Conservation of transposon structures in soil bacteria


*Corresponding author. Tel.: +44 (151) 794 3620; Fax: +44 (151) 794 3655


The presence of Class II transposon genes related to Tn21 and Tn501, and their structural arrangements have been determined in a collection of 124 mercury resistant Gram-negative bacteria. Seventy-five of the 124 isolates contained a tnpA (transposase) gene related to Tn21 and Tn501 and in all 64 isolates that contained both tnpA genes and plasmids, the tnpA gene was plasmid borne. The relative orientation of the tnp genes and the mer operon (encoding mercury resistance) was also studied and revealed the presence of two distinct structural groups. The merC gene was present in 44 isolates. Five isolates were found to carry integrase genes and these contained inserted gene cassettes varying in size from 1.1 kb to 4.5 kb. The structural arrangement of the tnpA and tnpR (resolvase) genes within the isolates was determined. Sixty-nine of the 75 tnpA containing isolates had an arrangement of tnpA and tnpR genes similar to that found in the Tn21 subgroup of transposons. Four strains did not produce a PCR product using tnpR primers. The remaining two isolates had undetermined arrangements of tnpA and tnpR genes. No Tn3-like arrangements of tnpA and tnpR genes were present in these isolates, despite being detected in DNA extracted directly from the isolation sites. This suggests that Tn3-like arrangements of tnpA and tnpR genes are not commonly associated with mercury resistance genes in these environments. It was also apparent that the recombination events which have previously been observed in these strains have not significantly affected the diversity of the transposon structures within the isolates.


The study of genes contained within the indigenous bacterial community in the natural environment is important as it provides an opportunity to assess the effects of selective pressure, e.g. due to pollutants, on bacterial gene diversity and therefore facilitates the development of predictive tools [1–5]. A wide range of mechanisms exist that can alter genetic diversity. These include mutation, recombination, transposition, transformation and conjugation [1,6–10]. In natural environments, transposition mediated transfer of DNA represents a major mechanism for increased mobility of genes contained within the transposon and also a potential mechanism for genetic rearrangement [7,10,11].

There are four classes of transposable elements [11], amongst which the Class II transposons have been extensively studied [4,7,11–13]. The archetypal members of this class contain two genes, encoding functions involved in transposition: the transposase gene, tnpA, and the resolvase gene, tnpR[14–18]. A resolution (res) site is present in most Class II transposons, at which site specific recombination occurs in order to resolve the cointegrate transposon structures which arise during the transposition process. Several transposon structures have been described which have more diverse transposition mechanisms, including Tn4652 and Tn4430 which have different resolution systems, Tn2610 which has two functional copies of the tnpR gene and Tn5271 which has no resolution system [19–22].

Amongst the most widely studied Class II transposons are the mercury resistance elements Tn21 and Tn501[7,13,14,16]. Other members of this class differ from Tn21 and Tn501, not only in the transposon associated genes, but also in the structural arrangement of the transposition genes, most notably the orientation of the tnpA and tnpR genes, and the position of the res site [7]. The arrangement of these genes has important implications for the study of the evolutionary relationships between different transposons.

The structures of four transposons, Tn21, Tn501, Tn3 and Tn5036, are shown schematically in Fig. 1. The major difference lies in the arrangement of the tnpA and tnpR genes. The tnpA and tnpR genes in Tn3-like transposons are in a different orientation to that found in the Tn21 subgroup [7]. The res site of Tn3 lies between the tnpA and the tnpR gene, whereas in the Tn21-like elements the res site is outside both genes. The Tn3-like arrangement is also found on transposons such as Tn2501 and Tn1331[23,24]. Tn4556 is believed to contain a different arrangement of genes, with the res site being present between the tnpA and tnpR genes which are transcribed in the same direction [25].

Figure 1.

Diagrammatic representation of Tn501, Tn21, Tn3 and Tn5036 and approximate binding position of primers mercP, mercA, mercD, 501R1/C and 501R2/C (not to scale).

The potential for transposon mobility is increased if the transposon is present on a conjugative or mobilisable plasmid. Transposons found on plasmids include Tn21, Tn501, Tn917, Tn1721, Tn3926 and Tn5422[14,16,26–29].

