ABSTRACT
Mycoplasma pneumoniae adherence to host cells is a multifactorial process that requires the cytadhesin P1 and additional accessory proteins. The hmw gene cluster consists of the genes p30, hmw3, and hmw1, the products of which are known to be essential for cytadherence, therpsD gene, and six open reading frames of unknown function. Putative transcriptional terminators flank this locus, raising the possibility that these genes are expressed as a single transcriptional unit. However, S1 nuclease protection and primer extension experiments identified probable transcriptional start sites upstream of thep32, p21, p50, and rpsDgenes. Each was preceded at the appropriate spacing by the −10-like sequence TTAAAATT, but the −35 regions were not conserved. Analysis of the M. pneumoniae genome sequence indicated that this promoter-like sequence is found upstream of only a limited number of open reading frames, including the genes for P65 and P200, which are structurally related to HMW1 and HMW3. Promoter deletion studies demonstrated that the promoter-like region upstream ofp21 was necessary for the expression of p30 and an hmw3-cat fusion in M. pneumoniae, while deletion of the promoter-like region upstream of p32 had no apparent effect. Analysis by reverse transcription-PCR confirmed transcriptional linkage of all the open reading frames in thehmw gene cluster. Taken together, these findings suggest that the genes of this locus constitute an operon expressed from overlapping transcripts.
The cell wall-less prokaryoteMycoplasma pneumoniae causes atypical pneumonia and tracheobronchitis in older children and young adults.Mycoplasma colonization of host respiratory epithelium (cytadherence) is mediated largely by a differentiated polar attachment organelle. This tip structure is a membrane-bound extension of the mycoplasma cell and contains an electron-dense core that enlarges to form a terminal button at the tip of the cell (3, 18, 19). The adhesin protein P1 is primarily found densely clustered on the surface of the attachment organelle (27). However, additional proteins, including the high-molecular-weight proteins HMW1, HMW2, and HMW3 are required for cytadherence (18, 20).
HMW1, HMW2, and HMW3 are encoded by two unlinked genetic loci in theM. pneumoniae chromosome (6, 15, 21). Thehmw2 gene is part of the P65 operon, also known as the cytadherence regulatory locus (crl [14]). Spontaneous frameshifts in poly(A) stretches in hmw2(7) or transposon insertional inactivation ofhmw2 (21) results in accelerated turnover of HMW1, HMW3, and P65 (24). It is not clear, however, whether this accelerated proteolysis reflects strictly a housekeeping activity or possibly a regulatory mechanism functioning improperly in the mutant. The hmw1 and hmw3 genes are located approximately 160-kbp from the P65 operon, in what is designated thehmw gene cluster (6, 18). This locus also includes the gene for P30, a putative cytadhesin that is required for normal cell development (2, 4, 28), and six open reading frames (ORFs) of unknown function (6) (Fig.1). Like P30, HMW1 and HMW3 have important roles in the architecture and assembly of the attachment organelle. HMW1 is found along the leading and trailing extensions of the mycoplasma cell and is essential for proper development of the tip structure (12, 31), while HMW3 is a major component of the terminal button of the electron-dense core (32).
Map of the hmw gene cluster of M. pneumoniae. The number of the first nucleotide relative to the published genome sequence of M. pneumoniae (15) is given to the left, and the scale in kilobase pairs is shown below the map. The genes for p30, hmw3,hmw1, serine tRNA (t), and rpsD are indicated, as are ORFs encoding predicted proteins of unknown function. Stem-loop structures predicted to function as terminators are designated by solid circles, while a possible attenuator is indicated by the open circle. The promoter-like sequences evaluated by primer extension are indicated, with those underlined by arrows yielding primer extension products. The predicted overlapping transcripts, based upon S1 nuclease protection, primer extension, and RT-PCR, are shown below the map. B,BamHI; E, EcoRI; K, KpnI; X,XbaI.
