Comparative analysis of antibiotic resistance gene markers in Mycoplasma genitalium: application to studies of the minimal gene complement

Mycoplasmas belong to the class Mollicutes, a wide group of micro-organisms closely related to the Gram-positive bacteria. Mycoplasmas have small circular genomes, with Mycoplasma genitalium (517 genes) being the free-living organism with the smallest gene complement so far described (Fraser et al., 1995). They have the fewest metabolic pathways described (Pollack et al., 1997) and also exhibit a very reduced biosynthetic capacity which forces them to obtain almost all metabolites from the external environment. Therefore, mycoplasmas are parasites of a wide range of hosts and usually their culture is fastidious in liquid as well as on solid media.

The genetic manipulation of mycoplasmas is hindered by the limited number of genetic tools available. The use of replicative plasmids in mycoplasmas is restricted to Mycoplasma pulmonis (Cordova et al., 2002), Mycoplasma capricolum (Lartigue et al., 2003) and Mycoplasma mycoides (Bergemann et al., 1989; King & Dybvig, 1992). Transposons have become commonplace in mycoplasma genetics. However, there are several mycoplasmas that remain refractory to transformation by transposons. Transposons available for mycoplamas are Tn4001 and Tn916, which were originally isolated from the Gram-positive bacteria Staphylococcus aureus (Lyon et al., 1984) and Enterococcus faecalis (Franke & Clewell, 1981), respectively. The usefulness of Tn4001 (4·7 kb) was first demonstrated in Mycoplasma pneumoniae (Hedreyda et al., 1993) and was then successfully tested in other Mycoplasma species including Mycoplasma gallisepticum (Cao et al., 1994) and Mycoplasma genitalium (Reddy et al., 1996). In contrast to Tn916 (18 kb), Tn4001 is small enough to be used as a routine cloning vector. However, one of the problems with the use of native transposons is their instability and dynamism once transposed, which makes clear interpretation of the results difficult. This problem was solved by the construction of a minitransposon based on Tn4001, which places the tnp gene outside the inverted repeats (Pour-El et al., 2002).

The tetM and the aac(6')-aph(2'') genes from Tn916 and Tn4001, respectively, have been used as selectable markers in different Mycoplasma species (Dybvig & Voelker, 1996). The tetM gene confers tetracycline (Tc) resistance upon all tested Mycoplasma species. However, in studies focused on M. genitalium, the aac(6')-aph(2'') gene is the only selectable marker used for transformation (Dhandayuthapani et al., 1999, 2001). The aac(6')-aph(2'') gene confers resistance to the aminoglycosides gentamicin (Gm), kanamycin and tobramycin. AAC(6')-APH(2'') is a unique bifunctional enzyme with two different domains: the N-terminal domain has acetyl-CoA-dependent aminoglycoside acetyltransferase activity, whereas the C-terminal domain exhibits aminoglycoside kinase activity (Culebras & Martinez, 1999). It has been suggested that AAC(6')-APH(2'') can also phosphorylate several eukaryotic and prokaryotic protein kinase substrates, modifying in this way signal transduction or regulatory pathways (Daigle et al., 1999).

M. genitalium has been proposed as a suitable model to achieve an in-depth understanding of the biology of a free-living organism (Roberts, 2004). Fulfilling this goal depends on the development of new genetic tools and, in particular, on the identification of new selectable markers for M. genitalium genetic studies. In this work we describe the construction of a modified tetM gene (tetM438) that can be used in M. genitalium. Comparison between the tetM438 and aac(6')-aph(2'') genes clearly shows a negative effect of aac(6')-aph(2'') on the growth of M. genitalium. Moreover, the use of tetM438 increases dramatically the number of transformants obtained and reduces considerably the time needed for colonies to become visible. Using tetM438 in conjunction with a minitransposon derived from Tn4001 (MTn4001), we have been able to obtain insertions in genes which were previously considered to be necessary for M. genitalium growth in laboratory conditions (Hutchison et al., 1999).