Variation also exists amongst the transposon associated genes which encode non-essential functions that may confer a selective advantage to the host bacterium. Tn21 and Tn501 are members of a large group of related transposons that can confer resistance to mercurial compounds, and the mercury resistance (mer) operons of these elements show considerable variation, whilst retaining a number of common features [30,31]. Some genes (merR, merP, merT and merA) are present in the majority of mer operons which have been described to date [30,31]. However some genes are not, including merB (organomercurial lyase), merC (transport protein) and merF (function unclear) [30,31]. Fig. 1 shows that Tn21 has a merC gene inserted between the merP and merA genes [16]. The merF gene is also commonly found inserted at this point. In addition, integron elements have also been identified in a number of transposons. These structures insert and excise specific gene cassettes into a recombination hot spot (rhs) contained within the integron structure, effectively providing a means by which the transposon can acquire novel genetic material [10,32]. The number of bacteria in estuarine environments which contain integrons has been estimated at 5% (Young, H.K. and Rosser, S.J., personal communication). This represents an enormous potential for gene transfer in the natural environment.

Previous studies on a diverse collection of 39 Gram-negative mercury resistant bacteria ('93 isolates) have concentrated on the study of sequence diversity. The diversity of tnpA, tnpR and merR genes has been studied by RFLP analysis and by DNA sequencing [1,3,4,33]. A significant conclusion from these studies was that recombination between transposon genes and between transposon and mer genes, was common. Prior to this study no information was available regarding the structural diversity of the transposon genes carried in these isolates. The presence of plasmids had not been studied and as such, the location of the transposons contained within these isolates was also not known. In this study, the relationship between the transposition genes and their associated mer operons has been investigated further. The number of strains studied was expanded to include 85 new isolates (′96 isolates). These mercury resistant strains were cultured from the same sampling sites as the ′93 isolates using the same extraction protocol and were also Gram-negative. The identity of the ′93 isolates has been determined by Osborn et al. [33] using API. A range of species were identified including Enterobacter cloacae, Alcaligenes faecalis, Acinetobacter calcoceticus, Klebsiella oxytoca, Agrobacterium radiobacter, Aeromonas spp and Pseudomonas spp[33]. The identity of the ′96 isolates was not determined in this study. The presence of transposon associated genes has been studied in all 124 isolates, and PCR reactions have been carried out to determine the presence of integrons, and the size of any genes present within the rhs of those integrons. The presence and structural organisation of the mer genes in the isolates has been studied, as have the relative orientations of the tnp genes and mer operons in the isolates. The presence and approximate size of plasmids present within the isolates was also determined to ascertain whether any correlation existed between the structural diversity of the transposon and the plasmid diversity.

The primary aim of this study was to determine the presence and structural diversity of the tnpA and tnpR genes contained within the isolates and to seek evidence, or otherwise, of genetic rearrangements and/or recombination occurring in the transposon associated region.

2Materials and methods

2.1Extraction of isolates from soil

The 39 mercury resistant ′93 strains used in this study were previously isolated by Osborn et al. [33] from both polluted and pristine sites; these strains were selected by their ability to hybridise to a merRTΔP probe. All new isolates have been designated ′96 in order to distinguish them from the ′93 isolates. For the ′96 collection, bacteria were isolated from soil and sediment samples as previously described [33], and were selected for mercury resistance. Unlike the ′93 isolates, they were not selected for hybridisation to the merRTΔP probe. They were isolated from the following sites:

′96 SO strains were isolated from soil at a mercury polluted site at Fiddlers Ferry, Merseyside [33]. Sediment at this site was used to isolate the ′96 SE strains. ′96 SB strains were isolated from soil at pristine site at Salterbrook Bridge [33].

2.2Isolation of plasmid DNA

Plasmids were isolated from bacteria according to the method described by Olsen et al. [34]. DNA was visualised on agarose gels (0.7%) using TAE buffer (40 mM Tris, 20 mM acetic acid, 1 mM EDTA, pH 8.3), run at 40 V for approximately 24 h, followed by visualisation using ethidium bromide staining (1 μg ml−1).


