OUP user menu

Determining the antibiotic resistance potential of the indigenous oral microbiota of humans using a metagenomic approach

Martha L. Diaz-Torres, Aurelie Villedieu, Nigel Hunt, Rod McNab, David A. Spratt, Elaine Allan, Peter Mullany, Michael Wilson
DOI: http://dx.doi.org/10.1111/j.1574-6968.2006.00221.x 257-262 First published online: 1 May 2006

Abstract

Studies of the prevalence and identity of genes encoding resistance to antibiotics in a microbial community are usually carried out on only the cultivable members of the community. However, it is possible to include the as-yet-uncultivable organisms present by adopting a metagenomic approach to such studies. In this investigation, four metagenomic libraries of the oral microbiota were prepared from three groups of 20 adult humans and screened for antibiotic-resistant clones. Clones resistant to tetracycline and amoxycillin were present in all four libraries while gentamicin-resistant clones were found in three of the libraries. The genes encoding tetracycline resistance in the clones were identified and found to be tet(M), tet(O), tet(Q), tet(W), tet37 and tet(A). However, only the first three of these were detected in all three groups of individuals investigated.

Keywords
  • metagenomics
  • antibiotic resistance
  • tetracycline
  • oral microbiota
  • resistome

Introduction

The oral cavity of humans harbours a large and diverse microbiota and any of more than 700 bacterial taxa may be present in these communities (Wilson, 2005). It is apparent that, as in many natural microbial communities, a large proportion (up to 50%) of the microbes present have not yet been cultured (Tanner et al., 1994). To circumvent this problem, there is, therefore, considerable interest in using molecular approaches, such as metagenomics, for investigating the composition and function of the oral microbiota (Diaz-Torres et al., 2003; Duncan, 2003; Sakamoto et al., 2005). One important attribute of the oral microbiota is its ability to act as a reservoir of antibiotic-resistant organisms (Walker, 1996; Roberts, 1998; Ready et al., 2003). Oral bacteria can easily reach other body sites (by swallowing and via the bloodstream) and also spread to other individuals (coughing, kissing etc.). Therefore, antibiotic-resistant oral bacteria have the opportunity for rapid dissemination through the community and to transfer their resistance genes to other bacterial species. Virtually all investigations of the prevalence of antibiotic-resistant oral bacteria are based solely on the cultivable microbiota and so take no account of the large number of as-yet-uncultured bacteria present in these complex communities (Preus et al., 1995; Walker, 1996; Roberts, 1998; Ready et al., 2003). To overcome this problem, we have used a metagenomic approach to identify the presence of genes encoding antibiotic resistance in the oral microbiota of adult humans. This involved the generation of expression libraries from the total DNA present in saliva and plaque from these individuals and screening the resulting clones for antibiotic resistance.

Materials and methods

Patients and sampling

Saliva and plaque samples were obtained from three groups of healthy adults attending the Orthodontic Dentistry Clinic at the Eastman Dental Hospital. Each of the groups consisted of 20 individuals. All of the individuals were healthy and were not suffering from any oral diseases. Furthermore, the individuals had not received any antibiotics during the previous 3 months. A saliva sample (2 mL) was collected in a sterile container. Plaque samples were collected with a sterile swab, and gingival margin plaque samples were collected from four different sites with sterile paper points; both samples were pooled into 4 mL of Ringer's solution and mixed with the saliva sample. The plaque/saliva samples from each of the 20 individuals in each group were pooled to provide three combined sets of plaque and saliva.

