Summary should be limited to 1-2 pages in length, double spaced, and 12pt Times New Roman or Calibri font.

Journal of Microbiological Methods 83 (2010) 149–152

Contents lists available at ScienceDirect

Journal of Microbiological Methods

journal homepage: www.elsevier.com/locate/jmicmeth

Improved performance of the modified Hodge test with MacConkey agar for screening carbapenemase-producing Gram-negative bacilli

Kyungwon Lee a, Chang Ki Kim a, Dongeun Yong a, Seok Hoon Jeong a, Jong Hwa Yum a,1, Young Hee Seo a, Jean-Denis Docquier b, Yunsop Chong a,⁎ a Department of Laboratory Medicine, Yonsei University College of Medicine, Seoul, South Korea b Department of Molecular Biology, Section of Microbiology, University of Siena, Siena, Italy

⁎ Corresponding author. Department of Laboratory College of Medicine, 250 Seongsanro, Seodaemungu, Se +82 2 2228 2446; fax: +82 2 313 0908.

E-mail address: [email protected] (Y. Chong). 1 Present address: Department of Clinical Laboratory

Busan, South Korea.

0167-7012/$ – see front matter © 2010 Elsevier B.V. Al doi:10.1016/j.mimet.2010.08.010

a b s t r a c t

a r t i c l e i n f o

Article history: Received 22 April 2010 Received in revised form 3 August 2010 Accepted 9 August 2010 Available online 27 August 2010

Keywords: Modified Hodge test Carbapenemase screening MacConkey agar

The detection of carbapenemases in Gram-negative bacilli is important for optimal patient treatment and to control spread of the resistance. The modified Hodge test can detect carbapenemase-producing Gram- negative bacilli. In this study, we compared the performance of MacConkey agar and Mueller-Hinton agar for metallo-β-lactamase (MBL) and OXA carbapenemase screening. Overall, the performance of Hodge test was better with MacConkey agar due to enhanced release of β-lactamases from the cells in the presence of bile compounds. Concomitant use of the modified Hodge test could resolve most of the problems with uncertain double-disk synergy tests in MBL detection.

Medicine, Yonsei University oul 120-752, South Korea. Tel.:

Science, Dong-eui University,

l rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

The detection of metallo-β-lactamases (MBL) and OXA-type carbapenemases in clinical isolates of Gram-negative bacilli (GNB) is important, because most of them are resistant to virtually all β-lactams and because the resistance can be transferred horizontally (Walsh et al., 2005; Walther-Rasmussen and Høiby, 2006). However, the CLSI publication (2010) does not contain any guideline for phenotypic detection of these carbapenemases. For detection of MBL-producing GNB, inhibitor based tests, i.e., the E-test MBL (Lee et al., 2005), double- disk synergy (DDS) tests (Arakawa et al., 2000; Lee et al., 2003), and a combined-disk test (Yong et al., 2002), were reported, but all of these tests showed either low sensitivity or specificity (Lee et al., 2003; Scoulica et al., 2004; Yan et al., 2004; Franklin et al., 2006). A recent evaluation confirmed that the performance of the DDS test depends on β-lactam substrate, MBL inhibitor, and bacterial species tested (Picao et al., 2008). Also a DDS test with ceftazidime disk may fail to detect MBLs when the isolates are AmpC hyperproducers (Lee et al., 2003). Franklin et al. (2006) used a DDS test together with a combined-disk test, to improve performance, and reported a sensitivity of 100% in detecting 84 MBL-producing isolates, but a specificity of 98% with 52 MBL-negative isolates. These studies indicate that development of novel phenotypic tests not dependent on MBL inhibitors is needed.

