The soil microbiota harbors a diversity of carbapenem-hydrolyzing β-lactamases of potential clinical relevance

(1)

The Soil Microbiota Harbors a Diversity of Carbapenem-Hydrolyzing

␤-Lactamases of Potential Clinical Relevance

Dereje Dadi Gudeta,a_{Valeria Bortolaia,}a_{Greg Amos,}b_{Elizabeth M. H. Wellington,}b_{Kristian K. Brandt,}c_{Laurent Poirel,}d Jesper Boye Nielsen,e_{Henrik Westh,}e,f_{Luca Guardabassi}a

Department of Veterinary Disease Biology, Faculty of Health and Medical Sciences, University of Copenhagen, Frederiksberg, Denmarka_{; School of Life Sciences, University} of Warwick, Coventry, United Kingdomb_{; Department of Plant and Environmental Sciences, Faculty of Science, University of Copenhagen, Frederiksberg, Denmark}c_; Medical and Molecular Microbiology Unit, Department of Medicine, Faculty of Science, University of Fribourg, Fribourg, Switzerlandd_{; Department of Clinical} Microbiology, University of Copenhagen, Hvidovre Hospital, Hvidovre, Denmarke_{; Institute of Clinical Medicine, Faculty of Health and Medical Sciences, University of} Copenhagen, Frederiksberg, Denmarkf

The origin of carbapenem-hydrolyzing metallo-

␤-lactamases (MBLs) acquired by clinical bacteria is largely unknown. We

inves-tigated the frequency, host range, diversity, and functionality of MBLs in the soil microbiota. Twenty-ﬁve soil samples of

differ-ent types and geographical origins were analyzed by antimicrobial selective culture, followed by phenotypic testing and

expres-sion of MBL-encoding genes in Escherichia coli, and whole-genome sequencing of MBL-producing strains was performed.

Carbapenemase activity was detected in 29 bacterial isolates from 13 soil samples, leading to identiﬁcation of seven new MBLs in

presumptive Pedobacter roseus (PEDO-1), Pedobacter borealis (PEDO-2), Pedobacter kyungheensis (PEDO-3),

Chryseobacte-rium piscium (CPS-1), Epilithonimonas tenax (ESP-1), Massilia oculi (MSI-1), and Sphingomonas sp. (SPG-1). Carbapenemase

production was likely an intrinsic feature in Chryseobacterium and Epilithonimonas, as it occurred in reference strains of

differ-ent species within these genera. The amino acid iddiffer-entity to MBLs described in clinical bacteria ranged between 40 and 69%.

Re-markable features of the new MBLs included prophage integration of the encoding gene (PEDO-1), an unusual amino acid

resi-due at a key position for MBL structure and catalysis (CPS-1), and overlap with a putative OXA

␤-lactamase (MSI-1).

Heterologous expression of PEDO-1, CPS-1, and ESP-1in E. coli signiﬁcantly increased the MICs of ampicillin, ceftazidime,

cef-podoxime, cefoxitin, and meropenem. Our study shows that MBL producers are widespread in soil and include four genera that

were previously not known to produce MBLs. The MBLs produced by these bacteria are distantly related to MBLs identiﬁed in

clinical samples but constitute resistance determinants of clinical relevance if acquired by pathogenic bacteria.

S

oil is an important reservoir of antibiotic resistance

determi-nants (

1 ,

2 ). Several studies have shown that speciﬁc antibiotic

resistance genes of high clinical relevance may have originated

from environmental bacteria (

3–6

). However, the origins of

resis-tance genes encoding carbapenem-hydrolyzing metallo-

␤-lacta-mases (MBLs) are largely unknown. MBLs are currently the most

critical

␤-lactamases in clinical settings because they are

insensi-tive to

␤-lactamase inhibitors and confer resistance to

carbapen-ems, a

␤-lactam class of last-resort drugs used for treatment of

severe bacterial infections (

7 ).

Based on amino acid sequences, MBLs are classiﬁed as

molec-ular class B

␤-lactamases and are further divided into B1, B2, and

B3 subclasses (

8 ). Subclass B1 and B3 MBLs bind two zinc ions

(Zn1 and Zn2) in their active sites, employ different metal binding

amino acids (for the Zn2 ligand), and exhibit broad-spectrum

activity (

9 ). Subclass B2 MBLs are mono-Zn enzymes whose

ac-tivity is inhibited upon binding the second Zn (

10 ) and have

strong preferences for carbapenem substrates (

11 ). B1 and B3

en-zymes exist across different bacterial genera, whereas B2 enen-zymes

have been described only in members of the genera Serratia and

Aeromonas (

10 ,

11 ).

The most common MBLs in clinical bacteria are NDM, VIM, and

IMP (

12 ). In addition, new types of MBLs are continuously emerging

in pathogens from unknown reservoirs (

13 ,

14 ). Identiﬁcation of

un-known MBL-encoding genes occurring in the environment is

impor-tant for assessment of the human health risks associated with

envi-ronmental development and transfer of carbapenem resistance (

15 ,

16 ). The aim of this study was to investigate the frequency, host range,

diversity, and functionality of MBLs in soil bacteria. Phenotypic and

genotypic characterization of the MBLs detected in the soil

microbi-ota revealed the presence of an environmental reservoir of new

car-bapenem-hydrolyzing MBLs of potential clinical relevance.

MATERIALS AND METHODS

Soil samples. A total of 25 soil samples recovered from six different

geo-graphical locations (Algeria, United Kingdom, Germany, Denmark, Nor-way, and Spain) and soil uses (agricultural and nonagricultural) were

analyzed, including 10 samples collected in previous studies (17–23) (

Ta-ble 1). Fifteen samples were collected in this study from 0- to 15-cm depth

using sterile plastic bags and stored at ⫺20°C. All soil samples were

thawed at room temperature before processing.

Selective isolation of carbapenem-resistant Gram-negative bacte-ria. A combined approach, including direct plating and selective enrichment

in media with different nutrient contents, was used for isolation of

carbap-Citation Gudeta DD, Bortolaia V, Amos G, Wellington EMH, Brandt KK, Poirel L, Nielsen JB, Westh H, Guardabassi L. 2016. The soil microbiota harbors a diversity of carbapenem-hydrolyzing␤-lactamases of potential clinical relevance. Antimicrob Agents Chemother 60:151–160.doi:10.1128/AAC.01424-15.

Address correspondence to Luca Guardabassi, lg@sund.ku.dk. Published in $QWLPLFURELDO$JHQWVDQG&KHPRWKHUDS\±

which should be cited to refer to this work.

(2)

enem-resistant bacteria. All media (here deﬁned as “selective” media) were

supplemented with 8␮g/ml of vancomycin for inhibition of Gram-positive

bacteria, 100␮g/ml of cycloheximide for inhibition of fungi, and 4 ␮g/ml of

meropenem (MP) for selection of carbapenem-resistant bacteria (24). After

sieving and dispersion by stirring and sonication (25), soil samples were

seri-ally diluted up to 10⫺4in sterile distilled water. These suspensions were used

for direct plating of 200-␮l aliquots per plate on selective brain heart infusion

agar (BHI-A) (nutrient-rich medium) and on selective diluted nutrient broth agar (DNB-A) optimized for growth of taxonomically diverse oligotrophic

soil bacteria (25). In parallel, selective enrichment of meropenem-resistant

bacteria was performed by adding aliquots of 1 ml undiluted soil bacterial suspension separately to 99 ml of selective brain heart infusion broth (BHI-B) and 99 ml of selective diluted nutrient broth (DNB-B). After 48 h of incuba-tion at 25°C under shaking at 150 rpm, the enriched cultures were serially

diluted up to 10⫺4, and each dilution was plated as described above. All the

agar plates were incubated at 25°C and read daily for 2 weeks. Subsequently, all presumptive carbapenem-resistant isolates displaying unique colony mor-phologies (relative to size, surface texture, color, and margins) and different microscopic appearances (relative to shape and size) were subcultured on

selective BHI-A or DNB-A prior to storage in 15% glycerol at⫺80°C.

