function.¹² F3 was found to process peptides with acidic residues.¹⁴ With a limited biochemical and structural data, it is hard to conclude if sequence based classification made above correspond to these functional classes. However, we believe that our sequence-based classification can serve as a starting point to characterize new M1 family proteins.

Data presented in this study not only confirm that E121-K319 salt bridge is essential for the activity but also cannot be interchanged to other pairs like Glu-Asn or Gln-Asn as in F3 and LTA4H proteins, respectively.

Materials and Methods

The wild type ePepN encoding plasmid used in this study is same as that was reported earlier.⁴ All the substrates and analytical grade chemicals otherwise mentioned were purchased from Sigma-Aldrich (St. Louis). Molecular biology grade enzymes and reagents were obtained from Fermentas International.

Site directed mutagenesis

Site-directed mutagenesis was performed by polymerase chain reaction (PCR). Nine different site specific mutations were made from wild type ePepN using specific primer sets. Primers for the construction of the variants E121A, E121Q, K319A, K319N, and K319Q are adopted as described elsewhere.²² To introduce a second mutation in E121A or E121Q mutants at K319 position (K319A, K319N, and K319Q) PCR was performed using the established respective mutant plasmid DNA as a template. Mutations were confirmed by DNA sequencing analysis.

Expression and purification of the wild type and mutant proteins

Protein expression was performed in E. coli strain BL21 (DE3) cells in LB medium containing 100 μg mL⁻¹ of ampicillin. The cells were grown at 37°C until the OD₆₀₀ reached 1.4–1.6. To induce protein expression, 1 mM IPTG (final concentration) was added to the culture medium and continued to grow overnight at 25°C with constant shaking at 150 rpm. Cells were harvested by centrifugation at 8000g for 15 min, and cell pellets were stored at −70°C until further use. For protein purification, all the steps were performed at 4°C, unless otherwise mentioned. The cell pellet was resuspended in +T/G buffer (50 mM Hepes, pH 8.0, 0.5M NaCl, 10% glycerol, 0.1% Triton X-100) containing 1 μg mL⁻¹ protease inhibitor cocktail. After adding 5 mM MgCl₂ and 1 μg mL⁻¹ of DNase I, cells were lysed by constant cell disruption system further cell lysate was cleared by centrifugation at 17,000g for 30 min. The supernatant was applied to a pre-equilibrated talon column (cobalt affinity resin; Clontech Laboratories, Mountain View, CA). The column was further washed with +T/G buffer and −T/G buffer (50 mM Hepes, pH 8.0, 0.5M NaCl, 5% glycerol, 5 mM imidazole). Bound protein fraction was eluted with 150 mM imidazole in −T/G buffer and dialyzed four times in 1 L buffer containing 25 mM Hepes, pH 7.5, 150 mM NaCl, and 5% glycerol. Purity of the protein was assessed by 10% SDS-PAGE. For further applications, protein was concentrated to at least 10 mg mL⁻¹ using 30 kDa cutoff Centricon® tubes (Millipore, Ireland), 100 μL aliquots were flash frozen in liquid nitrogen and stored at −80°C until further use.

Activity assays

Activity of the wild type and all variants was determined in buffer (100 mM NaCl, 10 mM Tris-HCl, pH 8.0) containing appropriate substrate in 96-well plates at 37°C. The product (p-nitroaniline) obtained on cleavage of the substrate was measured on Infinite M200 Pro multimode reader (Tecan). In brief, after incubation of the enzyme for 15 min at 37°C in reaction buffer, 100 μM of Leu-pNA or 100 μM of Ala-pNA was added to 100 μL of final reaction volume. Formation of free pNA was monitored by continuous absorption method at 405 nm. The linear portion of the progress curve was used to calculate the specific activity and kinetic parameters of the enzymes. All the reactions were performed in triplicates, ± SD values are reported.

Effect of pH, substrate, and metal ion concentration on enzyme activity

The enzyme activity was tested as described above on each mutant and the wild type by varying the reaction pH between 4 and 9 [citrate (pH 4.0 and 5.0), phosphate 6.0, and Tris (pH 7.0, 8.0, and 9.0) buffers] with each substrate (Ala-pNA, Met-pNA, Arg-pNA, Lys-pNA, and Glu-pNA). In separate reactions, extra 100 μM ZnCl₂ was supplemented. All the reactions were performed in triplicates and ± SD values are reported.

Determination of K_m and k_cat for wild type and mutant ePepN enzymes

The K_m and k_cat are determined for the Leu-pNA as substrate against the wild type ePepN. Substrate concentration was varied between 30 and 700 μM with the fixed enzyme concentration (75 nmol) in the 100 μL volume of 10 mM Tris-HCl, pH 8.0, 100 mM NaCl buffer at 37°C in triplicates. For mutants, the enzyme concentration was increased up to 20 μmol and the substrate up to 5 mM (Table II).

Table II.