Culture conditions, plasmids and primers.
Escherichia coli strain XL-1 Blue was used for plasmid amplification. It was grown at 37 °C in 2YT broth or LB agar plates containing 75 µg ampicillin ml¹ with X-Gal (40 µg ml¹) and IPTG (24 µg ml¹) when needed. All plasmids and primers used in this work are summarized in Table 1 and Table 2, respectively.

Table 1. Plasmids

Table 2. Primers Bold indicates the restriction sites introduced at the 5' end of selected primers. Italic indicates the 22 nt from the putative mg438 gene promoter region. Underlining indicates the three stop codons present in the inverted repeats to prevent translation of the disrupted genes (Byrne et al., 1989).

Wild-type M. genitalium strain G37 was grown in SP-4 medium (Tully et al., 1979) at 37 °C under 5 % CO₂ in tissue culture flasks (TPP). Gm- and Tc-resistant strains were selected on SP-4 agar plates supplemented with Gm 100 µg ml¹ (Invitrogen) or Tc 2 µg ml¹ (Roche). We previously determined that growth of strain G37 was completely abolished at Tc concentrations above 0·5 µg ml¹. Efforts were made to minimize light exposure when Tc-supplemented SP-4 medium was used.

Transformation of M. genitalium.
This was performed as described by Reddy et al. (1996), with a few modifications. Briefly, M. genitalium strain G37 was grown to mid-exponential phase in 75 cm² tissue culture flasks. Attached mycoplasmas were scraped off and passed through a 0·45 µm low protein binding filter (Millipore). The filtered cells were then recultured for 24 h in 40 ml SP-4 medium in 150 cm² tissue culture flasks. Attached mycoplasmas were washed three times with electroporation buffer (8 mM HEPES pH 7·2, 272 mM sucrose), scraped off and resuspended in this buffer at a concentration of approximately 10⁹ cells ml¹. Then 90 µl mycoplasma cell suspension was mixed with 5 µg plasmid DNA previously dissolved in 20 µl electroporation buffer. The mixture was transferred to a 2 mm gapped electroporation Plus BTX cuvette, kept on ice for 15 min and then electroporated (2·5 kV, 129 Ω) using an electro cell manipulator 600 (BTX). After 15 min on ice, 900 µl SP-4 was added and the cells were incubated at 37 °C for 2 h. Aliquots of 200 µl were spread onto Gm- or Tc-supplemented SP-4 agar plates. Isolated colonies were picked, propagated in 5 ml cultures and stored at 80 °C. Several 20 µl drops from serial dilutions of the electroporated cells were spotted on SP-4 agar plates to determine the number of viable cells. The same procedure was used to determine the number of transformant cells, except that SP-4 agar plates were supplemented with Gm or Tc. Colonies from eight spots on each of two different plates were counted to determine the number of viable or transformant cells.

DNA manipulations.
General DNA manipulations were performed according to Sambrook & Russell (2001). Plasmid DNA was obtained by using the Fast Plasmid Mini Eppendorf Kit. All PCR products were previously cloned into EcoRV-digested pBE and then excised with the corresponding restriction enzyme (Roche). PCR products and digested fragments were purified from agarose gels using the EZNA gel extraction Kit (Omega Bio-tek).

Plasmid pBSKII+ was used for the construction of pMTn4001. First, the tnp gene was amplified from pIVT-1 by PCR using primers tnp5' and tnp3', which include BsiWI and SpeI restriction sites at the 5' and 3' ends of the PCR product, respectively. Then, the 1·3 kb PCR product was digested with BsiWI/SpeI and included in a ligation mixture containing Acc65I/ApaI-digested pBSKII+ and 50 pmol of both IRO-1 and IRO-2 oligonucleotides. In this way, the annealed oligonucleotides recreate an outer inverted repeat adapter. The construction obtained (pRP1) was digested with NotI and SacI and included in a ligation mixture containing 50 pmol of both IRI-1 and IRI-2 oligonucleotides to reconstitute an inner inverted repeat adapter. One clone with the expected restriction pattern was sequenced to confirm that no mutations were introduced by these procedures. Most of the initial restriction sites of the pBSKII+ multicloning site (from ApaI to NotI) remain in pMTn4001.