Hodge et al. (1978) made a simple modification of the cloverleaf test (Østravik and Ødegaard, 1971) to detect penicillinase-producing gonococci by inoculating a plate with indicator organism, Escherichia coli, in a manner similar to the current disk diffusion susceptibility testing. We modified the method of Hodge et al. to screen MBL- producing GNB (Lee et al., 2001). Anderson et al. (2007) evaluated the test and reported a 100% sensitivity and specificity in detecting a class A carbapenemases, KPC, in Enterobacteriaceae isolates. Our original modified Hodge test using Mueller-Hinton agar (MHA) occasionally gave negative or equivocal results in detecting MBL-producing Pseudomonas spp. and Acinetobacter spp. Galani et al. (2008) reported that 4.2% of MBL-producing Enterobacteriaceae isolates gave false- negative Hodge test. Although our preliminary study showed that the use of MacConkey agar could improve the Hodge test, the effect on screening MBL-producing GNB and on OXA-type carbapenemase- producing Acinetobacter has not been investigated.

The aims of this study were to identify the component of MacConkey agar and other factors that improve the Hodge test performance in the screening of MBL and OXA-type carbapenemases, and to determine whether the improvement is due to increased production of β-lactamases or enhanced release of them.

2. Materials and methods

2.1. Bacterial strains

Clinical isolates of GNB producing MBLs, OXA carbapenemases, and other β-lactamases were used (Table 1). One strain each of Serratia marcescens with SME-1, and Pseudomonas aeruginosa with VIM-1 were

Table 1 Influence of media, carbapenem disks and inoculum density on the modified Hodge test for 19 carbapenemase-producing isolatesa.

Disks (no. of isolates tested)

Inoculum (McFarland)

Mean distortion in mm (test) Comparison and p by Student's t-test

MacConkey MHA Medium Disk Inoculum

Imipenem (18) No. 0.5 7.2 (A) 2.7 (E) A vs. E=0.024 A vs. C=0.384 A vs. B=0.752 No. 0.05 4.6 (B) 3.1 (F) B vs. F=0.027 B vs. D=0.021 C vs. D=0.249

Ertapenem (18) No. 0.5 5.0 (C) 4.2 (G) C vs. G=0.254 E vs. G=0.042 E vs. F=0.187 No. 0.05 5.6 (D) 5.1 (H) D vs. H=0.375 F vs. H=0.007 G vs. H=0.027

MHA = Mueller-Hinton agar. a Number of isolates is shown in parenthesis. P. aeruginosa w/VIM-1 (1 isolate), P. aeruginosa w/VIM-2 (1), K. pneumoniae w/VIM-2 (2), P. aeruginosa w/IMP-1 (3), Acinetobacter

genomospecies 13TU w/IMP-1 (1), P. aeruginosa w/IMP-6 (2), A. baumannii w/SIM-1 (1), K. pneumoniae w/KPC-2 (1), S. marcescens w/SME-1 (1), Acinetobacter genomospecies 3 w/ OXA-23 (1), A. calcoaceticus w/ OXA-58 (1), A. baylyi w/OXA-72 (1), A. baumannii w/OXA-109 (1), and A. baumannii w/OXA-115 (1).

150 K. Lee et al. / Journal of Microbiological Methods 83 (2010) 149–152

kindly provided by David M. Livermore. E. coli BL21(DE3) transfor- mants carrying pET9a with PCR-cloned blaTEM-106 or blaTEM-107 gene were used to compare β-lactamase production and release into the culture supernatant. E. coli ATCC 25922 was used as an indicator organism for the Hodge test.

2.2. Media and chemicals

Agar plates were prepared from Mueller-Hinton II agar and MacConkey agar (Becton Dickinson, Sparks, MD). Oxgall was obtained from Difco (Detroit, MI). The main ingredients used to prepare ZYP- 5052 medium (Studier, 2005; [email protected]) were Tryptone and Yeast extract (Becton Dickinson). Ampicillin (Sigma Chemical, St. Louis, MO) and kanamycin (Dong-A Pharmaceutical, Seoul, Korea) were added to ZYP-5052 medium at 50 μg/mL. Ampicillin and imipenem (Merck Sharp & Dohme, Rahway, NJ) were used for spectrophotometric determination of hydrolysis.

