Dissecting the telomere-inner nuclear membrane interface formed in meiosis

Abstract

Tethering telomeres to the inner nuclear membrane (INM) allows homologous chromosome pairing during meiosis. The meiosis-specific protein TERB1 binds the telomeric protein TRF1 to establish telomere-INM connectivity and is essential for mouse fertility. Here we solve the structure of the human TRF1-TERB1 interface to reveal the structural basis for telomere-INM linkage. Disruption of this interface abrogates binding and compromises telomere-INM attachment in mice. An embedded CDK-phosphorylation site within the TRF1-binding region of TERB1 provides a mechanism for cap exchange, a late-pachytene phenomenon involving the dissociation of the TRF1-TERB1 complex. Indeed, further strengthening this interaction interferes with cap exchange. Finally, our biochemical analysis implicates distinct complexes for telomere-INM tethering and chromosome-end protection during meiosis. Our studies unravel the structure, stoichiometry, and physiological implications underlying telomere-INM tethering, thereby providing unprecedented insights into the unique function of telomeres in meiosis.

Figure 1. Structural and biochemical dissection of the TRF1^TRFH - TERB1^TBM interface

(a) Domain diagram of human TERB1 is shown along the connectivities from the telomere to the inner nuclear membrane. Alignment of the TBM motifs of human and mouse, TERB1 and TIN2, are shown below the domain diagram. (b) Pull down of TRF1^TRFH using GST-TERB1^TBM peptide (done in duplicate) on glutathione (GSH) sepharose beads as bait. Experiment was performed twice. (c) Overall structure of TRF1^TRFH - TERB1^TBM is shown with the TRF1 subunits (green) and the two bound TERB1^TBM peptides (orange) rendered in ribbon and stick representations, respectively. (d) Comparison of the interactions of the (I/F)X residues of TERB1 (left) and TIN2 (center) with TRF1^TRFH; the right panel shows a superposition of the two structures. (e) Overlay of TIN2^TBM (yellow) and TERB1^TBM (orange) bound TRF1^TRFH (green) structures showing a similar conformation adopted by the LXP residues of TERB1 and TIN2. (f) Overlay of TERB1^TBM and TIN2^TBM bound TRF1^TRFH structures showing the different conformation adopted by the tri-arginine stretch. (g) Binding of TERB1 to TRF1^TRFH places the CDK-phosphorylation site TERB1 T648 in close proximity to the negatively charged TRF1 E106; the red sphere indicates a water molecule bridging these residues using H-bonding. (h) Overlay of the TERB1^TBM versus TIN2^TBM peptide backbone segments from their respective TRF1^TRFH-bound structures. (i) Direct association of Alexa Fluor 488-labeled GST-TERB1^TBM or GST-TIN2^TBM with biotin-labeled TRF1^TRFH on streptavidin beads was scored with a flow cytometer. The fluorescence signal was background corrected with binding reactions using unbound streptavidin beads. Mean of technical duplicate is plotted. Mean and s.e.m. values are indicated. Experiment was performed twice. (j-k) Fluorescence-based competition experiments using Alexa Fluor 488-labeled GST-TERB1^TBM (30 nM) pre-bound to biotin-labeled TRF1^TRFH on streptavidin beads titrated with varying concentrations of the indicated unlabeled TERB1^TBM peptides (j) or TERB1^TRFB – TERB2^1-107 complexes (k). Mean of technical duplicate is plotted. Mean and s.e.m. of two biological replicates (of technical duplicates) are shown in Table 2.

Figure 2. Weakening the TRF1-TERB1 interface compromises telomere-INM tethering

(a) Schematic showing telomere-INM attachment defect in Terb1^-/- pachytene cells that can be rescued with GFP-TERB1 WT expression. (b) Equator images of Terb1^-/- pachytene-like mouse spermatocytes expressing GFP-TERB1 WT, GFP-TERB1 I645E, GFP-TERB1 I645F, GFP-TERB1 L647E, GFP-TERB1 T648D, GFP-TERB1 P649E, or GFP-TERB1 R651E-R652E (green) stained with TRF1 (red) and SYCP3 (blue) antibodies. Red arrowheads indicate internal TRF1 foci. (c) Graph showing the number of internal TRF1 foci (black dot) in each cell with the median indicated (red dot). Only cells expressing a relatively strong GFP-TERB1 signal intensity were scored in this assay. Statistical significance (P-values) was assessed by one-way analysis of variance (ANOVA) with Tukey-Kramer's multiple comparisons test relative to GFP-WT. The numbers below indicate the number of cells observed.

