Molecular Characterization and In Silico Analysis of Naturally Occurring TEM Beta-Lactamase Variants among Pathogenic Enterobacteriaceae Infecting Indian Patients

Cephalosporin resistance, particularly due to bla TEM encoded β-lactamases, among Enterobacteriaceae is, though, an increasing public health problem in India; their circulating genetic variants remain unknown. The present study deals with determination of bla TEM variants among 134 pathogenic Enterobacteriaceae of Indian origin. Their resistance profile against 3rd generation cephalosporins was determined. The presence of bla TEM variants among the bacterial plasmids was characterized by PCR followed by sequencing. Intergenic relations among the variants was determined by phylogenetic analysis. bla TEM protein were modeled by Modeller9v5 and verified. The catalytic pockets were characterized, and their interaction with cephalosporins was analyzed using AutoDock tools. More than 87% of isolates showed cephalosporin resistance with ESBL production among 57.8% of Escherichia coli and 50.6% of klebsiella pneumoniae. bla TEM-1 (84.21%), bla TEM-1 like (3.94%), bla TEM-33 (3.94%), bla TEM-116 (3.94%), bla TEM-169 (3.94%), and bla TEM-190 (7.89%) were detected in 76 isolates. Four variants, namely, bla TEM-1like, bla TEM-33, bla TEM-169, and bla TEM-190, coexisted in 3 isolates. The largest catalytic pocket of bla TEM-33 explained its expanded activity towards β-lactam-β-lactamase inhibitor combinations. Molecular docking indicated differential resistance pattern of bla TEM variants.


Introduction
Microbial resistance, especially resistance to cephalosporins, is an ever increasing public health problem worldwide among pathogenic Enterobacteriaceae, namely, Escherichia coli and Klebsiella pneumoniae [1]. Compared to other countries, frequent overuse of antibiotics and crowding of patients with high levels of disease acuity in relatively small, specialized hospital areas in Indian subcontinent have been the selective force to drive this epidemic like antimicrobial resistance development among all major microbial pathogens. Pathogenic bacteria acquire this resistance property due to acquisition of class A -lactamase that hydrolyze -lactam ring of cephalosporins.
The most commonly encountered class A -lactamase is encoded by plasmid mediated bla TEM (Temoniera) gene, which was first isolated from E. coli of blood-infected patient [1]. During the catalytic process, this enzyme first uses its active site Ser-70 residue as the key functional group, followed by involvement of Glu-166, Lys-73, Lys-234, and Ser-130 residues to hydrolyze the -lactam ring [2]. However, the substrate binding process is most likely the rate limiting step. So far 204 types of bla TEM have been identified that differ at 87 amino acid positions-thus having different protein stability [1]. There is growing biological evidence that increased protein stability (indicated by decrease in Gibbs free energy upon folding, ΔΔG) can lead to protein malfunction and hence diseases [3]. The bla TEM variants demonstrate considerable variety with respect to their range of substrate preferences and their levels of hydrolytic activity [1]. Substitutions critical for expanding substrate profile or increasing hydrolytic activities are efficiently selected under selection pressure of -lactam use. During multifocal and multidirectional evolution, two clinically important TEM -lactamases were generated: expanded spectrum TEM (ESBL-TEM) and inhibitor resistant TEM (IRT) -lactamases [4]. ESBL-TEM has expanded substrate spectrum towards oxymino--lactams, whereas IRT have expanded their substrate profile to the -lactam-lactamase inhibitor combinations [1].
Despite reports of higher level of 3rd generation cephalosporin resistance and ESBL production in approximately 100% and 87% of pathogenic Enterobacteriaceae infecting Indian patients, no study has been done to analyze the involvement of different bla TEM variants for development of these properties among pathogenic bacteria [5,6]. Our earlier study detected 3rd generation cephalosporin resistance among 97% of pathogenic K. pneumoniae infecting eastern Indian patients, with presence of bla TEM genes among 52% of them [7]. The present study provides molecular insight into the types of bla TEM variants circulating among pathogenic Enterobacteriaceae of this region, their evolutionary relationship, and their role in development of broad spectrum/inhibitor resistance or ESBL production among these bacteria. This study also attempted to analyze any differences in structure and catalytic activity of these bla TEM variants that might have significant implications for development of better antibiotics.