The aac(6')-aph(2'') gene was amplified by PCR from pIVT-1 by using aac-aph-5'3' as a single primer. The resulting 2·5 kb fragment was digested with BamHI, purified and ligated into BamHI-digested pMTn4001, creating plasmid pMTnGm. For the construction of pMTnORFTet, the tetM coding region was PCR amplified from pAM120 by using primers ORFTc-5' and ORFTc-3', which include BamHI sites at the ends. The 2 kb PCR product was cloned into BamHI-digested pMTn4001. Finally, to construct pMTnTetM438 the tetM coding region was amplified from pAM120 using primers TAGAATet-5' and ORFTc-3'. The TAGAATet-5' oligonucleotide creates a fusion of the tetM coding region and the 22 bp region located upstream of the mg438 translational start codon. The 2 kb PCR product was cloned into EcoRI/BamHI-digested pMTn4001 to obtain plasmid pMTnTetM438.

Assay to compare the growth of MTnGm and MTnTetM438 transformants.
To monitor and compare the growth of MTnGm- and MTnTetM438-transformed cells, we quantified both colony number and size at different days after electroporation. Because colony size is strongly reduced when plates are crowded, a low number of MTnGm- and MTnTetM438-transformed cells (100500 c.f.u. per plate) was spread to make colony size measurements comparable. Thus, while cells electroporated in the presence of pMTnGm were spread undiluted onto several Gm-supplemented SP-4 agar plates, a 10² dilution of the cells electroporated in the presence of pMTnTetM438 was spread onto Tc-supplemented SP-4 agar plates. To reduce inter-experiment variability, electroporation of either minitransposon derivative was performed sequentially by using aliquots of the same batch of cells and the same medium stock. Additionally, duplicate electroporations were performed to reduce intra-experimental variability. As Tc is readily inactivated in the presence of light, plates were not returned to the incubator once examined. Thus, the same transformation was spread on different plates and a single plate from each transformation experiment was removed from the incubator for examination at 10, 12, 14 and 16 days after electroporation. Several pictures were taken of each plate by using a LeicaMZFLIII microscope and a LeicaDC500 camera. The diameter of 100 colonies was measured by using the Scion Image processing and analysis program to obtain the colony diameter average (CDA).

Assay to determine the effect of Gm or Tc on the colony size of MTnGm or MTnTetM438 transformants.
M. genitalium cells were electroporated in the presence of pMTnGm or pMTnTetM438 and the resultant transformants were grown in liquid medium supplemented with Gm or Tc respectively. Then, transformants were scraped off, passed through a 0·22 µm filter and plated on SP-4 agar with or without antibiotic. The CDA of each sample was obtained after 14 days of incubation. Wild-type cells were also plated to obtain a reference CDA.

Southern blots.
Eight MTnTetM438 transformants were selected and cultured in 20 ml Tc-supplemented SP-4 medium. Cells were then scraped off and genomic DNA was isolated by using the EZNA Bacterial DNA Kit (Omega Bio-tek). Genomic DNAs were digested with HindIII and probed by Southern blot hybridization by using the Dig DNA Labelling and Detection Kit (Roche). A 2 kb BamHI fragment from pMTnORFTet containing the tetM coding region was used as a probe.

RNA manipulation.
Total RNA was isolated from 20 ml cultures using TRI Reagent (Invitrogen), following the recommendations of the manufacturer. For RT-PCR assay, total RNAs were retrotranscribed by using random hexamers and the SuperScript First-Strand Synthesis system (Invitrogen). PCRs were then performed by using the primers listed in Table 2.

Primer extension of mg438 was performed by annealing 2 pmol 5'-Cy5-labelled MG438PE primer with 5 µg total RNA. First-strand synthesis was carried out as described above. Eight microlitres of the preceding reaction were analysed in an ALF DNA sequencer (Pharmacia Biotech). The product of the standard sequencing reactions, using the same 5'-Cy5-labelled primer and pENT438 plasmid as DNA template, was included in the same sequencing gel.

