Hydrolysis and Inhibition Profiles of β-Lactamases from Molecular Classes A to D with Doripenem, Imipenem, and Meropenem

Doripenem is a parenteral carbapenem with broad-spectrum activity against many aerobic and anaerobic pathogens. Doripenem MICs against gram-negative clinical isolates are frequently ≤0.5 μg/ml, even in Enterobacteriaceae expressing extended-spectrum β-lactamases (ESBLs) or overproduced AmpC (7, 12, 14). Resistance to doripenem and other carbapenems is observed in isolates producing metallo-β-lactamases (MBLs) or class A or class D carbapenemases (12, 13). In this study, we evaluated the hydrolysis of doripenem by a spectrum of β-lactamases from almost every functional group (3) and compared the doripenem kinetic profiles to those obtained for imipenem and meropenem.

(These data were presented in part as poster P909 at the 18th European Congress of Clinical Microbiology and Infectious Diseases, Barcelona, Spain, 2008.)

The broth microdilution methodology was used to determine MICs (6). Doripenem was from Shionogi & Co., Ltd. (Hyogo, Japan). Benzylpenicillin and cephaloridine were from Sigma (St. Louis, MO). Ceftazidime, imipenem, and meropenem were from U.S. Pharmacopeia (Rockville, MD).

MICs are shown in Table 1 for the strains used as sources of β-lactamases. The carbapenem MICs for isolates producing broad-spectrum, extended-spectrum, and AmpC β-lactamases were ≤2 μg/ml across the represented bacteria. Doripenem and meropenem MICs for the non-carbapenemase-producing isolates were generally four- to eightfold lower than those obtained for imipenem. Reduced susceptibility was apparent in isolates that expressed β-lactamases with known carbapenem hydrolysis profiles.

To compare the hydrolytic profiles of doripenem, imipenem, and meropenem, enzymes from freeze-thaw lysates were purified from lysates to >90% homogeneity by fast protein liquid chromatography, except for the OXA enzymes (∼50% purity) (15). Proteins were separated on Superdex 100 gel filtration and HiTrap SP cation- and Q anion-exchange columns (GE Healthcare, Piscataway, NJ). Columns and buffers were chosen based on the β-lactamase isoelectric point. Purity was assessed on NuPAGE 10% BT gels stained with colloidal blue (Invitrogen, Carlsbad, CA), and for protein quantitation, we used the Micro BCA assay (Pierce, Rockford, IL).

Initial hydrolysis rates were measured at 25°C in 50 mM phosphate buffer (pH 7.0) by using a Shimadzu UV-1601 spectrophotometer (15, 18). Reactions with MBLs contained 50 μM ZnCl₂, and those with OXA enzymes contained 10 mM NaHCO₃. K_m and V_max calculations used the Hanes plot. For Pseudomonas aeruginosa AmpC and CMY-2, hydrolysis was too slow to determine K_m; a 50% inhibitory concentration obtained graphically using nitrocefin as a substrate was used to determine apparent K_i values (5). Extinction coefficients were as follows: Δε₂₉₅ = 11,500 M⁻¹ cm⁻¹ for imipenem, Δε₂₉₇ = 10,940 M⁻¹ cm⁻¹ for meropenem, and Δε₂₉₇ = 11,460 M⁻¹ cm⁻¹ for doripenem. In general, substrates were tested on at least two separate days with variations of ≤20% of the average value reported in the tables.

Hydrolysis rates, K_m values, and hydrolytic efficiencies of noncarbapenemases are shown in Table 2. The class A β-lactamases from gram-negative bacteria included TEM-1 and SHV-1 (broad-spectrum β-lactamases), and CTX-M-15, K1, and TEM-26 (ESBLs). The class C β-lactamases were represented by chromosomal AmpC enzymes from Enterobacter cloacae and P. aeruginosa and a plasmid-mediated AmpC enzyme, CMY-2. The OXA-10 (PSE-2) enzyme represented a class D noncarbapenemase. OXA-10 and OXA-23 were obtained at ∼50% purity; therefore, V_max values are listed in Table 2 instead of k_cat values. The k_cat values and hydrolytic efficiencies for the carbapenems were generally at least 2 orders of magnitude lower than those for the standard substrates, benzylpenicillin or cephaloridine. K_m or K_i values were in the low micromolar or nanomolar range for CMY-2 and the AmpC enzyme of P. aeruginosa, indicating high carbapenem affinity of AmpC type β-lactamases. However, hydrolysis was very inefficient due to the low k_cat values. The k_cat values for imipenem ranged from 0.002 s⁻¹ for the TEM-26 enzyme to 0.2 s⁻¹ for CTX-M-15. Regardless of β-lactamase class, doripenem and meropenem k_cat values were often at least 10-fold lower than imipenem hydrolysis rates for the enzymes exhibiting measurable hydrolysis (Table 2).

Table 3 shows the kinetic parameters for enzymes with carbapenemase activity. Serine carbapenemases of functional group 2f were represented by the prevalent KPC-2 enzyme and the uncommon SME-3 β-lactamase. Both enzymes demonstrated similar k_cat values for doripenem and meropenem, ranging from 0.55 s⁻¹ to 3.6 s⁻¹, while the imipenem k_cat values were at least ninefold higher for KPC-2 and 100-fold higher for SME-3. Hydrolytic efficiencies for all three carbapenems were similar for the KPC-2 enzyme, whereas the hydrolytic efficiencies for doripenem and meropenem with the SME-3 carbapenemase were 10 to 20% of the rates observed with imipenem and cephaloridine.

The class B MBLs were represented by IMP-1, VIM-2, and SPM-1. The IMP-1 and SPM-1 MBLs demonstrated robust hydrolysis of the carbapenems, with k_cat values ranging from 26 to 190 s⁻¹. Hydrolysis by VIM-2 was slower, with carbapenem k_cat values of 2 to 20 s⁻¹. For both IMP and VIM, imipenem was the most labile substrate, but the pattern was reversed for SPM, for which doripenem and meropenem were hydrolyzed approximately fourfold faster than was imipenem. Hydrolysis of meropenem was previously observed to be faster than that of imipenem in kinetic studies of SPM-1 (11). The IMP-1 enzyme demonstrated the most efficient hydrolysis, with k_cat/K_m values that were 2.5- to 11-fold higher than the values obtained for SPM-1 and 18- to 128-fold higher than those obtained for VIM-2.

The OXA-type carbapenemases were represented by OXA-23, one of the most prevalent OXA carbapenemases found in Acinetobacter baumannii (19). These enzymes typically demonstrate weak hydrolysis of imipenem and meropenem (20); supporting these observations, OXA-23 had very low relative V_max values for all the carbapenems compared to those for penicillin, the preferred substrate. Although biphasic, or burst, kinetics are published for some substrates with oxacillinases (9), only imipenem demonstrated this behavior in our study.

In summary, doripenem was stable to hydrolysis by many β-lactamases of classes A, C, and D, including ESBLs, with high affinity demonstrated for the AmpC cephalosporinases. As expected, the class B metallo-β-lactamases, and the serine carbapenemases of classes A and D, demonstrated hydrolysis of all the carbapenems tested. For these enzymes, with the exception of SPM-1, doripenem and meropenem were hydrolyzed more slowly than was imipenem. These data are consistent with the potent microbiological activity observed with doripenem, especially against most gram-negative aerobic pathogens.

• American Society for Microbiology