Duplication of the attachment organelle is believed to precede cell division in M. pneumoniae (3). The tip structure is thought to function in chromosome partitioning and, therefore, one might expect the synthesis of attachment organelle components to be regulated in a manner that is coordinated with cell division (19). However, transcriptional control is poorly understood in mycoplasmas and, based on the presence of a single sigma factor and the lack of predicted two-component or other typical transcriptional regulatory systems, mechanisms for controlling M. pneumoniaegene expression may be limited (15). In the current study we have undertaken a detailed analysis of transcription of thehmw gene cluster of M. pneumoniae. S1 nuclease protection and primer extension were used to define two probable transcription start sites near the beginning of the gene cluster, as well as two additional sites much farther downstream, indicating likely overlapping transcripts. The nucleotide sequences upstream of each putative transcriptional start site exhibited homology to the consensus prokaryotic Pribnow-Schaller box (25). The frequency and distribution patterns for this promoter-like sequence in the M. pneumoniae genome were likewise consistent with promoter function. Analysis of reporter gene expression after deletion of each putative promoter near the 5′ end of the locus identified a region that is essential for the expression of p30 and hmw3 inM. pneumoniae. Finally, data from reverse transcription-PCR (RT-PCR) analysis were consistent with coexpression of the genes in thehmw locus.
MATERIALS AND METHODS
Bacterial strains and culture conditions. M. pneumoniae strains used in this study included wild-type strain M129 (broth passage 17) and two noncytadhering mutants derived from M129, I-2 and II-3 (2, 20). The relevant phenotypes of each are summarized in Table 1. Mycoplasmas were cultured at 37°C in Hayflick medium (13) until the mid-logarithmic phase and then harvested as described previously (10). Cultures were plated on PPLO agar (23), incubated at 37°C for 6 to 9 days, and visualized by hemolytic plaques for the isolation of individual colonies or the enumeration of CFUs (20). Gentamicin was included at a concentration of 18 μg/ml for selection and culturing of mycoplasma transformants.Escherichia coli Sure (Stratagene, La Jolla, Calif.) grown in Luria broth was prepared as competent cells for transformation and used for plasmid preparation by standard techniques (29). Plasmid DNA was purified by using pZ523 columns (5′→3′, Inc., Boulder, Colo.) according to the manufacturer’s protocol.
Relevant phenotype of the M. pneumoniaestrains used in this study
RNA preparation.All solutions and plasticware were rendered RNase-free by diethyl pyrocarbonate treatment (29). RNA was extracted from mid-logarithmic-phase cultures of M. pneumoniae as described previously (24), except that nucleic acids were incubated at 68°C for 10 min to enhance resuspension prior to treatment with RNase-free DNase I (30 U; Boehringer Mannheim, Indianapolis, Ind.) for 1 h at 37°C in 0.25 ml of 100 mM sodium acetate–5 mM MgSO4 (pH 5.0) containing 39 U of RNasin (Promega, Madison, Wis.).
S1 nuclease protection and primer extension analyses.S1 nuclease protection analysis was conducted to localize the 5′ end of RNA transcripts extending into the hmw3 gene. The double-stranded DNA templates used for nuclease protection spanned from either the EcoRI site or the KpnI site upstream of the hmw gene cluster to the first BamHI site in the hmw3 gene (Fig. 1) and were labeled at the 5′ end with [γ-32P]ATP. Hybridization and nuclease protection were carried out by using standard techniques (29) at hybridization temperatures based upon the G+C content of the predicted DNA-RNA hybrids.