Genomic DNA sequencing.
Genomic DNA from 30 independent MTnTetM438 transformants was isolated as described above. Sequencing with fluorescent dideoxynucleotides was performed by using the Big Dye 3.0 Terminator Kit (Applied Biosystems) and Tc upstream and Tc downstream primers, following the recommendations of the manufacturer, and analysed in an ABI 3100 Genetic Analyser (Applied Biosystems).

tetM expression analysis in M. genitalium
Since the usefulness of the tetM marker has been demonstrated for several mycoplasma species, we assayed the functionality of this gene in M. genitalium. For this purpose, several Tn4001T tranformants were obtained by electroporation in the presence of pIVT-1. This plasmid bears both the aac(6')-aph(2'') and the tetM markers (Dybvig et al., 2000). Transformants were selected in the presence of Gm. The resulting Tn4001T transformants were then spread onto SP-4 agar plates supplemented with different concentrations of Tc; no growth was observed even at the lowest Tc concentration. To test whether the absence of Tc resistance was due to a defect in transcription of the tetM, total RNA from several Tn4001T transformants was isolated and RT-PCR assays were performed (Fig. 1a). As a RT-PCR positive control we used a 0·5 kb internal fragment from the mg438 (Fig. 1b), a gene of constitutive expression recently characterized in our laboratory (data not shown). As expected, amplification of the aac(6')-aph(2'') cDNA (1·5 kb) was clearly detected. However, no amplification was observed for the tetM cDNA (2 kb), suggesting the absence (or a very low level) of tetM transcription in M. genitalium. To further assess the functionality of tetM in M. genitalium, a construct based on the tetM coding region under the control of the mg438 putative promoter (tetM438) was developed.

(19K): Fig. 1. Transcriptional analysis by RT-PCR of the aac(6')-aph(2''), tetM and mg438 genes in M. genitalium. (a) Total RNA was isolated from a Tn4001T transformant culture. Lanes 1 and 4 are the RT-PCR negative controls for aac(6')-aph(2'') and tetM, respectively. Lanes 2 and 5 are the RT-PCRs performed for aac(6')-aph(2'') and tetM, respectively. Lanes 3 and 6 are the respective PCR positive controls performed with the corresponding genomic DNAs. (b) Total RNA was isolated from a wild-type strain culture. Lanes: 1, mg438RT-PCR negative control; 2, the RT-PCR performed for mg438; 3, the PCR positive control performed with genomic DNA.

Determination of the mg438 transcriptional start point
To identify the mg438 promoter region, we determined its transcriptional start point by primer extension (Fig. 2). For this purpose, the primer MG438PE, which anneals 200 bp downstream of the mg438 translational start codon, was used. In accordance with previous transcriptional studies performed in M. pneumoniae (Weiner et al., 2000) and M. genitalium (Musatovova et al., 2003), heterogeneous transcriptional start points were observed. Bases at the mRNA 5' end were A, A, T and A, which were located at 3, 5, 6 and 7 bp upstream from the translational start codon, respectively. Analysis of the region located immediately upstream of the determined transcriptional start points allowed us to identify two putative 10 boxes: TAGTAT and TAGAAT. These 10 boxes are in agreement with the consensus previously described for M. pneumoniae (Weiner et al., 2000). Neither a 35 box nor a canonical RBS could be identified. The lack of a 35 box seems a major feature of mycoplasma promoters (Weiner et al., 2000). Transcription of genes lacking a 35 box has been reported in other bacteria (Sabelnikov et al., 1995). However, these genes have an extended 10 region that can not be found in the region immediately upstream of the mg438 coding region. The presence in M. genitalium of regulatory elements other than 10 boxes remains to be investigated.

(26K): Fig. 2. Determination of the mg438 transcriptional start points. (a) Standard ALF output corresponding to the sequencing reaction. The sequence shown is the reverse and complementary to the mg438 coding sequence. (b) Chromatogram obtained with the primer extension sample. The two profiles are aligned according to their respective running time. The two putative 10 regions found are boxed and the translational start codon is underlined.

