INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The cell wall-less prokaryote Mycoplasma 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 the M. pneumoniae chromosome (6, 15, 21). The hmw2 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 of hmw2 (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 the hmw 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).

View larger version (16K): [in this window] [in a new window]   FIG. 1.   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. pneumoniae gene expression may be limited (15). In the current study we have undertaken a detailed analysis of transcription of the hmw 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 in M. pneumoniae. Finally, data from reverse transcription-PCR (RT-PCR) analysis were consistent with coexpression of the genes in the hmw locus.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

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.

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 [-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.

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 the SmaI site of pISM2062 (17) to yield pKV193, pKV197, and pKV198 respectively (see Fig. 5).

RT-PCR. Northern blot analysis of hmw1- and hmw3-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

Top Abstract Introduction Materials and Methods Results Discussion References

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 G (7.4 kcal) and nucleation with an AT pair (33). The second follows rpsD, with a calculated 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 p32 consisting of a stem and loop of 16 and 6 nt, respectively. Hairpin nucleation begins with a GC pair, and the calculated 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.

View larger version (81K): [in this window] [in a new window]   FIG. 2.   Primer extension analysis upstream of ORF p32. 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).

View larger version (34K): [in this window] [in a new window]   FIG. 3.   (A) Primer extension analysis upstream of ORF p21. 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).

View larger version (54K): [in this window] [in a new window]   FIG. 4.   (A) Primer extension analysis upstream of ORF p50. 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).

Promoter deletion analysis. The recombinant wild-type p30 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).

View larger version (17K): [in this window] [in a new window]   FIG. 5.   Analysis of p30 and hmw3-cat expression 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.

Transcript size. Northern blot hybridization analysis of hmw1- 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 the hmw 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 ORF p32 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).

View larger version (94K): [in this window] [in a new window]   FIG. 6.   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

Top Abstract Introduction Materials and Methods Results Discussion References

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 (Table 3). 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.

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 ⁷⁰ 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.

Deletion studies were conducted to assess the requirement for each promoter-like region in the expression of p30 and hmw3. 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 encompassing p30 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 of rpsD 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- or hmw3-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, and hmw3 is consistent with coordinated function and suggests that the putative proteins encoded by ORFs p32, p21, p43, p36, p50, and p33 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 the hmw 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.