mug20 knockout

For the construction of diploid mug20 deletion strains, haploid strains of opposite mating types carrying complementing ade6-M210 and ade6-M216 alleles were transformed with a linearized knockout plasmid (Gregan et al. 2005), kindly provided by Juraj Gregan (see http://mendel.imp.univie.ac.at/Pombe/ for the design of the construct and the knockout procedure). Transformation resulted in the replacement of a region from 171 bp upstream to 107 bp downstream of the ORF SPBC36B7.06c with a nourseothricin resistance gene. Single colonies from clonNAT plates were screened by colony PCR for correct integration of the knockout cassette. Diploid mutant strains were produced by mating and maintained by growth in medium lacking adenine.

Mug20 tagging and antibody construction

Mug20-GFP was created by fusing the GFP gene to the C terminus of the mug20 ORF. A ca. 500 bp C-terminal fragment (excluding the stop codon) of the gene was PCR-amplified from genomic DNA using primers for creating flanking unique ApaI and XhoI restriction sites (5′-atatatgggcccttcaaccgtattccacc-3′ and 5′-atatatctcgagaaaattgtcgagaatagctttatg-3′; restriction sites underlined). The fragment was ligated to a GFP plasmid (Lindner et al. 2002) carrying a G418 resistance marker. The construct was tested for correct sequence, linearized with BlpI, and used for transforming S. pombe strains by homologous recombination. Transformants were grown on solid selective medium, and the correct integration of the tagging construct was checked in positive candidates by colony PCR. Haploid strains of opposite mating types carrying complementing ade6-alleles were produced and mated to create the diploid strain. Normal progression through meiosis indicated that the tagged protein was functional.

Polyclonal anti-Mug20 antibodies were raised in a rabbit against the internal LVQHRRNSQNKLKC and the C-terminal KMIETSTHKAILDNF peptide by a commercial provider (Eurogentec, Seraing, Belgium).

Yeast two-hybrid assay

cDNA of candidate genes was amplified by PCR using a meiotic cDNA library (kindly provided by Hiroshi Nojima, Osaka University) as template. The products were cloned and sequenced and inserted in frame in the assay vectors pGBKT7 and pGADT7, carrying the binding and activation domains (Matchmaker GAL4 Two-Hybrid System 3, Clontech, Mountain View, CA). S. cerevisiae strains Y187 and AH109 (Matchmaker 3) were transformed with the two plasmids, respectively. The strains carrying the different constructs were mated and a single diploid cell of each was grown to a colony. Clones were grown to saturation in rich liquid medium, and plates with rich medium and selective-Ade and -His medium were inoculated with 1:10-dilution series of cell suspensions. Growth of cultures of different cell densities on the different plates was evaluated after 2–3 days. Experiments were repeated three times, haploid strains were mated freshly for each repeat.

Immunoblotting

Samples from meiotic cultures were taken at 1-h intervals. Cells were broken up by glass beads in a multi-beads shocker (Yasui Kikai, Osaka, Japan) and protein extracts were prepared by trichloroacetic acid precipitation (see Spirek et al. 2010). 10 μl of extracts were run on 10% SDS-PAGE gels and blotted to Hybond-P PVDF membrane (Amersham Biosciences, GE Healthcare, Pittsburgh, PA).

Mug20-GFP was detected by incubating the blot over night at 4°C with goat anti-GFP antibody (1:2,000, Rockland, Gilbertsville, PA) in TBST (20 mM Tris pH 7.5, 140 mM NaCl, 0.05% Tween 20) + 3% dry milk. Membranes were washed and incubated for 2 h with HRP-conjugated anti-goat antibody (1:5,000, Pierce, Thermo Fisher Scientific, Rockford, IL), washed, then incubated with chemiluminescent reagent (Thermo Fisher Scientific), and exposed to X-ray film. Untagged Mug20 was detected by incubating the membranes with rabbit polyclonal antibody (1:1,000) and HRP-conjugated anti-rabbit antibody (1:10,000, Dako, Glostrup, DK), followed by a chemiluminescent reaction as above.