Preparation and screening of the metagenomic libraries

The bacteria present in each of the three pooled plaque/saliva samples were separately harvested by centrifugation at 3500 g, for 10 min at 4°C. The three resulting bacterial pellets were washed gently in sterile saline. Each of the three batches of cells was then processed as described below. Two aliquots of each batch of cells were prepared, and DNA was extracted from one of these aliquots using the Puregene Gram-positive DNA isolation protocol and from the other using the Puregene Gram-negative DNA isolation protocol (Gentra Systems) according to the manufacturer's instructions. The extracted DNA was subsequently pooled. To prepare the DNA for library construction, 0.2 mL of DNA (at a concentration of 250 ng mL−1) was sonicated for 5 s on ice at 80% power using an ultrasonic homogenizer (IKA-WERKE). The ends of the DNA were repaired by treating them with 2 U mg−1 of mung bean nuclease (Promega) in a final volume of 100 μL at 37°C for 1 h to produce blunt ends. The resulting DNA fragments were separated by agarose gel electrophoresis, and fragments between 800 and 3000 bp were cut from the gel and purified using the Qiagen agarose purification kit. To generate 3′-A overhangs before cloning into a TOPO-XL vector, the DNA fragments were incubated at 75°C for 1 h with 1 U mg−1 of Taq DNA polymerase (Bioline) in a final volume of 100 μL of 1 ×Taq buffer containing 2 mM dATP. Before ligation into the vector, the DNA was purified using the Qiagen PCR purification kit. Ligation of DNA into TOPO-XL and subsequent transformation into Escherichia coli TOP10 cells were performed according to the manufacturer's protocol. Each of these libraries was then screened for clones exhibiting resistance to either tetracycline, gentamicin or amoxycillin. Colonies were selected on Luria–Bertani (LB) agar plates containing 50 μg mL−1 kanamycin (encoded by the vector) and either tetracycline, gentamicin or amoxycillin at concentrations of 10, 15 and 10 μg mL−1, respectively Plasmid DNA was isolated from the resistant E. coli colonies using the Qiagen Miniprep kit. The plasmids were retransformed into E. coli and screened to confirm that the insert DNA was responsible for resistance.

Three metagenomic libraries (one from each batch of samples) were constructed in TOPO-XL. A fourth library was constructed from the third batch of samples using an EpiFOS Fosmid kit (Epicentre) – this allows larger insert sizes and therefore the potential to identify the organism from which the resistance determinant arose. EpiFOS utilizes a novel strategy of cloning randomly sheared, end repaired DNA. Therefore, 2.5 μg of genomic DNA was sheared in a Hydroshear device (GeneMachine USA) which uses hydrodynamic shearing forces to fragment DNA strands into c. 40 kb fragments. Confirmation of DNA shearing was performed by running pulsed field gel electrophoresis (PFGE; Bio-Rad FIGE Mapper Electrophoresis System) with voltage and ramp times recommended by the manufacturer for separation of 10–100 kb DNA. To generate blunt-ended, 5′-phosphorylated DNA inserts, the end-repair reaction was carried out using End Repair Enzyme Mix and 10 × Buffer, dNTPs and ATP supplied by Epicentre. The ligation reaction and fosmid packaging were performed according to the manufacturer's protocols. The fosmid library was based on the titer of the packaged fosmid clones and subsequently the diluted phage particles were mixed with the EPI cells (Epicentre) in the ratio of 100 μL of cells for every 10 μL of diluted phage particles following adsorption at 37°C for 20 min. The fosmid clones were selected on Luria–Bertani (LB) agar plates containing 12 μg mL−1 chloramphenicol and incubated at 37°C overnight. The chloramphenicol-resistant clones were screened on LB plates containing tetracycline, gentamicin or amoxycillin at concentrations of 10, 15 and 10 μg mL−1, respectively.

The four libraries were replicated into sets of 96-well microtiter plates with freezing medium and stored at −80°C.

Minimum inhibitory concentrations (MICs) of tetracycline-resistant clones

The MIC of each tetracycline-resistant clone was determined using an agar dilution technique as follows. The inoculum was standardized using a 0.5 Macfarland standard in accordance with NCCLS recommendations and inoculated using a multi-point inoculator (Mast Diagnostics, Bootle, UK) on to IsoSensitest agar plates (Oxoid) supplemented with 5% defibrinated horse blood and containing a range of tetracycline concentrations. The highest dilution of tetracycline that inhibited visible growth after 24 h aerobic incubation at 37°C was taken to be the MIC. Staphylococcus aureus NCTC 6571 and IsoSensitest broth (Oxoid) were included on the plates as positive and negative controls, respectively.

