The meiotic TERB1-TERB2-MAJIN complex tethers telomeres to the nuclear envelope

Abstract

During meiotic prophase I, telomeres attach to and move on the nuclear envelope (NE), regulating chromosome movement to promote homologous pairing. Meiosis-specific proteins TERB1, TERB2 and MAJIN play a key role in this process. Here, we report the crystal structures of human TERB1-TERB2 and TERB2-MAJIN subcomplexes. Specific disruption of the TERB1-TERB2 or the TERB2-MAJIN interaction in the mouse Terb2 gene abolishes the telomere attachment to the NE and causes aberrant homologous pairing and disordered synapsis. In addition, depletion of SUN1 also partially disrupts the telomere-NE connection. We propose that the telomere-TRF1-TERB1-TERB2-MAJIN-NE interaction network and the telomere-LINC complex connection are likely two separate but cooperative pathways to stably recruit telomeres to the NE in meiosis prophase I. Our work provides a molecular model of the connection between telomeres and the NE and reveals the correlation between aberrant synapsis and the defective telomere attachment to the NE.

Fig. 1

Characterization of the interactions among TERB1, TERB2, and MAJIN. a Illustration of different fragments of TERB1 (upper panel). Bimolecular fluorescence complementation (BiFC) data with full-length TERB2 and varies TERB1 fragments (lower panel). Cells expressing TERB2-VC only (blue, as a negative control) or TERB2-VC with -VN vector (green, as a negative control) or TERB2-VC with different TERB1-VN fragments (red) were subjected to fluorescence-activated cell sorting (FACS) analysis. b Illustration of various fragments of TERB2 (upper panel). BiFC data with full-length TERB1 and TERB2 fragments (lower panel). Cells expressing TERB1-VC only (blue, as a negative control) or TERB1-VC with -VN vector (green, as a negative control) or TERB1-VC with different TERB2-VN truncations (red) were subjected to FACS analysis. c Equator images of U-2 OS cells expressing SUN1-GFP (green) together with mCherry-MAJIN (red) (up) or Flag-TERB2 (magenta) (down). SUN1-GFP was used to indicate the nuclear envelope localization. Cells were stained with 4,6-diamidino-2-phenylindole (DAPI; blue) to show the nuclear position. Scale bars, 10 μm. d Schematics of various TERB2 fragments (upper panel). Equator images of U-2 OS cells expressing mCherry-MAJIN (red) and Flag-TERB2 truncations. Cells were stained with Flag antibody (magenta) and DAPI (blue). Scale bars, 10 μm. e Gel filtration chromatography profile of the TERB2_MBM–MAJIN_NTD complex. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis image of the peak in the gel filtration chromatography profile is shown

Fig. 2

Crystal structure of the TERB1_T2BM–TERB2_NTD complex. a Domain organization of TRF1, TERB1, TERB2, and MAJIN. Shadings indicate the interactions among these proteins. The interacting domains are labeled and highlighted by different colors. This diagram illustrates the TRF1–TERB1–TERB2–MAJIN interaction network connecting telomeres to the nuclear envelope. Acidic acidic motif, TRFH TRF homology domain, Myb myeloblastosis domain, ARM armadillo repeats, CC coiled coil, TM trans-membrane motif. b Ribbon diagrams of two orthogonal views of the TERB1_T2BM–TERB2_NTD complex. TERB1_T2BM is colored in yellow and TERB2_NTD in green. c, e, f Detailed interactions at the TERB1_T2BM–TERB2_NTD interface. The color scheme is the same as in b. Residues important for the interaction are shown in stick models. Salt bridges and hydrogen-bonding interactions are shown as magenta dashed lines. The Zn ion is shown as a gray sphere and residues chelating this Zn ion are shown in stick model in c. d Electrostatic surface potential of TERB1_T2BM-binding site of TERB2_NTD. Positive potential, blue; negative potential, red

Fig. 3

Crystal structure of the TERB2_MBM–MAJIN_NTD complex. a Ribbon diagrams of two orthogonal views of the TERB2_MBM–MAJIN_NTD complex. TERB2_MBM is colored in magenta and MAJIN_NTD in blue. b Electrostatic surface potential of TERB2_MBM-binding site of MAJIN_NTD. Positive potential, blue; negative potential, red. c, d Detailed interactions at the TERB2_MBM–MAJIN_NTD interface. The color scheme is the same as in a. Residues important for the interaction are shown in stick models. Salt bridges and hydrogen-bonding interactions are shown as magenta dashed lines. e Ribbon diagram shows the potential heterotetramer mode of the TERB2_MBM–MAJIN_NTD complex revealed in the crystal structure. f Structural model of the connection between telomeres and the NE through the TRF1–TERB1–TERB2–MAJIN interaction network

