Yeast Pch2 promotes domainal axis organization, timely recombination progression, and arrest of defective recombinosomes during meiosis

Abstract

We show that, during budding yeast meiosis, axis ensemble Hop1/Red1 and synaptonemal complex (SC) component Zip1 tend to occur in alternating strongly staining domains. The widely conserved AAA+-ATPase Pch2 mediates this pattern, likely by means of direct intervention along axes. Pch2 also coordinately promotes timely progression of cross-over (CO) and noncross-over (NCO) recombination. Oppositely, in a checkpoint-triggering aberrant situation (zip1Δ), Pch2 mediates robust arrest of stalled recombination complexes, likely via nucleolar localization. We suggest that, during WT meiosis, Pch2 promotes progression of SC-associated CO and NCO recombination complexes at a regulated early–midpachytene transition that is rate-limiting for later events; in contrast, during defective meiosis, Pch2 ensures that aberrant recombination complexes fail to progress so that intermediates can be harmlessly repaired during eventual return to growth. Positive vs. negative roles of Pch2 in the two situations are analogous to positive vs. negative roles of Mec1/ATR, suggesting that Pch2 might mediate Mec1/ATR activity. We further propose that regulatory surveillance of normal and abnormal interchromosomal interactions in mitotic and meiotic cells may involve “structure-dependent interchromosomal interaction” (SDIX) checkpoints.

Meiosis involves a temporally and spatially regulated program of interactions between homologous DNA sequences and their underlying structural axes. Local and global interactions at both levels ultimately yield chiasmata, which mediate homolog segregation. At pachytene, homolog axes are linked along their entire lengths via the SC, which forms during the preceding stage, zygotene. In budding yeast, the major SC central region component is Zip1 (1). Two meiosis-specific axis components, Red1 and Hop1, load prominently just before SC formation (2, 3) (see below).

Zygotene and pachytene also include progression and finalization of recombination. Recombination initiates earlier, during leptotene, via programmed double-strand breaks (DSBs) (4). Post-DSB recombination complexes are axis-associated and mediate rough juxtaposition of homolog axes (5). At the leptotene/zygotene transition, a subset of DSBs is designated for eventual maturation into CO recombination products; remaining DSBs are fated to become NCO products (6–8). During zygotene, at CO-designated sites, DSBs progress to single-end invasion (SEI) intermediates with concomitant nucleation and extension of SC formation (7–12). During pachytene, SEIs proceed to double Holliday junctions (dHJs) and then to COs; NCOs arise in another way (refs. 6 and 7 and N. Hunter and G.V.B., unpublished work). At early pachytene, all recombinosomes are SC-associated. CO complexes retain SC association through product formation; NCO complexes are released from the SC at early/mid pachytene, likely before product formation (13, 14). Pachytene exit and the first meiotic division (MI) then ensue.

Spatial and temporal coordination among these diverse events is ensured in part by direct physical contacts among involved components (3, 15), with progression from stage to stage involving specific regulatory transitions. The leptotene-to-zygotene transition is one major regulatory point (7, 10). Early/midpachytene could be another: It includes onset of post-SEI recombination stages (8), release of NCO complexes from the SC (12, 13), and release of Rec102 protein (16).

This study investigates midprophase events in WT and zip1Δ mutant meiosis, and the roles in these events of Pch2, a widely conserved AAA+-ATPase (17–19), originally implicated in checkpoint surveillance during mutant meiosis (17) and recently shown to also be important for recombination in WT meiosis in yeast (18, 20) and mouse (21).

Results

Zip1 and Hop1 Occur with Differential Hyperabundance Along Pachytene Chromosomes.

Zip1 localizes along pachytene chromosomes with uneven abundance. Illumination by Zip1-GFP (Fig. 1A) or indirect immunofluorescence (Fig. 1Ca), reveals regions of strong Zip1 staining (filled arrowhead) interrupted by regions of weak Zip1 staining (open arrowhead). Similar patterns were seen previously (but not emphasized) (7, 11, 22). Nonuniform abundance is peculiar, because the SC defined by electron microscopy is highly uniform along its length (e.g., ref. 23). Apparently, Zip1 occurs at low levels all along chromosomes (in the SC) and hyperabundantly in certain regions.

Fig. 1.