Identification of genes encoding tetracycline resistance

To identify the tetracycline resistance genes, PCR was used as described previously (Aminov et al., 2001; Villedieu et al., 2003). The primers used for identifying the various genes are shown in Table 1. For the recently identified tet(37) gene, the following specific primers were used in standard PCRs: Tet37F, 5′-AGGGATATTGGTTGGAGA-3′; Tet37R, 5′-ATCAGTCTCATATTTCGACA-3′. All PCR products were sequenced with forward primer XLF-2 (5′-CGC CAG TGT GAT GGA TAT-3′) and reverse primer XL-2R (5′-TAG AAT ACT CAA GCT ATG C-3′). The Fosmid clones were partially sequenced using the pCC1/pEpiFOS RP-2 Reverse Sequencing primer (5′ TACGCCAAGCTATTTAGGTGAGA 3′).

View this table:
1

PCR primers used in the identification of the various genes encoding resistance to tetracycline

PrimerClass targeted5′ to 3′ Sequence
TetA-FWtet(A)GCT ACA TCC TGC TTG CCT TC
TetA-RVtet(A)CAT AGA TCG CCG TGA AGA GG
TetB-FWtet(B)TTG GTT AGG GGC AAG TTT TG
TetB-RVtet(B)GTA ATG GGC CAA TAA CAC CG
TetC-FWtet(C)CTT GAG AGC CTT CAA CCC AG
TetC-RVtet(C)ATG GTC GTC ATC TAC CTG CC
TetE-FWtet(E)AAA CCA CAT CCT CCA TAC GC
TetE-RVtet(E)AAA TAG GCC ACA ACC GTC AG
TetK-FWtet(K)TCG ATA GGA ACA GCA GTA
TetK-RVtet(K)CAG CAG ATC CTA CTC CTT
TetL-FWtet(L)TCG TTA GCG TGC TGT CAT TC
TetL-RVtet(L)GTA TCC CAC CAA TGT AGC CG
TetM-FWtet(M)GTG GAC AAA GGT ACA ACG AG
TetM-RVtet(M)CGG TAA AGT TCG TCA CAC AC
TetO-FWtet(O)AAC TTA GGC ATT CTG GCT CAC
TetO-RVtet(O)TCC CAC TGT TCC ATA TCG TCA
TetS-FWtet(S)CAT AGA CAA GCC GTT GAC C
TetS-RVtet(S)ATG TTT TTG GAA CGC CAG AG
TetQ-FWtet(Q)TTA TAC TTC CTC CGG CAT CG
TetQ-RVtet(Q)ATC GGT TCG AGA ATG TCC AC
TetT-FWtet(T)AAGGTTTATTATATAAAAGTG
TetT-RVtet(T)AGGTGTATCTATGATATTTAC
TetW-FWtet(W)GAGAGCCTGCTATATGCCAGC
TetW-RVtet(W)GGGCGTATCCACAATGTTAAC

Sequencing reactions were performed using the ABI PRISM BigDye Terminator cycle sequencing protocol on either an Applied Biosystems model 310 genetic analyzer (Applied Biosystems, Foster City, CA) or a model 373 DNA sequencer (AB Biosystems) according to the manufacturer's instructions. DNA sequences were analyzed with the DNAMAN version 5.2.2 program (Lynnon Biosoft). Similarity analysis was carried out with the Advance Blast program of GenBank (National Center for Biotechnology Information, National Institutes of Health, Washington, DC), and alignments were performed using the CLUSTAL W program service at the European Bioinformatics Institute (http://www.ebi.ac.uk/).

