Transcription factor Mts1/Mts2 (Atf1/Pcr1, Gad7/Pcr1) activates the M26 meiotic recombination hotspot in Schizosaccharomyces pombe

RESULTS

The mts1 and mts2 Genes Encode the Heterodimeric M26 Hotspot Binding Protein.

To conduct further biochemical and genetic studies of the M26 hotspot, we cloned and sequenced the genes encoding the Mts1/Mts2 protein (Fig. 3). The predicted polypeptides contain basic leucine zipper (bZIP) motifs characteristic of dimeric, sequence-specific DNA binding proteins. Both proteins have extensive sequence identity (≈50%), restricted to the bZIP domain, to transcription factors of the activating transcription factor/cAMP response element binding protein (ATF/CREB) family, suggesting that the Mts1/Mts2 heterodimer is a transcription factor. To confirm that the cloned genes encoded M26 recombination hotspot DNA binding proteins, we produced Mts1 and Mts2 in E. coli, purified the polypeptides, and characterized their binding affinity and specificity (Fig. 4). Heterodimeric protein bound specifically to DNA bearing the M26 site; it bound poorly to control DNA molecules that differed by a single base pair substitution within the M26 site (Fig. 4). The individual protein subunits also displayed hotspot-specific DNA binding activity, but with much lower affinity (data not shown). Thus, proteins expressed from the cloned mts genes exhibited DNA binding specificity similar to that of protein purified from S. pombe (15), confirming that the mts1 and mts2 genes encode the M26 hotspot-specific DNA binding proteins.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Cloning, sequence analysis, and disruption of the mts1 and mts2 genes. (A and D) The genes encode proteins with bZIP dimerization and DNA binding motifs (shaded boxes) ≈50% identical to those in ATF/CREB proteins. R, EcoRI; N, EcoNI; B, BalI; H, HincII; P, HpaI; A, AatII; D, NdeIII; S, BstEII. Standard procedures were used to disrupt the genes. (B and E) Confirmation of gene disruption by PCR. Lanes: 1, λ DNA digested with HindIII; 2, PCR product of wild-type DNA; 3, PCR product of disrupted DNA (mts1-D15::ura4⁺ or mts2-D1::his3⁺); 4, PCR product of positive control DNA (plasmids used to obtain gene targeting fragments). Primers were as shown by inward-pointing arrows in A and D. (C and F) Confirmation of gene disruption by Southern blot analysis. Lanes: 1, wild-type DNA; 2, disrupted DNA (mts1-D15::ura4⁺ or mts2-D1::his3⁺). Restriction maps are shown in A and D, and the expected positions of altered restriction fragments are indicated (∗). DNA was digested with EcoRI and probed with the 4,077-bp EcoRI–EcoRI fragment for mts1 constructs or digested with HincII and probed with the 3,421-bp HincII-HincII fragment for mts2 constructs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Hotspot-specific DNA binding activity of Mts1/Mts2 heterodimer purified from E. coli. (A) Coomassie-stained, 12% SDS/PAGE analysis of proteins. Lanes: 1, purified Mts1; 2, whole-cell lysate containing Mts1; 3, whole-cell lysate containing Mts2; 4, purified Mts2; 5, molecular mass standards. (B) Gel mobility shift assay by using a mixture of renatured Mts1 and Mts2 polypeptides. Position of M26-specific complex is indicated (∗). Gel shift conditions and probes were as described (15) and contained ≈5 nM probe, 25 nM Mts1, and 25 nM Mts2.

Mts1, But Not Mts2, Is Required for Osmotic Stress Response in Mitotic Cells.