Localization of Zip1, Hop1, and Red1 along meiotic chromosomes in WT and pch2Δ meiosis. (A) Zip1-GFP localization in WT (SEY674). Filled and open arrowheads indicate regions of greater or lesser abundance, respectively. (B) Hop1 and Red1 localization in WT (NKY3330). (C) Zip1 and Hop1 staining in WT and pch2Δ. Pachytene in WT (a–d) and pch2Δ (e–h); zygotene in WT (i–k), and pch2Δ (m–o); leptotene in WT (l) and pch2Δ (p). WT (VBY338), pch2Δ (VBY1026). (D) Degree of overlap between strong Hop1 and Zip1 signals during zygotene and pachytene in WT (Left) and pch2Δ (Right). Zygotene (WT, t = 5, 6, and 7 h; n = 21, 12, and 11, respectively; pch2Δ, t = 5, 6, and 7 h; n = 26, 17, and 23, respectively) pachytene (WT, t = 5, 6, 7, and 8.5 h; n = 11, 18, 11, and 4, respectively; pch2Δ, t = 5, 6, 7, and 8.5 h; n = 6, 16, 12, and 25, respectively). (E) Abundance of continuous Hop1 lines in WT (black) and pch2Δ (red).

Axis components Hop1 and Red1 also exhibit regions of intense staining linked by regions of less intense staining identically for the two proteins (Fig. 1B), implying that they, too, occur all along chromosome axes but more abundantly in some regions than others. This pattern is documented molecularly for Red1 (3) and seen in previous cytological studies in two strain backgrounds (2, 24, 25).

Covisualization of Zip1 and Hop1 staining, at equivalent intensity levels, further reveals that the two proteins exhibit distinct and often complementary staining patterns [Fig. 1C a–d; see Visualization of Zip1 Domain Structure Is Evident in Minimally Adjusted Images in supporting information (SI) Text]. Usually, in any given region, one protein or the other strongly predominates, as seen by the tendency in overlap images for predominantly red or green coloring rather than yellow (Fig. 1C c–d). Often, one component stains prominently, whereas the other is barely visible, reciprocally in both directions. Published images reveal the same tendency for Zip1 vs. Red1 (2, 24, 25) (see Complementary Staining of Red1 and Zip1 in SI Text).

Analysis at various time points of meiosis in a synchronous culture further shows that (i) at leptotene, when Zip1 is present in only a few foci, Hop1 is already loaded abundantly, and discontinuously, along all chromosomes (Fig. 1Cl); (ii) The pattern of alternating Zip1/Hop1 localization emerges during zygotene, concomitant with SC assembly, and is stable throughout pachytene (Fig. 1D and SI Fig. 6). A pattern in which zygotene and pachytene nuclei exhibit regions enriched for either Hop1 (green) or Zip1 (red) compared with “yellow” regions of similar loading levels (Fig. 1C a–d and i–k) is seen throughout the period when zygotene/pachytene nuclei occur (t = 5–8.5 h; Fig. 1D Left) and therefore is not a characteristic of an early or late pachytene substage. Thus, the pachytene pattern arises by differential loading of Zip1 in response to an already-established Hop1(Red1) pattern.

Differential Hop1/Zip1 Hyperabundance Depends on Pch2.

In a pch2Δ mutant, Hop1 and Zip1 both occur abundantly along the entire lengths of pachytene chromosome axes: The full complement is readily visible by staining for either component and overlap images exhibit primarily yellow coloring (Fig. 1C e–h). This qualitatively altered pattern is apparent in all pachytene nuclei, even when they first appear, and in the majority of all nuclei where both proteins are present (Fig. 1D Right). Thus, Pch2 is required for the differential hyperabundance pattern of Hop1 vs. Zip1 seen along WT pachytene chromosomes. In contrast, Hop1 localization along leptotene pch2Δ chromosomes exhibits the same domainal hyperabundance seen in WT (Fig. 1Cp). Thus, the mutant defect arises after leptotene, concomitant with Zip1 loading/SC formation; moreover, this defect includes not only uniform loading of Zip1 but additional, promiscuous loading of Hop1 (as shown directly by high levels of nuclei with continuous, uniform Hop1 lines in pch2Δ, but not in WT; Fig. 1E). Experiments were performed at 33°C (rationale in ref. 7). Identical patterns occur at 23°C (SI Fig. 7 B and C).

Absence of Pch2 Coordinately Delays Progression of CO and NCO Recombination During Pachytene.