Results

Three separate libraries containing DNA from bacteria in the pooled saliva/plaque samples were produced in the TOPO-XL vector. The average insert size in these libraries was 3 kb, and the largest insert was c. 7 kb. A fourth library was also prepared from the third batch of samples in a fosmid in order to obtain larger insert sizes (mean size=40 kb). Clones resistant to tetracycline and amoxycillin were detected in each library whereas gentamicin-resistant clones were detected only in libraries #1, #2 and #4 (Table 2). Tetracycline-resistant clones were the most common resistance type in each library. The MICs of the individual tetracycline-resistant clones ranged from 4 to 30 μg mL−1 whereas the MIC of tetracycline for the Escherichia coli carrying only TOPO-XL was 1 μg mL−1. In all four libraries, the genes encoding tetracycline resistance included tet(M), tet(O) and tet(Q) (Table 3). In library #1, a novel gene encoding resistance to tetracycline (tet37) was identified (Gillespie et al., 2002) whereas in libraries #3 and #4 tet(W) and tetA(D) were also detected. In addition, library #3 also contained tet(A).

View this table:
2

Antibiotic-resistant clones in four metagenomic libraries prepared from oral bacteria obtained from three groups of 20 adults

Library#1Library#2Library#3Library#4
Total number of transformants analysed450412398600
Number of clones resistant to tetracycline (%)18 (4)17 (4)16 (4)7 (1)
Number of clones resistant to amoxycillin (%)13 (3)8 (2)5 (1)6 (1)
Number of clones resistant to gentamicin (%)7 (1)6 (1)0 (0)1(1)
  • Libraries #1, #2 and #3 were from the three different groups of adults, whereas library #4 was from the third group of individuals.

View this table:
3

Identity of genes encoding tetracycline resistance in the tetracycline-resistant clones from the four libraries

LibraryGenes encoding tetracycline resistance
#1tet(M), tet(O), tet(Q), tet37
#2tet(M), tet(O), tet(Q)
#3tet(M), tet(O), tet(Q), tet(W), tet(A)
#4tet(M), tet(O), tet(Q), tet(W)

Table 4 shows the identities of the tetracycline resistance determinants detected together with closest sequence homologies for the tetracycline-resistant clones from all four libraries.