Clinical
Isolates. Enterobacteriaceae specimens ( = 134; E. coli = 57, K. pneumoniae = 77) were collected from different non duplicate clinical isolates of unrelated patients visiting Calcutta School of Tropical Medicine, Kolkata, India, from June 2011 to October 2012. The bacterial isolates were obtained from patient's urine, blood, throat swab, and wound pus.

Third Generation Cephalosporin
Sensitivity, ESBL Production, and MIC Determination. Antibiotic sensitivity and ESBL property of these isolates were determined according to Kirby-Bauer disc diffusion method [8]. The following -lactam antibiotics were used ( g/disc) ceftazidime (30), ceftazidime and clavulanic acid (30/10), cefotaxime (30), cefotaxime and clavulanic acid (30/10), and cefpodoxime (10) (HiMedia Lab Ltd., India). The diameter of zone of inhibition produced by each antibiotic disc was measured, and results were interpreted according to guidelines of Clinical and Laboratory Standards Institute (CLSI). If the zone diameter increased by 5 mm or more, when clavulanate is added compared to the antimicrobial alone, the isolate was considered as ESBL producing. E. coli ATCC 352183 was used as positive strain harboring bla TEM gene [9]. The minimum inhibitory concentrations (MICs) of ceftazidime and cefotaxime were determined by macrobroth dilution technique, according to CLSI guidelines [10]. Both antibiotics were tested in two-fold dilutions in the range of 16-512 g/mL.

Bacterial Plasmid Isolation.
Plasmid DNA was extracted from the Enterobacteriaceae isolates by alkaline lysis method [11]. Briefly, 3 mL overnight culture of patient isolated Enterobacteriaceae was lysozyme and RNase treated followed by alkaline lysis, phenol/chloroform extraction, and isopropanol precipitation of the plasmid DNA. Integrity of the plasmids was checked by 1% agarose gel electrophoresis.

Molecular Characterization and Phylogenetic
Analysis of bla TEM Genes. Amplification of bla TEM gene was performed in 96-well thermal cycler (Applied Biosystems, USA) with primers: F: 5 -ATGAGTATTCAACATTTTCGTC-3 ; R: 5 -TTACCAATGCTTAATCAGTGAG-3 , generated by using Primer3 (version 0.4.0) [12]. Briefly, each reaction was carried out in 20 L reaction volume using 1x PCR buffer (Fermentas, USA), 20 pmol of primers (Integrated DNA Technologies, USA), 1 mM of each dNTPs, 1 unit of Taq DNA polymerase (Fermentas, USA), 1.5 mM MgCl 2 , and 100 ng plasmid DNA. Thermocycling parameters were as follows: initial denaturation at 94 ∘ C for 60 s, 30 cycles of denaturation at 94 ∘ C for 30 s, primer annealing at 55 ∘ C for 30 s, and extension at 72 ∘ C for 60 s. PCR products were separated on 2% agarose gels and visualized under UV transilluminator (Ultraviolet/Laboratory Products, USA). PCR product of 862 bp size was purified with QIAquick PCR purification kit (Qiagen, USA) and sequenced using BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, USA) in 3100-Avant Genetic Analyzer (PE Applied Biosystems Inc, USA). The sequencing was done using both forward and reverse primers for each sample. The nucleotide sequences and deduced protein sequences were analyzed by BLAST and ClustalW programs of the European Bioinformatics Institute (http://www.ebi.ac.uk/) [13]. The sequences were deposited in GenBank database of National Center for Biotechnology Information. Chromatograms were visually inspected for double peaks as signs of presence of different bla TEM variants in the same PCR product.
Phylogenetic analysis of our bla TEM variants and their evolutionary relationship with those reported in Lahey's mutation database was derived by maximum likelihood method using MEGA software version 5 [14]. The Jones-Taylor-Thornton nucleotide substitution model, selected with the Model Test program according to the Akaike Information Criterion, corrected (AICc), was used as evolutionary model and included a gamma distribution (G) with six rate categories and a fraction of invariant (I) sites to account for substitution rate heterogeneity among sites.
Amino acid positions in the gene clusters that were presumably subjected to positive selection were identified by application of HyPhy selection test using joint maximum likelihood reconstructions of ancestral states under a Muse-Gaut model of codon substitution and Felsenstein 1981 model of nucleotide substitution as implemented in MEGA version 5. According to the test, codon-by-codon ratio ( ) of number of nonsynonymous substitutions per nonsynonymous site (dN) to that of synonymous substitutions per synonymous site (dS), ( = dN/dS) was calculated; values of > 1 indicated positive selection pressure, whilst values of < 1 indicated purifying selection pressure, and values of = 1 represented neutral evolution.