DNA events of recombination were examined at the HIS4LEU2 hot spot via standard constructs and physical assays (7) (see SI Fig. 8A]. In pch2Δ, COs are prominently delayed and reduced in level by 15–25% (Fig. 2A). This defect arises primarily after SEI formation (and thus during pachytene): (i) DSB formation occurs normally in pch2Δ [in a background where DSBs accumulate; data not shown (18)]; (ii) DSB turnover is normal, or nearly so, with little or no accumulation (Fig. 2A). Slightly delayed DSB turnover reported in ref. 20, also seen here, may partially reflect persistence of the “noninvaded ends” of SEIs. SEIs and dHJs accumulate to much higher than normal levels and persist at later than normal times in pch2Δ than in WT (Fig. 2A) with life spans increased ≥4-fold and ≥6-fold, respectively (see Calculation of DSB, SEI, dHJ, and SC Life Spans in SI Text).

Fig. 2.

Recombination at the HIS4LEU2 recombination hotspot in WT and pch2Δ. (A) (Left) Quantitative analysis of recombination intermediates and products. DSBs and COs were quantitated from 1D gels; SEIs, interhomolog (IH)-dHJs, and intersister (IS)-dHJs were quantitated from 2D gels. (Right) Representative time points from 2D gels. # indicates long joint molecules (26) (see SI Fig. 8 C and D for complete gel images). Cells that had completed meiosis I (MI±MII) were identified by determining the number of cells with two to four DAPI-staining nuclei. WT (VBY310), pch2Δ (VBY311). (B) (Upper) Southern blot analysis of a 1D gel after digestion with XhoI and MluI. (Lower) Quantification of blot shown in Upper. (Lower Left) Absolute amounts of CO-1 and NCO-1 recombination products (SI Fig. 8A). (Lower Right) The same data plotted as a percentage of maximum levels to compare kinetics of CO and NCO formation. WT (VBY338); pch2Δ (VBY1026).

NCO products occur at somewhat increased levels and with a significant delay in pch2Δ vs. WT (Fig. 2B). CO and NCO formation are delayed by exactly the same amount, ≈90 min (Fig. 2B; SI Fig. 8E). This coordinate effect is intriguing: CO/NCO differentiation occurs at leptotene, before pch2Δ has any discernible effect on recombination or axis composition, Pch2 may thus modulate the two pathways coordinately after their divergence.

Pachytene Exit and Meiosis I Are Delayed to the Same Extent as Recombination.

Leptotene and zygotene nuclei appear and disappear with very similar kinetics in WT and pch2Δ (Fig. 3A). In contrast, exit from pachytene is strongly delayed in pch2Δ: Pachytene nuclei persist at high levels even at t = 11 h, well after they have disappeared from WT nuclei (Fig. 3A).

Fig. 3.

Pachytene exit in WT and pch2Δ. (A) Kinetics of Zip1 loading and unloading in WT (black) and pch2Δ (red). More than 150 nuclei per time point were classified as leptotene, zygotene, and pachytene as shown in Fig. 1C a, i, and l. Leptotene and zygotene occur in pch2Δ with kinetics similar to WT. Nuclei with lines of Zip1 appear and disappear in a timely manner in the WT, yet accumulate and persist in pch2Δ at high levels; WT (VBY338), pch2Δ (VBY1026). (B) Relative timing of crossing over, pachytene exit, and first meiotic division. All data are from parallel WT and pch2Δ time courses, respectively (time course 84). Cross-overs and MI±MII are plotted as percentage of maximum levels. The cumulative curve for pachytene exit was calculated by processing primary pachytene curves from Fig. 3A as described in ref. 11. Ninety percent of cells were assumed to enter pachytene in both WT and pch2Δ, and pachytene was assumed to end at 24 h in pch2Δ, consistent with absence of nuclei containing Zip1 staining in pch2Δ at t = 24 h. (C) Zip1 and Hop1 loading kinetics in WT and pch2Δ. Nuclei (n ≥ 30) were classified in surface-spread nuclei immunodecorated with appropriate antibodies at the indicated time points based on levels of Hop1 and Zip1 staining. Nuclei with strong Hop1 and Zip1 loading (yellow) start decreasing in the WT at 6 h but accumulate and persist in pch2Δ. Nuclei with only Zip1 or Hop1 are rare in the WT at late time points, yet a class carrying only strong Hop1 staining (green) reappears at late time points in pch2Δ, suggesting delayed unloading of Hop1.