Primer extension analysis was utilized to identify precisely the 5′ end of RNA transcripts for the hmw gene cluster. Mycoplasma total RNA was reverse transcribed by using the avian myeloblastosis virus (AMV) Reverse Transcriptase Primer Extension System (Promega) according to the manufacturer’s protocol. Primers corresponded to regions within 150 nucleotides (nt) of the expected transcription initiation sites (Table 2). Oligonucleotide primers (Life Technologies, Grand Island, N.Y.) were end labeled by incubating 10 pmol of primer with 10 U of T4 polynucleotide kinase, 30 μCi of [γ-32P]ATP (3,000 Ci/mmol; NEN Research Products, Boston, Mass.), 50 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 5 mM dithiothreitol, and 0.1 mM spermidine at 37°C for 10 min. Samples were heated to 90°C for 2 min to inactivate the T4 polynucleotide kinase, and nuclease-free water was then added to bring the final primer concentration to 100 fmol/μl. Radiolabeled primer (100 fmol) was annealed with 35 μg of total mycoplasma RNA in 50 mM Tris-HCl (pH 8.3), 50 mM KCl, 10 mM MgCl2, 10 mM dithiothreitol, a 1 mM concentration of each deoxynucleoside triphosphate (dNTP), and 0.5 mM spermidine at 58°C for 20 min. Annealed samples were cooled at room temperature for 10 min, and primers were extended by RT for 30 min at 41.5°C in 50 mM Tris-HCl (pH 8.3), 50 mM KCl, 10 mM MgCl2, 10 mM dithiothreitol, 1 mM each dNTP, 2.8 mM sodium pyrophosphate, 0.5 mM spermidine, and 1 U of AMV reverse transcriptase. After extension, an equal volume of loading dye (98% formamide, 10 mM EDTA, 0.1% xylene cyanol, 0.1% bromophenol blue) was added to each sample.
Nucleotide sequence of putative promoter regions from M. pneumoniae
Extension products were analyzed by polyacrylamide gel electrophoresis (PAGE). Approximately 20% (6 to 8 μl) of each primer extension sample was loaded onto a 6% polyacrylamide sequencing gel containing 8 M urea, 89 mM Tris base, 110 mM boric acid, and 2 mM EDTA. Cloned mycoplasma DNA corresponding to the region being analyzed by primer extension was used as a template for DNA sequencing reactions with the same primers used for primer extension analysis. The fmoleDNA Sequencing System (Promega) was used for all sequencing reactions according to the manufacturer’s protocol. Sequencing gels were dried and exposed to film overnight.
Promoter deletion analysis.Two putative promoter regions upstream of the p30 gene were evaluated in M. pneumoniae. First, a promoterless copy of a Staphylococcus aureus cat gene (11) was amplified by PCR so as to engineer flanking BamHI sites and an appropriate upstream stop codon. The PCR product was digested with BamHI and cloned into the BamHI site of pGEM7Zf(+) (Promega) to create pKV170. The sequence of the cat insert in pKV170 was confirmed by automated DNA sequencing (Molecular Genetics Instrumentation Facility, University of Georgia, Athens, Ga.). Plasmid pKV112 (28), which contains the 3.215-kb region of the HMW gene cluster spanning XbaI to BamHI (Fig. 1), was digested with BamHI, and the cat fragment was excised from pKV170 with BamHI and cloned into the corresponding site of pKV112 to yield pKV181. Orientation of the insert in the same direction as hmw3 was confirmed by restriction endonuclease digestion and sequencing. For deletion of the first putative promoter region, plasmid pKV181 was transformed into E. coli GM33, a damA strain (kindly provided by K. Dybvig, University of Alabama at Birmingham), so that a BclI site near the 5′ end of ORF p32 was no longer methylated and was cleavable by that restriction endonuclease (see Fig. 5). After digestion with BclI and Csp45I to remove a 233-bp fragment, the DNA was blunt ended with Klenow fragment of DNA polymerase and religated to yield pKV187 (29). The promoter region upstream of ORF p21 was deleted by digestion of pKV181 with MluI and SpeI to remove a 292-bp fragment, making the ends blunt with Klenow, and religating it to yield pKV184. Promoter deletions in each case were confirmed by DNA sequencing. The hmw3-cat fusions from pKV181, pKV184, and pKV187 were excised with EcoRV and BglII and blunt ended with the Klenow fragment. These were then ligated into theSmaI site of pISM2062 (17) to yield pKV193, pKV197, and pKV198 respectively (see Fig. 5).
Promoter function for each deletion mutant was assessed on the basis of chloramphenicol acetyltransferase (CAT) reporter activity in wild-type and mutant II-3 backgrounds and on the basis of production of recombinant P30 in the mutant II-3 background, as measured by Western blotting and complementation of the wild-type hemadsorption phenotype. Mycoplasma transformants were isolated and expanded as described previously (14). The presence of the transposon and twop30 alleles (resident and recombinant) was demonstrated for each transformant by Southern blot hybridization (29). The activity of the CAT reporter was measured as the MIC and by using the FAST CAT Green (deoxy) CAT Assay Kit (Molecular Probes, Eugene, Oreg.), which was carried out as described previously (11). P30 production was assessed by discontinuous sodium dodecyl sulfate-PAGE (12% polyacrylamide separating gel [10]), followed by Western immunoblotting with P30-specific antibodies (28). Hemeadsorption assays were conducted as described previously (20).
