No one had seen one of these huge red stars wink out of existence with so little fuss before.
Image credit: Shutterstock/Genevieve de Messieres.
New Research In
Physical Sciences
Featured Portals
Articles by Topic
Biological Sciences
Featured Portals
Articles by Topic
-
Edited by Nancy Kleckner, Harvard University, Cambridge, MA, and approved March 12, 2007 (received for review July 20, 2006)
Abstract
In meiotic prophase, telomeres associate with the nuclear envelope and accumulate adjacent to the centrosome/spindle pole to form the chromosome bouquet, a well conserved event that in Saccharomyces cerevisiae requires the meiotic telomere protein Ndj1p. Ndj1p interacts with Mps3p, a nuclear envelope SUN domain protein that is required for spindle pole body duplication and for sister chromatid cohesion. Removal of the Ndj1p-interaction domain from MPS3 creates an ndj1Δ-like separation-of-function allele, and Ndj1p and Mps3p are codependent for stable association with the telomeres. SUN domain proteins are found in the nuclear envelope across phyla and are implicated in mediating interactions between the interior of the nucleus and the cytoskeleton. Our observations indicate a general mechanism for meiotic telomere movements.
Homologous chromosome pairing and recombination during meiotic prophase are required to orient chromosomes for disjunction and haploidization during the meiotic divisions. Early in meiotic prophase, telomeres attach to and move along the nuclear envelope to concentrate transiently in one sector of the nucleus, generally adjacent to the centrosome, forming the chromosome bouquet, a widely conserved arrangement (1–7). Bouquet formation may promote homologous chromosome pairing and also may help to untangle chromosomes, to regulate recombination at telomeres, to regulate crossover distribution, to coordinate or synchronize chromosomal events by propagating signals from the telomeres, and/or to facilitate synaptonemal complex formation (see refs. 1, 5, and 7–10). Despite the wide conservation of bouquet formation, the molecular mechanisms responsible for telomere attachments and movements in meiosis are poorly understood.
In Saccharomyces cerevisiae, the meiotic bouquet resembles that of multicellular organisms (11–14). Ndj1p is a meiosis-specific protein that accumulates at telomeres in S. cerevisiae (15, 16) and is required for bouquet formation (17). Deletion of NDJ1 delays axial element formation, homolog pairing, synapsis and onset of the first meiotic division, causes an elevated frequency of nondisjunction and of ectopic recombination, and reduces spore formation and viability (15–21). Early recombination intermediates form between homologs with wild-type kinetics in ndj1Δ, suggesting that some aspect of bouquet formation is required for the normal coupling of the molecular and cytological events of pairing (22).
A large-scale two-hybrid screen (23) identified an interaction between NDJ1 and MPS3/NEP98 (24, 25). Mps3p is an essential, integral membrane protein that is concentrated at the spindle pole body (SPB), the S. cerevisiae centrosome equivalent, and is present at lower levels throughout the nuclear membrane. Mps3p is required for duplication of the SPB (24, 25) and also functions in karyogamy and in sister chromatid cohesion (25, 26). The present study demonstrates a critical requirement for the Ndj1p-Mps3p interaction for bouquet formation in budding yeast. Mps3p is a member of the SUN family of proteins which have been identified in the nuclear envelope of yeasts, worms and mammals, and have been implicated in bridging the nuclear envelope to connect the nuclear contents with the cytoskeleton in mitosis and in meiosis (27–37). Identification of Mps3p as a conserved component of the machinery that moves chromosomes in meiotic prophase will lead to further components of the machinery and to new, testable models for the regulation and function of these movements.
Results
Mps3p Amino Acids 2–64 Are Sufficient for Interaction with Ndj1p.