Homology Modeling, Model Validation, and Active Site
Identification of bla TEM Variants. The PSI-BLAST (Position Specific Iterated-Basic Local Alignment Search Tool) search with default parameters was performed with predicted protein sequences of our bla TEM variants against those available in Protein Data Bank (PDB) to find suitable templates for their homology modelling using modelling package MOD-ELLER9v7 [13,15]. The stereochemical qualities, compatibility of atomic models, and quality factors of these models were verified by PROCHECK, Verify3D, and ERRAT programs of Structural Analysis and Verification Server (SAVES), respectively, and quality of the models was also validated by ProSA server, a web server for Protein Structure Analysis [16].

Variant Stability Prediction.
The PyMOL program was used for measuring distance between reactive Ser-70 of the catalytic site and altered amino acid residue [17]. Effect of the alteration on protein stability was predicted by calculating change in free folding energy (ΔΔG) at that site using CUP-SAT (Cologne University Protein Stability Analysis Tool) server [18]. The prediction model used amino acid atom potentials and torsion angle distribution to assess the change in solvent accessibility and secondary structure specificity at the mutation site. According to experimental ΔΔ value, each mutation was placed in any of the following three classes: (i) destabilizing mutation: ΔΔ < −0.5 Kcal/mol; (ii) stabilizing mutation: ΔΔ > 0.5 Kcal/mol; (iii) neutral mutation: −0.5 < ΔΔ < 0.5 Kcal/mol [19]. SuperPose (version 1.0) was used to calculate Root-Mean-Square-Deviation (RMSD) values for stability change of bla TEM variants compared to wild type ( TEM-1 ), where dissimilarity cutoff >3.0Å indicated significant change in overall structural stability [20].

Catalytic Pocket Prediction and Measurement.
Possible ligand binding pockets containing catalytic residues within the generated models were predicted using Computer Atlas of Surface Topology of Protein (CASTp) server [21]. The catalytic pocket's solvent accessible volume (SV), surface area (SA), and pocket mouth area (MA) were also predicted.

Binding Mode Prediction of bla TEM Variants with 3rd
Generation Cephalosporins. Mol2 structures of 3rd generation cephalosporins, namely, ceftazidime (ZINC ID: 03830469), cefotaxime (ZINC ID: 04468780), and cefpodoxime (ZINC ID: 14235259), were retrieved from ZINC database [22]. The coordinates of bla TEM variants to be used for docking were retrieved using MetaPocket 2.0 server [23]. To study the nature of interactions of bla TEM protein models with these cephalosporins, docking was carried out using Autodock 4.0 tools [24]. Energy grid was built within a cubic box of dimensions 60 × 60 × 60Å grid points and 0.375Å spacing using the Autogrid program. Docking was performed based on Lamarckian Genetic Algorithm. Grid points were generated around the catalytic pocket to cover the entire ligand binding site, such that the compound to be docked can move freely within it. Docking simulations were performed using Lamarckian Genetic Algorithm (LGA). The docking parameters set to perform each docking experiment were derived from 100 different runs that were set to terminate after a maximum of 2,500,000 energy evaluations, elitism of 1, mutation rate of 0.02, cross-over rate of 0.8, and local search rate of 0.06. The population size was set to 150. Best run coordinates of the docked complex were analyzed and visualized through python molecule viewer and PyMol molecular graphics system for analysis of their mode of interaction with binding site residues. To analyze the molecular interaction between bla TEM variants and cephalosporins, complexes were generated using AutoDock tools. LIGPLOT analysis was run for the complexes to understand the hydrogen bonding and hydrophobic interaction within the docked complexes [25].

Molecular Characterization of bla
Maximum likelihood phylogenetic analysis of our bla TEM variants along with those reported in Lahey's database generated a dendrogram that showed different bla TEM clusters (