The Mts1 and Mts2 proteins are implicated in several biochemical pathways, including response to stress (Fig. 2), but the mutants exhibit no mitotic phenotype under normal laboratory culture conditions (data not shown) (16–20). We therefore tested the mutants for mitotic phenotypes under a variety of adverse conditions. Cells that were mts1 mutant were highly sensitive to osmotic stress (Fig. 5). However, the mts2 mutants were resistant to osmotic stress. We and others have observed similar differential responses to other cytotoxic and genotoxic agents (data not shown) (16, 18–20). The differential response to stress indicates that Mts1 and Mts2 have nonoverlapping functions in some pathways that sense environmental, cytotoxic, and genotoxic signals. Because hotspot activation requires a heterodimer of Mts1 and Mts2 (Table 2), we infer that the mechanism of hotspot activation is distinct from the mechanisms of developmental regulation (transcriptional transactivation of target genes) that require only the individual Mts proteins (Fig. 5).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Sensitivity of mts⁻ mutants to osmotic stress. The indicated number of cells from healthy, log-phase cultures were spotted onto supplemented NBA minimal medium containing 1 M NaCl and were incubated for 3 days at 32°C before being photographed.

DISCUSSION

We have cloned the genes, mts1 and mts2, encoding the heterodimeric protein that binds the M26 recombination hotspot (15). Disruption of either gene, or both, abolishes M26 hotspot activity (Table 2), as predicted (15). While our genetic studies were underway, two laboratories described S. pombe genes identical to mts1 (called atf1 and gad7) and mts2 (called pcr1), both isolated through different lines of research (16–18). Genetic analyses revealed that Mts1 (Atf1, Gad7) and Mts2 (Pcr1) proteins have roles in sexual development and stress response and the mutants exhibit altered induction patterns of several genes involved in those pathways (16–20). These data suggest a possible relationship between transcription and recombination. However, hotspot activity is not a consequence of increased transcription because there is no detectable difference in the level of ade6 mRNA during meiosis in strains exhibiting normal (M375) vs. hotspot (M26) recombination (29) or in cells having or lacking the Mts proteins (data not shown). Furthermore, the factors required for stress response (Fig. 5) and hotspot activation (Table 2) are genetically distinct: Mts2 is dispensible for the stress response (Fig. 5) but not for hotspot activation (Table 2). [This separability has also been inferred for hotspots in Saccharomyces cerevisiae: some promoter mutations that alter gene expression do not alter recombination hotspot activity (30, 31)]. The Mts1 and Mts2 proteins acting together or separately therefore have multiple biological roles: one in hotspot activation, one in regulating genes involved in sexual development, and one in regulating genes during stress conditions. In each case, the proteins are constitutively present (15) but apparently inactive during normal mitotic growth, permitting rapid induction of activity in times of crisis. Activation of the Mts1 protein for transcription involves signal transduction via mitogen-activated protein (MAP) kinase and cAMP-dependent kinase pathways, and perhaps others (Fig. 2) (16–20). This allows the cells to rapidly respond to environmental changes without requiring further protein synthesis. The posttranslational modifications of Mts1 and Mts2 that trigger the decision toward meiotic development (16–20) may overlap with those changes that result in meiotic hotspot activation.

Mts1/Mts2 heterodimer binding the M26 site is required for hotspot activation. Although both Mts1 and Mts2 are involved in transcriptional regulation (16, 18, 20), the relevant cis-acting DNA sites have not been defined. Nor is it known whether Mts1 and Mts2 act together, or as homodimers, or as heterodimers with other bZIP proteins when regulating transcription. We expect that different homodimers and heterodimers have different related binding sites and different biological functions (Fig. 2). The use of the same factors for two different processes during meiosis, developmental gene regulation and M26 recombination hotspot activation, provides a good example of the economy of nature.

That some transcription factors may activate recombination hotspots has also been inferred from studies with S. cerevisiae (32–34). For example, full activity of the HIS4 hotspot requires elements of the promoter region that include binding sites for Bas1, Bas2, and Rap1 transcription factors (32, 35). A deletion that includes all the binding sites for these transcription factors abolishes hotspot activity, and replacement with Rap1 binding sites restores high hotspot activity, suggesting that Rap1 protein binding can activate the hotspot (32, 35, 36). A direct demonstration that Rap1 is required for hotspot activation has not been possible however, because the Rap1 protein is essential for viability. Similarly, a direct demonstration of a discrete nucleotide sequence defining the DNA site(s) has not been reported for the natural HIS4 hotspot, but has been for the M26 hotspot (14). The Rap1 protein, like Mts1/Mts2, is a multifunctional transcription factor that is constitutively expressed in mitotic and meiotic cells. Thus, for two hotspots in two different organisms, specific transcription factors have additional roles in meiotic hotspot activation. For other hotspots no single, discrete activating factors have yet been identified, but this could be caused by redundancy of sites clustered at hotspots or limitations of the experimental approaches used (11).

