Trinucleotide GAA Repeats Dictate pMGA Gene Expression in Mycoplasma gallisepticum by Affecting Spacing between Flanking Regions

Bacterial strains and culture conditions. M. gallisepticum PG31 (ATCC 19610) was propagated in modified Frey broth or agar as described elsewhere (7). Plasmid constructions were performed in Escherichia coli strain INVαF' (Invitrogen, Carlsbad, Calif.) grown in Luria-Bertani medium.

Modifications in the pMGA-lacZ reporter. E. coli plasmid pCR2.1-M9.lacZ containing the pMGA-lacZ reporter inserted in the cloning vector pCR2.1 has been described previously (7). The reporter consists of a 336-bp pMGA gene fragment extending from 104 bp 5" of the GAA repeats into the region coding for the first 22 amino acids of the pMGA protein that is fused in frame with lacZ (Fig. 1). HindIII and Csp45I cleave plasmid pCR2.1-M9.lacZ at single sites within the pMGA-lacZ reporter (Fig. 1 and 2). To excise the nucleotides upstream of the GAA repeats, pCR2.1-M9.lacZ was digested with HindIII and Csp45I. The ends of the reporter-containing HindIII-Csp45I DNA fragment were made flush by incubation with T4 DNA polymerase, and SphI linkers were attached. The fragment was then inserted into the transposon Tn4001 portion of plasmid pISM2062.S, which is a modified version of pISM2062 in which the SmaI cloning site was replaced with SphI (7). Plasmid constructs were screened by restriction mapping to identify those that had the modified reporter oriented in Tn4001 as shown in Fig. 2. The resulting plasmid is referred to as pISM-M9.lacZ.del.

• Open in new tab
• Download powerpoint

FIG. 1.

Nucleotide sequence of the 5" end region of the unmodified pMGA-lacZ fusion gene. Sequences corresponding to the oligonucleotide primers used for PCR amplification and DNA sequencing are underlined. Also underlined are the GAA repeats, the amino acids encoded by lacZ, and the −10 and −35 regions of the promoter. The putative transcription start site (asterisk) was identified based on sequence similarity to the previously determined start site of pMGA1.1 (5). Restriction enzyme recognition sites are in bold.

• Open in new tab
• Download powerpoint

FIG. 2.

Schematic diagram of the construction of the modified pMGA-lacZ reporters and their insertion into the chromosome of M. gallisepticum. (A) Deletion of nucleotides to generate pISM-M9.lacZ.del and insertion of SalI linkers to generate pISM-M9.lacZ.ins10/12. White regions denote sequences derived from E. coli plasmid vectors. Regions in black, dark gray, and light gray denote the pMGA gene fragment, sequences originating from Tn4001, and lacZ, respectively. Arrows in plasmids show the direction of transcription. (B) A representation of Tn4001 containing pMGA-lacZ inserted into the chromosome of M. gallisepticum. Thin lines denote mycoplasmal chromosomal DNA flanking the transposon. Shaded regions are as in panel A. The leftmost HindIII site is the mycoplasmal HindIII site most proximal to the left end of the transposon. The XmnI site in the pMGA promoter region is replaced with SalI in the transposon of pISM-M9.lacZ.ins10 and pISM-M9.lacZ.ins12. The Csp45I site and upstream pMGA sequences were deleted in the transposon of pISM-M9.lacZ.del.