2.3. Modified Hodge test

An overnight culture of indicator organism E. coli ATCC 25922 was adjusted to a turbidity of McFarland No. 0.5 and No. 0.05 and these were used to swab inoculate the surface of the agar plates. After drying the surface, test organisms were heavily streaked from the center to the periphery of the plate using an inoculating loop and a 10- μg imipenem or ertapenem disk (Becton Dickinson) was placed at the center, and incubated overnight. The Hodge test is interpreted as positive by the presence of distortion of the inhibition zone (Hodge et al., 1978), but in this study, for the purpose of comparison, we measured the distortion of inhibition zone of the indicator organism along the streaked growth of a test organism.

2.4. Cell density and β-lactamase activity

After an overnight shaking culture in ZYP-5052 medium, bacterial cell density was measured at 600 nm, and the protein content, and ampicillin- and imipenem-hydrolyzing activities of the culture supernatant and cell sonicate were determined using a spectropho- tometer (Shimadzu, Tokyo, Japan). One unit of enzyme activity was defined as the amount of enzyme which hydrolyzed 1 μmol of a substrate per minute per mg of protein at 30 °C in 0.05 M phosphate buffer (pH. 7.0).

3. Results and discussion

3.1. Performance of MacConkey agar and MHA

Only one brand of MacConkey agar (Becton Dickinson) was used in this study, because one local brand of plated medium performed poorly. Therefore, we recommend testing a new brand of MacConkey agar prior to use, and also at each time of routine testing. As a control strain, an OXA-type carbapenemase-producing A. baumannii isolate is

suitable, because it gives smaller distortion zone than an MBL- producing isolate. Table 1 shows the Hodge test results obtained without repetition by a technologist who was unaware of carbape- nemases production. The mean distortions of the inhibition zones obtained for 18 isolates ranged from 2.7 mm to 7.2 mm depending on the test conditions, and in general the distortions were larger and more distinct with MacConkey agar (Table 1; Fig. 1). When the results were analyzed separately according to enzyme type, the mean distortions observed on MacConkey agar and MHA were 4.9 mm and 3.7 mm, respectively, for 11 MBL-producing isolates, and 4.1 mm and 1.9 mm for five OXA-type carbapenemase-producing isolates (data not shown). The single available isolates of KPC-2-producing K. pneumoniae, and SME-1-producing S. marcescens showed mean distortions of 6 mm with MacConkey agar and 9 mm with MHA. Although further study with a large number of KPC-producing isolates is required, it has previously been reported that the Hodge test with MHA showed 100% sensitivity and specificity for detection of KPC in 31 isolates of Enterobacteriaceae (Anderson et al., 2007).

In a study, OXA-type carbapenemase genes were detected by PCR in 105 of 107 (98%) Hodge test-positive and 2 of 9 (22%) Hodge test- negative isolates using MacConkey agar (Lee et al., 2009). This highly specific Hodge test result for OXA carbapenemase may be due to the scarcity of Acinetobacter spp. with other imipenem resistance mechanisms at the time of testing, and further study in different settings is required. In Acinetobacter spp. smaller distortion of inhibition zone could suggest production of OXA carbapenemase rather than MBL.

In general, the performance of imipenem disks was inferior to that of ertapenem disks, particularly when MHA was used (Table 1; Fig. 1). Indeed, the CLSI (2010) recommends use of an ertapenem or meropenem disk for detection of KPC enzyme.

The CLSI (2010) recommends the use of a 1:10 dilution of McFarland No. 0.5 turbidity of indicator organism for detection of KPC β-lactamase, as we initially suggested for detection of MBL (Lee et al., 2001). However, it was shown in our present study that the lower density did not consistently improve the performance of the Hodge test (Table 1). Therefore, we considered that if MacConkey agar is used the tedious dilution step could be eliminated.