We tested fragments of Mps3p for two-hybrid interaction with Ndj1p bait and found that Mps3p residues 2–64 are sufficient for interaction with Ndj1p (Fig. 1 A). This interaction is consistent with data that orients Mps3p with the amino terminus protruding into the nucleoplasm (25). Overexpression of Ndj1p causes a block in meiosis (16) which is suppressed by mps3 Δ2–64, supporting the 2-hybrid indication of an interaction between these proteins in vivo, during meiosis (data not shown). The significance of a second interaction with Ndj1p of Mps3p residues 240–361, which presumably lie within the lumen of the nuclear envelope, is being examined but deletion of this region does not give rise to the ndj1Δ-like phenotype seen in mps3 Δ2–64 [Fig. 1 B–E; supporting information (SI) Fig. 6].
Fig. 1.
Deletion of a region of Mps3p that interacts with Ndj1p causes a delay in homologous chromosome pairing during meiotic prophase. (A) Two-hybrid assays demonstrate that two regions of Mps3p, residues 2–64 and ≈240–361, each are sufficient for interaction with Ndj1p. Protein sequence motifs indicated on the map include an acidic region, the transmembrane domain (TM), a potential Walker P-loop (P), two coiled-coil regions (cc), a hydrophobic region(h'phob), a polyglutamine stretch (pQ), and a SUN domain. (B) Entry into anaphase I is delayed in ndj1Δ, mps3 Δ2–64, and ndj1Δ mps3 Δ2–64. Cells were scored at the fluorescence microscope as having single/round, elongated, two or more nuclei by using DAPI to label the DNA. Strains are wild-type (MCY506 × 507), ndj1Δ (MCY1405 × 1409), mps3 Δ2–64 (MCY1385 × 1389), and ndj1Δ mps3 Δ2–64 (MCY1397 × 1400). (C–E) Integrated arrays of lacO tagged with lacI-GFP appear as two spots (“unpaired”) or as one spot (“paired”) in strains with homozygous integrations in the middle of the left arm of chromosome VII (C), at the left telomere of chromosome III (D), and at the right telomere of chromosome IV (E). Two hundred nuclei were scored at each time point, and results were averaged for three or four experiments. Large variations in initial telomere pairing between different cultures occurred for all of the genotypes and account for the large error bars. However, in all experiments, the mutants were delayed to a similar extent relative to wild-type. Strains used are as follows: wild-type (MDY2256 × 2196), ndj1Δ (MDY2254 × 2198), and mps3 Δ2–64 (MDY2263 × 2201) (C); wild-type (MDY2366 × 2370), ndj1Δ (MDY2379 × 2371), mps3 Δ2–64 (MDY2381 × 2393), and ndj1Δ mps3 Δ2–64 (MDY2384 × 2390) (D); and wild-type (MDY2426 × 2266), ndj1Δ (MDY2431 × 2279), mps3 Δ2–64 (MDY2434 × 2283), and ndj1Δ mps3 Δ2–64 (MDY2436 × 2304) (E). Error bars indicate standard deviation.
MPS3 and NDJ1 Function in the Same Pathway in Meiosis.
Residues 2–64 of Mps3p are not essential for viability (25), facilitating examination of the meiotic phenotype of the mps3 Δ2–64 allele. Delays in meiotic prophase (Fig. 1 B), defects in ascus formation and reductions in spore viability (SI Fig. 6) are similar in ndj1Δ, mps3 Δ2–64 and the double mutant ndj1Δ mps3 Δ2–64.
Ndj1p and Mps3p Facilitate Homologous Chromosome Pairing.