SalI linkers were inserted into pCR2.1-M9.lacZ at the XmnI site located immediately downstream of the GAA repeats. The plasmid has a second XmnI site in the ampicillin resistance determinant. Therefore, a partial digestion of pCR2.1-M9.lacZ with XmnI was performed. After XmnI digestion, the samples were subjected to agarose gel electrophoresis, and full-length, linear pCR2.1-M9.lacZ molecules were isolated from the gel. SalI linkers of either 10 or 12 bp were ligated to the linear plasmid. After digestion with SalI to generate cohesive ends, the plasmid molecules were circularized by ligation and transformed into E. coli. The restriction map of the plasmid isolated from the transformants was determined to identify plasmids in which the XmnI site at the GAA repeats had been converted to SalI. The SalI-modified reporter was excised from the pCR2.1 vector by digestion with HindIII, the ends of the liberated DNA fragment were made flush by digestion with T4 DNA polymerase, and SphI linkers were attached. As described above, the reporter gene was inserted into the Tn4001 portion of pISM2062.S, screening for plasmid constructs that had the reporter oriented as depicted in Fig. 2. The resulting plasmids are referred to as pISM-M9.lacZ.ins10 (10-bp linker) and pISM-M9.lacZ.ins12 (12-bp linker).

Transformation of M. gallisepticum and pedigree analysis by blue/white color selection.The plasmids used in this study do not replicate in M. gallisepticum. Plasmids containing the wild-type gene or one of the modified pMGA-lacZ genes in pISM2062.S were transformed into M. gallisepticum by electroporation as described previously, resulting in transformants in which Tn4001, containing the pMGA-lacZ reporter, had transposed into the mycoplasmal chromosome (7). Transformants were selected in medium supplemented with gentamicin (50 μg/ml) and X-Gal (160 μg/ml) to monitor pMGA-lacZ expression on the basis of blue/white color selection. Incubation of M. gallisepticum in 10% CO₂ improved the quality of the blue/white color. Pedigree analysis was performed by filter cloning procedures as described previously (7).

DNA analysis of subclones.The promoter region of the reporter gene from representative blue and white M. gallisepticum colonies was amplified by PCR and sequenced. For the wild-type reporter and for the modified reporters containing SalI linkers, the primers used for PCR amplification (5" M9 and the lacZ primers) and DNA sequencing (3" M9 primer) were as depicted in Fig. 1 and as described previously for amplification and sequence analysis of the wild-type reporter (7). For analysis of the promoter region of the modified reporter in pISM-M9.lacZ.del, a new primer was required for PCR amplification because the binding site for the 5" M9 primer had been deleted. The replacement primer 5"-TAGCCATTTCGTTGACTTTGTCGG-3" binds to Tn4001 sequences located 282 bp 5" of the GAA repeats. The PCR products were directly sequenced at the Iowa State University DNA Sequencing and Synthesis Facility, Ames. Sequence analysis was performed with the Sequencher (version 3.0; Gene Codes Corporation, Ann Arbor, Mich.) and MacVector (version 6.5.3; Oxford Molecular Group Inc., Beaverton, Oreg.) software packages. The lacZ probe and Southern analysis procedures for monitoring transposition of Tn4001 in M. gallisepticum were as described previously (7).

Nucleotides upstream of the GAA repeats are required for pMGA-lacZ gene expression. M. gallisepticum was transformed with pISM-M9.lacZ.del and assayed for LacZ activity. All of the initial transformants were LacZ⁻. Five transformants were selected for further study. Essentially 100% of the progeny from three of the transformants were LacZ⁻, but about 5% of the progeny from the other two transformants were LacZ⁺. The nucleotide sequence of the promoter region of the pMGA lacZ reporter from both LacZ⁻ and LacZ⁺ clones was determined. No nucleotide differences were detected, and 12 tandem GAA repeats were present. Because the LacZ phenotype was independent of the sequence of the promoter region, we examined the pMGA-lacZ reporter by Southern analysis to determine whether a change in phenotype might correlate with transposition of Tn4001 to an alternative chromosomal site. The lacZ probe hybridized with a HindIII fragment of the same size from both LacZ⁻ and LacZ⁺ clones (Fig. 3A). The available data do not explain why some clones are LacZ⁺, but it is apparent that deletion of pMGA sequences upstream of the GAA repeats abolished GAA-dependent gene regulation.

• Open in new tab
• Download powerpoint

FIG.3.

