Trim66's paternal deficiency causes intrauterine overgrowth

Abstract

The tripartite motif-containing protein 66 (TRIM66, also known as TIF1-delta) is a PHD-Bromo-containing protein primarily expressed in post-meiotic male germ cells known as spermatids. Biophysical assays showed that the TRIM66 PHD-Bromodomain binds to H3 N-terminus only when lysine 4 is unmethylated. We addressed TRIM66's role in reproduction by loss-of-function genetics in the mouse. Males homozygous for Trim66-null mutations produced functional spermatozoa. Round spermatids lacking TRIM66 up-regulated a network of genes involved in histone acetylation and H3K4 methylation. Profiling of H3K4me3 patterns in the sperm produced by the Trim66-null mutant showed minor alterations below statistical significance. Unexpectedly, Trim66-null males, but not females, sired pups overweight at birth, hence revealing that Trim66 mutations cause a paternal effect phenotype.

Figure 1.. TRIM66 is a spermatid-specific PHD-Bromo protein that recognizes unmethylated lysine 4 of histone H3.

(A) Diagram of domain organization of murine TRIM66. The longest isoform containing a RING domain at the N-terminus is shown (NP_001164383.1). (B) Trim66’s expression in mouse adult testicular cells, as shown by Uniform Manifold Approximation and Projection of scRNA-seq data (GSE142585). The color scale represents the level of the expression of Trim66 that is only detected at the post-meiotic stages, namely, round and elongated spermatids. (C) mRNA levels of Trim66 in Transcripts Per Million in round and elongated spermatids and oocytes, and at key stages of preimplantation development. Embryo mRNA-seq data are from GSE66582 (Wu et al, 2016), and from this study for elongated and round spermatids. Individual biological replicates are shown in red and the average in black. The whiskers represent the mean Transcripts Per Million value plus or minus the SD. (D) Isothermal titration calorimetry measurements showed that TRIM66 PHD-Bromodomain binds to an unmodified histone H3 N-terminus tail (K_d = 3.4 μM). A similar binding affinity is observed when lysine 9 is tri-methylated, when lysines 18, 23, or 27 are acetylated, or when lysine 9 is tri-methylated in combination with acetylated lysine 18. Tri-methylation of lysine 4 dramatically decreases the binding affinity of the TRIM66 PHD-Bromodomain to an H3 N-terminus tail. Combining the tri-methylated lysine 4 with acetylated lysine 18 only decreases the binding affinity by a factor of 7.1 when comparing to binding to an unmodified H3 histone tail, indicating a contribution of lysine 18 acetylation to the binding. (E) Representative confocal images of the indicated histone marks (yellow) and TRIM66 (purple) in the seminiferous epithelium. In elongated spermatids, TRIM66 is present both within and outside the chromocenter. The DNA is stained with Hoechst 33342. The post-translational modifications that influence TRIM66’s binding to histone H3 (e.g., K4me3, K9me3, and K18Ac) show distinct patterns in elongated spermatids. The overlapping staining of TRIM66 and H3K18Ac and H3K9me3 is consistent with occupancy of H3K9me3, K18Ac nucleosomes by TRIM66. The scale bar is 25 μm.

Figure S1.. Restricted expression of TRIM66 in mouse and human round and elongated spermatids.

(A) Trim66 RNA levels in various mouse tissues show predominant expression in the testis. It is noteworthy that Trim66 mRNA is not detectable in the ovary. The dataset used to generate the plot is the FANTOM5 mouse tissue expression database. (B) Uniform Manifold Approximation and Projection representation of single-cell RNA-seq data from human (GSE142585, PubMed: 32504559) shows that TRIM66 is exclusively expressed in the testis in cells identified as round and elongated spermatids. (C) Murine Trim66 gene isoform exonic structure. Three different isoforms have been reported in Ensembl genome browser v109. We performed 5′ RACE on mouse testis RNA and identified a previously annotated transcript that contains an exon downstream of exon 1 of Trim66-202. We named this isoform Trim66-205; it corresponds to the GenBank predicted transcript XM_036153179.1.

Figure 2.. Creation of two Trim66’s loss-of-function murine alleles.