RT-PCR.Northern blot analysis of hmw1- andhmw3-specific mRNA failed to yield a clear indication of transcript size or to establish the cotranscription of these genes (data not shown). Therefore, RT-PCR was used to evaluate transcriptional linkage in the hmw gene cluster and to identify the likely 5′ and 3′ ends of the transcripts. Primer pairs for RT-PCR were chosen to enable synthesis of a PCR product that was 400 to 1,500 nt in size that would span each intergenic region (Table 1). Mycoplasma RNA was reverse transcribed by using the AMV Reverse Transcriptase Primer Extension System as described above, except that no radiolabel was necessary. After the extension step, 180 μl of 10 mM Tris–10 mM EDTA (pH 7.5) was added to each sample to stop T4 polynucleotide kinase activity (as modified from Ausubel et al. [1]). Nucleic acids were extracted with phenol, phenol-chloroform, and chloroform and then precipitated with ethanol and suspended in 20 μl. Then, 10-μl volumes were added to Easystart 50 reaction tubes (Molecular Bio-Products, San Diego, Calif.) containing 2 mM MgCl2, 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 0.2 mM dNTP, and wax to overlay the reaction. The PCR primers (2 μM) and Taq DNA polymerase (2.5 U; Promega) were added, and samples were immediately heated to 95°C for 5 min. The PCR cycle consisted of 95°C for 1 min (denaturation), 55°C for 1 min (annealing), and 72°C for 1 min (extension), and this was repeated 30 times. PCR products were visualized after agarose gel electrophoresis and ethidium bromide staining.
Computer analysis.Sequence analysis was performed by using the Wisconsin Package version 9.0 (Genetics Computer Group) through the Research Computing Resource at the University of Georgia. The FindPatterns program was used to locate possible promoter sequences in the M. pneumoniae genome, and a script written in the “Practical Extraction and Report Language” (PERL) was used to compare those matches with the annotated genome (15). A total of 696 genes, including predicted ORFs and RNA-encoding sequences, were considered.
RESULTS
Identification of potential transcription initiation sites.Previous studies identified two possible Rho factor-independent terminators downstream of the hmw3 gene (6) (Fig.1). The first immediately precedes hmw1 and probably functions as an attenuator, given its weak Δ G (−7.4 kcal) and nucleation with an AT pair (33). The second followsrpsD, with a calculated Δ G of −10.6 kcal and hairpin nucleation beginning with a GC pair. Examination of the sequence upstream of hmw3 in the same manner revealed a possible transcriptional terminator upstream of ORF p32consisting of a stem and loop of 16 and 6 nt, respectively. Hairpin nucleation begins with a GC pair, and the calculated Δ G is −10.2 kcal (nt 467491; 5′-TAAAAAAAGCACATCCCCCAAAAGGTGTGCTTTTTTAA). Based upon the location of these terminator-like sequences, the hmw operon is predicted to span from ORF p32 through rpsD, constituting potentially a single transcriptional unit.
The likely start site for transcription of this gene cluster was evaluated relative to the hmw3 gene by S1 nuclease protection analysis. Two faint bands with estimated lengths of 2.6 and 2.0 kb were detected in S1 nuclease protection studies by using a 5′-end-labeled, double-stranded DNA probe extending from the firstBamHI site in hmw3 upstream to theKpnI site preceding ORF p32 (Fig. 1 and data not shown). These sizes are consistent with transcription initiation downstream of the predicted terminator preceding ORF p32. Furthermore, the presence of two bands suggests overlapping transcripts with different start sites.