Media and antibiotics. Media and antibiotics were purchased from

Difco (Le Pont-de-Claix, France) and Sigma-Aldrich (Steinheim, Ger-many), respectively.

Phenotypic and genotypic characterization of carbapenem-resis-tant bacteria. The modiﬁed version of the CarbaNP test proposed by

Dortet et al. (26) was used for detection of carbapenemase activity in

presumptive carbapenem-resistant isolates from soil. Isolates not lysing in

Tris-HCl buffer (Thermo Scientiﬁc, Rockford, IL, USA) were sonicated

(two cycles of 30 s at 40% amplitude) in 200␮l of Tris-HCl lysis buffer

using a Sonopuls Ultrasonic homogenizer (Bandelin Electronic GmbH & Co. KG). Imipenem-hydrolyzing isolates were identiﬁed by 16S rRNA gene sequencing using universal primers (see Table S1 in the supplemen-tal material). Isolates were assigned to bacterial species based on the 97% cutoff value of nucleotide identity proposed by Stackebrandt and Goebel

(27). MBL phenotypic detection was done using MBL MP/MPI Etest

strips (bioMérieux, Marcy I’Etoile, France) containing MP at one extrem-ity and MP plus EDTA (MPI) at the other extremextrem-ity and under conditions reported below. Synergy between meropenem and EDTA was considered indicative of MBL production.

Shotgun cloning. Identiﬁcation of MBL-encoding genes was

at-tempted for all the isolates displaying carbapenemase activity, except those belonging to species or genera in which MBLs have been

de-scribed previously (28). Genomic DNA (Easy-DNA kit; Invitrogen,

Carlsbad, CA, USA) from the MBL producers was used for shotgun cloning in Escherichia coli TOP10 (Invitrogen, Carlsbad, CA, USA) as

described previously (28). We used the high-copy-number vector

pZE21MCS (3) or the low-copy-number vector pCF430 (29) to maximize

the chance for cloning efﬁciency (pCF430 was used when cloning with

pZE21MCS did not yield any MBL-producing clone). A total of 2␮l of

ligation mixture was transformed into One Shot TOP10 Electrocomp E.

coli (Invitrogen, Carlsbad, CA, USA) with a Gene Pulser II (Bio-Rad, CA,

USA) according to the manufacturer’s guidelines. Libraries were screened on Luria-Bertani agar (Difco, Le Pont-de-Claix, France) plates containing

30␮g/ml of amoxicillin and 5 ␮g/ml of tetracycline (pCF430 resistance

TABLE 1 Description of soil samples and metallo-␤-lactamase-producing isolates analyzed in this study

Soil

identiﬁer Country Soil type Collection site Reference Metallo-␤-lactamase-producing isolate(s)a

1 Denmark Nonagricultural Amagerfælled This study ND

2 Denmark Nonagricultural Amagerfælled This study S. maltophilia (n⫽ 3)

P. roseus strain SI-33 (n⫽ 1; PEDO-1)

3 Denmark Agricultural Copenhagen (unfertilized) 17 S. maltophilia (n⫽ 3)

4 Denmark Nonagricultural Hygum (control site) 18 J. lividum (n⫽ 1)

5 Denmark Agricultural Holeby 19 ND

6 Denmark Agricultural Lundgaard 20 ND

7 Denmark Agricultural Højgaard (Askov) 20 ND

8 UK Nonagricultural Warwickshire (Stockton) 21 C. piscium strain Stok-1(n⫽ 1; CPS-1); E. tenax

strain Stok-2 (n⫽ 1; ESP-1)

P. kyungheensis strain Stok-3 (n⫽ 1; PEDO-3)

9 Germany Agricultural Dossenheim (control site) 23 S. maltophilia (n⫽ 1)

10 Germany Agricultural Dossenheim (control site) 23 ND

11 Germany Agricultural Dossenheim (plantomycin treated) 23 ND

12 Algeria Nonagricultural Djelfa 22 ND

13 Spain Agricultural Abanilla This study ND

14 Spain Agricultural Abanilla This study M. oculi strain SB1-3 (n⫽ 1, MSI-1)

15 Spain Agricultural Abanilla This study S. maltophilia (n⫽ 1)

23 Norway Nonagricultural Ossian Sars Nature Reserve This study Sphingomonas sp. strain ALS-13. (n⫽ 1; SPG-1)

S. maltophilia (n⫽ 1) J. lividum (n⫽ 1)

P. borealis strain ALS-14 (n⫽ 1; PEDO-2)

24 Norway Nonagricultural Ny-Ålesund This study J. lividum (n⫽ 1)

25 Norway Nonagricultural Ny-Ålesund This study J. lividum (n⫽ 4)

S. maltophilia (n⫽ 3)

a

n, number of isolates; ND, not detected.

(3)

marker) or 50 ␮g/ml of kanamycin (pZE21MCS resistance marker). Amoxicillin was used for screening libraries to avoid false-negative results

(30). Amoxicillin-resistant transformants were screened for

carbapen-emase production by the modiﬁed CarbaNP test as described above. DNA inserts from recombinant plasmids harbored by carbapenemase-produc-ing transformants (recombinant clones) were ampliﬁed by PCR (see Ta-ble S1 in the supplemental material) and Sanger sequenced (Macrogen, Seoul, Republic of Korea).

WGS and bioinformatics analysis. Whole-genome sequencing

(WGS) was performed on soil MBL-producing strains belonging to species for which MBLs have not been described. DNA was extracted using a DNeasy blood and tissue kit (Qiagen, Düsseldorf, Germany), and normalization of the DNA concentration was performed using Qubit (Invitrogen, Paisley, United Kingdom). Libraries were gener-ated using a Nextera XT DNA Sample Preparation kit (Illumina, CA, USA), and sequencing was performed on an Illumina MiSeq benchtop sequencer (Illumina, CA, USA) using a MiSeq Reagent kit v2 (300 cycles) and two 150-bp paired-end reads. Reads from sequencing were

assembled using Velvet v1.0.11 (http://www.ebi.ac.uk/~zerbino

/velvet/) and VelvetOptimiser (http://bioinformatics.net.au/software .velvetoptimiser.shtml). Putative MBL-encoding genes in the WGS

data were detected by ARG-ANNOT (31). The MBL

three-dimen-sional (3D) structure was predicted with the protein-modeling server

Phyre2_{v2.0 (}₃₂_{). In the contigs containing the putative MBL-encoding}

genes, open reading frames (ORFs) were predicted and annotated us-ing CLC main workbench v6.8.2 (Qiagen, Aarhus, Denmark), and each predicted protein was used as a query sequence (BLASTP) to search similar conserved domains in the NCBI sequence database. MBLs displaying maximum identity to the query sequences were used for amino acid alignments (ClustalW2) and phylogenetic-tree

con-struction (FigTree v1.4.2 [http://tree.bio.ed.ac.uk/software/ﬁgtree/]).