Pch2 is also required for the timely release of Hop1 from chromosomes during pachytene exit. In WT, Hop1 is present during the time of pachytene (Figs. 1E and 3C). Moreover, at t ≥ 4 h, chromosomes exhibit either abundant amounts of both proteins (yellow) or lack both proteins (gray). At t = 7–8.5 h (corresponding to pachytene exit), nuclei with only one protein or the other (red or green) are very rare and equally represented, suggesting that Hop1 and Zip1 are lost at the same time from WT chromosomes (Fig. 3C). In contrast, at late times in pch2Δ meiosis, many nuclei lack extensive Zip1 staining but exhibit strong Hop1 staining (Fig. 3C (green) t = 11 h; SI Fig. 6C), implying persistence of Hop1 after Zip1 disappears. In another study, Hop1 staining decreased earlier (2), apparently because of a strain background difference. Absence of Pch2 also affects progression through MI. Approximately 30% of pch2Δ nuclei never carry out this division; ≈70% carry out MI but after a substantial delay (Fig. 2A). Both defects are dependent on recombination initiation [SI Fig. 8F (18)].

Interestingly, in pch2Δ, the delay in pachytene exit, estimated by cumulative curve analysis (see Calculation of DSB, SEI, dHJ and SC Life Spans in SI Text), is very similar to the delay of CO appearance (Fig. 3B). Further, the subset of nuclei that undergo MI do so with a delay identical to that observed for COs/NCOs/pachytene exit (Fig. 3B). These relationships point to a linear chain of dependencies in which a recombination-related step is rate limiting for pachytene exit, which in turn is rate limiting for MI onset.

pch2Δ delays in turnover of dHJs, pachytene exit, and MI/spore formation are seen also at 23°C but are less pronounced than at 33°C (SI Fig. 7 A and D).

In zip1Δ, Absence of Pch2 Permits Increased Progression of Recombination.

In zip1Δ, formation of COs, but not NCOs, is dramatically compromised because of a leptotene/zygotene transition defect, with arrest of meiosis before MI (7) (Fig. 4A). Absence of Pch2 alleviates this arrest, partially (Fig. 4A; ref. 18) or completely (17) depending on strain background and conditions.

Fig. 4.

Effect of pch2Δ on recombination and meiotic progression in zip1Δ. (A) Meiotic divisions are partially restored in pch2Δ, with half of the cells progressing through meiosis with an ≈3-h delay, and the other half still exhibiting arrest (Left). Cross-over formation is partially restored in the zip1Δpch2Δ mutant (Right). (B) Quantitation of DSBs, SEIs, IH-dHJs, and IS-dHJs in WT, zip1Δ, and zip1Δpch2Δ (time course #90). WT (VBY310), zip1Δ (VBY1050), zip1Δpch2Δ (VBY312). (C) Ratios of IS/IH-dHJs in time course #90. See also SI Fig. 9.

Pch2 is relevant to recombination in zip1Δ: Absence of Pch2 significantly increases the level of COs formed in zip1Δ (Fig. 4A). Similarly, pch2Δ increases the ratio of intersister dHJs to interhomolog dHJs in the zip1Δ background, an effect not observed in WT background (compare zip1Δpch2Δ vs. zip1Δ and pch2Δ vs. WT in Fig. 4C; SI Fig. 9). Thus, without Pch2, stalled recombination complexes deteriorate such that DNA events now progress, with at least some events redirected to sister chromatids.

Interestingly, however, unlike in the WT background, pch2Δ has little or no effect on SEIs, or dHJs in zip1Δ (Fig. 4B). This likely reflects the fact that, in zip1Δ, only a small number of DSBs progress to SEIs and dHJs, as shown by the fact that a zip1Δ mutation drastically reduces SEI and dHJ levels in two situations where these species accumulate: pch2Δ (maximal dHJ levels of ≈2% total DNA in zip1Δpch2Δ vs. ≈7% in pch2Δ; Figs. 2A and 4B) and analogously in ndt80Δ (N. Hunter, personal communication).

Absence of Sir2 Mimics Absence of Pch2 in zip1Δ but Not in WT.