The 5′ ends of the hmw3 transcripts identified by nuclease protection were localized more precisely by primer extension, as summarized in Fig. 1. The 5′ end of one transcript corresponded to the adenine nucleotide 12 bp upstream from the putative ATG of ORFp32 (Fig. 2A). The 5′ end of a second transcript was identified as an adenine nucleotide 22 bp upstream from the putative ATG of ORF p21 (Fig.3A). Both 5′ ends were confirmed by using a different oligonucleotide primer for primer extension (data not shown) and were consistent with findings obtained by S1 nuclease protection. Finally, the primer extension products obtained with RNA from noncytadhering mutants I-2 and II-3 were indistinguishable from those obtained with RNA from wild-type M. pneumoniae (Fig.2B and 3B). This finding is consistent with previous findings that thehmw3 and hmw1 genes are transcribed at wild-type levels in mutant I-2 (24).
Primer extension analysis upstream of ORFp32. The 5′ end of the oligonucleotide primer (5′-GACAAACTGTTGCTCGGAAAAATTGACCTG; sense strand) corresponds to nt 467836 of the M. pneumoniae genome (15). (A) The primer extension product (PE) obtained with wild-type M. pneumoniae RNA as a template is indicated by the arrow. The sequencing ladder obtained with the same primer and cloned hmw gene cluster DNA as template is given, with the corresponding sequence (sense strand) shown below. The asterisk indicates the +1 site, and a likely −10 box is underlined. (B) Primer extension products with equal amounts of RNA from wild-type M. pneumoniae (w-t) and cytadherence mutants I-2 and II-3. Control reactions lacking RNA template or reverse transcriptase yielded no primer extension product (data not shown).
(A) Primer extension analysis upstream of ORFp21. The 5′ end of the primer (5′-CTTCCTTAGAGAGAAAAAACGGTTAAACACATCCATTG; sense strand) corresponds to nt 468897 of the M. pneumoniae genome (15). The primer extension product (PE) obtained with wild-type M. pneumoniae RNA as a template is indicated by the arrow. To the left is the sequencing ladder, as described in the legend of Fig. 2. The asterisk indicates the +1 site, while a likely −10 box is underlined. (B) Primer extension products obtained with equal amounts of RNA from wild-type M. pneumoniae (w-t) and cytadherence mutants I-2 and II-3. Control reactions lacking RNA template or reverse transcriptase yielded no primer extension product (data not shown).
Six sites having the consensus −10-like (Pribnow-Schaller box) sequence TANANT were identified within the region spanning from the Ser-tRNA genes upstream of ORF p32 to thehmw3 gene (Fig. 1). A promoter-like sequence preceded the predicted start codon of ORF p32 by 19 nt (Fig. 1) and the 5′ end of one transcript by 6 nt (Fig. 2A). A second promoter-like sequence was identified 28 nt upstream of the predicted start of ORFp21 (Fig. 1) and 6 nt upstream of the 5′ end of the secondhmw3 transcript (Fig. 3A). No primer extension products were identified corresponding to the other four possible −10-like sequences upstream of hmw3 that were tested (data not shown).
Two potential transcription start sites were also identified by primer extension downstream of the hmw3 gene (Fig. 1). The first was located approximately 75 nt upstream of ORF p50 and was confirmed by using a second oligonucleotide for primer extension. In each case, two primer extension products were identified; they differed in length by 3 nt and began 3 and 6 nt downstream of a −10-like consensus sequence (Fig. 4A). These primer extension products were likewise comparable in intensity regardless of whether wild-type or mutant I-2 RNA was used (Fig. 4B). Finally, the probable 5′ end of the transcript for rpsD(Fig. 1) was identified only 4 nt upstream of this gene and 6 nt downstream from a consensus −10-like sequence (data not shown).
(A) Primer extension analysis upstream of ORFp50. The 5′ end of the primer (5′-CACGAAGGTCTTTAGCAAATG; sense strand) corresponds to nt 474144 of the M. pneumoniae genome (15). The primer extension products (PE) obtained with wild-type M. pneumoniae total RNA as a template are indicated by arrows. The sequencing ladder is shown to the left, as described in the legend of Fig. 2. The sequence (sense strand) is given below, with asterisks indicating the +1 nucleotides, and a likely −10 box is underlined. (B) Primer extension products synthesized with equal amounts of RNA from wild-type M. pneumoniae (w-t) and cytadherence mutants I-2 and II-3. Control reactions lacking RNA template or reverse transcriptase yielded no primer extension product (data not shown).
