The mechanism of ceftazidime and cefiderocol hydrolysis by D179Y variants of KPC carbapenemases is similar and involves the formation of a long-lived covalent intermediate

ABSTRACT

Klebsiella pneumoniae carbapenemase (KPC) variants have been described that confer resistance to both ceftazidime–avibactam and cefiderocol. Of these, KPC-33 and KPC-31 are D179Y-containing variants derived from KPC-2 and KPC-3, respectively. To better understand this atypical phenotype, the catalytic mechanism of ceftazidime and cefiderocol hydrolysis by KPC-33 and KPC-31 as well as the ancestral KPC-2 and KPC-3 enzymes was studied. Steady-state kinetics showed that the D179Y substitution in either KPC-2 or KPC-3 is associated with a large decrease in both k_cat and K_M such that k_cat/K_M values were largely unchanged for both ceftazidime and cefiderocol substrates. A decrease in both k_cat and K_M is consistent with a decreased and rate-limiting deacylation step. We explored this hypothesis by performing pre-steady-state kinetics and showed that the acylation step is rate-limiting for KPC-2 and KPC-3 for both ceftazidime and cefiderocol hydrolysis. In contrast, we observed a burst of acyl-enzyme formation followed by a slow steady-state rate for the D179Y variants of KPC-2 and KPC-3 with either ceftazidime or cefiderocol, indicating that deacylation of the covalent intermediate is the rate-limiting step for catalysis. Finally, we show that the low K_M value for ceftazidime or cefiderocol hydrolysis of the D179Y variants is not an indication of tight binding affinity for the substrates but rather is a reflection of the deacylation reaction becoming rate-limiting. Thus, the hydrolysis mechanism of ceftazidime and cefiderocol by the D179Y variants is very similar and involves the formation of a long-lived covalent intermediate that is associated with resistance to the drugs.

INTRODUCTION

Resistance to carbapenems can be mediated by the production of Klebsiella pneumoniae carbapenemase (KPC), which belongs to the family of class-A serine β-lactamases. This enzyme was first found in 1996 in the United States (1) and is now a common carbapenem resistance mechanism among Enterobacterales (2). The most frequently encountered variants worldwide are KPC-2 and KPC-3 (which differ from KPC-2 due to the H274Y mutation) (3). They demonstrate a broad substrate profile, provide high-level bacterial resistance to most standard β-lactams, and are weakly inhibited by classical β-lactamase inhibitors (clavulanic acid, tazobactam, and sulbactam). New inhibitors, such as avibactam, vaborbactam, and relebactam, have been developed that act against KPC β-lactamases (4). The combination of ceftazidime-avibactam demonstrates good in vitro activity and is associated with a decreased mortality rate in treated patients (5). Unfortunately, since its introduction in 2015, many new KPC variants have emerged that provide resistance to ceftazidime-avibactam (3). Another therapeutic alternative is cefiderocol, which is a recently approved siderophore cephalosporin antibiotic sharing similar side chains with ceftazidime (Fig. 1) and cefepime. It also contains a siderophore-like chlorocatechol group that facilitates entry into bacteria through iron transport pathways. Cefiderocol is poorly hydrolyzed by most β-lactamases, including KPC enzymes (6, 7).

Fig 1.

Structure of β-lactams used in this study: (A) ceftazidime and (B) cefiderocol. (C) Kinetic model for serine-active site β-lactamases. Minimal scheme for the reaction with E as a free enzyme, S as substrate, ES as enzyme-substrate complex, EAc as acyl-enzyme intermediate, and P as a product. (D) Equations for k_cat, K_M, and k_cat/K_M based on the kinetic scheme shown in C. A simplified model is also shown based on the assumption that k₋₁ >> k₂.

Recently, KPC variants have been described as providing cross-resistance to both ceftazidime-avibactam and cefiderocol, which are both last resort molecules (8–11). Of these, KPC-33 and KPC-31, which are derived from the KPC-2 and KPC-3 β-lactamases by the addition of the D179Y substitution, are among the most frequent (3).