View this table:
4

DNA sequence analysis of tetracycline-resistant clones

Tetracycline-resistant cloneTetracycline resistance determinantIdentity (%)Closest match by nucleotide sequence comparison (Accession number)
Library #1
T1-S180Hypothetical protein (P41421)
T2-S1TetO100Campylobacter jejuni, TetO (M18896)
T6-S195Insertion element, 1/5/6 protein insB (P03830)
T8-S1Tet37100This study (AF540889)
T10-S1No homologies
T12-S1TetM100Enterococcus faecalis, TetM (X56353)
T15-S1TetQ100Prevotella intermedia, TetQ (U73497)
T16-S1TetM100Enterococcus faecalis, TetM (M85225)
T9-S1TetO100Campylobacter jejuni, TetO (M18896)
T37-S196Hypothetical protein, Salmonella typhimurium plasmid NTP16 (JQ1541)
T41-S1100Streptoalloteichus hindustans, phleomycin-bleomycin binding protein (X52869)
T68-S1No homologies
T50-S1100Synthetic construct plasmid, pSN6 (AY159365.1)
T35RcS195Transposon delivery vector, pSC189 (AY115560)
TS1-T7100Autographa californica, nucleopolyhedrovirus, Polyhedrin protein (K01149)
3FS195Synthetic construct Canis familiaris his-tagged-multidrug resistance glycoprotein gene (AF269224)
Library #2
12TS2TetM100Enterococcus faecalis, TetM (M85225)
T3F-S295Citrobacter amalonaticus, class I integron integrase gene and streptomycin 3′ adenyltransferase AadA2 (AF486817)
T41-2sqf99Shuttle vector, pBSV2 (AY187276)
T7496Prevotella sp. oral clone BS041, 16S rRNA gene (AF385558)
85TS2TetM100Enterococcus faecalis TetM (M85225)
T3F-S295Synthetic construct, Canis familiaris his-tagged-multidrug resistance glycoprotein gene (AF269224)
TG3-38(S2)95Synthetic construct plasmid, pSN6 alkaline phosphatase-like protein, LacI, and β-lactamase genes (AY159365)
T3F-S295Citrobacter amalonaticus, class I integron integrase gene and streptomycin 3′ adenyltransferase AadA2 (AF486817)
D2-M2197Activation tagging vector, pSKI074 (AF218466)
D8-M2496Uncultured soil fungus clone f95-2, 16S rRNA gene (AF515414)
4TS2TetQ100Prevotella intermedia, TetQ (U73497)
D2-M19100Expression vector, pYPX251 (AY178046)
13TS2TetO100Campylobacter jejuni, TetO (M18896)
D6-M21100Cloning vector, pPGKneo-II (AF335420.3)
D8-M22100Envelope protein Envelope expression vector, pVpack 10A1 (AAK01731)
Library #3
T4-S3TetQ100Prevotella intermedia, TetQ (U73497)
T6-S399Bacillus subtilis transposable element DNA in the region before tetBS908 (X58999)
A1-M1-3100Uncultured Nitrosomonas Al-7K small subunit ribosomal RNA gene, partial sequence (AF043136)
T3595Transposon delivery vector, pSC189 (AY115560)
T1-S3TetM100Enterococcus faecalis, Tet(M)916 (X56353)
f6-m-S3TetO97Campylobacter jejuni, TetO (M18896)
10-M5-d(3)TetA(D)100Shigella flexneri SH4, TetA (AF467078)
A5-M398Cloning vector pUCP26, Escherichia-Pseudomonas shuttle vector with tetracycline efflux protein and LacZ α peptide (lacZ α) genes (U07168)
T5-S3TetM100Enterococcus faecalis, Tet(M)916 (X56353)
D6-M198Activation-tagging vector pSKI015 (AF187951)
h2-t8.3m13r97Prevotella sp. oral clone BS041, 16S rRNA gene (AF385558)
E5-M5100Citrobacter freundii, aminoglycoside adenyltransferase (AAL59387)
T10-S3TetW100Roseburia sp. A2-183, TetW gene (AJ421625)
T50-S3100Phellinus ferrugineofuscus, 28S ribosomal RNA gene (AF311031)
Library #4
1FTetM100Staphylococcus aureus strain 2952 tetracycline resistance protein TetM (tetM) gene, partial cds (AY057894.1)
20FTetO100Campylobactyer jejuni TetO (M18896)
22FTetM100Staphylococcus aureus strain 2952 tetracycline resistance protein TetM (tetM) gene, partial cds (AY057894.1)
29F100Maltophilia smeF genes component of a multidrug efflux system (AJ252200.1)
32FTetW100Roseburia sp. A2-183, TetW gene (AJ421625)
51FTetQ100Prevotella intermedia TetQ (U73497)
  • Those clones found to contain human DNA are excluded.

Discussion

The purpose of this study was to determine whether a metagenomic approach could be used to ascertain the presence of genes encoding resistance to representative members of three different classes of antibiotics in the oral microbiota of a large number of adults. A further objective was to identify those genes that were responsible for conferring resistance to tetracycline. Cloning and expressing the genomic DNA of a microbial community has great potential for assessing and accessing all of the genetic capabilities of a community – including those residing in the as-yet-uncultivable taxa present (Handelsman, 2004). Although a number of studies have reported the application of a metagenomic approach to identifying the antibiotic-producing capabilities of a microbial community, few have used this technique to identify the antibiotic resistance inherent in a community (Gillespie et al., 2002; Moreira et al., 2004). Riesenfeld . (2004) detected nine clones expressing resistance to aminoglycoside antibiotics and one expressing tetracycline resistance in a metagenomic library prepared from soil samples. Furthermore, Diaz-Torres . (2003) reported the identification of a novel tetracycline resistance determinant, tet(37), in a metagenomic library prepared from the oral microbiota of humans.