PCR reactions were carried out under a variety of conditions. Typically 50 μl reactions were used containing 5 μl of 10×PCR buffer, 1.5 mM MgCl2 and 1.25 U of Taq DNA polymerase (GIBCO BRL). DNA template was prepared by the boiling method previously described [33]. Twenty pmol of each primer was added to the reaction mix along with dNTPs (Pharmacia Biotech) at a final concentration of 1 mM. Reactions were brought up to 50 μl using sterile distilled water and then overlaid with mineral oil (Sigma). PCR reactions were carried out in a Perkin Elmer 480 thermal cycler and were typically comprised of 4 min at 95°C, followed by 30 cycles of 95°C for 1 min, 62°C for 1 min and 72°C for 2 min. A final extension step of 10 min at 72°C was carried out before fragments were visualised on agarose gels. Different annealing temperatures were used depending on the primers used in each reaction.

To amplify regions of DNA longer than 3 kb, a long template PCR system was used (Expand PCR, Boehringer Mannheim) in accordance with the manufacturers' instructions [35].


The primers used in this study are shown in Table 1. The position of the primer within the accession number is also shown. All accession numbers listed are for Tn21 sequences, except tn3A and tn3R, which are shown for Tn3, and pos2000 and pos2400, which are shown for Tn4430.

Table 1.  Oligonucleotide primers used in this study
PrimerDNA sequenceAccession no.Position (bp)
  1. aThese four primers have been previously described [4], having been designed to complement the DNA sequences of the Tn21 and Tn501 tnpR and tnpA sequences respectively.

  2. bThis primer was designed to correspond to the sequence approximately 930 bp from the start codon of the tnpA gene of Tn501/Tn21.

  3. cint21A/int21B were used as an indicator of integrase gene presence and 4127/4128 allow the size of the gene cassettes within an integron to be determined (Young, H.K. and Rosser, S.J., personal communication).

  4. dThese primers have been designed to the tnpA and tnpR sequences of Tn3 respectively.

  5. eThese primers complement the DNA sequence in merP and merA respectively allowing the nature of any genes present inbetween merP and merA to be determined.

  6. fThis primer was designed to be homologous to merD Tn21/501 sequence.

  7. gThese primers were designed to consensus sequences from the tnpA genes of Tn917, Tn1546, Tn4430 and Tn5422[20,26,29,36].

  8. hThese primers were designed to consensus sequences from the tnpA genes of Tn1, Tn3, Tn21, Tn501, Tn1000, Tn1721, Tn2501, Tn3926, Tn5036 and Tn5401[14–16,23,27,28,38–41].

501R1/Ca5′-GTT CAG CA[GC] CTT CGA CCA G-3′X01298539–557
501R2/Ca5′-TA[CG] AGG GTT TC[GC] CG[AG] CTG AT-3′X012981026–1045
1406a5′-TGC GCT CCG GCG ACA TCT GG-3′X048911462–1481
2638a5′-TCA GCC CGG CAT GCA CGC G-3′X048912676–2694
950b5′-[TC]CT GGA ACT GCT GCT GAT GCT T-3′X04891987–1008
int21Ac5′-GTC AAG GTT CTG GAC CAG TTG C-3′M336331090–1111
int21Bc5′-ATC ATC GTC GTA GCG ACG TCG G-3′M33633219–240
4127c5′-TGA TCC GCA TGC CCG TTC CAT ACA G-3′M33633709–733
4128c5′-GGC AAG CTT AGT AAA GCC CTC GCT AG-3′X128702321–2346
tn3Ad5′-GTA TCA GCG CTG CAT GCT CAC-3′V006132721–2741
tn3Rd5′-CCC TGC ATC TTT GAG CGC TCT-3′V006133270–3290
mercPe5′-CCC GAT CAC [AT]GT CAA G[AC]A [ACG]GC-3′K030891075–1095
mercAe5′-CGC TCG ATC AGC G[AT]G AC[ACG] [CT]G-3′K030892116–2135
mercDf5′-GTT CGT CGA GCG TCG GCG-3′K030893733–3750
pos2000g5′- GGA ATG AAT ATT GTT CTT ACC AAA ATG-3′X134813466–3493
pos2400g5′-CAG TAT AGC CAG CTG TGT CTG-3′X134813445–3466
2501h5′-CAT TGG GAC GAG ATG ATG CGG-3′X049812490–2501
2850h5′-GCT CCA TAT ACA CCG TGT TCC-3′X049812829–2851

2.5Southern blotting

Both Southern blots and dot blots were utilised in this study, depending on the nature of the DNA sample being studied. Dot blots were carried out using a Bio-Rad Bio-Dot vacuum manifold in accordance with manufacturers' instructions. DNA samples for both dot blots and Southern blots were transferred onto positively charged membranes (Appligene) in accordance with manufacturers' instructions. Overnight transfer of DNA was set up using 0.4 M NaOH as the transfer buffer. Southern blots of plasmid DNA differed in that the transfer time was extended to 48 h. Prehybridisation of the membrane in both cases was carried out in accordance with the manufacturers' instructions.