Each putative transcriptional start site was preceded at the appropriate spacing by the sequence (TTAAAATT), corresponding to a consensus-like −10 region. The sequence TAAAAT matches five of six residues in the consensus prokaryotic ς70-dependent −10 site (TATAAT; Table 2). Furthermore, this putative −10 sequence is identical to that identified for the P65 operon of M. pneumoniae(21), but it differs from −10 sequences previously described in this species (Table 2). No pattern was evident in the putative −35 region of the sequence. Nevertheless, we hypothesize that these −10-like sites function in M. pneumoniae in RNA polymerase recognition, and we refer to these sites as PromP32, PromP21, PromP50, and PromRpsD.
Promoter deletion analysis.The recombinant wild-typep30 allele restores P30 to normal levels in P30 mutants when introduced by transposon delivery (28). We reasoned that it should be possible to test the requirement for PromP32 and PromP21 in the expression of recombinant p30 by engineering promoter deletion mutants and examining the consequences on P30 production (Fig.5). P30 mutant II-3 cells transformed with recombinant p30 lacking the PromP21 (pKV197) promoter produced very little P30 that was detectable by Western blotting. However, deletion of PromP32 (pKV198) had no effect on P30 levels (data not shown). Wild-type M. pneumoniae has a hemadsorption-positive phenotype, while mutant II-3 is hemadsorption negative. M. pneumoniae mutant II-3 transformed with pKV197 remained hemadsorption negative, but when transformed with pKV193 or pKV198, it was restored to a hemadsorption-positive phenotype (Fig. 5).
Analysis of p30 and hmw3-catexpression with deletion of the putative promoter regions. Construction of pKV193 in Tn4001mod (17) and deletion of PromP32 and PromP21 are described in detail in the text. The scale is in kilobase pairs. PromP32 and PromP21 are indicated above the restriction map, with arrows underlining the predicted −10 region.M. pneumoniae transformants for each construct were evaluated for P30 synthesis, hemadsorption (HA), and growth on chloramphenicol. B, BamHI; Bc, BclI; Bg,BglII; C, Csp45I; M, MluI; R,EcoRV; S, SpeI; Sm, SmaI.
The effect of promoter deletion on expression of an hmw3-cattranscriptional fusion was examined in parallel studies. The S. aureus cat gene is expressed in M. pneumoniae, conferring resistance to chloramphenicol at a concentration of more than 50 μg/ml (11) compared to an MIC of 3.2 to 12.5 μg of chloramphenicol/ml for untransformed wild-type and mutant II-3M. pneumoniae. Similar levels of chloramphenicol resistance were seen for transcriptional fusions with a promoterlesscat cloned into the BamHI site of hmw3(MIC of 50 to 100 μg of chloramphenicol/ml). Deletion of PromP32 had no effect on chloramphenicol resistance, while deletion of PromP21 rendered the transformants chloramphenicol sensitive at levels comparable to the untransformed controls (Fig. 5). Chloramphenicol resistance in these promoter deletion transformants correlated with CAT activity as measured by the FAST CAT Green (deoxy) assay (data not shown).
Transcript size.Northern blot hybridization analysis ofhmw1- and hmw3-specific mRNA failed to establish the cotranscription of these genes, probably due to the large transcript size predicted (up to 13.5 kbp) (data not shown). Therefore, RT-PCR was used to evaluate transcriptional linkage in thehmw gene cluster and to identify the likely 5′ and 3′ ends of the transcripts (Table 3). Primer pairs for RT-PCR were chosen to enable synthesis of PCR products 400 to 1,500 bp in length that would span each intergenic region and to test whether putative terminators flanking this gene cluster actually coincided with the 5′ and 3′ ends of the operon. RT-PCR products of the predicted length were generated for all primer pairs spanning intergenic regions from ORFp32 to rpsD (Fig.6A; Table 1). Importantly, control reactions containing total RNA without reverse transcriptase enzyme or else containing reverse transcriptase but not RNA yielded no PCR products. As expected from the location of the predicted transcriptional terminators, no RT-PCR products were observed extending upstream from ORF p32 or downstream from rpsD(Fig. 6B). In both cases the expected PCR product was observed when cloned DNA corresponding to each region was included in the control samples (Fig. 6B).