Construction and functionality of the tetM438 marker in M. genitalium
To avoid the genomic instability of Tn4001, we constructed a minitransposon named pMTn4001 (Fig. 3) with a structure very similar to that described by Pour-El et al. (2002). First, we developed a pMTn4001 derivative (pMTnGm) harbouring the aac(6')-aph(2'') marker to test the functionality of MTn4001. A second construction (pMTnTetM438), harbouring the tetM438 marker, was also developed. The tetM438 marker is a fusion between the 22 bp region located immediately upstream of the mg438 translational start codon (including the two putative promoters identified) and the tetM coding region. A third construction (pMTnORFTet) based on pMTn4001 harbouring only the tetM coding region was designed to exclude the presence in pMTn4001 of any sequence able to promote transcription of the tetM coding region in M. genitalium. Initial electroporation studies were performed using the three plasmids described above. Transformation efficiencies obtained with pMTnGm and pMTnTetM438 were 5x10⁵ and 1x10³ per viable cell, respectively. As expected, no transformants were obtained when M. genitalium cells were electroporated in the presence of pMTnORFTet. This result confirms that the 22 bp selected region promotes transcription of the tetM438 marker in the MTnTetM438 minitransposon and also the effective translation of the resulting transcript. Thus, this result suggests the presence in this 22 bp region of a RBS different from the canonical one complementary to 16S RNA or alternatively the possibility that mycoplasma ribosomes could initiate translation through a 5' mRNA end in a way similar to eukaryotes, as has been already proposed (Weiner et al., 2000).

(18K): Fig. 3. Schematic representation of plasmid pMTn4001, showing the multicloning site with the available restriction enzyme sites. The plasmid skeleton is based on the pBSKII+ (bla, β-lactamase; f1(+) ori, single-stranded replication origin; ColE1 ori, vegetative replication origin; tnp, transposase from Tn4001). The boxes show the marker genes cloned in the multicloning site to obtain pMTnGm, pMTnTetM438 and pMTnORFTet respectively.

In addition to the surprisingly higher transformation efficiency of MTnTetM438, colonies derived from this minitransposon were also noticeably larger than those derived from MTnGm. To monitor these differences, a new set of electroporation experiments was devised to carefully quantify both colony number and size at different days after electroporation. The transformation efficiency obtained for each plasmid in such a set of experiments is shown in Table 3. The earliest colonies derived from MTnTetM438 transformants were observed 10 days after plating, while no colonies derived from MTnGm transformants were detected until day 12 (Fig. 4). The number of MTnTetM438 transformants obtained at days 12, 14 and 16 was 50-, 35- and 25-fold higher, respectively, than for MTnGm transformants. This result confirms that the number of transformants obtained with pMTnTetM438 is significantly higher than that obtained with pMTnGm.

Table 3. Transformation efficiencies obtained for pMTnTetM438 and pMTnGm Values represent the mean of multiple platings (see Methods) of one representative transformation assay.

(36K): Fig. 4. Colonies derived from MTnTetM438 (top row) or MTnGm transformants (bottom row) after 1016 days of growth. Bar,1 mm.

The average diameter of colonies derived from each minitransposon at different days after electroporation was determined. The CDA of MTnTetM438 transformants was always higher than that of MTnGm transformants and increased from day 10 to day 16 (Fig. 4). The CDA of MTnGm transformants also clearly increased between days 12 and 16 (Fig. 4). These results show that although the aac(6')-aph(2'') gene can be used as an effective selectable marker, undesired effects are also evident on the growth of MTnGm-transformed M. genitalium cells. The slower growth of MTnGm transformants could possibly be due to a residual negative effect of Gm on mycoplasma cell growth. However, when MTnGm transformants were plated without antibiotic, the CDA was as low as that obtained when Gm was present (Table 4), and the CDA of MTnTetM438 transformants was very similar to that exhibited by the wild-type colonies. These data suggest that the negative effects described above do not result from an inefficient Gm inactivation by AAC(6')-APH(2'') but are a direct consequence of aac(6')-aph(2'') expression in M. genitalium. These results also show that the use of tetM438 marker has no detrimental effect on growth of M. genitalium.

Table 4. CDA of wild-type strain and MTnTetM438 or MTnGm transformants CDAs were obtained after 14 days incubation in presence or absence of antibiotic. Values are means±SE of diameters of 100 colonies from two separate experiments.