Screening of MBL production in reference strains. MBL production

was screened in 17 validly described reference strains displaying close 16S

rRNA gene identity (⬎90%) to selected MBL producers isolated from soil

in order to determine whether it was an intrinsic property within mem-bers of the same genus. Pedobacter, Sphingomonas, Chryseobacterium, and

Massilia strains were purchased from the CCUG strain collection

(Gothenburg, Sweden) and Epilithonimonas strains from the DSMZ strain collection (Braunschweig, Germany). Reference strains were screened by PCR using primers (see Table S1 in the supplemental mate-rial) targeting the MBL-encoding genes identiﬁed through cloning and whole-genome sequencing. The MBL activity and MIC of meropenem were determined as described above. MBL production in reference strains

displaying meropenem MIC values of⬍0.5 ␮g/ml was conﬁrmed by using

CarbaNP test II (33).

MIC determination. The MIC of meropenem was determined by

Etest (bioMérieux, Marcy I’Etoile, France) in MBL-producing soil iso-lates, reference strains, and recombinant clones. The Etest was performed according to the manufacturer’s instructions, with the exception of the incubation temperature, set at 25°C for MBL-producing soil isolates and

reference strains. The MICs of a broader range of␤-lactams were

deter-mined in the recombinant clones by broth microdilution using the Sen-sititre ESBL plate format (Trek Diagnostic Systems, OH, USA) according

to standard procedures (http://www.eucast.org/).

Nucleotide sequence accession numbers. The nucleotide sequences

described in this study were deposited in the GenBank nucleotide

se-quence database with accession numbersKP109675toKP109681and in

the Comprehensive Antibiotic Research Database (CARD [http://aac.asm

.org/content/57/7/3348.full]) to enhance early detection in clinical iso-lates by WGS, which is becoming a routine tool for strain typing in clinical microbiology.

RESULTS

Frequency and classiﬁcation of carbapenemase-producing

bac-teria isolated from soil. Meropenem-resistant Gram-negative

bacteria were isolated from all 25 soil samples analyzed in this

study, leading to a total of 130 isolates. Carbapenemase activity

was detected by CarbaNP test in 29 of these isolates, originating

from 13 soil samples, including both agricultural and

nonagricul-tural soils (

Table 1

). These carbapenemase-producing isolates

be-longed to the genera Pedobacter, Epilithonimonas, Sphingomonas,

Massilia, Chryseobacterium, Janthinobacterium, and

Stenotroph-omonas (

Table 1

). All carbapenemases were classiﬁed as MBLs

based on synergy between meropenem and EDTA. Based on

16S rRNA gene sequence identity, seven strains were assigned

to species or genera that were not known to produce MBLs

prior to this study: Pedobacter roseus strain SI-33 (n

⫽ 1),

Pedo-bacter borealis strain ALS-14 (n

⫽ 1), Pedobacter kyungheensis

strain Stok-3 (n

⫽ 1), Chryseobacterium piscium strain Stok-1 (n ⫽

1), Epilithonimonas tenax strain Stok-2 (n

⫽ 1), Massilia oculi

strain SB1-3 (n

⫽ 1), and Sphingomonas sp. strain ALS-13 (n ⫽ 1)

(Table 2). The Sphingomonas strain was not assigned to any

spe-cies, as the closest species (Sphingomonas hakookensis) displayed

16S rRNA gene sequence identity below the cutoff value of 97%

(95.71%). The remaining 22 strains were related to

Stenotroph-omonas maltophilia and Janthinobacterium lividum, which are

known MBL producers (

28 ).

Genetic context of MBL-encoding genes in soil bacteria .

New MBL-encoding genes were identiﬁed in presumptive C.

pis-cium strain Stok-1 (annotated as bla

CPS-1

), E. tenax strain Stok-2

(bla

ESP-1

), and P. roseus strain SI-33 (bla

PEDO-1

) by using shotgun

cloning in E. coli TOP10 cells, and their genetic context was

fur-ther examined by WGS (see Table S2 in the supplemental

mate-rial). The bla

CPS-1

gene was located between two genes encoding

putative pirin-C protein (Orfa2) and NADPH-dependent FMN

reductase (Orfa3) (

Fig. 1A

). The bla

ESP-1

gene was identiﬁed on an

operon encoding a putative diphosphomevalonate decarboxylase

(Orfb1), two hypothetical proteins (Orfb2 and Orfb3), and a

pu-tative bleomycin resistance protein (Orfb4) (

Fig. 1B

; see Table S3

in the supplemental material). The bla

PEDO-1

gene was located

between 275-bp (Nr1) and 440-bp (Nr2) noncoding DNA

se-quences (

Fig. 1C

). An operon encoding putative phage tail sheath

proteins (Ptp1 and Ptp2), phage tail tube protein (Pttp), and

phage base plate (Pbp) was found upstream of Nr1. Nr2 contained

16-bp inverted repeats located 57 bp downstream of the bla

PEDO-1

stop codon, followed by two genes encoding putative YVTN

␤-propellin repeat-containing proteins (YVTN1 and YVNT2). An

additional operon encoding acriﬂavine resistance protein (AcrB),

resistance nodulation cell division protein (RND), and RND

fam-ily efﬂux transporter MFP subunit (MFP) was identiﬁed

down-stream of the

␤-propellin repeats (

Fig. 1C

; see Table S3 in the

supplemental material).

MBL-encoding genes from presumptive P. borealis strain

ALS-14 (bla

PEDO-2

), P. kyungheensis strain Stok-3 (bla

PEDO-3

),

Sphingomonas sp. strain ALS-13 (bla

SPG-1

), and M. oculi strain

SB1-3(bla

MSI-1

) were identiﬁed by WGS, since the shotgun

clon-ing approach was unsuccessful with these strains. The bla

PEDO-2

gene was ﬂanked by genes encoding a putative nuclease (Orfc3)

and a putative transcriptional regulator (Orfc4) (

Fig. 1D

; see

Ta-ble S3 in the supplemental material), whereas bla

PEDO-3

was

lo-cated between two genes encoding a putative GNAT family

acetyl-transferase (Orfd3 and Orfd2) (

Fig. 1E

). The bla

SPG-1

gene was

ﬂanked by genes encoding a putative phosphopantothenate

syn-thase (Orfe1) and a hypothetical protein (Orfe2) (

Fig. 1F

; see

Ta-ble S3 in the supplemental material). The 3

= end of bla

MSI-1

(4)

lapped the 5

= end of a gene encoding a putative OXA ␤-lactamase,

which was annotated as bla

MSI-OXA

(

Fig. 1G

). Genes encoding a

putative hypothetical protein and a putative argininosuccinate

synthase were detected upstream of bla

MSI-1

and downstream of

bla

MSI-OXA

, respectively (data not shown).

Sequences of the new MBLs identiﬁed in soil bacteria. The

predicted amino acid sequences of the seven new MBLs from soil

were analyzed, and their evolutionary relationship with previously

described MBLs was shown by phylogenetic-tree analysis (

Fig. 2

).

All zinc-coordinating amino acid residues of subclass B3 MBLs

were conserved in PEDO-1, PEDO-2, SPG-1, CPS-1, MSI-1, and

ESP-1, with the exception of a His-to-Gln substitution at position

116 (standard numbering system) in

␤-lactamases PEDO-1,

PEDO-2, CPS-1, and ESP-1, which was previously reported in

GOB

␤-lactamases (

34 ) (

Fig. 3

). Similarly to all GOB

␤-lactamases

(GOB-1 to GOB-18), a Met residue was present at position 221 in

PEDO-1, PEDO-2, and ESP-1

␤-lactamases instead of the Ser

res-idue found at this position in other subclass B3

␤-lactamases (

8 ).