Meiotic chromosome pairing generally culminates in the formation of synaptonemal complexes (SCs) which hold homologous chromosome regions in close proximity, at distances near or below the resolution limits of standard fluorescence microscopy. Cytological assays of homolog proximity, or of SC formation per se, have suggested that pairing is delayed in ndj1Δ (15–17). We assayed pairing cytologically in intact cells of diploid strains, using chromosomes marked by lacI-GFP binding to tandem arrays of lacO (38). These markers were integrated in the middle of the left arm of chromosome VII, at the left telomere of chromosome III or at the right telomere of chromosome IV, positions representing an interstitial site (VII) and telomere-proximal sites of short (III) and long (IV) chromosome arms. In strains with homozygous spots, the numbers of distinct fluorescent spots indicate whether the marked chromosome regions are unpaired (2 spots) or paired (1 spot) (39, 40). An interstitial locus on VIIL was paired in 40–50% of wild-type cells upon entry into sporulation (Fig. 1 C), similar to results published by others (40–42). The fraction of paired loci then declined slightly before increasing to a maximum value at ≈6.5 h, just before the first meiotic division. In ndj1Δ, mps3 Δ2–64 and the double mutant ndj1Δ mps3 Δ2–64, pairing decreases further than in wild-type before increasing slowly to reach levels comparable with wild-type at 7.5–8 h, just before onset of the (delayed) meiotic divisions. Pairing also is delayed at the telomeres of IIIL (Fig. 1 C) and IVR (Fig. 1 D). To estimate the probability that two telomere signals might coincide in our assay because of random proximity, we analyzed cells heterozygous for two telomere markers, one on IIIL and one on IVR. These signals overlapped in 20–25% of cells at all time points (data not shown), indicating that the majority of the observed associations between homologous loci are not accounted for by random associations or by heterologous associations mediated by bouquet formation.
The delay in pairing arises mainly after entry into meiotic prophase. The kinetics of premeiotic DNA synthesis are similar in wild-type, ndj1Δ and mps3 Δ2–64 (SI Fig. 7A ; and for ndj1Δ, see ref. 22). Additionally, meiotic DNA double-strand break formation occurs with only slight delays in ndj1Δ (22) and in mps3 Δ2–64 (A. Shinohara, personal communication). Finally, initiation of meiotic recombination in ndj1Δ and mps3 Δ2–64, as measured by a return-to-growth assay of gene conversion of heteroalleles, is similar in timing although slightly reduced in final levels as compared with wild-type (ref. 16 and SI Fig. 7B ). Thus Mps3p, like Ndj1p, functions to promote timely pairing of telomeric and of nontelomeric loci in meiotic prophase.
Sister chromatid cohesion defects have been reported in mps3 mutants (26, 43). Similarly, unlike the case for wild-type cells, separated sister spots at the IVR telomere are present in nonanaphase cells during vegetative growth in mps3 Δ2–64, increase in frequency after the shift into sporulation medium, then fuse into a single spot before the first meiotic division (SI Fig. 8). A small but reproducible fraction of ndj1Δ cells also have separated sister spots in meiosis (SI Fig. 8). Separation was not seen for the interstitial VIIL sister spots and was intermediate in frequency for the IIIL telomere sister spots, where it was rarely seen in ndj1Δ and reached maximum levels of ≈1/2 that seen at the IVR telomere in mps3 Δ2–64 and ndj1Δ mps3 Δ2–64 strains (data not shown). It is not clear how or whether this sister separation defect is related to the other meiotic defects observed in ndj1Δ and in mps3 Δ2–64.
Mps3p and Ndj1p Are Mutually Dependent for Stable Accumulation at Telomeres.
In living and in whole, fixed wild-type and ndj1Δ cells, prominent patches of Mps3-YFP are seen in the meiotic prophase nuclear envelope, away from the SPB (Fig. 2, compare A with B and C). These non-SPB accumulations are not visible in mps3 Δ2–64 (Mps3Δ2–64-YFP in Fig. 2 D) but are present in ndj1Δ (Mps3-YFP in Fig. 2 C), indicating an Ndj1p-independent function for the 2–64 region. To localize Mps3-YFP with respect to telomeres, cells were marked with Rap1-CFP, which concentrates at telomeres, and with Spc42-dsRed, which identifies the SPB. Mps3-YFP at the SPB is visualized as a small, discrete spot (Fig. 3 A). Away from the SPB, Mps3-YFP appears in less discrete spots that generally are found in association with the Rap1-CFP signal, hence, at telomeres (Fig. 3 B–D). Even in bouquet-stage nuclei, the Mps3-YFP that associates with the Rap1-CFP signal generally appears separate from the Mps3-YFP signal in the SPB (arrowhead in Fig. 3 D). Mps3p and Ndj1p association with telomeres is sufficiently stable to persist in spread, flattened preparations of nuclei from wild-type cells [Fig. 4 A–C and G–I; (15, 16)], indicating that parts of the nuclear envelope remain associated with the chromosome ends in these preparations (as do nuclear pores; SI Fig. 9). In ndj1Δ, although patches of Mps3p still are seen along the nuclear membrane away from the SPB (Fig. 2 C), Mps3p no longer is associated with telomeres in the spread preparations (Fig. 4 J–L), suggesting that stability of the telomere-Mps3p interaction requires Ndj1p. Similarly, in spread preparations of mps3 Δ2–64, Ndj1p is occasionally seen in patches but no longer appears associated with the telomeres (Fig. 4 D–F), suggesting that Ndj1p and Mps3p each stabilizes telomere association of the other.