Aminoglycoside antibiotics such as Gm are readily inactivated by AAC(6')-APH(2'') phosphorylation (Boehr et al., 2004). It has been reported that aminoglycoside phosphotransferase enzymes are also serine protein kinases (Daigle et al., 1999). Since there is evidence of protein modification by phosphorylation in mycoplasmas (Dirksen et al., 1994; Platt et al., 1988), we hypothesize that AAC(6')-APH(2'') may phosphorylate some M. genitalium proteins, probably thereby modifying the intracellular signalling. The relationship between the possible phosphorylation of mycoplasma proteins by AAC(6')-APH(2'') and the reduction of colony size remains to be investigated. However, the negative effect of aac(6')-aph(2'') gene expression has to be considered when designing gene knock-out experiments by either transposition or homologous recombination.

Random insertion and stability of MTnTetM438 in the M. genitalium chromosome
Genomic DNA isolated from eight independent clones selected from the MTnTetM438 transformants obtained was analysed by Southern blotting (data not shown). As expected, the presence of a single band of different size in each clone showed that MTnTetM438 is randomly inserted in the M. genitalium genome. Since no additional band was detected, we conclude that MTnTetM438 inserts as a single copy and remains stable once transposed.

We also addressed the question of whether insertions in genomic locations not previously described when using aac(6')-aph(2'') could be obtained with pMTnTetM438. We determined the transposon insertion site of 30 randomly selected MTnTetM438 transformants (Table 5). We found 28 different transposon insertion sites and two transposon insertions in the same location of the mg269 and mg281 coding regions. Four transposon insertions located in the intergenic regions were in sequences belonging to a family of repetitive DNA elements known as MgPa islands (Peterson et al., 1995). Surprisingly, transposon insertions were detected in two coding regions, mg298 and mg358, in which no transposon insertions were previously reported in the global transposon mutagenesis analysis (Hutchison et al., 1999). Moreover, both transposon insertions are expected to disrupt the gene function (Fig. 5a, b). The mg298 gene encodes P115, a protein with unknown function, and the mg358 gene is currently annotated as ruvA, which is involved in the resolution of Holliday intermediates. The ruvA gene has been previously shown to be dispensable in E. coli (Sharples et al., 1990). No transcriptional defects derived from the transposon insertions were detected by RT-PCR in the downstream genes mg297 and mg359 (Fig. 5c), which have also been considered as essential. However, disruptive insertions in the mg298 and mg358 coding regions show that M. genitalium genes which have not been previously considered as dispensable under laboratory growth conditions could be knocked out by our pMTnTetM438 construction; thus the list of non-essential M. genitalium genes could be longer than previously described.

Table 5. Transposon insertion sites of 30 randomly selected MTnTetM438 transformants

(30K): Fig. 5. (a, b) Schematic representation showing the precise insertion points of MTnTetM438 in the mg298 (a) and mg358 (b) genes. Both transposon insertions are centred in the target gene and are expected to disrupt the gene function. (c) Transcriptional analysis by RT-PCR of the mg297 and mg359 genes in the MTnTetM438 transformant clones 2 and 22, respectively. Total RNA was isolated from cultures of these clones and the wild-type strain. Lanes 2 and 4 correspond to the mg359 RT-PCR products obtained from total RNA isolated from the wild-type and clone 2, respectively. Lanes 6 and 8 are the mg297 RT-PCR products obtained from total RNA isolated from the wild-type and clone 22, respectively. Lanes 1, 3, 5 and 7 correspond to the respective RT-PCR negative controls.

This work was supported by grant BFU2004-06377-C02-01 to E. Q. R. B. acknowledges an FPU predoctoral fellowship from the Ministerio de Educación y Ciencia. O. Q. and R. P. acknowledge a predoctoral fellowship from CeRBa (Centre de Referència en Biotecnologia). Plasmids pIVT-1 and pAM120 were the kind gifts of Dr K. Dybvig and Dr D. B. Clewell, respectively. We thank Dra Oxana Musatovova (UTHSCSA) for her valuable advice on performing primer extension and Joan Ruiz (UAB) for performing primer extension ALF analysis. We also thank Anna Barceló (Servei de Seqüenciació UAB) for DNA sequencing.

Footnotes

^†, Raul Burgos †These authors contributed equally to this work.