Fig. 4

Mutational analyses of the TERB1–TERB2 and TERB2–MAJIN interactions. a, c Detailed interactions around Tyr56^TERB2 in the TERB1_T2BM–TERB2_NTD complex (a) and around Phe192^TERB2 in the TERB2_MBM–MAJIN_NTD complex (c). b, d Co-immunoprecipitation (Co-IP) of human TERB1 (b) or MAJIN (d) with wild-type (WT) or mutant Flag-tagged human TERB2 in 293T cells. The levels of each protein in the input and IP samples were analyzed by immunoblotting with the indicated antibodies. Input contains 5% of the input whole-cell lysate used for IPs. e Multiple sequence alignment of the N-terminal (top) and C-terminal (bottom) regions of TERB2 required for its association with TERB1 and MAJIN, respectively. Sequence positions are shown based on the human TERB2. The residues of Tyr56 (Try56 in mouse) and Phe192 (Phe188 in mouse) are indicated by blue arrowheads. f, g Co-IP of mouse TERB1 (f) or mouse MAJIN (g) with WT or mutant Flag-tagged mouse TERB2 in 293T cells. The levels of each protein in the input and IP samples were analyzed by immunoblotting with the indicated antibodies. Input contains 5% of the input whole-cell lysate used for IPs. Source data are provided as a Source Data file

Fig. 5

Disruption of the TERB1–TERB2–MAJIN complex results in defective spermatogenesis in mice. a Testes from 6-week-old wild-type, Terb2^Y56E, and Terb2^F188R mice. Scale bars, 5 mm. b Body weight (top) and testis weight (bottom) of 6-week-old male mice of the indicated genotypes. c Hematoxylin and eosin-stained histological cross-sections of testes. Arrowheads indicate abnormal spermatocyte-like cells in mutant seminiferous tubules. Asterisks show vacuolated seminiferous lacking spermatozoa and spermatids. Scale bars, 100 μm. d Immunofluorescence (IF) staining for SOX9 (green) of testis sections from adult mice. DNA was stained by 4,6-diamidino-2-phenylindole (DAPI; blue). Asterisks show seminiferous containing a single layer of Sertoli cells and very few spermatogonial cells at the tubule periphery. Scale bars, 100 μm. e TUNEL staining of testis sections from 6-week-old mice. DNA was stained with DAPI (blue) and apoptosis was detected by IF terminal deoxinucleotidyl transferase-mediated dUTP-fluorescein nick end labeling (TUNEL) (green). Many TUNEL-positive cells were observed in both Terb2^Y56E and Terb2^F188R testes. Scale bars, 1 mm. f Population of TUNEL-positive tubules shown as mean and s.d. n = 3 mice of each indicated genotype. Statistical significances (*P < 0.05, **P < 0.01, ***P < 0.001) were assessed by two-tailed t tests. g Representative spermatocyte nuclei from 6-week-old mice at different stages stained with DAPI and antibodies of SYCP3 and SYCP1. Pac-like Pachytene-like. h Quantification of the frequencies of meiotic stages shown in g. Data are from three independent experiments of different mice and error bars indicate s.d. More than 100 nuclei from each mouse were counted. Lep Leptotene, Zyg Zygotene, Pac-like Pachytene-like, Pac Pachytene, Dip Diplotene, Dia Diakinesis. Source data are provided as a Source Data file

Fig. 6

Disruption of the TERB1–TERB2–MAJIN complex results in aberrant synapsis. a Chromosome spreads from adult spermatocytes were immunostained with anti-SYCP3 (green) and anti-TRF1 (red). b Quantification of TRF1 foci number per nucleus (n > 40 cells for each genotype) shown in a. Data are from three independent experiments of three different mice. Note that there are 40 chromosomes (2N) in murine cells and the number of telomere foci ranges from 80 (fully unpaired) to 41 (fully paired) in wild-type spermatocytes. c Fluorescent images of spermatocyte spreads labeled with anti-SYCP3 (green), anti-SYCP1 (red), and chromosome 5 painting (magenta). DNA was stained by 4,6-diamidino-2-phenylindole. Representative images of Y56E spreads show various disordered synapsis of chromosome 5. Insets are enlarged figures of the framed regions. SYCP3 filaments of chromosome 5 were indicated by sketches. d Quantification of the frequencies of different status of chromosome 5. Error bars indicate s.d. of three independent experiments with different mice. For each mouse, >50 pachytene-like spreads were counted for quantification. e Immunofluorescence–fluorescence in situ hybridization (FISH) analysis by stimulated emission depletion of telomere (FISH), SYCP1, and SYCP3 on spermatocyte spreads from adult testes. The white arrowhead indicates self-pairing, the light blue arrowhead shows a Y-shaped partial pairing, the red arrowhead points to a non-pairing chromosome, the magenta arrowhead indicates an apparently fully paired chromosome, the zoom-in view shows the bend on the top of the apparently fully paired chromosomes, and the purple arrowhead shows a bridge-shaped partial pairing. Source data are provided as a Source Data file