The basis for the ceftazidime-avibactam resistance provided by KPC-31 and KPC-33 has been characterized in biochemical and structural studies (12–15). From a biochemical point of view, what emerges is that the D179Y mutation in KPC-2 confers a decrease in K_M with a variable change in the k_cat/K_M ratio for ceftazidime hydrolysis. This apparent increase in affinity for ceftazidime is accompanied by a reduced inhibition potency for avibactam, which is consistent with the phenotypic resistance to ceftazidime-avibactam (14, 15). However, both k_cat and K_M are macroscopic parameters that are governed by microscopic rate constants. Moreover, K_M is often interpreted as a direct reflection of the affinity between the enzyme and its substrate but it can be impacted by microscopic rate constants, particularly the deacylation rate (Fig. 1) (16, 17).

Here, we examined the mechanism of ceftazidime hydrolysis by the KPC-33 and KPC-31 variants through the determination of microscopic rate constants in the reaction scheme. Furthermore, because of the structural similarity between ceftazidime and cefiderocol, we hypothesized that their biochemical mechanisms of hydrolysis would be similar. In an attempt to better understand this atypical phenotype among KPC-β-lactamases and to understand this complex substrate specificity relationship, the catalytic mechanism of ceftazidime and cefiderocol hydrolysis from KPC variants harboring the D179Y mutation from KPC-33 and KPC-31 as well as their ancestor (KPC-2 and KPC-3) enzymes was studied.

RESULTS

Steady-state enzyme kinetic measurements

Comparison of steady-state kinetic measurements for KPC-2 and the D179Y mutant KPC-33

We and others have shown that the KPC-2 β-lactamase poorly hydrolyzes ceftazidime (18, 19). Consistent with these findings, it was not possible to accurately determine k_cat for ceftazidime with KPC-2 due to a high K_M value (>400 µM); however, it is >0.4 s⁻¹. k_cat/K_M was determined by fitting progress curves, yielding a value of 1,000 M⁻¹ s⁻¹. The addition of the D179Y substitution to KPC-2 to create KPC-33 resulted in a >28-fold reduction in K_M to 14 µM and k_cat was 0.015 s⁻¹, which represents a >26-fold reduction compared to KPC-2. Despite the decrease in k_cat for ceftazidime hydrolysis, the catalytic efficiency (k_cat/K_M ratio) did not change significantly for KPC-33 compared to KPC-2 due to the compensating effect of the lower K_M (Table 1; Fig. 2).

TABLE 1.

Steady-state and pre-steady-state enzyme kinetics results for KPC enzymes

	kcat (s−1)	KM (µM)	kcat/KM (M−1 s−1)	k2 (s−1)	k3 (s−1)	KD (µM)
Ceftazidime
KPC-2	>0.4	>400	1,000 ± 40	–	–	–
KPC-33	0.015 ± 0.0006	14 ± 3	1,000 ± 250	0.5 ± 0.1	0.017 ± 0.003	150 ± 60
KPC-3	>3	>400	7,300 ± 70	–	–	–
KPC-31	0.02 ± 0.0006	6 ± 1.5	3,300 ± 800	>0.6	0.02 ± 0.003	>200
Cefiderocol
KPC-2	>0.03	>500	50 ± 2	–	–	–
KPC-33	0.006 ± 0.00009	3 ± 0.4	2,000 ± 250	>0.08	0.006 ± 0.001	>250
KPC-3	>0.13	>500	250 ± 6	–	–	–
KPC-31	0.008 ± 0.0004	12 ± 3	660 ± 160	0.1 ± 0.01	0.008 ± 0.001	80 ± 25

Fig 2.

Rates of ceftazidime and cefiderocol hydrolysis plotted versus substrate concentration for KPC-2 (A and B, respectively), KPC-33 (KPC-2 D179Y) (C and D), KPC-3 (E and F), and KPC-31 (KPC-3 D179Y) (G and H). For KPC-2 and KPC-3, it was not possible to reach saturating substrate concentrations. The second-order rate constant at steady-state, k_cat/K_M, was determined by fitting the progress curves to the equation v = k_cat/K_M [E][S], where [S] << K_M. For KPC-33 and KPC-31, the steady-state parameters were obtained by a non-linear least squares fit of the data to the Michaelis-Menten equation v = k_cat[S]/(K_M + [S]).