Homology Modelling and Validation of bla TEM Proteins.
The 3D model of proteins provided invaluable insights into the structural basis of its function. Search of PDB confirmed presence of templates for each of our bla TEM variants; namely, protein with PDB ID: 1AXB (at resolution 2.00Å) was selected for TEM-116 and that with PDB ID: 1ZG4 (at resolution 1.55Å) was selected for rest of the bla TEM variants. Our bla TEM variants showed 93-94% sequence identity with the selected templates (Table 2). Ramachandran plots confirmed good stereochemical quality of the models-as evidenced by the number of residues in most favoured, additional allowed and generously allowed regions, respectively, and no residue in the disallowed region. The overall quality factors of 3D models predicted by ERRAT were above 94.553 in all the models, except for TEM-1 with ERRAT score of 80.292. Verify 3D server predicted that 80% of the residues of our models had an average 3D-1D, so the models were also verified by Verify 3D. Z-score values indicated that the input structures were within the range typically found for similar sized native proteins. Overall stability change, calculated using both atom potential and torsion angle distribution, was found to increase due to four alterations (W165G, M69L, A184V and N276D) and decrease due to V84I change ( Table 3). The RMSD ( carbon) values of bla TEM variants with respect to TEM-1 (wild type) were 2.24Å ( TEM-1 like), 0.43Å ( TEM-116 ), 2.58Å ( TEM-169 ), 2.46Å ( TEM-190 ), and 3.02Å ( TEM-33 ).  (Table 2). Approximately 12 times increase in SV, 9.5-fold increase in SA, and 3-fold increase in MA were noticed in inhibitor resistant TEM-33 compared to that of  , one of the broad spectrum variants. The most potential catalytic residue containing ligand binding pocket of bla TEM variants was also validated by metaPocket 2.0.   docking energy (Δ 0 value) of docked complexes indicated strong favorable interaction between bla TEM variants and 3rd generation cephalosporins, whereas low inhibition constant (K value) indicated high affinity of -lactamase for these antimicrobials (Table 2, Figure 4). Considering Δ 0 and K values of cephalosporin--lactamase interaction, two patterns were observed among the detected bla TEM variants. The TEM-1 like, TEM-116 ,  , and TEM-33 (Group I) might interact with cephalosporins in the following order: cefotaxime > ceftazidime > cefpodoxime, whereas TEM-1 and TEM-169 (Group II) might interact with cephalosporins in the order: ceftazidime > cefotaxime > cefpodoxime. However, for both Groups I & II bla TEM s, distance between carbonyl carbon atom of -lactam ring of cephalosporin and Ser-70 of bla TEM was lowest for cefotaxime.