Remarkably, CPS-1

␤-lactamase displayed a previously

unob-served Leu residue at position 221 (

Fig. 3

). Based on multiple

amino acid sequence alignment, PEDO-1, PEDO-2, CPS-1, and

ESP-1

␤-lactamases showed maximum identity (56 to 75%) to

GOB-1

␤-lactamase from Elizabethkingia meningoseptica

(for-merly Chryseobacterium meningosepticum) (

34 ) (see Table S4 in

TABLE 2 Phenotypic and genotypic traits of metallo-␤-lactamase-producing soil strains and closely related reference strains

Soil strain

Meropenem

MIC (␮g/ml)

Reference strain with closest 16S rRNA gene identity to soil straina

Name (accession no.)

16S rRNA identity (%) CarbaNP test result Meropenem MIC (␮g/ml)

P. roseus strain SI-33 8 P. roseus CL-GP80T₍_DQ112353₎ _99.78 _ND _ND

P. kyungheensis THGT17 (JN196132) 98.40 ND ND

P. borealis DSM19626 (EU030687) 98.13 ND ND

P. agri PB92T₍_EF660751₎ _98.00 _Negative ₈

P. heparinus CCUG 12810 (AJ438172) 96.29 Negative 0.125

P. africanus CCUG 39353 (AJ438171) 95.40 Negative 0.5

P. boryungensis CCUG 60024 (HM640986) 95.32 Negative 0.125

P. saltans CCUG 39354 (AJ438173) 91.31 Negative 0.064

P. kyungheensis strain Stok-3 ⬎32 P. kyungheensis THGT17 (JN196132) 98.16 ND ND

P. roseus CL-GP80T₍_DQ112353₎ _97.46 _ND _ND

P. borealis DSM19626 (EU030687) 97.31 ND ND

P. borealis strain ALS-14 ⬎32 P. borealis DSM19626T₍_EU030687₎ _98.94

P. kyungheensis THGT17T₍_JN196132₎ _97.85

P. roseus CL-GP80T₍_DQ112353₎ _97.46

E. tenax strain Stok-2 4 E. tenax DSM 16811 (AF493696) 98.24 Positive 3

E. lactis DSM19921 (EF204460) 98.00 Positive 3

C. piscium strain Stok-1 4 C. piscium CCUG 51923 (AM040439) 98.58 Positive 3

C. balustinum CCUG 13228 (M58771) 98.58 Positive 1.5

C. scophthalmum CCUG 33454 (AJ271009) 98.07 Positive 3

Sphingomonas sp. strain ALS-13 ⬎32 S. hankookensis CCUG 57509 (FJ194436) 95.71 Positive ⬎32

S. aquatilis CCUG 48825 (AF131295) 94.93 Positive 8

S. adhaesiva CCUG 27819 (D16146) 94.57 Positive ND

S. paucimobilis CCUG 6518 (D16144) 94.21 Negative 0.047

S. jaspsi CCUG 53607 (AB264131) 94.20 Positive ND

S. echinoides CCUG 2870 (AB021370) 94.13 Positive ND

M. oculi strain SB1-3 0.5 M. oculi CCUG 43427AT₍_FR773700₎ _98.94

M. timonae CCUG 45783 (U54470) 97.14 Negative 0.25

a_{ND, not determined, as the strain was not available or did not grow on Mueller-Hinton medium.}

(5)

the supplemental material). SPG-1 showed the highest amino acid

identity (47%) to CAU-1

␤-lactamase from Caulobacter crescentus

(

35 ), and MSI-1

␤-lactamase displayed approximately 40%

iden-tity to SMB-1 and AIM-1

␤-lactamases from Serratia marcescens

(

13 ) and Pseudomonas aeruginosa (

36 ), respectively (see Table S4

in the supplemental material). MSI-1 (class B

␤-lactamase

homo-logue) and MSI-OXA (class D

␤-lactamase homologue) displayed

the highest amino acid identity (64 to 66%) to putative domains in

Massilia timonae strain CCUG 45783, where the two domains

were fused to form a putative bifunctional MBL-OXA enzyme.

The two domains in the reference strain were annotated as MT-1

and MT-OXA (

Fig. 1H

). MT-1 displayed a predicted 3D structure

similar to that of SMB-1 from S. marcescens, whereas MT-OXA

and MSI-OXA resembled FUS-1 (OXA-85), a narrow-spectrum

class D

␤-lactamase identiﬁed in Fusobacterium nucleatum subsp.

polymorphum (

37 ). All metal-coordinating amino acid residues of

subclass B1 MBLs were conserved in PEDO-3 (data not shown).

PEDO-3 showed the highest amino acid identity (55%) to

JOHN-1 from Flavobacterium johnsoniae (

38 ) (see Table S5 in the

supplemental material).

Carbapenem resistance in reference strains. Among the 17

purchased reference strains, carbapenemase activity was detected

in ﬁve out of six Sphingomonas species and in all Chryseobacterium

(n

⫽ 3) and Epilithonimonas (n ⫽ 2) species tested. All these

ref-erence strains produced class B

␤-lactamase, as demonstrated by

inhibition of carbapenemase activity by EDTA using CarbaNP test

II. In contrast, no carbapenemase activity was found in the ﬁve

Pedobacter species and the M. timonae reference strain (

Table 2

).

The MICs of meropenem were determined for only 14 strains,

since three Sphingomonas reference strains did not grow on

Mu-eller-Hinton agar. The MICs ranged between 1.5 and

⬎32 in

strains displaying carbapenemase activity, whereas lower MICs

(0.047 to 0.5

␮g/ml) were observed for strains without

carbapen-emase activity, with the exception of Pedobacter agri PB92, for

which the MIC was 8

␮g/ml.

PCR screening of the reference strains using primers targeting

the seven new MBL-encoding genes revealed the presence of (i)

bla

CPS-1

in Chryseobacterium balustinum and Chryseobacterium

scophthalmun but not in the C. piscium reference strain CCUG

51923; (ii) bla

ESP-1

in both E. tenax and Epilithonimonas lactis

ref-erence strains; and (iii) bla

MSI-1

in the M. timonae reference strain

CCUG 45783. The remaining four genes (bla

PEDO-1

, bla

PEDO-2

,

bla

PEDO-3

, and bla

SPG-1

) were not detected in any reference strain.

Heterologous expression of MBL-encoding genes in E. coli.

The effects of heterologous expression of ESP-1, CPS-1, and

PEDO-1 on

␤-lactam susceptibility were assessed by comparing

the MICs of a broad range of

␤-lactams between the three

carbap-enemase-producing recombinant clones and the wild-type E. coli

TOP10. Expression of these MBL-encoding genes conferred

resis-tance to ampicillin, cefoxitin, cefpodoxime, and ceftazidime alone

FIG 1 Genetic organization of metallo-␤-lactamase-encoding genes from presumptive C. piscium strain Stok-1 (A), E. tenax strain Stok-2 (B), P. roseus strain SI-33 (downstream inverted repeats are underlined) (C), P. borealis strain ALS-14 (D), P. kyungheensis strain Stok-3 (E), Sphingomonas sp. strain ALS-13 (F), M.

oculi strain SB1-3 (G), and M. timonae CCUG 45783 with putative bifunctional (PBF)␤-lactamase (H).

(6)

or combined with clavulanic acid and determined an increase in

the MICs of cefazolin, cefotaxime (with or without clavulanic

acid), and meropenem below the clinical breakpoints (

Table 3

).

The MICs of ceftriaxone and piperacilltazobactam were

in-creased by expression of ESP-1, but not by expression of CPS-1

and PEDO-1. None of the MBLs isolated from soil bacteria

af-fected the levels of susceptibility to cefepime.

DISCUSSION

This study shows that carbapenem-hydrolyzing MBLs are

wide-spread in soil, as indicated by the recovery of

carbapenemase-producing bacteria in 52% of the samples tested. Seven new

car-bapenemases were detected in a wide range of bacterial hosts

ranging from Bacteroidetes (Pedobacter, Chryseobacterium, and

Epilithonimonas) to Proteobacteria (Sphingomonas and Massilia).