Fig. 2.
Mps3p patches in the nuclear membrane, away from the spindle pole body, are more prominent in meiotic than in mitotic cells in wild-type and ndj1Δ but are absent in mps3 Δ2–64. Single focal plane images of fields of whole cells labeled with YFP-tagged Mps3p reveal mainly SPB localization (the small, bright dots) in vegetative cells (A), larger accumulations of Mps3p in the nuclear membrane away from the SPB in meiotic prophase in wild-type (B) and in ndj1Δ (C) but little or no extra-SPB accumulation in meiotic prophase in mps3Δ2–64 p-YFP (D). Strains used are as follows: MCY1443 × 1469 (A); MCY1443 × 1469 (B); MCY1449 × 1450 (C); and MCY1472 × 1474 (D). (Scale bar: 4 μm.)
Fig. 3.
Mps3p colocalizes with Rap1p in telomere clusters during meiotic prophase. Maximum intensity projection images of meiotic prophase nuclei from whole wild-type cells taken at 5 h in sporulation, marked with Mps3-YFP, with Rap1-CFP, which concentrates at telomeres and with Spc42-dsRed to visualize the SPB (A) or stained with DAPI to label DNA (B–D). Mps3-YFP at the SPB (white arrowheads) generally is distinct from Mps3-YFP associated with patches of Rap1-CFP. Strains used are as follows: MDY2877 × 2879 (A); MCY1824 × 1826 (B–D). (Scale bar: 2 μm.)
Fig. 4.
Ndj1p and Mps3p are mutually dependent for stable association with telomeres in meiotic prophase. Spread preparations of nuclei from mid-meiotic prophase were labeled with DAPI (DNA; A, D, G, and J), anti-HA antibody (Ndj1p in NDJ1-HA strains; B and E), and/or anti-fluorescent protein antibody (Mps3p in MPS3-YFP strains; H and K) in wild-type (A–C and G–I), mps3 Δ2–64 (D–F) and ndj1Δ (J–L) backgrounds. Ndj1p and Mps3p localize to telomeres in wild-type strains (C and I, respectively) and can accumulate in patches that seem to connect telomeres of different bivalents [Mps3p in I; Ndj1p, see Fig. 3 A; (16)]. Ndj1p and Mps3p accumulations appear separate from telomeres in mps3 Δ2–64 (F) and ndj1Δ (L) strains, respectively, presumably marking the location of patches of nuclear envelope that were not removed during the spreading process. Strains are as follows: wild-type (MCY1460 × 1462), mps3 Δ2–64 (MCY1461 × 1463), and ndj1Δ (MCY1648 × 1649). (Scale bar: 4 μm.)
Mps3p Is Required for Bouquet Formation.
Telomere clustering and proximity to the SPB was assayed in whole cells by using signals from DNA (DAPI), the SPB (Tub1-GFP or Spc42-dsRed) and telomeres (Rap1-CFP; Rap1p remains associated with the telomeres in ndj1Δ and mps3 Δ2–64; SI Fig. 10). Nuclei identified subjectively as “tight” bouquets, where a single prominent telomere cluster lies immediately adjacent to the SPB, peaked in frequency between 3 and 6 h in sporulation in wild-type populations, i.e., during meiotic prophase (Fig. 5 A). Bouquet frequencies in mps3 Δ2–64 are reduced to ndj1Δ levels, indicating that mps3 Δ2–64 is defective in forming or maintaining telomere clusters and/or in bringing telomeres to the SPB. “Loose” bouquets, where telomere clusters are limited to the half of the nucleus containing the SPB but are not immediately adjacent to it (as in Fig. 3 C), and where telomere signal can be spread over more than one cluster, similarly appear more prevalent in wild-type populations (data not shown).