PCR products were used as probes. These were prepared for hybridisation by an initial electrophoresis step using low melting point agarose (NuSieve GTG) after which the band was excised using a scalpel and the probe resuspended in sterile distilled water. Probes were then radiolabelled with 32P using a random prime labelling kit (Boehringer Mannheim) and the probe purified on a Sephadex G50 column. After overnight hybridisation at 65°C, the membrane was washed twice in 2×SSC for 5 min, twice in 2×SSC/1% SDS for 15 min and once in 0.1×SSC for 15 min. Signal detection was carried out on a Molecular Dynamics STORM 860 phosphoimager.

2.6Restriction digests

The merC probe used in this study was prepared by digestion of the Tn21 mercP/mercA PCR product with HaeII and StyI to produce a fragment of 500 bp. This fragment represents the merC gene and 80 bp of extra sequence at the 5′ end of the gene. Both restriction enzymes (GIBCO BRL) were used in accordance with manufacturers' recommendations.

3Results and discussion

3.1Presence of tnpA genes

One hundred and twenty-four Gram-negative mercury resistant strains were examined for the presence of tnpA genes using two sets of primers: pos2000/pos2400 and 2501/2850. Primers pos2000 and pos2400 were designed to allow the detection of the tnpA genes of Tn917, Tn1546, Tn4430 and Tn5422, which were originally characterised in Gram-positive bacteria [20,26,29,36]. No strain produced a PCR product using these primers, despite allowing tnpA amplification from control strains. Given that these primers were designed to amplify Gram-positive transposon sequences, the lack of amplification from Gram-negative strains is perhaps unsurprising. PCR reactions carried out on DNA extracted directly from the soil (SO) and sediment (SE) sites yielded products using these primers (data not shown) [37]. The presence of Gram-positive tnpA sequences at these sites was confirmed by DNA sequencing (data not shown).

Primers 2501 and 2850 were designed to allow the detection of the tnpA genes of Tn1, Tn3, Tn21, Tn501, Tn1000, Tn1721, Tn2501, Tn3926, Tn5036 and Tn5401[14–16,23,27,28,38–41]. Seventy-five strains produced a PCR product using these primers (Table 2). The presence of tnpA genes in 22 out of 30 ′93 SO, SE and SB isolates had been previously demonstrated by hybridisation to tnpA probes from both Tn21 and Tn501[4]. Those isolates previously identified as containing tnpA sequences by probe hybridisation, also produced PCR products using primers 2501 and 2850 [4]. To allow the comparison of the ′93 and ′96 isolates, PCR reactions were carried out on all 124 isolates using the primers 1406 and 2638, used in the characterisation of the ′93 isolates [4]. All 75 strains produced a PCR product using these primers (Table 2). These PCR products hybridised to the corresponding tnpA PCR products from both Tn21 and Tn501.

Table 2.  Presence of tnpA genes and arrangements of tnpA and tnpR genes
Isolation groupNumber in isolation grouptnpA genetnpA/tnpR arrangement
   Tn21/Tn501No tnpRUnknown
′93 SO1010910
′93 SE106600
′93 SB106600
′93 T299900
′96 SO30202000
′96 SE31221822
′96 SB242110

The high number of isolates in this study containing transposase genes compared to similar studies on marine and clinical isolates [12,13] may have been caused by the initial selection for resistance to HgCl2. The higher numbers of tnpA containing isolates observed in the ′93 isolates may be explained by their initial selection for hybridisation to a mer probe [33]. Forty-nine of the strains did not produce a tnpA PCR product, 22 of which were in the ′96 SB group. This may indicate sequence variation at the primer annealing sites or suggest the lack of this gene in these isolates.