Similar to the results with ceftazidime, the KPC-2 enzyme poorly hydrolyzes cefiderocol (Table 1; Fig. 2). k_cat could not be accurately determined due to a high K_M (>500 µM) but it is >0.03 s⁻¹. It is clear that cefiderocol is a very poor substrate for KPC-2 in that k_cat/K_M was determined to be 50 M⁻¹ s⁻¹, which is 20-fold lower than that observed with ceftazidime as substrate. Similar to observations with ceftazidime, the D179Y substitution had a strong impact on cefiderocol hydrolysis in the KPC-2 background, with a decrease in K_M from >500 µM to 3 µM (>170-fold), and a decreased k_cat of 0.006 s⁻¹. In contrast to the observation with ceftazidime, where the D179Y substitution did not change k_cat/K_M, the substitution resulted in a 40-fold increase in catalytic efficiency for cefiderocol hydrolysis, from 50 to 2,000 M⁻¹ s⁻¹ (Table 1; Fig. 2). Thus, although the trend of strongly decreased k_cat and K_M values is observed for the KPC-33 enzyme with both substrates, cefiderocol is hydrolyzed less efficiently than ceftazidime by KPC-2, and the D179Y substitution increases the hydrolysis of cefiderocol to a greater extent than ceftazidime.

Comparison of steady-state kinetic measurements for KPC-3 and the D179Y mutant KPC-31

The H274Y substitution in KPC-2 creates KPC-3, which has been reported to increase k_cat/K_M for ceftazidime hydrolysis ninefold relative to KPC-2 (18). Consistent with these results, we measured a k_cat/K_M value of 7,300 M⁻¹ s⁻¹, which is sevenfold higher than the value for KPC-2 (Table 1). As observed previously, k_cat could not be accurately determined due to a high K_M value but it is >3 s⁻¹ (Table 1; Fig. 2). Similar to the observations with KPC-2, the addition of the D179Y substitution to KPC-3 to create KPC-31 greatly reduced both k_cat (0.02 s⁻¹) and K_M (6 µM). Also, similar to the findings with KPC-2, the addition of the D179Y substitution to KPC-3 had only a modest effect (twofold reduction) on k_cat/K_M (Table 1).

As with ceftazidime, the addition of the H274Y substitution to create KPC-3 increased the k_cat/K_M value fivefold relative to KPC-2 for cefiderocol hydrolysis (250 M⁻¹s⁻¹) (Table 1). k_cat could not be determined for cefiderocol hydrolysis by KPC-3 due to a high K_M value but it is >0.13 s⁻¹. As observed with KPC-2, the introduction of the D179Y substitution into KPC-3 to create KPC-31 decreased the cefiderocol K_M at least 40-fold, from >500 µM to 12 µM. However, k_cat/K_M was only increased by 2.4-fold for KPC-31 relative to KPC-3 due to a large decrease in k_cat coincident with the decrease in K_M (Table 1; Fig. 2). Thus, the same trend of greatly decreased k_cat and K_M values due to the addition of the D179Y substitution to either KPC-2 or KPC-3 is observed for both ceftazidime and cefiderocol substrates.

Pre-steady-state enzyme kinetic measurements

In the kinetic scheme for β-lactamases, a decrease in k_cat coupled with a decrease in K_M is consistent with a reduced and rate-limiting deacylation step (17). To further explore this hypothesis, we performed pre-steady-state kinetics with 9 µM of KPC-2, KPC-3, KPC-33, and KPC-31 and multiple concentrations of ceftazidime (from 50 to 200 µM). A rate-limiting deacylation step predicts that the reaction of β-lactamase with an excess of substrate would exhibit burst kinetics whereby the progress curve is biphasic with a rapid step with an amplitude near the enzyme concentration followed by a slower, steady-state rate (20). The reaction of KPC-2 and KPC-3 with ceftazidime showed a progress curve that could be fit to a single exponential equation, i.e., no burst, suggesting that the acylation step is rate-limiting (or has a similar rate as the deacylation rate) and thus k₂ ≤ k₃ (Fig. 3).

Fig 3.