Cytological methods

Cells were taken from cultures 5–7 h after induction of sporulation, when LinEs of all types are most abundant in the wild type (Loidl 2006), and meiotic nuclei were prepared by a detergent-spreading method. In short, cells were spheroplasted by enzymatic digestion of cell walls, dropped onto a slide, membranes were solubilized by the addition of a detergent (Lipsol, Barloworld Scientific) and the spread nuclei were fixed by the addition of a 4% paraformaldehyde solution supplemented with 3.4% sucrose (for a detailed protocol see Loidl and Lorenz 2009). For immunofluorescence staining, slides were washed with PBS + 0.05% Triton X-100, incubated with primary antibodies, washed, incubated with secondary antibodies, washed, and mounted in antifading buffer (Vectashield, Vector Laboratories Inc., Burlingame, CA) supplemented with 0.5 μg/ml DAPI (see Loidl and Lorenz 2009). Primary antibodies were rabbit anti-Rec10 [1:200, (Lorenz et al. 2004)], chicken anti-GFP (1:1,000, Chemicon, Temacula, CA), rabbit anti-Mug20 (1:200), and monoclonal mouse anti-Rad51 (1:50, NeoMarkers, Fremont, CA). Secondary antibodies were anti-rabbit conjugated to CY3 (1:1,000, Amersham), anti-rabbit conjugated to FITC (1:200, Sigma, St. Louis, MO), anti-chicken conjugated to CY3 (1:200, Abcam, Cambridge, UK), anti-mouse conjugated to CY3 (1:400, Jackson Immuno Research, West Grove, PA), and anti-mouse conjugated to FITC (1:100, Sigma-Aldrich, St. Louis, MO).

Slides were inspected under a fluorescence microscope equipped with the appropriate filters, and separate color channels were recorded using a cooled monochrome CCD camera. Color channels were merged and assigned artificial colors using image-processing software.

LinEs do not elongate in mug20Δ

LinE development can be monitored by immunostaining of Rec10, one of the LinE’s major components (Fig. 1c). In the wild type, it begins as dots and seems to progress by elongation to threads and, later, their merging. Several such tracts can become laterally connected providing a reticular appearance, which is most clearly seen in the electron microscope (Bähler et al. 1993). In addition, during a later stage in LinE development, they may appear as thick rods which, in the electron microscope, have been seen to consist of bundles of LinEs (Bähler et al. 1993). In the mug20Δ mutant, only Rec10 dots were observed (Fig. 1d). This suggests that LinEs failed to elongate and Rec10 localization remained restricted to the initial loading sites.

The conserved strand exchange protein Rad51 associates with ssDNA during the strand invasion step in meiotic recombination (e.g., San Filippo et al. 2008) and hence it serves as a cytological marker for DSB processing. In the wild type, it localizes preferentially to LinEs (Lorenz et al. 2006). In the mutant, an average of 3.0 (SD = 1.5, n = 40 nuclei) Rad51 foci were counted per nucleus as compared to 22.5 (SD 6.8, n = 39 nuclei) foci in the wild type (Fig. 1e, f). This reduction is consistent with the observed reduction of genetic recombination and it suggests that DSB formation and/or processing is reduced in the mug20Δ mutant.

Mug20 localizes to LinEs

To study the expression and localization of Mug20, we constructed a GFP-tagged version, and we had antibodies produced against the protein. Immunodetection of Mug20 or Mug20-GFP on Western blots of whole cell extracts revealed that the protein was expressed from ~3 h after induction of meiosis (Fig. 2a), i.e., when LinEs begin to form (see Loidl 2006).