Fig. 5.
Mps3p and Ndj1p are required for bouquet formation and for accumulation in rec8Δ of large telomere clusters. (A) ndj1Δ and mps3 Δ2–64 are defective in bouquet formation. Three-color, 3D, deconvolved, merged images of whole cells marked with Tub1-GFP (which identifies the SPB), with Rap1-CFP and stained with DAPI (which labels the DNA), were scored as “tight” bouquets when all of the bright Rap1-CFP/telomere was adjacent to the Tub1-GFP/SPB. Strains are wild-type (MDY2101 × 2103), ndj1Δ (MDY2101 × 2104) and mps3 Δ2–64 (MCY1384 × 1391). (B) ndj1Δ and mps3 Δ2–64 suppress formation of the large telomere clusters in rec8Δ meiotic prophase arrest. Whole cells labeled and imaged as in A were assayed for the numbers of telomere clusters, defined as local maxima above a predefined threshold in the Rap1-CFP signal, by automated, objective scoring for each nucleus having a single SPB (see legend to SI Fig. 11 for methods details). Only small changes in distributions were seen between samples of a given genotype taken at 3, 4, 5, and 6 h in sporulation, so these data are pooled here. Filled bars, 1–3 telomere clusters; open bars, 4 or more telomere clusters. Total cells scored were as follows: wild-type, 665; ndj1Δ, 632; mps3 Δ2–64, 496; rec8Δ, 635; rec8Δ ndj1Δ, 708; rec8Δ mps3 Δ2–64, 565. Strains are as follows: wild-type (MDY2455 × 2513), ndj1Δ (MCY1570 × 1571), mps3 Δ2–64 (MCY1553 × 1554), rec8Δ (MDY2517 × 2534), ndj1Δ rec8Δ (MCY1533 × 1535), and mps3 Δ2–64 rec8Δ (MCY1531 × 1532).
NDJ1 and MPS3 Are Required for Formation of the Bouquet-Like Telomere Cluster in rec8Δ.
Given the transient nature of the bouquet stage and imperfect synchrony of the cultures, the bouquet is detected in only a fraction of the population. However, deletion of the meiosis-specific cohesin subunit Rec8p causes a block in meiotic prophase (44) in which telomeres accumulate in one or a small number of clusters, a configuration assumed to be related to bouquet formation even though the clusters frequently are not close to the SPB (14). To test directly whether the bouquet formation machinery is involved in the formation of these clusters, we examined telomere positioning in rec8Δ, rec8Δ ndj1Δ and rec8Δ mps3 Δ2–64 mutants. An automated approach to identify and count the numbers of discrete telomere clusters was developed to quantify the dispersal of the Rap1-CFP telomere signal (i) to determine its proximity to the SPB to independently confirm the subjective bouquet scores, (ii) to measure changes in the distribution of telomeres in the nucleus (e.g., homogeneously spread around the nucleus versus residing in a restricted region, even if in multiple clusters) and (iii) to ask whether telomeres are more clustered when in relatively close proximity to the SPB [SI Fig. 11(detailed methods are described in the legend)].
