Differential active site requirements for NDM-1 β-lactamase hydrolysis of carbapenem versus penicillin and cephalosporin antibiotics

New Delhi metallo-β-lactamase-1 exhibits a broad substrate profile for hydrolysis of the penicillin, cephalosporin and ‘last resort’ carbapenems, and thus confers bacterial resistance to nearly all β-lactam antibiotics. Here we address whether the high catalytic efficiency for hydrolysis of these diverse substrates is reflected by similar sequence and structural requirements for catalysis, i.e., whether the same catalytic machinery is used to achieve hydrolysis of each class. Deep sequencing of randomized single codon mutation libraries that were selected for resistance to representative antibiotics reveal stringent sequence requirements for carbapenem versus penicillin or cephalosporin hydrolysis. Further, the residue positions required for hydrolysis of penicillins and cephalosporins are a subset of those required for carbapenem hydrolysis. Thus, while a common core of residues is used for catalysis of all substrates, carbapenem hydrolysis requires an additional set of residues to achieve catalytic efficiency comparable to that for penicillins and cephalosporins.


Supplementary Figure and Tables
Differential active site requirements for NDM-1 β-Lactamase hydrolysis of carbapenem versus penicillin and cephalosporin antibiotics Sun et al.
Supplementary Figure 1. Relative fitness of NDM-1 mutants based on sequencing versus antibiotic resistance levels. The relative fitness of individual NDM-1 mutants from the antibiotic selected libraries ( " # ) was calculated based on the frequency of occurrence of the mutant and wild-type alleles from the deep sequencing data for the naïve library and antibiotic-selected libraries as described in Methods. The relative fitness of the mutants based on antibiotic resistance levels ( " # ) was calculated as the logarithm of the resistance level of mutants relative to that of wild type for each antibiotic. The " # value of each mutant is plotted versus " # of the corresponding mutant for ampicillin (A), cefotaxime (B) and imipenem (C) experiments and the results of linear regression analysis are shown indicating a correlation between relative fitness determined based on sequencing data versus that determined based on antibiotic resistance levels of individual mutants.
Supplementary Figure 2. In vivo steady-state protein levels of NDM-1 wild-type and mutants. Steady-state protein expression levels in E. coli of StrepII-tagged wild-type NDM-1 and mutants were determined by SDS-PAGE of whole cell lysates followed by immunoblotting with an anti-StrepII tag monoclonal antibody conjugated to horseradish peroxidase (HRP). Constitutively expressed DnaK (~70 kDa) was used as a loading control and probed with anti-DnaK monoclonal antibody and an HRP-conjugated secondary antibody. The hybridization signal for wild-type and mutant StrepII-tagged NDM-1 and DnaK was quantified by densitometry. The signal for NDM-1 was normalized to that for DnaK in the same sample. Protein expression levels of NDM-1 mutants are shown in the bar graph relative to that of wild-type NDM-1, which was set as 1. Quantification data are based on three independent experiments and a representative blot is shown. The error bars indicate the standard deviation of the protein expression levels based on the repeated experiments.
Supplementary Figure 3. pH profile for wild-type NDM-1 and the K224R/G232A/N233Q triple mutant. The pH dependence of k cat /K M (A), k cat (B), K M (C) of the wild type (closed circles) and the K224R/G232A/N233Q triple mutant (open circles) NDM-1 β-lactamase for ampicillin hydrolysis is shown. The error bars in the plots represent the standard deviations for each data point. Lys224 plays an important role in facilitating substrate binding to the active site through electrostatically interacting with the negatively charged C-3/4 carboxylate group of β-lactam antibiotics. Its substitution by arginine, which also has a positively charged side chain but with a different pKa (12.5) from that of lysine (10.5), may change the pH dependence of the charge status of their side chain groups and thus their interactions with β-lactam substrates. To test this hypothesis, the pH dependence of the wild type and K224R/G232A/N233Q mutant enzymes was examined for the hydrolysis of ampicillin. Fitting of the k cat and k cat /K M to a double ionization model produced bell-shaped curves for k cat /K M of both wild type and the mutant enzyme. Although the optimum pH for k cat /K M is similar for wild type and mutant enzymes, the mutant enzyme exhibited a narrower pH profile (pK1 = 6.7 and pK2 = 7.5) than the wild-type enzyme (pK1 = 5.9 and pK2 = 8.9). This difference is mainly attributed to the larger fluctuation of the K M of the mutant enzyme than that of the wild-type enzyme at pH 5.5-8. 5. This indicates that complex formation with ampicillin is more sensitive to pH variation versus the wild-type enzyme.
Supplementary Figure 4. Stereo view of the active site of the NDM-1 K224R/G232A/N233Q βlactamase structure. Electron density is shown of an 2Fo-Fc map contoured at 1 σ. Zn1 and Zn2 are shown as gray spheres with Zn1 at top. Residues are represented as stick models with carbons colored dark cyan and non-carbon colored by type (N, blue; O, red; S, yellow). The mutated residues, Arg224, Ala232, and Gln233, are shown with carbon atoms in magenta.
Supplementary Figure 5. Overlay of the four chains of the NDM-1 K224R/G232A/N233Q structure. Chains A, B, C, and D are represented as dark cyan, gray, pale green and salmon, respectively. Zinc ions are represented as gray spheres and the water molecule coordinated by the corresponding zinc ions is shown as a red sphere. Zinc-chelating residues (His116, His118, Asp120, His196, and His263), as well as residues Arg224, Ala232, and Gln233 are represented as stick model in all structures with non-carbon atoms colored by type (N, blue; O, red; S, yellow). The active site loop structures L3 and L10 are labeled.
Supplementary Figure 6. Hydrogen bond networks linking loops in the NDM-1 active site. A. The network of interactions involving Asp84 and Lys121 is shwon. Note that, for clarity, not all zinc ligand residues are shown. B. The network of interactions involving Asp199 is shown. Hydrogen bonds are depicted as black lines. Black lines ending at the cartoon tube represent hydrogen bonds to the main chain CO or NH groups. Zinc ions are colored gray and labeled. Carbon is shown in tan, nitrogen is blue and oxygen is red. Side chains of residues are shown and the main chain is shown in cartoon tube. Figure 7. Sequence logo comparison of the deep sequencing results for NDM-1 versus CphA β-lactamases. Deep sequencing results from single codon random libraries selected for imipenem resistance for the NDM-1 and CphA metallo-β-lactamases are compared using sequence logos. A. Essential residues for both enzymes. The wild-type residues dominate among sequences from populations selected for imipenem resistance. B. Residue positions where the sequence requirements for impenem resistance differ between NDM-1 and CphA β-lactamases. C. Non-essential residues, no specific residues strongly dominate after the imipenem resistance selection by either NDM-1 or CphA β-lactamase. The data for CphA is from reference 1 .