Pre-steady-state experiments. Enzymes were mixed with 50 µM ceftazidime (A and B) or cefiderocol (C and D) and absorbance at 260 or 259 nm, respectively, was monitored for 15 s. The sign of ΔConcentration (µM) was changed from negative to positive. (A) Progress curves of KPC-2 (blue) and KPC-33 (red) with ceftazidime. Black lines correspond, respectively, to fit burst equation and a single exponential fit for KPC-33 and KPC-2. (B) Progress curves of KPC-3 (blue) and KPC-31 (red) with ceftazidime fit to a single exponential and burst equation, respectively (black lines). (C) Progress curves of KPC-2 (blue) and KPC-33 (red) with cefiderocol fit to a single exponential and burst equation, respectively (black lines). (D) Progress curves of KPC-3 (blue) and KPC-31 (red) with cefiderocol fit to a single exponential and burst equation, respectively (black lines).

In contrast, for KPC-33 and KPC-31 with ceftazidime as substrate, we observed a burst of acyl-enzyme formation followed by a slower steady-state rate (Fig. 3 and 4; Fig. S1). These results suggest that deacylation is the rate-limiting step and thus k₂ > k₃, indicating the D179Y substitution changes the rate-limiting step from acylation to deacylation. Fitting these data with the burst equation provided a deacylation rate (k_ss = k₃) of 0.017 s⁻¹ for KPC-33, which is similar to the k_cat value obtained by steady-state kinetics (0.015 s⁻¹) (see Materials and Methods). A plot of the k_obs values obtained with different ceftazidime concentrations was fitted to a hyperbola to obtain an estimate of the acylation rate (0.5 s⁻¹) and binding constant (K_D) for ceftazidime (150 µM) (Fig. 4). The same analysis performed on KPC-31 kinetics data provided a similar deacylation rate of 0.02 s⁻¹ but we could not determine the acylation rate or binding constant as saturating ceftazidime concentrations could not be achieved, although k₂ must be greater than 0.6 s⁻¹ and the K_D > 200 µM (Fig. S1).

Fig 4.

Example of results obtained during pre-steady-state experiments. The change in ceftazidime concentration was monitored at 260 nM and plotted versus time (red points) for KPC-33 (KPC-2 D179Y) with ceftazidime concentrations of 50 µM (A), 100 µM (B), 150 µM (C), and 200 µM (D). The sign of ΔConcentration (µM) was changed from negative to positive. The data were fit to the burst equation (black line; see Materials and Methods). The individual k_obs values at each substrate concentration were then fit to a hyperbola to estimate K_D (dissociation constant) and k₂ (E) (see Materials and Methods).

Note that with KPC-33 and ceftazidime, where k₂, k₃, and K_D values were obtained, we substituted these values into the equations for k_cat and K_M using the simplified β-lactamase (k_-1 > k₂) model in Fig. 1D and obtained a k_cat of 0.016 s⁻¹ and K_M of 4.9 µM. These values are consistent with the experimentally determined k_cat and K_M values in Table 1, providing a check on the pre-steady-state values.

Similar experiments were performed with cefiderocol as substrate (Fig. 3; Fig. S2 and S3). As observed with ceftazidime as substrate, the progress curve for the reaction of KPC-2 and KPC-3 with cefiderocol did not show a burst and could be fit to a single exponential equation, suggesting that the acylation step is rate-limiting (or is similar to the deacylation rate) (Fig. 3). In contrast, the progress curves for the reaction of KPC-31 and KPC-33 with cefiderocol exhibited a burst, indicating deacylation is the rate-limiting step (Fig. 3; Fig. S2 and S3). Fitting the data to the burst equation yielded k₃ values of 0.006 s⁻¹ and 0.008 s⁻¹, respectively, for KPC-33 and KPC-31, similar to the k_cat values obtained with steady-state kinetics. The k_obs values for KPC-31 were plotted and fit to a hyperbola to yield an acylation rate (k₂) of 0.1 s⁻¹ and a K_D value of 80 µM for the binding of cefiderocol to KPC-31. We could not determine K_D for KPC-33 as a saturating cefiderocol concentration could not be achieved (Fig. S3). However, the results indicate that k₂ must be greater than 0.08 s⁻¹ and K_D > 250 µM.