Pch2 occurs in foci along chromosome arms and is abundant in the nucleolus (ref. 17 and N. Joshi and G.V.B., unpublished work). Sir2 is required for nucleolar localization of Pch2 but not chromosomal localization (17). Nucleolar Pch2 is specifically implicated as the mediator of MI arrest in zip1Δ: In zip1Δ (i) a sir2Δ mutation mimics the effect of pch2Δ and (ii) Pch2 is no longer detectable along the lengths of chromosomes (17). In contrast, absence of Sir2 has no significant effect on recombination or on Zip1/Hop1 loading in otherwise WT meiosis (Fig. 5 A and B). Thus, Pch2 roles in WT meiosis are independent of nucleolar Pch2 and hence likely executed by chromosomal Pch2. In contrast, in zip1Δ, absence of Sir2 not only alleviates MI arrest [Fig. 5C; (17)], but confers effects similar to those of the absence of Pch2 on recombination, e.g., increased COs (Fig. 5 C and D and SI Fig. 9 A and C), implicating nucleolar Pch2 also in recombination arrest.

Fig. 5.

Absence of Sir2 results in a pch2Δ-like mutant phenotype only in a zip1Δ mutant background. (A) (Left) Formation of CO products (solid lines) and meiotic progression (dashed lines) in WT (black) and sir2Δ (green). COs were determined at the HIS4LEU2 hotspot and quantitated from a 1D gel. (Right) Levels of DSBs are from a 1D gel; SEI, IH-dHJ, and IS-dHJs were determined from 2D gels. Increased IH-dHJ levels in sir2Δ are detected only at a single time point and are likely due to an associated increase in CO products. Data are from time course #71. WT (VBY310), sir2Δ (VBY945). (B) Hop1 and Zip1 localization in sir2Δ pachytene nuclei. See Fig. 1C a–h for images of pachytene images and classes. For sir2Δ, pachytene nuclei (n = 14) at t = 6 h were analyzed for colocalization of strong Hop1 and Zip1 signals (data for WT and pch2Δ from Fig. 1D). Pachytene nuclei at t = 6 h are observed at the following frequencies: 14% (WT), 22% (sir2Δ), 40% (pch2Δ). Hop1 lines per class d at t = 6 h are observed with the following frequencies: 4% (WT), <1% (sir2Δ), and 31% (pch2Δ; n > 150 for WT and mutants). (C) Meiotic progression and recombination in zip1Δ and zip1Δsir2Δ. (D) Ratios of IS/IH-dHJs in zip1Δsir2Δ and zip1Δ (time course #90).

Discussion

Differential Hyperabundant Localization of Hop1/Red1 vs. Zip1 Dependent on Pch2.

Molecular studies show that Red1 localizes hyperabundantly to GC-rich vs. AT-rich domains (3). Differential abundance of axis components in R vs. G bands is also seen in mammalian mitotic chromosomes (28). These effects could be dictated directly by underlying base composition. Indeed, Hop1 binds preferentially to GC-rich DNA in vitro (27). However, a more dynamic/interactive process could also be involved, e.g., with loading patterns dictated by CO-designation sites during late leptotene.

Because Hop1 loads before Zip1, indistinguishably in WT and pch2Δ, Pch2 acts as a zygotene “stringency factor,” to prevent aberrant loading of Zip1 and concomitant additional loading of Hop1. This effect likely involves chromosomal Pch2. Intriguingly, Pch2 is a member of the AAA+-ATPase family, which frequently mediate the remodeling of multicomponent complexes.

Pch2 Is Required for Progression of Recombination During Pachytene.

Absence of Pch2 dramatically delays progression along both CO and NCO pathways but only mildly affects final product levels. Thus, the primary role of Pch2 is as a timing/progression factor. Coordinate delay of CO- and NCO-formation points to a single common underlying effect. However, the two branches of the recombination pathway diverge at late leptotene, before any thus-far-detected Pch2 role. Interestingly, steps affected by Pch2 during CO formation, SEI-to-dHJs and dHJs-to-COs, both occur in SC-associated recombinosomes at midpachytene (8, 14), which is the same time that NCO-fated recombinosomes are eliminated from the SC (see Introduction). We propose that NCO-fated recombination complexes are normally arrested in their biochemical progression until they are released from the SC. If so, a Pch2-dependent transition at midpachytene could trigger both this release (and thus NCO formation) and onset of the two SC-associated events of CO recombination (and thus CO formation). In accord with this model, absence of Pch2 during mouse meiosis delays turnover of SC-associated NCO-fated recombination complexes (20). Also, in plants (29), NCO complexes are more weakly associated with the axis/SC than CO complexes such that single coordinate SC/axis transition might differentially release more weakly bound NCO complexes, whereas more strongly bound CO complexes retain robust SC association. Pch2 might be important because of modulation of global axis status and/or local axis/recombinosome interfaces (3, 12).