3.2Plasmid extraction

Plasmid extractions were carried out on the collection of 124 isolates, and the samples were analysed on agarose gels to determine the size of any plasmids present. Results are shown in Fig. 3. One hundred of the isolates were found to contain plasmids, with both large plasmids (>50 kb) and small plasmids (<20 kb) being identified. Of the 75 strains containing tnpA genes, 64 were found to contain plasmids, while 11 apparently did not. Whilst large and small plasmids were observed in both the ′93 and ′96 collections, the ′96 isolates contained a higher proportion of smaller plasmids (data not shown). No incompatibility group data are available for these plasmids, but a previous study indicates that mercury resistance plasmids isolated from the environment do not conform to existing incompatibility groupings [42].

Figure 3.

Presence of plasmids and location of tnpA genes. Numbers in brackets indicate the number of strains.

3.3Location of tnpA genes

The location of the tnpA genes contained within the 75 positively amplifying isolates was determined by Southern hybridisation (Fig. 3). In all 64 plasmid containing strains, the tnpA gene was located on a large plasmid. In the 11 apparently plasmid-free strains, the tnpA gene was detected in the chromosomal material. Chromosomally located tnpA genes were only found in those isolates which appeared to contain no plasmids.

3.4Structural arrangement of tnpA and tnpR genes

The structural arrangement of the tnpA and tnpR genes was determined by the use of PCR. Of the four possible arrangements of these genes indicated in Fig. 2, the most commonly encountered was that of the Tn21-like elements [7]. Initially, PCR reactions were carried out using 950 and 501R1/C primers [4], corresponding to the arrangement of tnpA and tnpR genes found on Tn21-like elements (Table 2). Of the 75 isolates that produced a tnpA PCR product, 69 of these yielded an arrangement of genes similar to that found in the Tn21 subgroup of transposons, this fragment being of the size expected from these transposons. The nature of these PCR products was verified by their hybridisation to tnpA/tnpR PCR products from Tn21 and Tn501.

Figure 2.

Schematic representation of possible arrangements of tnpR and tnpA genes and appropriate primer binding sites (not to scale).

The six strains which produced no PCR products were then examined using four combinations of primers to determine whether any other arrangements of genes were present. This was carried out using the following primer combinations: 501R1/C and 950, 501R1/C and 1406, 501R2/C and 950, and 501R2/C and 1406, thus covering all four possible arrangements of genes as indicated in Fig. 2. No PCR products were produced using these combinations of primers. The six strains were also subjected to PCR using primers tn3A and tn3R, which were designed to Tn3 tnpA and tnpR genes respectively. No strains had Tn3-like gene arrangements. PCR products were however amplified using primers tn3A and tn3R on DNA extracted from these soil (SO) and sediment (SE) sites indicating that Tn3-like elements were present (data not shown) [37]. These data suggest that Tn3-like arrangements of tnpA and tnpR genes are present in these environments but are not commonly associated with mer operons. However as the Tn3-like arrangement of tnpA and tnpR genes was detected in DNA extracted directly from the environment, it was not possible to determine the nature of the genes associated with this arrangement.

Subsequently the presence of tnpR genes in those isolates not producing any tnpA/tnpR PCR products was determined by PCR using primers 501R1/C and 501R2/C. Four strains produced no PCR products using these primers, suggesting that these strains either have no tnpR gene or that the specificity of the primers used is such that they did not allow the detection of diverse genes (Table 2). Two strains, ′96 SE19 and ′96 SE30 had both tnpR and tnpA genes, in undetermined configurations. Expanded PCR using all five sets of primers yielded no PCR products for these two strains. All strains producing a tnpA PCR product have had their tnpA/tnpR arrangement determined except for ′96 SE19 and SE30. These two isolates may contain transposon structures which are distinct in evolutionary terms from the Tn3 group of Class II transposons, despite having gene sequences which are similar to the other transposons studied.

3.5Presence of integrase genes and size of inserted gene cassettes

The presence of integrase genes serves as an indicator of the presence of integron elements within the transposon. PCR using primers int21A and int21B, designed to allow amplification from a wide range of integrase genes (Young, H.K. and Rosser, S.J., personal communication) was carried out on the collection of isolates. Five of the 124 strains produced PCR products of the correct size which hybridised to the corresponding PCR product from Tn21 (Table 3). The size of the gene cassettes inserted into the rhs of the integron elements was determined by PCR using primers 4127 and 4128 (Young, H.K. and Rosser, S.J., personal communication). Insert size varied between 1.1 kb and 4.5 kb. Although the nature of these inserts is currently undetermined, such cassettes commonly encode antibiotic resistance genes. With the exception of ′93 T2 37, all the isolates found to contain integron structures were members of the ′96 SE group, totalling 13% of those isolates. This is in contrast to the ′93 SE group in which no PCR products were seen using these primers.