As described above for KPC-33 and ceftazidime, we used the k₂, k₃, and K_D values for KPC-31 with cefiderocol for substitution into the equations for k_cat and K_M using the simplified β-lactamase (k_-1 > k₂) model and obtained a k_cat of 0.007 s⁻¹ and a K_M of 5.9 µM. These values are in range with the steady-state experimental values indicating the pre-steady-state determinations are consistent with steady-state results (Table 1).

DISCUSSION

KPC β-lactamases have spread worldwide, and the increasing reports of mutations conferring resistance to ceftazidime-avibactam are worrisome and illustrate their evolvability when facing new substrates (3). In 2021, a first report showed that the D179Y mutation present in either the KPC-2 or KPC-3 background could confer cross-resistance between ceftazidime-avibactam and cefiderocol (8). Here, we present a kinetic analysis of the catalytic properties of KPC-2, KPC-3, KPC-33, and KPC-31 β-lactamases with ceftazidime and cefiderocol as substrates.

Our steady-state kinetics results are in accordance with those of Shapiro et al. and Mehta et al. who found that the K_M for ceftazidime of KPC-3 was too high to allow substrate saturation to be obtained (12, 21). The catalytic efficiency (k_cat/K_M) determined here is twofold higher than the value reported by Shapiro et al. (12); however, it is similar to that reported by Mehta et al. (21). The D179Y mutation decreased k_cat/K_M for ceftazidime by twofold, which is lower but remains in the same range as the eightfold decrease found by Shapiro et al. (12).

In agreement with previous reports, the K_M for KPC-2 with ceftazidime was too high to be determined accurately (14, 15, 21). The catalytic efficiency determined here was very similar to that reported by Mehta et al. and by Tsivkoski et al. but approximately fourfold lower than that reported by Compain et al. (14, 15, 21). The D179Y mutation (KPC-33) lowered K_M to 14 µM with a k_cat of 0.015 s⁻¹ in our experimental conditions. However, the catalytic efficiency (k_cat/K_M) was similar to that of KPC-2, indicating that the large decrease in K_M was offset by the decrease in k_cat, which is different from the >10-fold increase reported by others (14, 15). Thus, from a biochemical point of view, our results suggest that the D179Y mutation does not improve catalytic efficiency for ceftazidime, and yet, it increases phenotypic resistance to both ceftazidime and ceftazidime-avibactam (Table 2). The decrease in k_cat coupled with a decrease in K_M is consistent with a reduced and rate-limiting deacylation step (17), which could slow down the enzymatic reaction, resulting in the trapping of ceftazidime in the acyl-enzyme form, which could contribute to resistance. Note, however, that although acyl-enzyme hydrolysis is rate-limiting, turnover still occurs. Thus, it is also possible that the lower K_M values for KPC-31 and KPC-33 lead to increased hydrolysis due to enzyme saturation and V_max being reached at lower antibiotic concentrations in the periplasm compared to KPC-2 and KPC-3.

TABLE 2.

Ceftazidime, ceftazidime-avibactam, cefiderocol, and imipenem MICs of KPC-2, KPC-33, KPC-3, and KPC-31

MIC (mg/L)
	Ceftazidime	Ceftazidime-avibactam	Cefiderocol	Imipenem
KPC-2	16	0.5	0.25	<0.5
KPC-33	>128	24	2	<0.5
KPC-3	64	1	0.12	2
KPC-31	>128	48	4	2

Concerning cefiderocol hydrolysis, the D179Y mutation (KPC-31) lowered the K_M and k_cat but only a modest increase in k_cat/K_M was observed (2.6-fold) when compared with KPC-3. This increase is not likely to explain 32-fold increase in cefiderocol MICs previously reported when the D179Y mutation was added to KPC-3 and expressed in Escherichia coli (Table 2) (8). Conversely, we observed a 40-fold increase in catalytic efficiency for cefiderocol hydrolysis by KPC-33 compared with KPC-2, suggesting that the large decrease in K_M value was not fully compensated by the decrease in k_cat. It is noteworthy that this increase in catalytic efficiency appears to have less impact (compared to KPC-31) on the cefiderocol MIC, which showed only an eightfold increase between KPC-2 and KPC-33 expressed in E. coli (8). k_cat and K_M are complex terms in the β-lactamase kinetic scheme. As the β-lactam hydrolysis reaction is divided into two steps, acylation and deacylation, the value of k_cat is constrained by the rate-limiting step, which can be acylation, deacylation or both if the rates are similar. The K_M does not necessarily reflect the binding affinity of the substrate, as a rate-limiting deacylation rate (k₃) can also lead to low K_M values (17).