Figure 3. Further strengthening the TRF1-TERB1 interface disrupts cap exchange

(a) Projection images of WT early pachytene (left, H1T-negative) and late pachytene (right, H1T-positive) spermatocytes stained with TRF1 antibody. Equator and peripheral images of telomere-localized TRF1 (white rectangles) are shown in magnified views. Double-ended arrows indicate the outer diameter of the TRF1 signal. (b) WT spermatocytes in late pachytene expressing GFP-TRF1 WT, GFP-TRF1 E93K, GFP-TRF1 E93A, or GFP-TRF1 F129A. Magnified images of synapsed telomeres at the nuclear peripheries (white rectangles) are shown. (c) WT spermatocytes expressing GFP-TERB1 WT, GFP-TERB1 I645E, or GFP-TERB1 I645F (green) stained with TRF1 (red) and SYCP3 (blue) antibodies at the indicated stage of pachytene. (d) Graph showing the diameter of GFP-TRF1 or endogenous TRF1 signals of each telomere (black dot) with the mean (red dot) and s.d. (red bar). For cells expressing GFP-TRF1, the outer diameter of the GFP signal was measured, and for control cells and cells expressing GFP-TERB1, the outer diameter of the endogenous TRF1 signal was measured. The numbers below indicate the number of cells observed. Blue lines indicate the average outer diameter of the TRF1 signals in early pachytene (EP) or late pachytene (LP) spermatocytes. P-values for the early/late pachytene pairs for each TRF1 or TERB1 construct are indicated above the plot. (e) Schematic shows method for determining the area of the TERB1 versus TRF1 signals. Quantitation shows that GFP-TERB1 I645F distributes to the wider area compared to WT protein presumably because of its persistent interaction with TRF1, which relocates to the surrounding area after cap exchange. Mean (red dot), s.d. (red bar), and P-values are indicated.

Figure 4. Defining protein complexes involved in telomere tethering versus chromosome end protection

(a) Fluorescence-based competition experiments using Alexa Fluor 488-labeled GST-TERB1^TBM pre-bound to biotin-tagged TRF1^TRFH on streptavidin beads titrated with varying concentrations of unlabeled full-length TIN2 and TERB1^TRFB – TERB2^1-107 proteins. Mean of technical duplicate is plotted. Error bars indicate s.e.m. The schematics alongside the data represent probable binding stoichiometry. The yellow stars indicate fluorophore-labeled TERB1^TBM peptides bound to the TRF1 dimer. Three biological replicates were performed. (b) Strategy for engineering intramolecular homodimers or heterodimers of human TRF1^TRFH containing WT and/or F142A sequence connected in tandem by a gly/ser linker; and the predicted number of binding sites for TIN2 and TERB1. (c) Amylose pull down with indicated intramolecular TRF1^TRFH dimeric constructs was performed to determine the stoichiometry of binding of TERB1 versus TIN2 for a TRF1 dimer. Quantitation of the coomassie-stained PAGE shown at the bottom was performed by dividing the appropriate band intensities by the molecular weight of the species followed by normalization against the molecular weight-normalized TRF1^TRFH dimer signal in that lane. See Supplementary Fig. 6d and e, for replicate data and statistics, respectively. (d) Ternary pull down experiment on amylose resin of MBP-tagged TERB1^TRFB – TERB2^1-107 and TRF1^TRFH, and TIN2. Experiment was performed twice. (e) Model showing separate TRF1-containing complexes for protecting chromosome ends and tethering telomeres to the INM during meiosis. (f) GST pull down of TRF2^TRFH using GST-TERB1^TBM peptide on glutathione (GSH) sepharose beads as bait. Experiment was performed twice. (g) Direct binding curves of biotin-tagged GST-TERB1^TBM with Alexa Fluor 488-labeled full-length TRF2 either in the absence or the presence of Rap1 protein. Mean of technical duplicate is plotted. Error bars indicate s.e.m. Two biological repeats were performed.