3.2. Factors improving the Hodge test

Addition to blood agar base of lactose or starch, which are components in MacConkey agar and MHA, respectively, did not improve the Hodge test results (data not shown). As MacConkey broth (Difco) contains 5 mg/mL Oxgall, we next tested MHA supplemented with this compound (Table 2). All seven MBL- producing isolates showed visually positive results on Oxgall-added MHA. Distinct distortion of ≥3 mm was observed in four isolates (mean 3.3 mm) when tested with 2 mg/mL Oxgall, but in all 7 isolates (mean 8.6 mm) with 20 mg/mL. It was suggested that the cloverleaf test can detect cell-bound enzymes due to cell lysis associated with very heavy growth (Reig and Baquero, 1984). The positive effect of

Table 3 Effects of Oxgall in the medium on optical density of culture, and on protein concentration and β-lactam hydrolytic activity of culture supernatant.

Test and species Cell density (OD at 600 nm)

Protein concentration (mg/mL)

Ampicillin hydrolysis (U/mL)a

ZYP-5052 with Oxgall

ZYP-5052 with Oxgall

ZYP-5052 with Oxgall

No Yes No Yes No Yes

Oxgall 0 mg/mL vs. 5 mg/mL P. aeruginosa with VIM-2 6.6 4.6 0.036 0.387 0.047b 0.060b

A. baumannii with OXA-23 5.5 4.5 0.023 0.230 0.016c 0.120c

E. coli BL21(DE3) with blaTEM-106

5.4 6.0 0.041 0.103 0.371 1.513

IPM disk (A) ETP disk (A) ETP disk (B) MAC MHAMAC MHA MAC MHA

Fig. 1. Comparison of performance of MacConkey agar (MAC) vs. Mueller-Hinton agar (MHA), and an imipenem (IPM) disk vs. an ertapenem (ETP) disk in the modified Hodge test. The inocula were McFarland No. 0.5 (A) and No. 0.05 (B). This figure shows the test only for recently-emerged carbapenemase-producing gram-negative bacilli in Korea. 1 and 2, P. aeruginosa w/IMP-6; 3, A. baumannii w/SIM-1; 4, A. baumannii w/OXA-23. The distortions were larger and more distinct with MAC and with an ETP disk, whereas an isolate of A. baumannii w/OXA-23 showed a negative result with MHA.

151K. Lee et al. / Journal of Microbiological Methods 83 (2010) 149–152

Oxgall on the Hodge test suggested that the compound either induce lysis or increase permeability of the cells.

3.3. Specificity of the Hodge test

In our present study on the specificity of the Hodge test with imipenem disk and MacConkey agar, none of the following strains with various non-carbapenemase enzymes showed positive results: E. coli with TEM-1 and TEM-8 (n=2), and Salmonella enterica with TEM- 52 (n=1); K. pneumoniae with SHV-3 and SHV-18 (n=2); E. coli with CTX-M-14 (n=1); E. coli with CMY-1 (n=1), and K. pneumoniae with CMY-10 (n=1) (data not shown). This high specificity was in contrast to a report that showed false-positive results with CTX-M- producing K. pneumoniae isolates (Carvalhaes et al., 2010).

3.4. Effect of Oxgall on β-lactamase release

We determined the effect of adding Oxgall to ZYP-5052 medium for one each of VIM-2-, SIM-1, OXA-23- and CTX-M-37-producing isolates, and E. coli transformants with pET9a plasmid carrying PCR- cloned blaTEM-106 or blaTEM-107 (Table 3). ZYP-5052 medium can autoinduce β-lactamase production by transformed E. coli BL21(DE3) which carry β-lactamase gene-inserted pET-9a plasmid (Studier, 2005). After incubation at 35 °C with shaking at 240 rpm for 23 h, the mean cell density at 600 nm was slightly lower for medium contain- ing 5 mg/mL Oxgall than for control medium (OD 6.6 vs. 6.2), but much higher values were obtained for mean protein content (0.221 mg/mL vs. 0.036 mg/mL), and ampicillin hydrolyzing activity (0.464 U/mL vs. 0.110 U/mL). Imipenem hydrolyzing activity of the culture supernatant of P. aeruginosa with VIM-2 was 0.103 U/mL with 5 mg/mL Oxgall compared with 0.011 U/mL without Oxgall. Imipe- nem hydrolyzing activity was not detectable in the culture superna-

Table 2 Effects of addition of Oxgall to Mueller-Hinton agar (MHA) on the imipenem disk Hodge test.