A current model for hotspot activation (10, 11, 34) proposes that the DNA within “open chromatin” becomes accessible to meiotic recombination enzymes that introduce a double-strand DNA break (37, 38). Alternatively, one or a few components that are usually associated with open chromatin, rather than open chromatin itself, enhance recombination (39). Because Mts1/Mts2 protein binding to M26 is required to activate the hotspot, the protein likely remodels chromatin, or recruits recombination enzymes, or both. Meiotically induced, M26 site-dependent open chromatin is found at the M26 hotspot (40) and is entirely dependent on the presence of the Mts1/Mts2 heterodimer (K. Ohta and W.P.W., unpublished observations). [The similarity to “architectural transcription factors” (41) and chromatin remodeling enzymes involved in gene regulation (42, 43) is striking. Perhaps some factors of those processes are also components of M26 hotspot activation.] The current evidence is consistent with open chromatin being either the cause or the effect of hotspot activation: Mts1/Mts2 might create meiotically altered chromatin that serves as an assembly point for meiotic recombination enzymes, or the chromatin alteration might simply be a consequence of Mts1/Mts2 activating recombination while bound to M26. Because the Mts proteins are also transcription factors (16–20), and hotspot activation is not the result of increased transcription (29) (data not shown), we infer that the Mts1/Mts2 heterodimer has separate roles in transcriptional regulation and hotspot activation. It remains to be seen whether Mts1/Mts2 directly recruits recombination enzymes via protein–protein interactions, or whether recombination enzymes assemble at chromatin conformationally altered by Mts1/Mts2 binding to M26 DNA, or both. In either case, binding of Mts1/Mts2 to M26 sites increases recombination above basal levels, indicating that the hotspot with its binding protein regulates where and when recombination occurs within the genome.

Recombination serves a surveillance role in mitotic cells (to repair certain types of DNA damage) and recombination rates are elevated during meiosis, presumably to increase genetic diversity or to insure homolog disjunction or both. Part of the meiotic enhancement is caused by induction of meiotic recombination genes (27) and part may be because of Mts1/Mts2 protein binding to M26 sites throughout the genome (15). Intergenic recombination in mts mutants is reduced as much as 50% in four intervals tested so far (N.K. and W.P.W., unpublished observations), and each of eight M26 sites created by mutagenesis (involving one or a few base pair changes) at five sites in two widely separated loci has hotspot activity (44). These findings support the view that some natural M26 sites are active in meiosis and may contribute to elevated meiotic recombination levels.

These findings permit a speculative calculation of the fraction of meiotic recombination attributable to Mts1/Mts2 heterodimer interacting with M26 sites. There are an estimated 600 M26 sites in the diploid S. pombe genome (15). If each M26 site is, on average, as recombinogenic as ade6-M26 and each promotes recombination in 5% of meioses [the level of gene conversion at ade6-M26 (5)], then ≈30 conversion events per meiosis would be due to M26 sites. About 65% of the conversion events of ade6-M26 are accompanied by reciprocal exchange (28). Thus, ≈20 reciprocal exchange events per meiosis could be because of M26 sites. Because the S. pombe meiotic map size is ≈2,200 cM (≈44 reciprocal exchanges) (45), the predicted 600 M26 sites in the diploid genome could be responsible for ≈50% of meiotic homologous recombination. However, because no study has yet reported whether or not any naturally occurring M26 sites are recombinogenic, this hypothesis remains unproven. The availability of mts mutants that abolish ade6-M26 hotspot activity will be useful for determining what fraction of genomic M26 sites are active.