In this study, we have prepared metagenomic libraries from the indigenous oral microbiota of three separate groups of 20 adult humans and screened them for the presence of antibiotic-resistant clones. Each of the libraries contained clones resistant to tetracycline and amoxycillin while three of them also had gentamicin-resistant clones. We have identified some of the genes encoding tetracycline resistance in these clones and found mainly genes encoding ribosomal protection proteins –tet(M), tet(O), tet(Q) and tet(W). Other genes detected were tet(A) and tet(37). There was a surprisingly high proportion of tetracycline-resistant clones in the libraries. However, sequencing of these clones revealed that many were genes encoding ribosomal proteins or human proteins which conferred tetracycline resistance on the host by mechanisms that remain to be established. The tetracycline resistance gene profiles of the oral microbiotas of the three groups investigated were very similar, with all groups harbouring tet(M), tet(O) and tet(Q). In a culture-based analysis of the tetracycline resistance genes present in the first group of individuals, a greater variety of genes were found (Villedieu et al., 2003). Hence, in addition to tet(M), tet(O) and tet(Q), the following genes were detected: tet(W), tet(S), tet(L), tet(A) and tet(K). This implies that there may be problems with the expression of certain tetracycline resistance genes in Escherichia coli. In fact, we have data showing this to be the case.

The production of ribosomal proteins that bind and protect the ribosomes from the action of tetracycline is very common as a mechanism of resistance to tetracycline in both Gram-positive and Gram-negative bacteria (Speer, 1992; Roberts, 1996). So far, nine classes of genes encoding ribosome protection proteins have been described: tet(M), tet(O), tetB(P), tet(Q), tet(S), tet(W), tet(T), tet(32) and tet(36), the most common of which is tet(M) (Chopra & Roberts, 2001). One of the reasons for the success of these genes is the fact that they are commonly contained within conjugative transposons, which have an extraordinarily broad host range (Roberts, 2005). Enzymatic inactivation of tetracycline is encoded by two determinants similar in their functions: tet (X) and tet (37). The tet(X) gene was isolated from the anaerobic intestinal Bacteroides transposons (Speer et al., 1991) whereas tet(37) was cloned from the oral metagenome (Diaz-Torres et al., 2003). Both gene products chemically modify tetracycline in the presence of both oxygen and nicotinamide adenine dinucleotide phosphate (NADPH). However, to date no surveys have been conducted to assess the distribution of these two genes. A third mechanism of resistance to tetracycline is through the production of efflux proteins. They are encoded by different tetracycline resistance genes, tet(A) to tet(I) found mostly in Gram-negative species and tet(K) and tet(L) widely distributed among Gram-positive organisms. The efflux genes from Gram-negative organisms are normally associated with large plasmids, most of which are conjugative, whereas efflux genes from Gram-positive are generally found on smaller transmissible plasmids (Chopra & Roberts, 2001; Roberts, 1996).

Studies of the cultivable oral microbiota of humans have shown that the most frequently-encountered genes responsible for tetracycline resistance are tet(M), tet(O), tet(Q) and tet(W) (Lacroix & Walker, 1995; Olsvik et al., 1994, 1995; Lancaster et al., 2003; Villedieu et al., 2003). Interestingly, the first three of these were the only tetracycline resistance-encoding genes detected in all four metagenomic libraries in the present study. The results of this study have shown that it is possible to use a metagenomic approach to determine the antibiotic resistance potential of a microbial community, including that arising from the as-yet-uncultured organisms present, and also to identify the genes responsible for such resistance.

Acknowledgement

This project was supported by Medical Research Council Grant G9900875.

References

View Abstract