Analysis of the hmw gene cluster by RT-PCR. The primers used and the predicted sizes of the products are listed in Table 3. (A) RT-PCR products from primers designed to span the intergenic regions between p32-p21, p21-p30,p30-hmw3, hmw3-p43, p43-p36,p36-p50, p50-p33-hmw1, and hmw1-rpsD(lanes b, c, d, e, f, g, h, and i, respectively). (B) RT-PCR reactions for primers spanning the regions between H08_orf1005-p32, and rpsD-H08_orf289 (15) (Fig. 1) are shown in lanes b and c, respectively. PCR products for the indicated primers with genomic DNA as a template are in the lanes marked “DNA”. The sizes of the DNA markers in lanes a are indicated in base pairs, and a plus (+) indicates that reverse transcriptase was included, while a minus (−) indicates that reverse transcriptase was omitted.
DISCUSSION
The hmw gene cluster of M. pneumoniae spans nearly 13 kbp from ORF p32 through rpsD, including the genes for the cytadherence proteins P30, HMW1, and HMW3, and is flanked by predicted stem-loop structures (Fig. 1). Washio et al. calculate, based upon free-energy changes around stop codons, that hairpin termination is evident, for example, in E. coli, but is not evident in several archeae genomes and is questionable in mycoplasma genomes (35). However, their calculations are averaged around stop codons over the entire genome and may not necessarily apply to any given ORF. Furthermore, our RT-PCR findings are consistent with a terminator function for the predicted stem-loops flanking the hmw operon (Table3). The ORFs of this gene cluster are oriented in the same direction and, based upon the data presented here, are transcribed in four overlapping transcripts. Little is known regarding the potential regulation of transcription in mycoplasmas, and this gene cluster is a good candidate for such studies, largely because of the availability of several mutants lacking products of this locus, making it possible to use complementation techniques to explore genetic elements necessary for expression.
Summary of RT-PCR analysis of the hmwgene clustera
Analysis by S1 nuclease protection indicated that hmw3 is expressed by overlapping transcripts beginning approximately 2.0 and 2.6 kbp upstream. Subsequent studies by primer extension identified the 5′ end of transcripts immediately upstream of ORF p32 and ORF p21. Each was preceded by a likely Pribnow-Schaller box (TAAAAT) 6 bp upstream of the 5′ end of the mRNA (Table 2). The sequence corresponding to the putative −35 region was not conserved. The putative −10 sites in the hmw operon were identical to that of the P65 operon (21) but differed from predicted −10 sequences for other M. pneumoniae genes (Table 2). We assessed the predicted and actual frequency of each putative −10 sequence in the M. pneumoniae genome, as well as the number of each found within 100 bases upstream of an ORF or RNA gene (Table 4). The ratio of the actual to the predicted number of ORFs preceded by the 8-base sequence identified here was even higher than that for the consensus ς70 promoter (Table 4). Perhaps more significant, however, was the finding that, among the genes downstream of a TTAAAATT promoter-like sequence was the gene encoding P200 (data not shown). This protein shares common deduced structural features with HMW1, HMW3, and P65, and all are associated with the mycoplasma cytoskeleton (26). The relative incidence of this 8-base sequence upstream of ORFs in the M. pneumoniae genome and its association with transcription of a family of structurally and functionally related proteins may reflect some form of regulation for this putative promoter.