Table 3.  Integron containing isolates
StrainIntegrase geneaIntegron insert sizeb
  1. a+, integrase gene present; −, no PCR product.

  2. bn.a., not applicable.

′93 T2 37+1.2 kb
′96 SE6+1.1 kb
′96 SE9+1.4 kb
′96 SE15+4.5 kb
′96 SE19+1.3 kb
Other 119 isolatesn.a.

3.6Relative arrangement of merD and tnpR genes

The arrangement of mer and tnp genes was studied in the 75 isolates that positively amplified for tnpA by PCR using primers mercD and 501R2/C, corresponding to the Tn21/501 orientation of the merD and tnpR genes. All strains failing to produce a PCR product were tested using primers mercD and 501R1/C in order to ascertain whether any structural diversity (i.e. inversion of tnpR) was present in the isolates. No such arrangements were observed (Table 4).

Table 4.  Results of merD/tnpR PCR and merC PCR
Isolation groupmerD/tnpR PCRmerC PCR
 2 kb1.3 kbNo productYesNoNo productUnknown
  1. merD/tnpR PCR was carried out on the 75 strains which positively amplified using tnpA primers.

  2. merC PCR was carried out on all 124 strains.

  3. a′96 SE9 produces both 2 kb and 1.3 kb PCR products.

′93 SO90110000
′93 SE0334051
′93 SB5100550
′93 T20455022
′96 SO002000300
′96 SE1a9a13162130
′96 SB00290150

Reactions using the Tn21/Tn501-like primers yielded two sizes of product from different isolates: a 2 kb product, which is the expected size from Tn501 and a 1.3 kb product. The PCR product from Tn501 hybridised to all the 2 kb PCR products indicating that the gene arrangement in these strains was similar to that of Tn501. The DNA sequence of the 1.3 kb PCR product from ′96 SE13 was determined (Fig. 4) and this sequence has been assigned accession number AF134211. This indicated the presence of a region of DNA corresponding to that found between the tnpR and merD genes of Tn3926, Tn5036, Tn5059 and pMER610, which contains three open reading frames of unknown function [28,40,43]. DNA sequence information was used to obtain a probe from the sequenced PCR product, corresponding to the region of DNA unique to these transposons. This was used to ascertain that the region of DNA between the tnpR and merD genes in all the strains producing a 1.3 kb PCR product was of a similar nature. One isolate, ′96 SE9 was seen to produce both sizes of fragment. This may represent two distinct mer operons contained within this strain.

Figure 4.

Nucleotide sequence of region spanning merD and tnpR genes in ′96 SE13.

The 44 strains which contain a tnpA gene, but which did not produce a PCR product using the tnpR/merD primers, may contain a mer operon which is not associated with the tnpA gene, or alternatively this may be due to sequence diversity at the primer binding sites.

3.7Presence of merC gene

Using PCR primers mercP and mercA, the presence and size of any genes contained within the merP/merA interval was determined for all 124 isolates. PCR reactions using these primers produced two sizes of product using Tn21 or Tn501 as templates. Tn21 yielded a 1 kb PCR product due to the presence of the merC gene, and Tn501 a 600 bp product.

PCR products were obtained from 54 of the isolates (Table 4) and all were found to hybridise to both Tn21 and Tn501 mercP/mercA PCR products. To distinguish between those strains which produce a PCR product containing merC and those which do not, PCR products of both sizes were hybridised to a merC gene probe. Forty-four of the 47 1 kb PCR products hybridised to the merC probe, i.e. they contained a merC gene. The three strains that did not contain merC genes, ′93 SE31, T2 37 and T2 38, apparently contain a gene between the merP and merA genes that is of similar size to merC, but which remains of an undetermined nature. Seven isolates were found to produce a 600 bp PCR product, which did not hybridise to the merC probe and were therefore assumed to be Tn501-like.