Our study is the ﬁrst to report MBLs in members of the genera

Pedobacter, Sphingomonas, Epilithonimonas, and Massilia.

Al-though the occurrence of presumptive MBLs has been

sporadi-cally reported in environmental bacteria (

3 ,

28 ,

39 ), previous

studies were based on either metagenomics, which does not allow

identiﬁcation of the host bacteria, or culture on nutrient-rich

me-dia, which are known to inhibit growth of a large proportion of

soil bacteria (

25 ). Moreover, most of the studies did not show the

hydrolytic activity of presumptive MBLs toward carbapenems.

Knowledge of the bacterial hosts, the genetic organization of

MBL-encoding genes, and the activities of the enzymes is useful to

predict the likelihood of horizontal gene transfer to clinically

rel-evant species, as well as to facilitate detection of the possible

emer-gence of new resistance genes in clinical settings (

15 ,

16 ).

Phylogenetic-tree and amino acid sequence analyses showed

that the seven MBLs detected in soil belonged to either subclass B1

(PEDO-1) or B3 (PEDO-1, PEDO-2, CPS-1, ESP-1, MSI-1, and

SPG-1). These new MBLs were distantly related to the most

fre-quently detected MBLs in clinical isolates, with predicted amino

acid identities ranging from 40 to 69%. Closely related Pedobacter

species produced different MBL subclasses, highlighting the

diver-sity of MBLs within the genus. MSI-1 in the M. oculi strain was

phylogenetically related to THIN-B, a resident carbapenemase

produced by J. lividum (

28 ). These two MBLs shared common

ancestors with acquired SMB-1 and AIM-1 carbapenemases

pro-duced by S. marcescens (

13 ) and P. aeruginosa (

35 ), respectively

(

Fig. 2

). This may suggest that bla

SMB-1

and bla

AIM-1

could have

FIG 2 Unrooted phylogenetic tree of subclass B1 (red lines) and subclass B3 (blue lines) MBLs. MBLs isolated from clinical and environmental samples are indicated by stars and triangles, respectively. GenBank/EMBL sequence database accession numbers for the previously known MBLs are as follows: IND-1,

AF099139; CGB-1,AF339734; EBR-1,AF416700; JOHN-1,AY028464; BlaB1,AF189298; IMP-1,ADI87504; SIM-1,AAX76774; KHM-1,BAH16555; GIM-1,

CAF05908; MUS-1,AF441286; TUS-1,AF441287; NDM-1,CAZ39946; VIM-1,CAC35170; SPM-1,CAD37801; FIM-1,JX570731; AIM-1,AM998375; THIN-B,

CAC33832; SMB-1,AB636283; CAU-1,AJ308331; BJP-1,NP_772870; L1,ABO60992; POM-1,EU315252; GOB-1,AAF04458; and FEZ-1,CAB96921.

(7)

FIG 3 Alignment of amino acid sequences of subclass B3 metallo-␤-lactamases newly described in this study with previously described MBLs. Metal-binding

amino acids (12) are indicated with arrowheads. Position 221 is indicated with a star. Leu 221 in CPS-1 is boxed. GenBank/EMBL sequence database accession

numbers are as follows: GOB-1,AAF04458; L1,ABO60992; FEZ-1,CAB96921; CAU-1,AJ308331; BJP-1, NP772870; and AIM-1, M998375.

(8)

been acquired from members of Oxalobacteraceae, to which both

Massilia and Janthinobacterium belong.

Remarkable features were observed in some of the new MBLs.

CPS-1 contains an unusual, not previously reported Leu residue at

position 221. Position 221 is critical for MBL structure and

catal-ysis, as previously shown by the decrease in resistance to

ampicil-lin, cefotaxime, and imipenem obtained by in vivo replacement of

Met221 by Cys, Ser, His, Gln, Asp, Glu, and Ile in the GOB enzyme

(

40 ). As CPS-1 is related (68%) to GOB enzymes, we hypothesize

that the Leu221 residue has an impact on the catalytic efﬁciency of

CPS-1 for

␤-lactam antibiotics. The bla

PEDO-1

gene was associated

with genes encoding putative phage proteins showing structural

and functional homology to phage tail assembly proteins of

bac-teriophage T4 (

41 ,

42 ). Moreover, the gene encoding this MBL

was located between strands of noncoding DNA regions

contain-ing inverted repeats, which have been suggested to mediate

pro-tein-primed replication of bacteriophages and rolling-circle

transposition (

43 ). Altogether, these observations strongly suggest

that bla

PEDO-1

is located on a prophage integrated into the

chro-mosome of P. roseus strain SI-33. This may have implications for

potential acquisition of bla

PEDO-1

by other bacterial species, as

phages are known to mobilize antibiotic resistance genes among

environmental bacteria (

44 ,

45 ). Finally, MSI-1 overlapped a

pu-tative OXA-type

␤-lactamase, and both predicted ORFs appeared

to be driven by a single promoter. A similar genetic organization

was detected by examining the published sequence of the M.

ti-monae CCUG 45783 reference strain. In this case, the MBL and

OXA domains were fused into a putative bifunctional MBL-OXA

enzyme (protein accession number

WP_005669365.1

), although

hydrolytic activity toward carbapenems associated with MBL

pro-duction was not observed. To our knowledge, fusion of class B and

class D

␤-lactamases has never been reported prior to this study,

although the existence of bifunctional enzymes for other classes of

␤-lactamases has been described previously (

38 ). Possible

acqui-sition of bifunctional

␤-lactamases by clinically relevant species is

a reason for concern due to the broad substrate preferences that

such enzymes may retain (

46 ).

Analysis of reference strains closely related to the

MBL-pro-ducing species identiﬁed in this study revealed that

carbapen-emase production might be intrinsic to Chryseobacterium and

Epi-lithonimonas, as it was detected in all species tested within the

genera. In contrast, all Pedobacter and single Sphingomonas and

Massilia reference strains did not show carbapenemase activity.

Most (ﬁve out of six) of the Sphingomonas reference strains and C.

piscium CCUG 51923 showed carbapenemase activity but did not

harbor the MBL-encoding genes found in soil Sphingomonas sp.

strain ALS-13 (bla

SPG-1

) and Chryseobacterium strain Stok-1

(bla

CPS-1

), respectively. This suggests heterogeneity of the MBLs

within these genera, likely including as-yet-undescribed MBL

types. Although none of the Pedobacter reference strains produced

carbapenemase and harbored bla

PEDO1-3

, P. agri PB92

T

showed a

high (8

␮g/ml) meropenem MIC, and analysis of available WGS

data (accession number

AJLG00000000

) indicated that it

har-bored a putative MBL (GenBank protein accession number

WP_010603234.1

, displaying 57% and 73% amino acid identity to

PEDO-1 and PEDO-2, respectively. Based on 3D structure

predic-tion (data not shown), this putative MBL showed signiﬁcant

sim-ilarity to the X-ray structure of FEZ-1 carbapenemase produced

by Fluoribacter (Legionella) gormanii (

47 ). However, no

carbapen-emase activity was detected in P. agri PB92

T

_{, suggesting that the}

putative MBL is not optimally expressed under the conditions of

the CarbaNP test. Alternatively, other mechanisms of resistance

may be responsible for the high meropenem MIC value observed

for the strain.

No mobile genetic elements were detected in the regions

adjacent to the MBL-encoding genes, except in bla

PEDO-1

.