Discussion
Our study indicated high degree of 3rd generation cephalosporin resistance among pathogenic Enterobacteriaceae infecting Indian patients. The most common cause of cephalosporin resistance is acquisition of plasmid encoded bla TEM gene variants that produce -lactamases with altered degree of cleaving capacity towards 3rd generation cephalosporins. As multiple plasmids were found in majority of our bacterial isolates, higher chance of -lactamase gene  TEM84  TEM151  TEM152  TEM78   TEM190 (KC699842)   TEM190   TEM104   TEM125  TEM158   TEM191  TEM82  TEM122  TEM192  TEM163  TEM171  TEM116 (JF973688)  TEM116  TEM157   TEM98  TEM99  TEM97   TEM162   TEM54  TEM145  TEM146  TEM67  TEM160  TEM2  TEM33  TEM33 (KC699843)   TEM108   TEM80  TEM81  TEM154  TEM189  TEM83  TEM169  TEM169 (KC699843)  TEM185  TEM186  TEM70  TEM198  TEM164  TEM148  TEM76  TEM90  TEM1like (KC699843)  TEM156  TEM1 (JN002395)  TEM96  TEM105  TEM183  TEM57  TEM201  TEM1  TEM141  TEM55  TEM166  TEM127  TEM128  TEM95  TEM178 0.005 TEM84 cluster TEM171 cluster TEM33 cluster mobilization existed in eastern Indian patient population indicating higher bacterial clonality for bla TEM carriage. The present study is the first to report presence of six bla TEM variants, namely, TEM-1 , TEM-1 like, TEM-116 , TEM-169 ,  , and TEM-33 , in patient-isolated E. coli and K. pneumoniae from India. Multiple bla TEM variants were suggested to coexist among 3 clinical isolates indicating extensive role of TEM -lactamases for resistance against cephalosporins in our patient isolates. Interestingly, our isolated bla TEM s encoded five types of broad spectrumlactamases and one inhibitor resistant TEM -lactamase. Though ESBL production was indicated in more than 50% of the analyzed bacteria, no ESBL type TEM was found in our study indicating limited role of TEM -lactamases and  presence of other ESBL genes within them, thus, imparting bacterial ESBL property. This is the first report from India to detect presence of multiple bla TEM variants, as only TEM-1 has been previously reported from southern India [26].
TEM-1 like variant has not been previously reported elsewhere. Selection pressure at 165th and 276th residues of our TEM-1 like, TEM-169 and TEM-190 proteins indicated positive selection pressures during natural evolution of these variants.
These bla TEM variants differed at five amino acid positions: M69L, V84I, W165G, A184V, and N276D. In case of M69L, distance measurement of M/L at position 69 from Ser-70 indicated no significant structural change due to this alteration. Moreover, both Met and Leu were good helix-forming residues, and ΔΔG value of M69L conversion indicated it to be a stabilizing mutation [27]. In case of V84I, bothbranched Val and Ile at position 84 had more bulkiness near the protein backbone, which might restrict the protein to adopt an alpha-helical conformation [28]. Also, significant change in solvent accessibility and negative ΔΔG value of V84I conversion indicated it to be less stabilizing mutation. Among all the substitutions, only W165G was located within the omega loop. Both Trp and Gly were nonpolar amino acids; hence, this conversion did not change the protein polarity. But due to side-chainless property of Gly, Glycontaining bla TEM s might have much more conformational flexibility which might play an important role to maintain the omega loop structure [27]. Earlier studies also reported that flexibility of the omega loop might allow the distance between two residues to be shortened during the acylation process of catalytic mechanism [29]. No change in distance of Ser-70 from Trp/Gly at position 165, very little change in solvent accessibility, and positive ΔΔG value of W165G indicated that no significant structural change was caused due to this substitution (Table 3). In case of A184V, Ala, one of the best helix-forming residues, was replaced bybranched Val, which was invariably a poor helix former and for this conversion on -helix-10, overall stability was decreased [27,28]. There was significant change in solvent accessibility due to A184V substitution, with Val-184 of the helix being totally exposed on the enzyme surface.
Solvent accessible pocket volume/area of these six types of bla TEM s indicated differential antibiotic spectrum among them (Table 2). Greatest SV, SA, and MA of TEM-33 , compared to other bla TEM variants, might explain its expanded substrate profile towards -lactam--lactamase inhibitor combinations. RMSD value of 3.02Å of TEM-33 , compared to TEM-1 , referred to significant change in its structural stability. Previous studies reported that minimum free energy of interaction or tight binding for an enzymeantimicrobial complex was regarded as an indicator of resistance against antimicrobials [29]. ΔG 0 and K values of the docked complexes indicated that for Group I bla TEM variants, cefotaxime might interact more efficiently than ceftazidime/cefpodoxime-thus imparting greater resistance towards cefotaxime than other two antibiotics; whereas in case of Group II bla TEM variants, ceftazidime might interact more efficiently than cefotaxime/cefpodoxime-thus showing higher ceftazidime resistance. However, resistance pattern of the analyzed bacteria towards 3rd generation cephalosporins could not be explained solely due to presence of bla TEM groups, as other -lactamase encoding genes were found to be present within them (data not shown). Hydrogen bonds and hydrophobic interaction played a critical role in stabilizing protein-ligand complexes and therefore contributed significantly to improving binding affinity and efficacy of the antimicrobial [30]. LIGPLOT analysis that revealed catalytic residues Ser-70 and Ser-130 might play critical role in stabilizing the docked complexes by hydrogen bonding (Table 2).
TEM-1 and TEM-169 were also stabilized by K234, another catalytic residue. Amino acids R244, S235, A237, and N132 are frequently hydrogen bonded with the antibiotics indicating their crucial role during molecular interactions. An increase in number of hydrophobic atoms in the active core of antimicrobial-target interface further increases the binding affinity between protein-antimicrobial interfaces. In case of Group II bla TEM variants, the number of amino acids showing hydrophobic interaction with ceftazidime was higher than that with cefotaxime-thus validating greater resistance of these variants towards ceftazidime. Similarly, in case of TEM-116 and TEM-1 like of Group I, a higher number of hydrophobic interactions were found in case of cefotaxime than ceftazidime-thus indicating their greater resistance property towards cefotaxime. However,  and TEM-33 did not follow this pattern.

Conclusion
Thus, the study indicated significant role of multiple TEMlactamases in imparting 3rd generation cephalosporin resistance among pathogenic bacteria infecting Indian patient population. In silico analysis predicted differential antibiotic resistance pattern among bla TEM variants. Hence, early detection of antibiotic resistant gene variants could guide the choice of optimal antibiotic therapy for successful treatment-thus improving the outcomes for patients with severe Enterobacterial infections.