Pre-steady-state kinetics was used here to determine the microscopic rate constants contributing to the steady-state kinetic parameters. We utilized burst kinetics experiments where the substrate is in excess of enzyme and a rate-limiting deacylation step is predicted to result in a burst of reaction with an amplitude approaching the concentration of enzyme used, followed by a slow steady-state rate. The burst reflects the accumulation of acyl-enzyme due to an acylation rate that is higher than the deacylation rate, which explains why the amplitude theoretically approaches the enzyme concentration. A burst was not observed for KPC-2 and KPC-3 with ceftazidime or cefiderocol, suggesting that the acylation step is rate-limiting or that the acylation and deacylation rates are similar. This is in contrast to a previous report of a burst for KPC-2 with ceftazidime (22). In that case, the burst was a small fraction of the enzyme concentration, and our findings match the conclusion of the previous study that k₂ is rate-limiting.

We observed a clear burst followed by a slower steady-state rate for the KPC-33 and KPC-31 enzymes containing the D179Y substitution with both ceftazidime and cefiderocol. This suggests a change in the rate-limiting step of the reaction. Thus, acylation is likely to be the rate-limiting step (or have the same rate as deacylation) for KPC-2 and KPC-3, while deacylation is the rate-limiting step upon the addition of the D179Y mutation in both genetic backgrounds. These experiments revealed that the deacylation rate constants (k_ss = k₃) were comparable to the corresponding k_cat values (Table 1) for both KPC-33 and KPC-31 with both ceftazidime and cefiderocol. This supports deacylation as the rate-limiting step, which determines the k_cat value. We noted that the burst amplitude is lower than the enzyme concentration, although these experiments were repeated several times and with different enzyme preparations. These results suggest the active enzyme concentration is lower than the measured enzyme concentration. We performed an active site titration using avibactam as an inhibitor to confirm this hypothesis. It is known from structural studies that the active site omega loop of the KPC-2 D179Y enzyme is disordered and samples many conformations (13). It is possible that many of these conformations are not catalytically competent, thus leading to the active enzyme concentration being lower than the measured concentration.

To determine the substrate binding affinity, the K_D can be obtained by fitting the decay rate of the burst for different substrate concentrations to a hyperbola, assuming that k_-1 >> k₂. By doing so, we obtained, respectively, a 10-fold and >33-fold higher K_D than the K_M for KPC-33 and KPC-31 with ceftazidime, and a >80-fold and sixfold higher K_D than K_M for KPC-33 and KPC-31, respectively, with cefiderocol, which demonstrates that the rate-limiting deacylation step lowers the K_M value, giving an impression of higher affinity. The experimental conditions with ceftazidime and KPC-31 as well as cefiderocol with KPC-33 did not allow saturation of the reaction, also suggesting a low substrate binding affinity (>200 µM).

The observed structural disorder of the omega-loop in KPC-2 and KPC-3 containing the D179Y mutation may enhance the binding of ceftazidime (13, 23). These structural changes might also impact the positioning and hydrolysis of cefiderocol but this remains to be demonstrated.

Taken together, our kinetics results indicate that the addition of the D179Y substitution to KPC-2 and KPC-3 to create KPC-33 and KPC-31 changes the rate-limiting step for the hydrolysis of ceftazidime and cefiderocol from the formation of the covalent intermediate (acylation) to deacylation of the ester intermediate. The low k₃ rate, in turn, results in a low K_M value for ceftazidime and cefiderocol. Note that, in this case, the low K_M value does not indicate a high affinity for ceftazidime and cefiderocol but rather is due to the slow deacylation reaction. We propose that the D179Y mutation increases ceftazidime and cefiderocol resistance through a change in the rate-limiting step of the enzymatic reaction that results in the trapping of the acylated substrate inside the active site, coupled with moderate binding affinity for the substrates. A trapping mechanism conferring resistance has previously been suggested for ceftazidime in some class A β-lactamases such as KPC-33, KPC-49 (R164S), or TEM-1 (13, 22–24). In this study, we arrive biochemically at a similar conclusion. However, it is also possible that the low KM values lead to enzyme saturation by substrate and increased hydrolysis. Finally, KPC-33 was reported to be less inhibited by avibactam, which can explain the resistance to ceftazidime-avibactam (14).