Imipenem-resistant organism (no. of isolates tested)

Mean distortion in mm (no. of isolates Hodge test positivea) with

Oxgall

None 2 mg/mL 20 mg/mL

All MBL-producing isolates (7) 1.7 b (2) 3.3b (4) 8.6b (7) P. aeruginosa with VIM-2 (4) 1.8 (1) 3.0 (2) 6.8 (4) Acinetobacter sp. with IMP-1 (3) 1.7 (1) 3.7 (2) 11.0 (3)

Non-MBL-producing P. aeruginosa (1) 0 (0) 0 (0) 0 (0)

a ≥3 mm was interpreted as positive. b p by Students t-test: MHA vs. MHA with Oxgall 2 mg/mL 0.1012; MHA, vs. MHA

with Oxgall 20 mg/mL 0.0016.

tant of A. baumannii with OXA-23, possibly due to the typically weak activity of the enzyme, although addition of Oxgall did increase protein concentration and ampicillin hydrolytic activity.

The slightly lower cell density of the culture with Oxgall was probably not due to inhibitory activity of the compound, as the Oxgall MICs determined by the agar dilution method were ≥8 mg/mL for various GNB (data not shown). Cell sonicates of E. coli BL21(DE3) with blaTEM-106 grown in ZYP-5052 and ZYP-5052 plus 5 mg/mL Oxgall showed similar protein content, 0.403 mg/mL and 0.461 mg/mL, respectively, and similar ampicillin hydrolyzing activities, 6.712 U/ mL and 7.482 U/mL, respectively, indicating that Oxgall did not increase production of β-lactamase (data not shown). To test the effects of Oxgall and sodium deoxycholate on non-growing bacteria, overnight cultures of P. aeruginosa with VIM-2, and Acinetobacter sp. with IMP-1 were adjusted to ca OD 2 at 600 nm, and then the compounds were added and observed after 24 h at 4 °C. An increased Oxgall concentration resulted in slightly lower ODs and much higher

E. coli BL21(DE3) with blaTEM-107

8.9 9.5 0.043 0.163 0.030 0.342

Mean (n=4) 6.6 6.2 0.036 0.221 0.110 0.464 Oxgall 0 mg/mL vs. 2 mg/mL

Acinetobacter sp. with SIM-1

6.4 7.0 0.040 0.120 1.920 3.490

E. cloacae with CTX-M-37 13.3 13.9 0.051 0.428 0.470 1.480 E. coli BL21(DE3) with blaTEM-107

5.3 6.0 0.070 0.109 NDd 0.084

Mean (n=3) 8.3 9.0 0.053 0.219 0.796 1.685 All mean (n=7) 7.34e 7.35e 0.043c 0.220c 0.4651c 0.987c

a One unit of enzyme activity was defined as the amount of enzyme which hydrolyzed 1 μmol of a substrate per minute per mg of protein at 30 °C.

b Imipenem hydrolyzing activities of culture supernatants were 0.103 U/mL with 5 mg/mL Oxgall, and 0.011 U/mL without Oxgall.

c Imipenem hydrolyzing activities were not detectable. d ND, not detectable. e p by Students t-test (ZYP-5052 vs. ZYP-5052 with Oxgall): optical density 0.9730,

protein concentration 0.0152, ampicillin hydrolysis 0.0306.