Open in a separate window

Fig. 2

Mug20 expression and localization in meiosis. a Western time courses of Mug20-GFP (left) and Mug20 (right). The signals correspond to the calculated sizes of the Mug20-GFP fusion-protein (47 kDa) and Mug20 (19 kDa). The asterisk denotes untagged Mug20 extract from timepoint t = 5 h. Tub. α-tubulin loading marker. b Immunostaining of Mug20 delineates dots and elongated structures typical of LinEs. c Mug20-GFP (red) co-localizes with Rec10 (green) along LinEs at all stages. In the absence of Rec10 (d), and in the rec10-155 mutant (e), Mug20 does not localize. f In the rec8Δ mutant, Mug20 highlights truncated Rec10 structures. g Mug20 localization to LinEs is independent of Rec12/Spo11. Examples in b–g were taken from cultures 6 h after induction of sporulation. Bar 5 μm

Cytological detection of both immunostained and GFP-tagged Mug20 demonstrated that in the wild type, Mug20 delineates structures resembling the LinEs (Fig. 2b, c). The co-localization of Mug20-GFP with Rec10 confirmed that it indeed decorates LinEs (Fig. 2c). In the absence of the LinE core component Rec10 or in the rec10-155 mutant (see Wells et al. 2006), Mug20 structures were not visible, indicating that Mug20 localization depends on intact Rec10 (Fig. 2d, e). Rec8 is a meiosis-specific cohesin subunit, and in its absence only few abnormal LinEs are formed (Molnar et al. 1995, 2003; Lorenz et al. 2004). In a rec8Δ strain, Mug20 co-localized with Rec10 on these structures (Fig. 2f). This confirms the absolute dependency of Mug20 on the presence of Rec10.

While the reduced nuclear localization of Rad51 in the absence of Mug20 (see above) suggests a dependency of DSBs on Mug20, it is still possible that Mug20 is loaded onto LinEs in the wake of (initial) DSB formation. We tested this possibility by detecting Mug20 in a rec12Δ mutant. We found that Mug20 co-localizes with Rec10 LinEs in a wild-type manner, hence its loading is independent of DSBs (Fig. 2g).

To further corroborate the association of Mug20 with LinEs, we set up yeast two-hybrid assays for testing interactions of Mug 20 with Rec10, Rec25 and Hop1 (Fig. 3). While an interaction of Mug20 with Rec10 is evident from their co-immunoprecipitation (Spirek et al. 2010), the two-hybrid tests failed to show a direct interaction between Rec10 and Mug20. However, we found an interaction of Mug20 with Rec25 and an interaction of the latter with Rec10 (Fig. 3). Thus it is possible that Mug20 is recruited to LinEs via Rec25. From our cytological and two-hybrid data, we conclude that Mug20 localizes to a platform provided by the LinE core components Rec10 and Rec25 (and possibly Rec27).

Open in a separate window

Fig. 3

Spot tests for detecting protein-protein interactions by yeast two-hybrid assays. A budding yeast strain carrying the bait plasmid expressing Mug20 linked to the Gal4 binding domain (BD) was mated to strains carrying the prey plasmid expressing the Gal4 activation domain (AD) linked to Rec10, Hop1, Mug20 or Rec25. As a control, the bait strain was mated to a strain carrying the empty prey plasmid (0-AD). In similar experiments, Rec10 and Hop1 were used as baits. Diploid colonies were applied in a 10× dilution series to rich medium (YSD-Leu-Trp) and to selective media YSD-Leu-Trp-His and YSD-Leu-Trp-Ade. Growth on selective media indicated Mug20–Rec25 interaction and Mug20 self-interaction, and also an interaction of Rec10 with Hop1

This work was supported by Grant no. P18186 from the Austrian Science Fund (FWF) and by Grant no. I031-B from the University of Vienna. We are grateful to Juraj Gregan (University of Vienna) for providing us with the mug20 knockout construct, to Jürg Kohli (University of Berne), Gerald R. Smith (Fred Hutchinson Cancer Research Center, Seattle) and Matthew Whitby (University of Oxford) for providing strains, and to Mario Spirek (Masaryk University, Brno) for his support during the early stage of this study.