As expected rec8Δ caused almost complete loss of cohesion during meiotic prophase (data not shown) and did not cause a bouquet arrest per se but did cause persistent meiotic telomere clustering which was maintained independent of association with the SPB (ref. 14 and SI Fig. 11). However, whereas rec8Δ ndj1Δ and rec8Δ mps3 Δ2–64 cells arrest in meiotic prophase (data not shown) and have concentrations of Ndj1p that colocalize at the telomeres (i.e., at Rap1p foci; SI Fig. 12), these mutants do not accumulate the telomere clusters seen in rec8Δ (Fig. 5 B). Nuclei with telomeres close to the SPB are depleted in ndj1Δ and in mps3 Δ2–64, both in REC8 (consistent with the visual bouquet scores) and in rec8Δ backgrounds (SI Fig. 11). It is not known whether the large telomere clusters seen in rec8Δ form in association with the SPB and are maintained after moving away, or form independent of association with the SPB (14) but, in either case, cluster formation requires Ndj1p and Mps3p, potentially to generate movements of telomeres toward each other and/or toward the SPB. Thus the clusters that accumulate in rec8Δ do require bouquet pathway genes.
Discussion
Bouquet formation requires several regulated events, including telomere attachment to the nuclear envelope, interaction with motors (or other force-generating mechanisms), movements that culminate in their collection at the base of the bouquet, and transient retention of telomeres in the bouquet arrangement before they are dispersed over the nuclear envelope. Various functions have been attributed to the bouquet (see introduction), and it is possible that the different bouquet-related events contribute individually to the different functions. For example, association with Ndj1p and Mps3p could stabilize telomere structure and, by contributing to anchoring at the nuclear envelope, bring about structural changes in the chromosomes that regulate the distribution of crossovers (see ref. 15), perhaps by means of tension (7, 45, 46). Alternatively, or in addition, telomere-promoted movements themselves may contribute to the normal pattern of crossovers, perhaps in part by promoting stable homologous interactions, pairing and synapsis. In ndj1Δ, mild delays in the earliest DNA transactions in meiotic recombination are followed by substantial delays in later steps (22). Although the final level of pairing in ndj1Δ and mps3 Δ2–64 is approximately wild-type, it is possible that the delay in synapsis accounts for the reduced recombination seen in ndj1Δ and mps3 Δ2–64, perhaps by prolonging exposure of recombination intermediates to activities that reduce recombination (see ref. 47). Finally, telomere movements could contribute to the dissolution of nonhomologous associations such as those between heterologous centromeres (48) and other ectopic interactions (18, 39), essentially editing out connections that might hinder rather than promote appropriate disjunction in the first meiotic division (49). The role played by transient accumulation of telomeres adjacent to the spindle pole, specifically, remains unclear.
We have demonstrated that Mps3p plays an essential role in the NDJ1 pathway for bouquet formation in S. cerevisiae, in addition to its roles in SPB duplication, karyogamy and sister chromatid cohesion (24–26, 43, 50). Our observations suggest a working model for Mps3p function in meiotic telomere behavior in which during meiotic prophase, Mps3p accumulates in the nuclear envelope away from the SPB and is stably attached to telomeres by its association with Ndj1p. The 2-hybrid interaction between Mps3p and Ndj1p, and the observation that Mps3p pulls down Ndj1p in coimmune precipitation experiments (A. Shinohara, personal communication), suggest that their association at telomeres is relatively direct. Association of Mps3p with further components at the nuclear membrane, perhaps Mps2p (50), then links telomeres across the nuclear envelope to motors that move the telomeres to congregate in a single cluster and to form the bouquet.
This model is derived in part from the observation that Mps3p contains a SUN domain, originally defined as a region of amino acid sequence similarity between Schizosaccharomyces pombe Sad1p and Caenorhabditis elegans Unc-84 (35). SUN domain proteins are present in eukaryotes from fungi to vertebrates, and SUN domains are proposed to be protein interaction domains (28, 29, 31, 32, 35, 51, 52). Unc-84, an inner nuclear membrane protein, is proposed to bridge the perinuclear space to interact with a second protein, Anc-1, of the outer nuclear envelope (32, 37, 53). Anc-1, in turn, interacts with the actin cytoskeleton to anchor the nucleus. A similar model has been elaborated for the role of a second C. elegans SUN domain protein, Sun1/Matefin, in mediating interaction of the nucleus with the centrosome (29, 33), and the mammalian Sun1 and Sun2 proteins also are implicated in bridging the nuclear envelope (54, 55). A role for the actin cytoskeleton in bouquet formation is suggested by the cessation of telomere clustering during meiotic prophase in S. cerevisiae after addition of latrunculin, which depolymerizes actin filaments (14).