All isolates in the ′93 SO, SE and T2 group which produced a PCR product (except the three strains with unknown inserts) contained a merC gene, whereas all the ′93 SB isolates were Tn501-like, containing no genes between merP and merA. The ′96 SO isolates produced no PCR products, which correlates well to the merD/tnpR data for this group (data not shown). This suggests that the observed mercury resistance of these isolates may not be due to archetypal Gram-negative mer genes, i.e. mercury resistance may be conferred by a non-mer operon system or by a mer operon with a sequence divergent from that of Tn21 and Tn501.

All nine of the ′96 SB strains which produced a PCR product contained the merC gene, as did all of the ′96 SE strains except ′96 SE26 and SE28. This is interesting as the ′96 SB group did not produce a merD/tnpR PCR product. This may be due to the mer genes being located at a position removed from the transposase gene or contained on non-transposon structures. There is also a difference in the SB strains in that the ′93 SB isolates which gave a PCR product all contained a Tn501-like mer operon while the ′96 SB isolates which gave a PCR product, all contained Tn21-like mer operons containing merC. The observed frequency of merC genes is higher than previously described and merC genes are seen here in a wider range of bacterial species and plasmids [44].


The majority of isolates in this study contained arrangements of tnpA and tnpR genes similar to that found in the Tn21 subgroup of bacteria. Those strains which differed from this basic gene arrangement were ′96 SE19 and SE30 which had undetermined arrangements of tnpA and tnpR genes, the four strains which may not contain a tnpR gene and those strains which did not produce PCR products with the primers used. The majority of mer operons contained in the isolates fell into two major structural groups, the Tn21/501-like structures and the shorter Tn3926-like structures. The differences observed between the ′93 and ′96 isolates may be due to the selection of the ′93 isolates by their ability to hybridise to a mer probe; such a selection was not carried out on the ′96 isolates. This may explain the higher numbers of strains in the ′96 groups which do not appear to contain mer operons related to those of Tn21-like transposons [33]. Such novel transposon genes may not be detectable using the PCR primers employed in this study.

This study shows that the predominant transposon gene structures contained in a collection of mercury resistant isolates were Tn21-like, whereas the region of DNA between the mer operon and the transposase genes fell into two structural groups. It is interesting to note that these isolates show a distinct lack of structural diversity compared to that which might be expected from such a study if genetic recombination and rearrangement were common place. The Tn3-like arrangement of tnpA and tnpR genes was not observed in any of the strains, despite being detected in DNA extracted directly from the soil (SO) and sediment (SE) sites, suggesting that Tn3-like structures are not associated with mercury resistance genes in this environment. However, if recombination is frequent within the bacterial community, it might be expected that recombination between Tn3-like transposons and Tn21-like transposons would give rise to a mercury resistance transposon containing a Tn3-like arrangement of tnpA and tnpR genes. The position of the res site in members of the Tn21 subgroup of transposons is such that it may allow recombination to occur with greater frequency between transposition genes and the genes with which they are associated, whereas the Tn3-like arrangement may favour recombination between the transposition genes themselves [7]. This may explain the high frequency of Tn21-like mercury resistance transposons in this environment.

The genetic diversity of the ′93 isolates has previously been studied in detail [1,3,4,33]. The sequence diversity of both tnpA and tnpR genes has been studied by RFLP and DNA sequencing [1,4]. The previous RFLP study carried out by Pearson et al., 1996, indicated the presence of three classes of tnpR genes and six classes of tnpA genes within the strains isolated from soil and sediment. There was no observed linkage between different classes of the two genes, suggesting that recombination is frequent between the tnpA and tnpR genes and between mer and tnp genes within the ′93 isolates [4]. This compared with the data presented in this study suggests that recombination may have occurred only between closely related transposons and that this has not significantly affected the actual structural arrangement of the genes.


This work was funded by NERC. R.J.H. is supported by a NERC Ph.D. studentship, (Ref No GT4/95/173/T) and K.D.B. by a NERC fellowship, (Ref No GT5/94/TLS). This work was also partly supported by a NERC grant awarded to P.S and D.A. Ritchie (Ref No GR3/09502). This work benefited from the use of the SEQNET facility, Daresbury. We would like to thank Dr Hillary K. Young for her assistance with the integron work, Dr Paul Eggleston for his critical reading of the manuscript, Angela Rosin for DNA sequencing and Prof. D.A. Ritchie for his continued support of this work.