Predicted and actual frequency of putative promoter sequences in the M. pneumoniae genome
Deletion studies were conducted to assess the requirement for each promoter-like region in the expression of p30 andhmw3. Recombinant transposons were engineered that contained a cat transcriptional fusion in hmw3 but that lacked specific putative promoters. These were transformed into wild-type M. pneumoniae and the cytadherence mutant II-3. Deletion of PromP32 had no effect on P30 production or CAT activity. However, deletion of PromP21 resulted in very little recombinant P30 and no chloramphenicol resistance (Fig. 5). Given the results from the nuclease protection studies, it is not clear why PromP32 was insufficient for the production of P30 or CAT. The deletion of PromP21 removed the last six codons from p32 and the first 60 codons from p21. Therefore, if P30 translation were coupled to that of P21, this might account for the inability to produce P30 with deletion of PromP21. The failure to product CAT with the deletion of PromP21 cannot be attributed directly to translational coupling, but the translation of P21 or P30 may be essential for RNA stability or ribosome-binding site accessibility. Preliminary analysis by RT-PCR detected cat-specific transcripts despite the absence of CAT activity (data not shown), a finding consistent with this scenario. Nevertheless, the role of the overlapping transcripts encompassingp30 and hmw3 is unclear. They may reflect a means for optimizing the stoichiometry of the gene products or perhaps a mechanism for the regulation of expression, for example, during cell growth and development of the tip structure. Additional studies employing site-specific mutagenesis in each putative promoter, or the introduction of nonsense mutations within the genes for P21 and P30, may provide more insight.
Analysis by primer extension identified two additional likely transcriptional start sites within the hmw gene cluster, one just upstream of ORF p50 and the second immediately upstream of rpsD. Both were preceded by a −10 sequence identical to that of PromP32 and PromP21 and appropriately spaced upstream from the +1 site of the mRNA. Like PromP32 and PromP21, PromP50 and PromRpsD exhibited no conserved −35 region. Given the sequence identity within the −10 regions identified here, we feel that it is highly probable that these represent M. pneumoniae promoter elements. The function of the overlapping transcripts generated is not known but may reflect the same requirements as with the putative promoters in the upstream region of this gene cluster.
The presence of a likely promoter immediately upstream ofrpsD was surprising. The rpsD in closely related gram-positive organisms is subject to autogenous control by the gene product, ribosomal protein S4, which binds to a region in the untranslated leader of the monocistronic rpsD transcript (8, 9). However, the very short untranslated leader observed in M. pneumoniae would seem to preclude this mechanism of regulation.
Attempts to determine the size of hmw1- orhmw3-specific mRNA by Northern blot hybridization have been unsuccessful. As an alternative means to assess transcript size, we evaluated whether each intergenic region in the hmw gene cluster could be spanned by RT-PCR. As expected, no RT-PCR products were observed with primers flanking the putative terminators on either side of the gene cluster. Furthermore, each intergenic region was bridged by RT-PCR, suggesting that these genes are cotranscribed as a unit. The cotranscription of p30, hmw1, andhmw3 is consistent with coordinated function and suggests that the putative proteins encoded by ORFs p32,p21, p43, p36, p50, andp33 may also participate in the assembly or structure of the attachment organelle (36). The predicted weak stem-loop structure upstream of hmw1 may have a regulatory role and help account for the apparent overlapping transcript in this region of the gene cluster. However, because of the sensitivity of PCR, we cannot rule out the possibility that the RNA detected may be produced at trace rather than biologically significant levels. In the absence of likely transcription terminators, however, we predict that the 3′ end of the transcript extends to the end of the gene cluster (Fig. 1). Finally, the hmw gene cluster is reminiscent of the superoperons described in E. coli containing genes for diverse cellular functions (34). Hence, interpretation of their probable cotranscription as an indication of coordinate function must be approached cautiously, especially since the rpsD gene, encoding a ribosomal protein, appears to be cotranscribed with thehmw genes. Additional studies are required to establish whether the products of each unknown ORF are required in tip assembly, cell division, and/or cytadherence.
Mycoplasmas continue to be paradoxical. Their greatly reduced genome, lack of obvious transcriptional regulators, parasitic lifestyle, and small size belie a complex subcellular structure and regulation. The studies described here reveal a complex transcriptional organization when a much simpler pattern was expected. A better understanding of the role of overlapping transcripts and multiple promoter-like sites is likely to increase our appreciation for this unusual microorganism.
ACKNOWLEDGMENTS
We thank April Varn and Hannah Andrianopolis for their technical assistance.
This work was supported by Public Health Service research grant AI23362 from the National Institute for Allergy and Infectious Diseases to D.C.K.
FOOTNOTES
- Received 1 March 1999.
- Accepted 7 June 1999.
- Copyright © 1999 American Society for Microbiology