The role of Ndj1p in promoting the association of telomeres with Mps3p in S. cerevisiae strikingly parallels the requirement for Bqt1p and Bqt2p to connect telomeres to Sad1p and for the subsequent accumulation of telomeres at the SPB during meiotic prophase in S. pombe (27). This similarity suggests conservation of the mechanism that actively positions telomeres in meiosis despite considerable divergence in subsequent events. In S. pombe, telomeres move to and are held at the SPB throughout meiotic prophase in a dynein-independent process, whereas the nucleus sweeps back and forth in the cell throughout meiotic prophase by a dynein-dependent mechanism that is required for normal meiotic recombination (56, 57). The long duration of this stage and the dramatic nuclear movements may be required to promote and stabilize homolog alignment because the synaptonemal complex is absent in S. pombe (58). In the more canonical meiosis of S. cerevisiae, the bouquet is transient (13), homolog pairing is stabilized by synaptonemal complex formation, the nucleus as a whole is relatively immobile whereas the telomeres move throughout meiotic prophase (ref. 14; M.N.C., C.-Y.L., and M.E.D., unpublished results), and dynein seems dispensable for a largely normal meiotic outcome (refs. 59 and 60; M.N.C., unpublished observation). Thus, in S. pombe the bouquet arrangement may be built through relatively static contacts between the components, as proposed (27), whereas in S. cerevisiae the associations of telomeres with each other and with the SPB are more dynamic, regulated by mechanisms that have not yet been defined. Understanding the role of Mps3p and of additional proteins in the NDJ1 pathway will provide a better understanding of the molecular mechanisms that position telomeres and will provide further insights into the role of the bouquet in meiotic chromosome metabolism.
Materials and Methods
DNA Constructs.
Ndj1p bait (amino acids 12–352) was cloned into pAS1-CYH2. MPS3 prey constructs were derived from PCR of pSJ140 (24) and were cloned into pGADT7. Two-hybrid assays were conducted in the AH109 strain background [Clontech (Mountain View, CA) Matchmaker 3 system (61)].
mps3 Δ2–64 was constructed by ligating PCR fragments that deleted codons 2 through 64. and was cloned into YIplac204 and YIplac211 to construct pMCB739 and pMCB740, respectively. Plasmids were digested with ClaI to target integration at MPS3 and excision events were selected on 5-fluoroorotic acid (5-FOA) for loss of URA3 (62) or 5-fluoroanthranilic acid for loss of TRP1 (63).
Rap1-CFP (pMCB414) was constructed by subcloning a 3.6 kb NheI-PvuII RAP1 fragment into pRS305 digested with XbaI-SmaI. CFP was PCR amplified from pDH3 (Yeast Resource Center, University of Washington, Seattle, WA) and inserted into the NruI site of RAP1, as described for Rap1-GFP (64). Digestion with SpeI targeted plasmid integration for 2-step gene replacement.
Mps3p was tagged at the C terminus with YFP by PCR (65). An SPC42-GFP plasmid (pIA29) provided by Ian Adams and John Kilmartin (66) was used to build fluorescent constructs that mark the spindle pole body. SPC42-CFP and -dsRed plasmids were constructed by replacing GFP with CFP or dsRed from pDH3 or pTY24 (Yeast Resource Center, University of Washington).
Strain Construction for Pairing Assays.
Chromosomal loci were marked by integration of 256 repeats of the lacO element and supplying lacI-GFP in trans by integration of pAFS152 at URA3 (38). To ensure production of lacI-GFP throughout meiosis, pMDE789, which contains lacI-GFP under control of the DMC1 promoter, was integrated at LYS2 (67). To insert the lacO array at the left telomere of chromosome III, a HindIII-SphI fragment of p784–1 (from C. Newlon, UMDNJ-New Jersey Medical School, Newark, NJ) was cloned into pAFS59 to construct pMDE1140, which was digested with BglII before transformation to target integration at a unique sequence in the subtelomeric region. To integrate the lacO array near the right telomere of chromosome IV, plasmid pMDE780 was used (67). To label the middle of the left arm of chromosome VII, a fragment adjacent to AMS1 (68) was cloned into pAFS59.