We also show that the cefiderocol hydrolysis occurs by a similar mechanism in the KPC-31 and KPC-33 variants, which could explain the cross-resistance observed (8, 9). In light of the high evolvability potential of the KPC enzyme, it is important to understand the catalytic steps of the enzymatic reaction, as it can facilitate modifying existing antibiotics to alter reaction rates as a possible path to develop drugs with lower potential for resistance.

MATERIALS AND METHODS

Bacterial strains, plasmids, and MIC determinations

The recombinant plasmids KPC-2-pET28a-TEV, KPC-3-pET28a-TEV, and KPC-33-pET28a-TEV were constructed by cloning bla_KPC genes of the mature proteins in a pET28a-TEV plasmid as previously described (25). The expression vector for KPC-31 was constructed by site-directed mutagenesis (26) using pET28a-TEV-KPC-33 as the template DNA and the primers

H274Y-F ^5’ACAAGGATGACAAGtACAGCGAGGCCGTCATCGCCGCT^3’;

H274Y-R ^5’CCTCGCTGTaCTTGTCATCCTTGTTAGGCGCCCGGGTGT^3’

This plasmid is a modified pET28a vector with a thrombin recognition sequence between the 6xHis tag and the protein of interest replaced by Tobacco Etch Virus (TEV) protease recognition sequence. E. coli SHuffle cells were used for overexpression experiments. The same KPC genes were also cloned in pBR322 plasmid and expressed in E. coli TOP10 for MIC determination as previously described (8, 27).

Expression and purification of KPC-2, KPC-3, KPC-31, and KPC-33 proteins wild-type and mutants

As previously described, for the overexpression of KPC-2, KPC-3, KPC-31, and KPC-33 proteins, the corresponding plasmid was transformed into E. coli SHuffle cells (25). Briefly, a few well-isolated colonies were inoculated in LB broth containing 25 µg mL⁻¹ kanamycin and incubated overnight at 37°C with vigorous shaking. Overnight cultures were 1:100 diluted in 1.5 L of LB broth containing 25 µg mL⁻¹ kanamycin and incubated at 30°C until OD₆₀₀ reached 0.8 when isopropyl ß-D-1-thiogalactopyranoside (IPTG) was added to the final concentration of 0.5 mM. Incubation with shaking was continued at 18°C for 16 hours before cells were pelleted by centrifugation at 5,000 rpm for 15 min.

For the purification of 6xHis-tagged KPC proteins, the cell pellet was resuspended in Lysis Buffer (20 mM sodium phosphate, pH 7.4, 500 mM NaCl, and 20 mM imidazole). Cells were lysed by sonication. The cell debris was removed by centrifugation at 12,000 g for 30 min, and the soluble fraction of the lysate was filtered by a 0.45-µm filter unit and loaded onto a 5 mL HisTrap FF column (GE Healthcare). After washing the column with lysis buffer, 6xHis-tagged recombinant KPC protein was eluted with a 20- to 500-mM imidazole gradient in the lysis buffer. Fractions containing the recombinant proteins were pooled and concentrated, and the buffer was exchanged to the lysis buffer with a 10-kDa cut-off Amicon concentrator unit (EMD Millipore). The 6xHis tag was removed by incubation overnight at 4°C with TEV protease at 1:50 (TEV:His-KPC protein) ratio, and the TEV protease was removed by incubation with Ni Sepharose Fast Flow beads (GE Healthcare). KPC proteins were further purified by gel-filtration chromatography with a Superdex 75 Increase (10/300) column (GE Healthcare) using (20 mM sodium phosphate, pH 7.4, 200 mM NaCl) as the running buffer. Fractions containing the recombinant proteins were pooled and concentrated with a 10-kDa cut-off Amicon concentrator unit (EMD Millipore). The purity of each protein was over 95% based on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis.