Southern analysis of mycoplasmal genomic DNA digested with HindIII and probed with lacZ. Controls showing that the lacZ probe does not hybridize with DNA from M. gallisepticum strain PG31 that lacks the pMGA-lacZ reporter were previously described (7). (A) Clone 2 is a LacZ⁻ transformant of M. gallisepticum transformed with pISM-M9.lacZ.del. Clones 2B1 and 2B4 are LacZ⁺ progeny of clone 2. (B) Clone 16 is a LacZ⁻ transformant of M. gallisepticum transformed with pISM-M9.lacZ.ins.10. The lineage and phenotype of LacZ⁺ and LacZ⁻ derivatives of clone 16 (clones 16B1, 16B1W1, 16B1W4, 16B1W1B2, and 16B1W4B2) are provided in Fig. 4. As determined by comparison to the mobility of radiolabeled HindIII fragments of bacteriophage lambda DNA, the hybridizing DNA fragments from clones 2 and 16 are about 15 and 5 kb, respectively. Similar experiments (data not shown) indicate that transposition of Tn4001 also did not occur during switching of LacZ production in M. gallisepticum transformed with pISM-M9.lacZ.ins.12.

Effect of adding SalI linkers 3" of the GAA repeats.Transformants of M. gallisepticum that contain the pMGA -lacZ reporter modified by the addition of 10- or 12-bp SalI linkers were monitored on X-Gal for β-galactosidase production. The initial transformants were LacZ⁻, but second-generation LacZ⁺ progeny were isolated at a high frequency. From the LacZ⁺ progeny, third-generation LacZ⁻ progeny were isolated, which in turn were subcloned to obtain LacZ⁺ fourth-generation progeny. The percentage of blue (LacZ⁺) progeny colonies for each of these successive generations of subclones is summarized in Fig. 4, revealing the phase-variable production of LacZ as described previously for the wild-type pMGA-lacZ reporter (7). Southern analysis of mycoplasmal DNA probed with lacZ revealed that switching of LacZ production did not correlate with transposition of Tn4001 to alternative sites in the chromosome (Fig. 3B).

• Open in new tab
• Download powerpoint

FIG. 4.

Pedigree analysis of M. gallisepticum containing the pMGA-lacZ reporter modified by insertion of 10- and 12-bp linkers. For each of the subclones, the percentage of progeny that were scored as blue, lacZ⁺ colonies is indicated.

The promoter region of the SalI-modified reporter from representative LacZ⁺ and LacZ⁻ subclones was PCR amplified and sequenced to determine whether switching in LacZ production correlated with changes in the number of GAA repeats. The insertion of the linkers into the XmnI site disrupted the terminal GAA repeat, resulting in 11 tandem repeats followed by a GA dinucleotide, the linker, and the final A nucleotide of the 12th trinucleotide repeat (Fig. 5). Thus, the initial LacZ⁻ transformants had 11 tandem GAA repeats. The numbers of tandem GAA repeats in the LacZ⁺ and LacZ⁻ progeny that were analyzed in successive generations are given in Fig. 4. The second-generation LacZ⁺ progeny all had eight tandem GAA repeats, regardless of whether the linker was 10 or 12 bp. No third-generation LacZ⁻ subclones had eight GAA repeats, but all fourth-generation LacZ⁺ subclones once again had eight repeats. Thus, in the SalI-modified reporters, eight tandem GAA repeats were associated with gene expression.

• Open in new tab
• Download powerpoint

FIG. 5.

Sequence alignment of the pMGA GAA repeat region of the unmodified promoter (w.t.) and promoters modified by the addition of 10-bp (SalI 10) and 12-bp (SalI 12) linkers. For the unmodified promoter, the XmnI site is boxed, with an arrow showing the site of XmnI cleavage. For the SalI-modified promoters, the SalI linker is boxed. Dashes refer to gaps in the nucleotide sequence introduced to realign the sequences 3" of the XmnI site or linker. (A) Alignment of the unmodified sequence to that of the original SalI-modified sequences that had 12 GAA repeats and gave rise to LacZ⁻ transformants; (B) alignment of sequences from LacZ⁺ clones.