Microscopy.
Yeast strains are listed in SI Table 1. Cells in liquid cultures (67) were harvested for spread preparations of nuclei for immunocytology (12) or were mixed with equal volumes of 4% paraformaldehyde at room temperature for 5 min, washed, stained by mixing briefly with an equal volume of 0.5 μg/ml DAPI in 30% ethanol, washed then mounted on a thin pad of 0.5% agarose to hold the cells adjacent to the coverslip for fluorescence microscopy of whole cells. Cells heterozygous for SPC42-CFP/SPC42-GFP were added to experimental samples to aid registration of stacks made at different wavelengths. Image stacks were acquired at NA1.4 with pixel–pixel spacing of 0.08 μm, at 0.3 μm between image planes, by using a Zeiss Axioplan 2ie equipped with a Quantix57 camera (Roper) and MetaMorph software (UIC). Image stacks were deconvolved by using AutoDeblur (AutoQuant). Visualization, registration and fusion of images made at different wavelengths, and quantitative analyses, were carried out by using custom software (OMRFQANT, M.E.D.).
Image Analysis.
Cells were scored visually for bouquet arrangement by using Rap1-CFP to mark clumps of telomeres, Tub1-GFP or Spc42-dsRed to mark the spindle pole body and DAPI to mark DNA/chromatin, in 3D image stacks where images of individual nuclei were extracted from larger image stacks, registered and fused in three-color 3D images that could be freely rotated and zoomed to aid viewing. Spc42-CFP/Spc42-GFP heterozygote cells were included in each field for fiducial markers for registration. Automated methods of analysis are detailed in the legend to SI Fig. 11.
Acknowledgments
We thank Ben Fowler (Oklahoma Medical Research Foundation Imaging Facility) and Anton Konovchenko for technical help, Mark Winey (University of Colorado, Boulder, CO) and Sue Jaspersen (Stowers Institute, Kansas City, MO) for strains and plasmids, Jose-Angel Conchello for help with the quantitative image analysis, and Dean Dawson for comments on the manuscript. This work was supported by National Science Foundation Grant MCB 98-08000 (to M.E.D.) and by Oklahoma Center for the Advancement of Science and Technology Grant HR98-019 (to M.N.C.).
Footnotes
- ‡To whom correspondence should be addressed. E-mail: dresserm{at}omrf.ouhsc.edu
-
Author contributions: M.N.C., C.-Y.L., and M.E.D. designed research; M.N.C., C.-Y.L., J.L.W., and M.E.D. performed research; M.N.C., J.L.W., and M.E.D. contributed new reagents/analytic tools; M.N.C., C.-Y.L., and M.E.D. analyzed data; and M.N.C. and M.E.D. wrote the paper.
-
The authors declare no conflict of interest.
-
This article is a PNAS Direct Submission.
-
This article contains supporting information online at www.pnas.org/cgi/content/full/0606165104/DC1.
- Abbreviation:
- SPB,
- spindle pole body.
- © 2007 by The National Academy of Sciences of the USA
References
MPS3 mediates meiotic bouquet formation in Saccharomyces cerevisiae
Sign up for Article Alerts
Jump to section
You May Also be Interested in
Kerry Sieh recounts the hunt for a meteorite impact crater in Southeast Asia.
Ocean carbon isotope and pH patterns observed after the K–Pg mass extinction indicate a 50% reduction in global marine primary productivity.
Image courtesy of Michael J. Henehan.
Wood fragments in sediment cores from the Bengal Fan, in the Bay of Bengal, were episodically deposited over the past 19 million years.
Image courtesy of Pixabay/Sarangib.
A tar-backed tool from the present-day North Sea reveals the use of complex technology by Neandertals.
Image courtesy of Paul R.B. Kozowyk.