Supplementary
Supplementary Figure 8. Structure alignment of active site residues of subclass B1 NDM-1 (PDB ID: 3SPU)(tan) versus subclass B3 AIM-1 β-lactamase (PDB: 4AWZ 2 ) (pink). Amino acid residues that are identical in sequence and position are labeled in black. Unique residues are labeled with the enzyme name and colored according to the structure. Note that AIM-1 residues Ser221 and Thr223 are positioned to replace the function of NDM-1 Lys224. Also, the side chain of AIM-1 Gln157 is positioned near that of NDM-1 Asn233. Finally, AIM-1 His121 serves as a zinc ligand in place of Cys221 of NDM-1, which is Ser221 in AIM-1. The zinc ions are shown in tan for NDM-1 and pink for AIM-1.  Supplementary Figure 13. Quality scores for Illumina DNA sequencing reads. The scores were generated with the Galaxy web server where the file containing sequencing data was uploaded onto https://usegalaxy.org/ and the FASTQ groomer was run to convert the FASTQ file to standard format. FastQC Read Quality reports were produced, which indicate per base sequence quality scores. The per base sequence quality is shown as a BoxWhisker type plot to indicate the average quality score at each position across all reads. In the graph, the x-axis shows the position in the sequence read and the y-axis shows the quality scores. Higher scores indicate higher confidence in the base call. The background of the graph divides the y axis into calls of very good quality (green), calls of reasonable quality (orange), and calls of poor quality (red). The yellow bars represent the inter-quartile range (25-75%) and the central red line represents the median value of the quality score. The upper and lower whiskers represent the 10% and 90% points of the quality scores, representatively. The blue line indicates the mean quality score.