Activity of Mutant and Wild Type Enzyme^a

	Km (μM)	kcat (s−1)	kcat/Km (s−1 μM−1)
Wild type
Leu-pNA	161.1 ± 35.2	13.5 ± 3.7	0.08
Ala-pNA	907.1 ± 45.5	185.7 ± 65.1	0.2
K319N
Leu-pNA	43.5 ± 2.3	8.7 × 10−6 ± 0.2	0.2 × 10−6
Ala-pNA	412.6 ± 40.7	0.56 × 10−3 ± 0.01	1.3 × 10−6

^aNo activity could be determined for E121A, E121Q, K319Q, K319A, E121Q-K319N, and E121Q-K319Q mutants.

Thermal shift assay to assess the stability of wild type and mutant ePepN

Thermal shift assays were performed for the wild type and the mutant enzymes using Applied Biosystems 7500 Real-Time PCR instrument. Equal amounts of enzyme (3 μg) were mixed with 2.5 μL of Sypro Orange (50× concentration S5692 Sigma-Aldrich) in a reaction volume of 20 μL of buffer (10 mM Tris pH 8.0, 100 mM NaCl). Samples were heat denatured from 25°C to 90°C at a ramp rate of 1°C min⁻¹. Protein unfolding curves were monitored by detecting changes in Sypro Orange fluorescence. The inflection point of the fluorescence-versus-temperature curves was identified by plotting the first derivative over the temperature (Table III). The temperature at which fluorescence minima observed is referred as the melting temperature (T_m).

Table III.

Thermal Denaturation Studies

Sample	Tm (°C)
Wild	51.5 ± 0.2
E121A	52.1 ± 0.4
E121Q	52.4 ± 0.4
K319N	47.8 ± 0.1
K319Q	48.3 ± 0.4
K319A	48.4 ± 0.1
E121Q-K319N	52.7 ± 0.5
E121Q-K319Q	48.9 ± 0.6
E121A-K319A	46.9 ± 0.5

Multiple sequence alignment analysis and generation of phylogenetic tree

HSSP is a database of multiple sequence alignments (MSAs) of every protein in the PDB.²³ MSA files were accessed from this database for ePepN, F3, and LTA₄H. Statistics given toward the end of each file were analyzed to understand the percentage of identity of an amino acid at a given position.

Three-dimensional coordinates of crystal structures of ePepN, LTA₄H, hErPepN, PfPepN, NmPepN, and F3 deposited in the PDB along with their sequences in the fasta format were downloaded, and alignment of structures was performed in coot.²⁴ Similarly, sequence alignment was performed using ClustalW.^{24, 25} Phylogenetic trees were plotted using the MEGA4 program (Fig. 2).²⁶

Protein crystallization, X-ray data collection, structure solution, and refinement

All the mutants were subjected to crystallization experiments at 293 K. Crystals for only E121Q single mutant were obtained. After optimization, E121Q mutant was crystallized from 1.0M ammonium sulfate and 50 mM Hepes pH 5.8. Crystals were frozen by dipping directly in the liquid nitrogen after incubation for 5 s in mother liquor that contained 25% glycerol.

X-ray data were collected at the Advance Light Source (ALS, Berkeley) synchrotron radiation on beam lines 8.2.1 at a fixed wavelength of 1.0000 Å. Data were processed using the HKL2000.²⁷ X-ray data statistics are given in Table IV. As E121Q data were isomorphous with the wild type, the R_free were transferred from the wild type for the apo structure (PDB code 2HPO), and the structure was directly refined in Refmac.²⁸ Molecular graphic program COOT was used to visualize the structures.²⁴ Pymol was used to generate all the structural figures.²⁹

Table IV.

Crystallographic Data for E121Q Mutant Aminopeptidase

Cell parameters N
a, b (Å)	120.76
c (Å)	170.84
Space group	P3121
Resolution range (Å) (Highest shell)	50–2.15 (2.06–2.15)
Rsym (%)	8.5 (39.7)
I/σI	25.5 (6.5)
Completeness (%)	100 (100)
Redundancy	6.3 (6.2)
Refinement
No. reflections (Rfree)	74,863 (3962)
Rwork (RFree) (%)	15.4 (18.7)
B-factor
Protein (Å2)	17.3
Ions (Å2)	23.7
Water (Å2)	27.9
No. atoms
Protein	6940
Ions	3
Water	752
R.m.s. deviation
Bond lengths (Å)	0.016
Bond angles (°)	1.427
PDB ID	3PUU

Glossary

Abbreviations

APN

human cell-surface membrane-associated aminopeptidase N

ePepN

E. coli aminopeptidase N

F3

tricon interacting factor F3

LTA₄H

leukotriene A₄ hydrolase

NmPepN

Neisseria meningitidis aminopeptidase N

PfPepN

Plasmodium falciparum aminopeptidase N

PSA

cytosolic puromycin-sensitive aminopeptidase.