For LacZ⁺ clones containing the 10-bp SalI linker, eight GAA repeats restores the pMGA promoter region to match as closely as possible the length of the promoter region found in LacZ⁺ clones containing the unmodified promoter (12 GAA repeats). The 10 bp that was added by addition of the linker was compensated by the deletion of three GAA repeats (9 bp), resulting in a net gain of a single base pair (Fig. 5). This result strongly suggests that the length of the GAA repeat region is critical for gene expression.

The results obtained for LacZ⁺ clones containing the 12-bp linker were something of a surprise. It had been anticipated that LacZ⁺ clones would have had seven GAA repeats, which would have restored the length of the promoter region to precisely that of the unmodified promoter. However, the insertion of the 12-bp linker was compensated by the loss of only three GAA repeats, resulting in a net gain of 3 bp (Fig. 5). Clones containing the 12-bp linker that had seven GAA repeats were not isolated in this study. Therefore, it is not known whether the phenotype of such clones would be LacZ⁻ or LacZ⁺. A finding of LacZ⁺ clones with (GAA)₇ would indicate some degree of flexibility in the number of repeats required for gene expression. A finding of LacZ⁻ (GAA)₇ clones would indicate that the overall length of the repeats is not the only consideration. The 12-bp SalI linker might have introduced structural changes to the pMGA promoter region independent of length, perhaps as a consequence of the linker consisting of a palindrome of high G+C content.

Because the LacZ⁺ subclones that contained the 12-bp linker had one more GAA repeat than had been anticipated, experiments were undertaken to determine whether the wild-type pMGA promoter might have more flexibility than previously realized with regard to the number of GAA repeats required for gene expression. Specifically, whether gene expression would occur when 13 GAA repeats were present was unknown. From a LacZ⁺ clone with the wild-type promoter sequence and 12 GAA repeats, 42 independent LacZ⁻ progeny were selected. The promoter region of each of these progeny was PCR amplified, and the nucleotide sequence was determined. Seven of the LacZ⁻ subclones had 13 GAA repeats. The other LacZ⁻ subclones had 14, 11, 10, 8, or 6 repeats. Thus, the wild-type promoter seems to strictly require 12 GAA repeats for gene expression, and 13 repeats, in particular, does not promote expression.

We propose the existence of a protein, hemagglutinin activator protein (HAP), that serves as a regulatory protein(s) that binds to the pMGA promoter region and stimulates transcription if 12 GAA repeats are present. Although negative regulators of pMGA transcription may also exist, they are not required to explain the data. Most pMGA gene family members in any given strain of M. gallisepticum are not transcribed and have a number of GAA repeats other than 12. Accordingly, it is our view that the pMGA promoter is insufficient to stimulate transcription in the absence of HAP. The pMGA genes lack a strong consensus sequence −10 and −35 promoter region upstream of the transcription start site (Fig. 1), which supports the contention that accessory factors are required. Because deletion of nucleotides upstream of the GAA repeats abolished GAA-dependent gene expression, we propose that one HAP-binding site is within these deleted sequences, as shown in Fig. 6. According to our model, a second HAP-binding site is located 3" of the GAA repeats, and changes in the number of repeats affect transcription by varying the spacing between the two HAP-binding sites. The GAA repeats might also serve as a HAP-binding site, which as discussed previously (7) may involve a triple helical structure with an unpaired single strand. Proteins that regulate transcription initiation in Mycoplasma species have not been reported. Thus, interactions between HAP and the pMGA genes would serve as a model for studying the regulation of mycoplasmal gene expression.

• Open in new tab
• Download powerpoint

FIG. 6.

Schematic illustration of the binding of HAP to sequences upstream and downstream of 12 GAA repeats. Numbers are distances (in nucleotides) from the indicated position to the transcription start site. Restriction enzyme sites: H, HindIII; C, Csp45I; X, XmnI.