Sim-ilarly, progenitors of genes encoding CTM-X and OXA-48

␤-lactamases, now widespread in clinical Gram-negative

iso-lates worldwide, were detected on the chromosome of

environ-mental isolates of Kluyvera and Shewanella in the absence of

mo-bile genetic elements (

5 ,

6 ). Interestingly, downstream of the

bla

ESP-1

gene from E. tenax Stok-2 there was a gene encoding a

putative bleomycin resistance protein that has been previously

reported downstream of bla

NDM-1

(

48 ,

49 ), suggesting a

possi-ble evolutionary relationship between ESP-1 and NDM-1. It

has been suggested that environmental opportunistic

patho-gens may facilitate transfer of resistance genes between the

en-vironment and clinical settings (

3 ). Based on the recent

resis-tome risk ranking proposal (

16 ), bla

PEDO-1

could qualify for

resistance readiness condition 2 (RESCon 2), i.e., potentially

TABLE 3 Effects of heterologous expression of metallo-␤-lactamase-encoding genes from E. tenax strain Stok-2 (bla_ESP-1), C. piscium strain Stok-1 (blaCPS-1), and P. roseus strain SI-33 (blaPEDO-1) on MICs of␤-lactams in E. coli TOP10

Antimicrobial agent

EUCAST clinical breakpoint

(␮g/ml)b

MIC (␮g/ml)

E. coli E. coli pESP-1 E. coli pCPS-1 E. coli pPEDO-1

Meropenema _⬎8 _0.032 _0.5 _0.094 _0.094 Ceftazidime _⬎4 _ⱕ0.25 32 4 16 Cefazolin _⫺ _ⱕ8 _⬎16 _ⱕ8 16 Cefepime _⬎4 _ⱕ1 _ⱕ1 _ⱕ1 _ⱕ1 Cefoxitin _⫺ 8 _⬎64 64 64 Cefpodoxime _⬎1 0.5 32 4 2 Cefotaxime _⬎2 _⬍0.25 2 0.5 0.5 Ceftriaxone _⬎2 _ⱕ1 2 _ⱕ1 _ⱕ1 Ampicillin _⬎8 _ⱕ4 256 64 64 Piperacillin-tazobactam _⬎16 _ⱕ4 16 _ⱕ4 _ⱕ4 Ceftazidime-clavulanic acid _⫺ _ⱕ0.25 16 4 8 Cefotaxime-clavulanic acid _⫺ 0.12 2 0.25 0.5 a

MIC was determined by Etest.

b_{⫺, data could not be obtained.}

(9)

transferable resistance determinants with known mechanisms

harbored by a nonpathogen. The remaining MBL-encoding

genes could be grouped as RESCon5, a group conferring

resis-tance to currently used antibiotics but not located on mobile

genetic elements.

The new metallo-

␤-lactamase-encoding genes isolated from

soil bacteria displayed poor activity against meropenem after

cloning into E. coli (

Table 3

). Previous studies have shown that

most of the clinically important MBLs, such as VIM (

50 ,

51 ) and

IMP (

52 ), confer reduced susceptibility, but not resistance, to

meropenem on E. coli laboratory strains despite their high

cata-lytic efﬁciency toward carbapenems (

7 ,

9 ). This could be due to

additional mechanisms (e.g., porin loss and efﬂux pumps) or

␤-lactamases that may correspond to the carbapenem resistance

phenotype in the original host (

53 ) or to the high permeability

coefﬁcient of carbapenems in E. coli (

54 ).

In conclusion, this study expands the current knowledge of

the occurrence, host range, diversity, and functionality of

MBL-encoding genes in the soil microbiota. MBL production

was observed for the ﬁrst time in members of the genera

Pedo-bacter, Epilithonimonas, Sphingomonas, and Massilia, revealing

the existence of seven new MBLs belonging to subclass B1 or B3

enzymes. Some of the MBLs produced by soil bacteria

dis-played remarkable features, such as integration of the

MBL-encoding gene (bla

PEDO-1

) into prophage, an unusual amino

acid residue at a key position for MBL structure and catalysis

(Leu221 in CPS-1), or overlap of bla

MSI-1

with a putative

OXA-like

␤-lactamase-encoding gene. These enzymes have not yet

emerged in clinical settings but constitute potential

carbap-enem resistance determinants in pathogenic bacterial species,

as demonstrated by their ability to confer resistance to

ampi-cillin and various cephalosporins, as well as reduced

suscepti-bility to carbapenems, once expressed in E. coli.

ACKNOWLEDGMENTS

K.K.B. acknowledges the Center for Environmental and Agricultural Mi-crobiology (CREAM) funded by the Villum Foundation. We thank Geir Wing Gabrielsen for providing us the soil samples from Svalbard, Nor-way; Rob Willems for providing imipenem/cilastatin; and Roh Seong Woon for the kind gift of Pedobacter agri PB92.

FUNDING INFORMATION

This study was supported by grant HEALTH-F3-2011-282004 (EvoTAR) from the European Union.

REFERENCES

1. Wright GD. 2007. The antibiotic resistome: the nexus of chemical and

genetic diversity. Nat Rev Microbiol 5:175–186.http://dx.doi.org/10.1038

/nrmicro1614.

2. D’Costa VM, McGrann KM, Hughes DW, Wright GD. 2006. Sampling

the antibiotic resistome. Science 311:374 –377.http://dx.doi.org/10.1126

/science.1120800.

3. Forsberg KJ, Reyes A, Wang B, Selleck EM, Sommer MO, Dantas G. 2012. The shared antibiotic resistome of soil bacteria and human

patho-gens. Science 337:1107–1111.http://dx.doi.org/10.1126/science.1220761.

4. Guardabassi L, Perichon B, van Heijenoort J, Blanot D, Courvalin P. 2005. Glycopeptide resistance vanA operons in Paenibacillus strains

iso-lated from soil. Antimicrob Agents Chemother 49:4227– 4233.http://dx

.doi.org/10.1128/AAC.49.10.4227-4233.2005.

5. Humeniuk C, Arlet G, Gautier VP, Grimont RL, Philippon A. 2002. Beta-lactamases of Kluyvera ascorbata, probable progenitors of some plas-mid-encoded CTX-M types. Antimicrob Agents Chemother 46:3045–

3049.http://dx.doi.org/10.1128/AAC.46.9.3045-3049.2002.

6. Poirel L, Heritier C, Nordmann P. 2004. Chromosome-encoded Ambler

class D beta-lactamase of Shewanella oneidensis as a progenitor of carbap-enem-hydrolyzing oxacillinase. Antimicrob Agents Chemother 48:348 – 351.http://dx.doi.org/10.1128/AAC.48.1.348-351.2004.

7. Cornaglia G, Giamarellou H, Rossolini GM. 2011. Metallo-

␤-lactamases: a last frontier for␤-lactams? Lancet Infect Dis 11:381–393.

http://dx.doi.org/10.1016/S1473-3099(11)70056-1.

8. Galleni M, Lamotte-Brasseur J, Rossolini GM, Spencer J, Dideberg O,

Frere JM, Metallo-beta-Lactamases Working Group. 2001. Standard

numbering scheme for class B beta-lactamases. Antimicrob Agents

Che-mother 45:660 – 663.http://dx.doi.org/10.1128/AAC.45.3.660-663.2001.

9. Queenan AM, Bush K. 2007. Carbapenemases: the versatile

beta-lactamases. Clin Microbiol Rev 20:440 – 458.http://dx.doi.org/10.1128

/CMR.00001-07.

10. Bebrone C, Delbruck H, Kupper MB, Schlomer P, Willmann C, Frere

JM, Fischer R, Galleni M, Hoffmann KM. 2009. The structure of the

di-zinc subclass B2 metallo-beta-lactamase CphA reveals that the second inhibitory zinc ion binds in the histidine site. Antimicrob Agents

Che-mother 53:4464 – 4471.http://dx.doi.org/10.1128/AAC.00288-09.