The proposed early/midpachytene transition could also be the rate-limiting step for pachytene exit that, in turn, could be the rate-limiting step for meiosis I, thus explaining why pachytene exit and MI are both delayed in pch2Δ to the same extent as CO/NCO formation during pachytene. Pachytene exit/MI linkage is sensible because SC disappearance is immediately followed by onset of spindle pole body separation, which is the triggering event of MI spindle formation.

Roles of Pch2 in Mutant Meiosis.

In zip1Δ, where progression of CO-designated DSBs along the “designated CO pathway” is blocked at the DSB-to-SEI step (7), absence of Pch2 results in deterioration of stalled recombination complexes such that many DNA events progress more efficiently to COs or sister chromatids, respectively. Thus, Pch2 normally acts to keep stalled recombination complexes intact. Absence of Pch2 also restores high levels of CO formation in a nonnull zip1 mutant that makes SC (and thus SEIs) but arrests in pachytene (30); thus, Pch2 could analogously ensure recombination blocks at a later stage. Such a role is not restricted to Pch2: Absence of Dot1 also alleviates zip1Δ arrest and results in progression of stalled recombination via a deteriorated process, apparently before SEI formation, with redirection to intersister interactions (31).

Blockage of defective recombination is easily rationalized. Yeast cells blocked in meiosis before becoming committed to the MI division (i.e., before pachytene exit) can return to a mitotic cell cycle if nutrient conditions improve. If stalled but defective CO-fated recombination complexes remain arrested, their DNA intermediates can potentially be redirected, at earlier stages via mitotic DSB repair off the sister chromatid (e.g., refs. 32 and 33) or at the dHJ stage via topoisomerase-mediated resolution to NCO products (34). In contrast, if recombination progressed, interhomolog COs could be deleterious during the first ensuing mitotic division: Because sister arm cohesion persists through anaphase onset in mitosis, COs would create interhomolog connections that compete with regular bipolar sister alignment and segregation (35).

Absence of Pch2 (or Dot1) in zip1Δ (or dmc1Δ) meiosis abrogates both recombinational block and prophase arrest. Pch2 could mediate its effects on mutant recombination as an additional consequence of the checkpoint response proposed to mediate prophase arrest (17) or, alternatively, via its standard roles, as during WT meiosis. The first model seems more likely because (i) these effects involve nucleolar Pch2, whereas chromosomal effects do not (see above); (ii) Pch2 does not occur detectably along chromosomes in zip1Δ (17); and (iii) Pch2 acts positively to promote recombinational progression in WT meiosis but negatively to inhibit recombinational progression in mutant meiosis (see above). Furthermore, it is simpler and biologically more sensible if Pch2 coordinately mediates recombinational arrest and MI progression arrest as two parallel, independent mutation-induced checkpoint effects. Return of a stalled meiotic nucleus to the vegetative program with a regular chromosome complement requires not only innocuous resolution of recombination intermediates (see above) but absence of an (aberrant) reductional MI division, which would scramble the chromosome complement.

We note that the mouse homolog of Pch2, Trip13, seems to play no role in blocking either recombinational progression or MI progression (20). This difference could reflect different biological imperatives of mammals vs. yeast. In yeast, the “return to growth” option may dictate checkpoint arrest of both processes. In mouse oocytes and spermatocytes, return to the mitotic program does not occur, so neither type of arrest is useful. Moreover, recombination-defective meiosis in mouse triggers complete elimination of pregamete cells via apoptosis; this is likely preferable to any sort of simple in-place checkpoint-mediated block or return to mitosis.

Different Roles of Pch2 in WT and zip1Δ Mutant Meiosis.

Pch2 is involved in regulatory progression in both WT and zip1Δ meiosis, but positively vs. negatively in the two cases. This seemingly paradoxical situation is, in fact, exactly analogous to modulation of (mitotic) S-phase by Mec1/ATR (36). In WT, Mec1/ATR positively promotes progression of replication forks through genetically determined “programmed pause sites;” oppositely, when replication is defective, Mec1/ATR blocks replication progression, actively preventing deterioration of stalled replication forks and concomitantly blocking occurrence of mitotic progression, the target of which is G₂/M progression (i.e., SPB separation). Moreover, although this point is not always appreciated, ATR/Mec1 plays exactly the same paradoxically opposite roles for meiotic recombination, promoting normal recombination in WT meiosis (37–39) and mediating robust arrest of both recombination and SPB separation/MI in dmc1Δ and zip1Δ mutants (40, 41). Perhaps Pch2 exerts its effects in both WT and aberrant meiosis by promoting the regulatory action of Mec1/ATR.