VIM-2

IMP-1

VIM-2

IMP-1

0.00

0.05

0.10

0.15

0.20

P ro

te in

( m

g /m

l)

1.5

1.6

1.7

1.8

1.9

2.0

0 1 2 4 8 0.5 1

0 1 2 4 8 0.5 1

O D

Oxgall (mg/ml) SD (mg/ml)

A

B

Fig. 2. Effect of Oxgall and sodium deoxycholate (SD) on non-growing P. aeruginosa with VIM-2, and Acinetobacter sp. with IMP-1. Overnight cultures in ZYP-5052 medium were adjusted to an OD ca 2 at 600 nm, and then Oxgall and SD were added and observed after 24 h at 4 °C. Higher Oxgall concentrations resulted in slightly lower OD (A), and much higher protein content (B). These tendencies were more pronounced with the addition of SD.

152 K. Lee et al. / Journal of Microbiological Methods 83 (2010) 149–152

protein content (Fig. 2), suggesting only partial lysis, but much increased permeability of the cells. It was reported that it is possible to release β-lactamase using an E. coli mutant with a plasmid carrying the tac promoter (Georgiou et al., 1988).

Although the increase in protein content and ampicillin hydrolyz- ing activity was less pronounced with 2 mg/mL Oxgall, we were able to precipitate protein from the culture supernatant and purify TEM- 106 and TEM-107 β-lactamases by chromatography using an AKTA purifier (Amersham Pharmacia Biotech Uppsala, Sweden). Further study is required to determine the effectiveness of Oxgall in the release of other β-lactamases.

In summary, for the screening of MBL- and OXA carbapenemase- producing GNB, the modified Hodge test performed better with MacConkey agar than with MHA, and with an ertapenem disk than with an imipenem disk when MHA was used, whereas the inoculum densities of the indicator organism, McFarland 0.5 and one tenth dilution of it, showed similar results. Bile compound in MacConkey agar was found to improve performance of the Hodge test. Concomitant use of the modified Hodge test and a DDS test could resolve most of the problems with uncertain DDS test results in detecting MBL.

Acknowledgement

We are grateful to Ms. M. S. Kim, Clinical Microbiology Laboratory, Severance Hospital, Seoul, Korea for isolating and identifying imipenem-resistant Gram-negative bacilli and for technical assistance.

References

Anderson, K.F., Lonsway, D.R., Rasheed, J.K., Biddle, J., Jensen, B., McDougal, L.K., Carey, R.B., Thompson, A., Stocker, S., Limbago, B., Patel, J.B., 2007. Evaluation of methods to identify the Klebsiella pneumoniae carbapenemase in Enterobacteriaceae. J. Clin. Microbiol. 45, 2723–2725.

Arakawa, Y., Shibata, N., Shibayama, K., Kurokawa, H., Yagi, T., Fujiwara, H., Goto, M., 2000. Convenient test for screening metallo-β-lactamase-producing gram-nega- tive bacteria by using thiol compounds. J. Clin. Microbiol. 38, 40–43.

Carvalhaes, C.G., Picão, R.C., Nicoletti, A.G., Xavier, D.E., Gales, A.C., 2010. Cloverleaf test (modified Hodge test) for detecting carbapenemase production in Klebsiella pneumoniae: be aware of false positive results. J. Antimicrob. Chemother. 65, 249–251.

Clinical and Laboratory Standards Institute, 2010. Performance Standards for Antimicrobial Susceptibility Testing; Twentieth Informational Supplement. M100-S20. Clinical and Laboratory Standards Institute, Wayne, PA.

Franklin, C., Liolios, L., Peleg, A.Y., 2006. Phenotypic detection of carbapenem- susceptible metallo-β-lactamase-producing gram-negative bacilli in the clinical laboratory. J. Clin. Microbiol. 44, 3139–3144.

Galani, I., Rekatsina, P.D., Hatzaki, D., Plachouras, D., Souli, M., Giamarellou, H., 2008. Evaluation of different laboratory tests for the detection of metallo-β-lactamase production in Enterobacteriaceae. J. Antimicrob. Chemother. 61, 548–553.

Georgiou, G., Shuler, M.L., Wilson, D.B., 1988. Release of periplasmic enzymes and other physiological effects of β-lactamase overproduction in Escherichia coli. Biotechnol. Bioeng. 32, 741–748.