11. Fonseca F, Bromley EH, Saavedra MJ, Correia A, Spencer J. 2011. Crystal structure of Serratia fonticola Sfh-I: activation of the nucleophile in

mono-zinc metallo-beta-lactamases. J Mol Biol 411:951–959.http://dx

.doi.org/10.1016/j.jmb.2011.06.043.

12. Nordmann P, Naas T, Poirel L. 2011. Global spread of

carbapenemase-producing Enterobacteriaceae. Emerg Infect Dis 17:1791–1798.http://dx

.doi.org/10.3201/eid1710.110655.

13. Wachino J, Yoshida H, Yamane K, Suzuki S, Matsui M, Yamagishi T,

Tsutsui A, Konda T, Shibayama K, Arakawa Y. 2011. SMB-1, a novel

subclass B3 metallo-beta-lactamase, associated with ISCR1 and a class 1 integron, from a carbapenem-resistant Serratia marcescens clinical isolate.

Antimicrob Agents Chemother 55:5143–5149.http://dx.doi.org/10.1128

/AAC.05045-11.

14. Pollini S, Maradei S, Pecile P, Olivo G, Luzzaro F, Docquier JD,

Rossolini GM. 2013. FIM-1, a new acquired metallo-beta-lactamase from

a Pseudomonas aeruginosa clinical isolate from Italy. Antimicrob Agents

Chemother 57:410 – 416.http://dx.doi.org/10.1128/AAC.01953-12.

15. Ashbolt NJ, Amezquita A, Backhaus T, Borriello P, Brandt KK,

Colli-gnon P, Coors A, Finley R, Gaze WH, Heberer T, Lawrence JR, Larsson DG, McEwen SA, Ryan JJ, Schonfeld J, Silley P, Snape JR, Van den Eede C, Topp E. 2013. Human health risk assessment (HHRA) for

environ-mental development and transfer of antibiotic resistance. Environ Health

Perspect 121:993–1001.http://dx.doi.org/10.1289/ehp.1206316.

16. Martínez JL, Coque TM, Baquero F. 2015. What is a resistance gene?

Ranking risk in resistomes. Nat Rev Microbiol 13:116 –123.http://dx.doi

.org/10.1038/nrmicro3399.

17. Poulsen PH, Al-Soud WA, Bergmark L, Magid J, Hansen LH, Sørensen

SJ. 2013. Effects of fertilization with urban and agricultural organic wastes

in a ﬁeld trial—prokaryotic diversity investigated by pyrosequencing. Soil

Biol Biochem 57:784 –793. http://dx.doi.org/10.1016/j.soilbio.2011.12

.023.

18. Berg J, Thorsen MK, Holm PE, Jensen J, Nybroe O, Brandt KK. 2010. Cu exposure under ﬁeld conditions co-selects for antibiotic resistance as determined by a novel cultivation-independent bacterial community

tol-erance assay. Environ Sci Technol 44:8724 – 8728. http://dx.doi.org/10

.1021/es101798r.

19. Brandt KK, Petersen A, Holm PE, Nybroe O. 2006. Decreased abun-dance and diversity of culturable Pseudomonas spp. populations with in-creasing copper exposure in the sugar beet rhizosphere. FEMS Microbiol

Ecol 56:281–291.http://dx.doi.org/10.1111/j.1574-6941.2006.00081.x.

20. Petersen SO, Henriksen K, Mortensen G, Krogh P, Brandt KK,

Sø-rensen J, Madsen T, Petersen J, Grøn C. 2003. Recycling of sewage sludge

and household compost to arable land: fate and effects of organic

contam-inants, and impact on soil fertility. Soil Till Res 72:139 –152.http://dx.doi

.org/10.1016/S0167-1987(03)00084-9.

21. Pontiroli A, Travis ER, Sweeney FP, Porter D, Gaze WH, Mason S,

Hibberd V, Holden J, Courtenay O, Wellington EMH. 2011. Pathogen

quantitation in complex matrices: a multi-operator comparison of DNA extraction methods with a novel assessment of PCR inhibition. PLoS One

6:e17916.http://dx.doi.org/10.1371/journal.pone.0017916.

22. Selama O, Amos GC, Djenane Z, Borsetto C, Laidi RF, Porter D,

Nateche F, Wellington EMH, Hacène H. 2014. Screening for genes

coding for putative antitumor compounds, antimicrobial and enzymatic activities from haloalkalitolerant and haloalkaliphilic bacteria strains of

(10)

Algerian Sahara soils. Biomed Res Int 2014:317524.http://dx.doi.org/10 .1155/2014/317524.

23. Tolba S, Egan S, Kalliﬁdas D, Wellington EMH. 2002. Distribution of streptomycin resistance and biosynthesis genes in streptomycetes

recov-ered from different soil sites. FEMS Microbiol Ecol 42:269 –276.http://dx

.doi.org/10.1111/j.1574-6941.2002.tb01017.x.

24. Nordmann P, Gniadkowski M, Giske CG, Poirel L, Woodford N,

Miriagou V. 2012. Identiﬁcation and screening of

carbapenemase-producing Enterobacteriaceae. Clin Microbiol Infect 18:432– 438.http:

//dx.doi.org/10.1111/j.1469-0691.2012.03815.x.

25. Janssen PH, Yates PS, Grinton BE, Taylor PM, Sait M. 2002. Improved culturability of soil bacteria and isolation in pure culture of novel mem-bers of the divisions Acidobacteria, Actinobacteria, Proteobacteria, and

Ver-rucomicrobia. Appl Environ Microbiol 68:2391–2396.http://dx.doi.org /10.1128/AEM.68.5.2391-2396.2002.

26. Dortet L, Bréchard L, Cuzon G, Poirel L, Nordmann P. 2014. Strategy for rapid detection of carbapenemase-producing Enterobacteriaceae.

An-timicrob Agents Chemother 58:2441–2445. http://dx.doi.org/10.1128

/AAC.01239-13.

27. Stackebrandt E, Goebel BM. 1994. Taxonomic note: a place for DNA-DNA reassociation and 16S rRNA sequence analysis in the present species

deﬁnition in bacteriology. Int J Syst Bacteriol 44:846 – 849.http://dx.doi

.org/10.1099/00207713-44-4-846.

28. Rossolini GM, Condemi MA, Pantanella F, Docquier JD, Amicosante

G, Thaller MC. 2001. Metallo-␤-lactamase producers in environmental

microbiota: new molecular class B enzyme in Janthinobacterium lividum.

Antimicrob Agents Chemother 45:837– 844. http://dx.doi.org/10.1128

/AAC.45.3.837-844.2001.

29. Newman JR, Fuqua C. 1999. Broad-host-range expression vectors that carry the L-arabinose-inducible Escherichia coli araBAD promoter and the

araC regulator. Gene 227:197–203. http://dx.doi.org/10.1016/S0378 -1119(98)00601-5.

30. Poirel L, Rodriguez-Martinez JM, Al Naiemi N, Debets-Ossenkopp YJ,

Nordmann P. 2010. Characterization of DIM-1, an integron-encoded

metallo-beta-lactamase from a Pseudomonas stutzeri clinical isolate in The

Netherlands. Antimicrob Agents Chemother 54:2420 –2424.http://dx.doi

.org/10.1128/AAC.01456-09.

31. Gupta SK, Padmanabhan BR, Diene SM, Lopez-Rojas R, Kempf M,

Landraud L, Rolain JM. 2014. ARG-ANNOT, a new bioinformatic

tool to discover antibiotic resistance genes in bacterial genomes.