Fig. 7

Recombination and double-strand break repair was severely impaired in mutant spermatocytes. a Representative spermatocyte nuclei from 6-week-old mice at different indicated stages immunostained with SYCP3 (green) and γ-H2AX (red). Pac-like Pachytene-like stage. b Quantification of the frequencies of meiotic stages shown in a. Data are from three independent experiments of different mice and error bars indicate s.d. More than 100 nuclei from each mouse were counted. Lep Leptotene, Zyg Zygotene, Pac-like Pachytene-like, Pac Pachytene, Dip Diplotene, Dia Diakinesis. c Immunofluorescence staining of spermatocyte chromosome spreads for SYCP3 (green) and DMC1 (red). d Quantification of the number of DMC1 foci on the SYCP3 filaments per nucleus. e Immunofluorescence staining of spermatocyte chromosome spreads for SYCP3 (green) and MSH4 (red). White arrows indicate chromosomes with MSH4 foci in the mutant nuclei. f Quantification of the number of MSH4 foci on the SYCP3 filaments per nucleus shown in e. g Immunofluorescence staining of spermatocyte chromosome spreads for SYCP3 (green) and MLH1 (red). h Quantification of the number of MLH1 foci on the SYCP3 filaments per nucleus shown in g. DNA were stained with 4,6-diamidino-2-phenylindole (a, c, e, g). Source data are provided as a Source Data file

Fig. 8

Intact TERB1–TERB2–MAJIN complex is required for telomere–nuclear envelope (NE) attachment. a Equator images of structurally preserved zygotene spermatocytes from wild-type (WT), Terb2^Y56E, Terb2^F188R, and Sun1^−/− immunostained for SYCP3 and Lamin B and subjected to fluorescence in situ hybridization of telomeric DNA probe (TelC). DNA was stained by 4,6-diamidino-2-phenylindole (DAPI). Scale bars, 5 μm. b Quantification of internal telomere foci number shown in a. c Electron micrographs showing telomeres (magenta arrowheads) and nuclear membrane (red arrows) in WT pachytene and mutant pachytene-like spermatocytes. Scale bars, 500 nm. d Populations of nuclei from spermatocyte sections showing at least one telomere–NE attachment. Only nuclei with stretches of synaptonemal complex (SC) observed were counted for the quantification. Data are from three independent experiments with different mice and error bars indicate s.d. About 100 nuclei containing SC stretches from each mouse were counted for quantification. n is the total number of nuclei with SC stretches used for counting. e Chromosome spreads from adult spermatocytes were immunostained with anti-SYCP3 (green) and anti-SUN1 (red). DNA was stained by DAPI. The zoom-in views show one chromosome end from each spread. Scale bars, 10 μm. f Dot blot showing population of chromosome ends with SUN1 foci in one nucleus. Source data are provided as a Source Data file

Fig. 9

Telomere localization of Speedy A and CDK2 is reduced in mutant nuclei. a Immunofluorescence–fluorescence in situ hybridization (IF-FISH) images of chromosome spreads from adult spermatocytes of the indicated genotypes. Chromosome spreads were immunostained for SYCP3 (green) and Speedy A (red) and subjected to telomere FISH (TelC-FISH) (magenta). DNA was stained by 4,6-diamidino-2-phenylindole (DAPI; blue). Zyg Zygotene, Pac Pachytene, En showing enhanced signals (red) of Speedy A with a lower cutoff. White arrowheads indicate sex chromosomes. Scale bars, 10 μm. b Quantification of the relative intensity of Speedy A foci at telomeres in wild-type (WT) or Terb2-mutant spermatocytes in a. c IF-FISH images of chromosome spreads from adult spermatocytes of the indicated genotypes. Chromosome spreads were immunostained for SYCP3 (green) and CDK2 (red) and subjected to telomere FISH (TelC-FISH) (magenta). DNA was stained by DAPI (blue). Zyg Zygotene, Pac Pachytene, En showing enhanced signals (red) of CDK2 with a lower cutoff. White arrowheads indicate sex chromosomes. Scale bars, 10 μm. d Quantification of the relative intensity of CDK2 foci at telomeres in WT or Terb2-mutant spermatocytes in c. Source data are provided as a Source Data file

Fig. 10

Speedy A and SUN1 exhibit congregation in Terb2 mutant spermatocytes. a, c Immunofluorescence–fluorescence in situ hybridization (FISH) showing distribution of Speedy A (a) and SUN1 (c) in spermatocytes. Sections of 2-week-old testes of different genotypes as indicated were immunostained for SYCP3 and Speedy A (a) or SYCP3 and SUN1 (c) and subjected to telomere FISH (TelC). DNA was stained by 4,6-diamidino-2-phenylindole. White arrowheads indicate congregation of Speedy A (a) or SUN1 (c). White arrows in a indicate Speedy A signals co-localized with telomeres. Scale bars, 5 μm. b Population of nuclei that display Speedy A congregation in a. Data are from three independent experiments with different mice and error bars indicate s.d. d Population of nuclei that display SUN1 congregation in c. Data are from three independent experiments with different mice and error bars indicate s.d. For each experiment, >100 spermatocytes from each indicated genotype were counted for the frequencies of nuclei showing congregation of Speedy A (b) or SUN1 (d). Source data are provided as a Source Data file