Protein concentration determinations

Molar absorption coefficients (Epsilon) were determined experimentally according to the Edelhoch method (28) giving Ʃ280 nM of KPC-31 = 44,805 M⁻¹ cm⁻¹, KPC-33 = 43,130 M⁻¹cm⁻¹, KPC-2 = 41,125 M⁻¹ cm⁻¹ (+4%) and KPC-3 = 43,104 M⁻¹ cm⁻¹. Protein concentrations were determined experimentally using absorbance spectroscopy with the determined molar absorption coefficients.

Steady-state enzyme kinetic measurements

The kinetics of β-lactamases hydrolyzing β-lactam substrates through a two-step reactions is represented in Fig. 1. The Michaelis-Menten kinetic parameters for KPC-2, KPC-3 and the variant enzyme-substrate pairs were determined at 25°C in 50 mM sodium phosphate buffer, pH 7.0, containing 0.1 mg/mL BSA using variable amounts of enzyme depending on the enzyme-substrate pair. The initial velocities of β-lactam hydrolysis were measured on a Beckman-Coulter spectrophotometer model DU-800 (Fullerton, CA) at, respectively, 260 nm and 259 nm for ceftazidime and cefiderocol using the following extinction coefficients: ceftazidime Δε260 = −8,660 M⁻¹ cm⁻¹ and cefiderocol Δε259 = −9,443 M⁻¹ cm⁻¹. GraphPad Prism 8.4.2 was used to obtain the steady-state parameters by non-linear least squares fit of the data to the Michaelis-Menten equation v = k_cat[S]/(K_M + [S]). When the velocity of substrate hydrolysis could not be saturated by measurable concentrations due to a high K_M, the second-order rate constant at steady-state, k_cat/K_M, was determined by fitting the progress curves to the equation v = k_cat/K_M [E][S], where [S] << K_M. Each set of measurements was repeated with at least two separate experiments.

The error for k_cat and K_M is the standard error of the fit of the initial velocity determinations to the Michaelis-Menten equation. The error on k_cat/K_M was determined by propagating the error of both k_cat and K_M by calculating the square root of the sum of the fractional errors of k_cat and K_M squared (21).

Pre-steady-state kinetics

Pre-steady-state kinetics were carried out on a Kintek SF-300X stopped-flow spectrophotometer. For the KPC-2, KPC-3, KPC-31, and KPC-33 proteins, we used 9 µM of enzyme and variable concentrations of substrate (ceftazidime or cefiderocol). These reactions were performed at 25°C in 50 mM phosphate buffer pH 7.0 containing 0.1 mg/mL BSA. For each kinetic experiment, 2,000 data points were collected, and at least five kinetic traces were collected to ensure high-quality data for times ranging from 1 to 60 s on two different protein preparations. The dead time of the instrument is 0.85 ms. The observed data were fit to either a one-phase decay (single exponential) equation or the burst equation using GraphPad Prism 8.4.2.

Burst equation:

$${\displaystyle Y=A\ast (1-\mathrm{e}\mathrm{x}\mathrm{p}(-{\mathrm{k}}_{obs}\ast x))+\mathrm{k}\mathrm{s}\mathrm{s}\ast x}$$

where A is the absorbance, k_obs is the apparent rate from the fast phase of the biphasic reaction, and k_ss is the slower, steady-state rate. The individual k_obs data at each substrate concentration were then fit to a hyperbola to determine K_D (affinity constant) and the mean of k_ss data was used to determine the k₃ value for KPC-33 (ceftazidime and cefiderocol) and KPC-31 (cefiderocol).

The acylation rate (k₂) and K_D were determined by plotting the concentration dependence of the observed decay rate (k_obs) of the burst phase to a hyperbolic function v = ((k_obsmax × [S])/(K_D + [S])) to give the maximum decay rate (k_obsmax) and the K_D for substrate binding, where [S] represents substrate concentration (Fig. 4E). The maximum burst rate in the β-lactamase mechanism (Fig. 2C) is the sum of k₂ + k₃. Since we know from the presence of the burst that k₃ << k₂, the maximum burst rate from the hyperbola represents k₂.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.01108-23.

Supplemental figures. aac.01108-23-s0001.pdf.

Fig. S1 to S3

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.