Anti-microb Agents Chemother 58:212–220. http://dx.doi.org/10.1128

/AAC.01310-13.

32. Kelley LA, Sternberg MJ. 2009. Protein structure prediction on the web:

a case study using the Phyre server. Nat Protoc 4:363–371.http://dx.doi

.org/10.1038/nprot.2009.2.

33. Dortet L, Poirel L, Nordmann P. 2012. Rapid detection of

carbapen-emase-producing Pseudomonas spp. J Clin Microbiol 50:3773–3776.http:

//dx.doi.org/10.1128/JCM.01597-12.

34. Bellais S, Aubert D, Naas T, Nordmann P. 2000. Molecular and bio-chemical heterogeneity of class B carbapenem-hydrolyzing beta-lactamases in Chryseobacterium meningosepticum. Antimicrob Agents

Chemother 44:1878 –1886. http://dx.doi.org/10.1128/AAC.44.7.1878

-1886.2000.

35. Docquier JD, Pantanella F, Giuliani F, Thaller MC, Amicosante G,

Galleni M, Frere JM, Bush K, Rossolini GM. 2002. CAU-1, a subclass B3

metallo-beta-lactamase of low substrate afﬁnity encoded by an ortholog present in the Caulobacter crescentus chromosome. Antimicrob Agents

Chemother 46:1823–1830. http://dx.doi.org/10.1128/AAC.46.6.1823

-1830.2002.

36. Yong D, Toleman MA, Bell J, Ritchie B, Pratt R, Ryley H, Walsh TR. 2012. Genetic and biochemical characterization of an acquired subgroup

B3 metallo-beta-lactamase gene, blaAIM-1, and its unique genetic context

in Pseudomonas aeruginosa from Australia. Antimicrob Agents

Che-mother 56:6154 – 6159.http://dx.doi.org/10.1128/AAC.05654-11.

37. Voha C, Docquier JD, Rossolini GM, Fosse T. 2006. Genetic and bio-chemical characterization of FUS-1 (OXA-85), a narrow-spectrum class D beta-lactamase from Fusobacterium nucleatum subsp. polymorphum.

An-timicrob Agents Chemother 50:2673–2679. http://dx.doi.org/10.1128

/AAC.00058-06.

38. Naas T, Bellais S, Nordmann P. 2003. Molecular and biochemical char-acterization of a carbapenem-hydrolysing beta-lactamase from

Flavobac-terium johnsoniae. J Antimicrob Chemother 51:267–273.http://dx.doi.org /10.1093/jac/dkg069.

39. Allen HK, Moe LA, Rodbumrer J, Gaarder A, Handelsman J. 2009.

Functional metagenomics reveals diverse␤-lactamases in a remote

Alas-kan soil. ISME J 3:243–251.http://dx.doi.org/10.1038/ismej.2008.86.

40. Lisa M, Morán-Barrio J, Guindón M, Vila AJ. 2012. Probing the role of

Met221 in the unusual metallo-_{␤-lactamase GOB-18. Inorg Chem 51:}

12419 –12425.http://dx.doi.org/10.1021/ic301801h.

41. Aksyuk AA, Leiman PG, Kurochkina LP, Shneider MM, Kostyuchenko

VA, Mesyanzhinov VV, Rossmann MG. 2009. The tail sheath structure of

bacteriophage T4: a molecular machine for infecting bacteria. EMBO J

28:821– 829.http://dx.doi.org/10.1038/emboj.2009.36.

42. Kostyuchenko VA, Leiman PG, Chipman PR, Kanamaru S, van Raaij

MJ, Arisaka F, Mesyanzhinov VV, Rossmann MG. 2003.

Three-dimensional structure of bacteriophage T4 baseplate. Nat Struct Biol 10:

688 – 693.http://dx.doi.org/10.1038/nsb970.

43. Chandler M, de la Cruz F, Dyda F, Hickman AB, Moncalian G,

Ton-Hoang B. 2013. Breaking and joining single-stranded DNA: the

HUH endonuclease superfamily. Nat Rev Microbiol 11:525–538.http:

//dx.doi.org/10.1038/nrmicro3067.

44. Colomer-Lluch M, Jofre J, Muniesa M. 2011. Antibiotic resistance genes in the bacteriophage DNA fraction of environmental samples. PLoS One

6:e17549.http://dx.doi.org/10.1371/journal.pone.0017549.

45. Balcazar JL. 2014. Bacteriophages as vehicles for antibiotic resistance

genes in the environment. PLoS Pathog 10:e1004219.http://dx.doi.org/10

.1371/journal.ppat.1004219.

46. Zhang W, Fisher JF, Mobashery S. 2009. The bifunctional enzymes of

antibiotic resistance. Curr Opin Microbiol 12:505–511.http://dx.doi.org

/10.1016/j.mib.2009.06.013.

47. Garcia-Sáez I, Mercuri P, Papamicael C, Kahn R, Frere J, Galleni M,

Rossolini GM, Dideberg O. 2003. Three-dimensional structure of FEZ-1,

a monomeric subclass B3 metallo-␤-lactamase from Fluoribacter

gorma-nii, in native form and in complex withD-captopril. J Mol Biol 325:651– 660.http://dx.doi.org/10.1016/S0022-2836(02)01271-8.

48. Dortet L, Nordmann P, Poirel L. 2012. Association of the emerging carbapenemase NDM-1 with a bleomycin resistance protein in

Enterobac-teriaceae and Acinetobacter baumannii. Antimicrob Agents Chemother 56:

1693–1697.http://dx.doi.org/10.1128/AAC.05583-11.

49. Poirel L, Dortet L, Bernabeu S, Nordmann P. 2011. Genetic features of

blaNDM-1-positive Enterobacteriaceae. Antimicrob Agents Chemother 55:

5403–5407.http://dx.doi.org/10.1128/AAC.00585-11.

50. Lauretti L, Riccio ML, Mazzariol A, Cornaglia G, Amicosante G,

Fon-tana R, Rossolini GM. 1999. Cloning and characterization of blaVIM, a

new integron-borne metallo-beta-lactamase gene from a Pseudomonas

aeruginosa clinical isolate. Antimicrob Agents Chemother 43:1584 –1590.

51. Poirel L, Naas T, Nicolas D, Collet L, Bellais S, Cavallo JD, Nordmann

P. 2000. Characterization of VIM-2, a carbapenem-hydrolyzing

metallo-beta-lactamase and its plasmid- and integron-borne gene from a

Pseu-domonas aeruginosa clinical isolate in France. Antimicrob Agents

Che-mother 44:891– 897.http://dx.doi.org/10.1128/AAC.44.4.891-897.2000.

52. Gibb AP, Tribuddharat C, Moore RA, Louie TJ, Krulicki W, Livermore

DM, Palepou MF, Woodford N. 2002. Nosocomial outbreak of

carbap-enem-resistant Pseudomonas aeruginosa with a new bla(IMP) allele,

bla(IMP-7). Antimicrob Agents Chemother 46:255–258. http://dx.doi

.org/10.1128/AAC.46.1.255-258.2002.

53. Nordmann P, Dortet L, Poirel L. 2012. Carbapenem resistance in

Enter-obacteriaceae: here is the storm! Trends Mol Med 18:263–272.http://dx

.doi.org/10.1016/j.molmed.2012.03.003.

54. Matsumura N, Minami S, Watanabe Y, Iyobe S, Mitsuhashi S. 1999. Role of permeability in the activities of beta-lactams against gram-negative bacteria which produce a group 3 beta-lactamase. Antimicrob Agents Chemother 43:2084 –2086.