Sensing the Status of Interchromosomal Interactions?

Yeast models suggest that surveillance of interhomolog interactions monitors the status of either (i) local events relating to the recombination complex and/or the recombinosome/axis interface or (ii) the global absence of SC along chromosomes (“synapsis”) (7, 18, 42). In Caenorhabditis elegans, SC can form in the absence of recombination, and situations that abrogate SC formation trigger an apparent Pch2-mediated checkpoint response, leading to the proposition that this organism has a “synapsis checkpoint” (43). However, in this case, the triggering defect is a local one, involving the aberrant status of an SC-nucleating “pairing site.” These and other observations can be reconciled as follows. In WT meiosis, a local interhomolog interaction would trigger a local change in axis status, which in turn would nucleate SC formation. The triggering interaction could be either recombination-dependent or recombination-independent according to the organism. Regulatory surveillance mechanisms (e.g., Mec1/ATR) would then monitor the second event, i.e., the local axis change, and thus neither recombination per se nor SC formation per se. An analogous process could potentially be involved outside of meiosis. During mitotic S-phase progression, sites of Mec1/ATR-mediated progression are preferential sites of condensin localization (44) and Mec1/ATR-mediated mitotic DSB repair involves cohesins (e.g., ref. 45). Thus, structure-dependent interchromosomal interaction (SDIX) checkpoints may be involved in many types of surveillance processes in both WT and aberrant situations.

Materials and Methods

Cytology.

Time courses and surface spreading of nuclei and immunodecoration with primary and secondary antibodies were performed as in ref. 7. Unselected nuclei were identified as DAPI signals and then assigned to appropriate Zip1 and Hop1 classes by analysis of the respective channels. Degree of overlap of strong signals was then determined by side-by-side comparison of the respective images. Immunodecoration used rabbit anti-Hop1 (F. Klein, University of Vienna, Vienna, Austria); mouse anti-Zip1 (Fig. 1 A–C and E; P. Moens, York University, Toronto, Canada), rabbit anti-Zip1 (Fig. 1D; S. Keeney, Memorial Sloan–Kettering Cancer Center, New York); or mouse anti-HA antibody (Covance) to detect Red1-HA (Fig. 1B). Secondary FITC and/or Texas-red antibodies were used in all cases. Zip1-GFP localization used strain SEY674, heterozygous for Zip1 containing GFP inserted at amino acid position 700 and WT Zip1, which is indistinguishable from an isogenic WT strain in all aspects of meiosis (S. Kameoka and N.K., unpublished work). In time courses, >150 nuclei were examined at each time point.

Strains.

VBY338 (a.k.a. NKY3639; ho::hisG/″, his4X. LEU2-(Mlu)-URA3/HIS4::LEU2-(NBam), ura3(Δsma-pst::hisG)/″, leu2::hisG/″); VBY1026 (same as VBY338, but pch2Δ::KanMX4/″); VBY310 (a.k.a. NKY3230; ho::hisG/″, his4X::LEU2-(NBam)-URA3/HIS4::LEU2-(NBam), ura3(Δsma-pst::hisG)/″, leu2::hisG/″). The following strains are isogenic to VBY310 except for the mutations indicated in parentheses: VBY311 (pch2Δ::KanMX4/″); VBY312 (pch2Δ::KanMX4/″, zip1Δ::KanMX4/″); VBY1099 (sir2Δ::KanMX4/″, zip1Δ::KanMX4/″, hml::HygroMX4/HML, HMR/hmr::HygroMX4); VBY1050 (a.k.a. NKY3624; zip1Δ::KanMX4/″); VBY945 (sir2Δ::KanMX4/″, hml::HygroMX4/HML, HMR/hmr::HygroMX4); VBY1179 (spo11Y135F::HygroMX4/″, pch2Δ::KanMX4/″); VBY1180 (spo11Y135F::HygroMX4/″); VBY1181 (pch2Δ::KanMX4/″); VBY1182 (SPO11/″, PCH2/″); NKY3330 (RED1-HA::URA3/″) (3); SEY674 (ZIP1-GFP-700aa-URA3@ZIP1/ZIP1).