Hodge, W., Ciak, J., Tramont, E.C., 1978. Simple method for detection of penicillinase- producing Neisseria gonorrhoeae. J. Clin. Microbiol. 7, 102–103.

Lee, K., Chong, Y., Shin, H.B., Kim, Y.A., Yong, D., Yum, J.H., 2001. Modified Hodge test and EDTA-disk synergy tests to screen metallo-β-lactamase-producing strains of Pseudomonas and Acinetobacter species. Clin. Microbiol. Infect. 7, 88–91.

Lee, K., Lim, Y.S., Yong, D., Yum, J.H., Chong, Y., 2003. Evaluation of the Hodge test and the imipenem-EDTA double-disk synergy test for differentiating metallo-β- lactamase-producing isolates of Pseudomonas spp. and Acinetobacter spp. J. Clin. Microbiol. 41, 4623–4629.

Lee, K., Yong, D., Yum, J.H., Lim, Y.S., Bolmström, A., Qwärnström, A., Karlsson, A., Chong, Y., 2005. Evaluation of Etest MBL for detection of blaIMP-1 and blaVIM-2 allele- positive clinical isolates of Pseudomonas spp. and Acinetobacter spp. J. Clin. Microbiol. 43, 942–944.

Lee, K., Kim, M.-N., Choi, T.Y., Cho, S.E., Lee, S., Whang, D.H., Yong, D., Chong, Y., Woodford, N., Livermore, D.M., the KONSAR Group, 2009. Wide dissemination of OXA-type carbapenemases in clinical Acinetobacter spp. isolates from South Korea. Int. J. Antimicrob. Agents 33, 520–524.

Østravik, I., Ødegaard, K., 1971. A simple test for penicillinase production in Staphylococcus aureus. Acta Pathol. Microbiol. Scand. B 79, 855–856.

Picao, R.C., Andrade, S.S., Nicoletti, A.G., Campana, E.H., Moraes, G.C., Mendes, R.E., Gales, A.C., 2008. Metallo-β-lactamase detection: comparative evaluation of double-disk synergy versus combined disk tests for IMP-, GIM-, SIM-, SPM-, or VIM-producing isolates. J. Clin. Microbiol. 46, 2028–2037.

Reig, M., Baquero, F., 1984. Limitations in the use of the clover leaf test in Bacteroides species. Eur. J. Clin. Microbiol. Infect. Dis. 3, 561–563.

Scoulica, E.V., Neonakis, I.K., Gikas, A.I., Tselentis, Y.J., 2004. Spread of blaVIM-1- producing E. coli in a university hospital in Greece. Genetic analysis of the integron carrying the blaVIM-1 metallo-β-lactamase gene. Diagn. Microbiol. Infect. Dis. 48, 167–172.

Studier, F.W., 2005. Protein production by auto-induction in high-density shaking culture. Protein Expr. Purif. 41, 207–234.

Walsh, T.R., Tolman, M.A., Poirel, L., Nordmann, P., 2005. Metallo-β-lactamase: the quiet before the storm? Clin. Microbiol. Rev. 18, 306–325.

Walther-Rasmussen, J., Høiby, N., 2006. OXA-type carbapenemases. J. Antimicrob. Chemother. 57, 373–383.

Yan, J.J., Wu, J.J., Tsai, S.H., Chuang, C.L., 2004. Comparison of the double-disk, combined disk, and Etest methods for detecting metallo-β-lactamases in gram-negative bacilli. Diagn. Microbiol. Infect. Dis. 49, 5–11.

Yong, D., Lee, K., Yum, J.H., Shin, H.B., Rossolini, G.M., Chong, Y., 2002. Imipenem-EDTA disk method for differentiation of metallo-β-lactamase-producing clinical isolates of Pseudomonas spp. and Acinetobacter spp. J. Clin. Microbiol. 40, 3798–3801.