(A) Schematic representation of the genomic structure of the murine Trim66 gene and location of the CRISPR-mediated insertions of the alleles Trim66^gfp and Trim66^phd (upper panel). The three coding mRNA isoforms are shown, as well as their translation (upper panel). The middle and bottom panels show splicing patterns as inferred from RNA-seq in sorted round spermatids. Two independent loss-of-function alleles were created: Trim66^gfp and Trim66^phd. Trim66^gfp consists of the insertion of a premature stop codon and poly-A signal in exon 3 (Trim66-202) resulting in truncated transcription as shown with the sashimi plot (lower panel). No alternative splicing event is detected in the homozygous mutant. The allele Trim66^phd was created by insertion of a premature stop codon in exon 15 (Trim66-201). The genomic coordinates of the insertions are indicated (GRCm39). (B) Western blot detection of TRIM66 in whole testis extracts and Neuro2A cell line (positive control) using polyclonal antibodies raised against murine TRIM66’s Bromodomain. Detection of TRIM66 in adult but not in juvenile testes is consistent with its expression in post-meiotic germ cells. TRIM66 protein is undetectable in the lysate from adult testes homozygous for Trim66^gfp, thus supporting a complete loss of function. (C) Western blot with the antibody anti-TRIM66 showing that TRIM66 is undetectable in adult testes homozygous for Trim66^phd, arguing that the insertion of a stop codon in exon 15 is a bona fide null mutation. Source data are available for this figure.

Figure S2.. Isothermal titration calorimetry of complex formation between TRIM66 PHD-Bromo and various H3 N-ter peptides (1–30).

(A) Thermodynamic parameters of complex formation between murine TRIM66 PHD-Bromodomain and various H3 (1–30) peptides. The values are the average of triplicate. *N (binding stoichiometry) value was fixed to 1 and the concentration of the peptide left to vary in the fitting parameter for the single-site binding model. The H3 (1–30) K23Ac and H3 (1–30) K27Ac peptide concentrations could not be accurately determined. (B) Representative isothermal titration calorimetry measurements of complex formation between murine TRIM66 PHD-Bromodomain and 30-mer peptides corresponding to the N-terminus of histone H3 bearing the indicated post-translational modifications. Top panel: raw data of the titration. Lower panel: integrated heat changes (symbols) and fitted binding models (lines).

Figure S3.. Transcription of egfp-mRNA but undetectable GFP-fluorescence in Trim66^gfp/gfp testicular cells.

To assess the expression of the gfp reporter, we quantified transcript and protein levels in round and elongating spermatids of WT and Trim66^gfp/gfp mice. (A) Box plot of the mRNA level estimated from RNA-seq. Gene expression is reported with FPKM (see the Materials and Methods section). Reads mapping to the eGFP sequence inserted in the mouse genome could be detected in testicular cells from Trim66^gfp/gfp mice but not from WT samples, indicating active transcription of the Trim66-eGFP gene. (B) Violin plot represents the flow cytometry analysis of the fluorescence emitted at 530 nm in WT (negative ctrl), D4/XEGFP (positive ctrl), and Trim66^gfp/gfp dissociated testicular cells. Analyses were performed with flowCore and ggcyto packages (https://www.bioconductor.org/packages//2.10/bioc/html/flowCore.html). The data were plotted with ggplot2. A dashed line is the cutoff for detectable fluorescence. The violin plot contains a box plot representing the median (horizontal line) and the outlying points (whiskers). Fluorescence intensity at 530 nm was undetectable in Trim66^gfp/gfp testicular cells, indicating that the gfp reporter is not functional.

Figure S4.. Normal splicing and the absence of TRIM66 protein in testes homozygous for the phd/phd mutation.

The genome view of the murine Trim66 gene (top panel). The sashimi plot showing Trim66 mRNA exon–exon junctions in WT (mid panel) and Trim66^phd/phd round spermatids (lower panel).

Figure 3.. Trim66-deficient males sire overweight progeny and produce functional spermatozoa.

(A) Trim66^gfp/gfp homozygous males sire progeny overweight at birth. The violin plots represent the weight distribution of pups at birth sired by Trim66^gfp/gfp homozygous males bred with WT females. The boundaries of the overlaid box plot show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, the rhombus indicates the mean, and the loose points show the data outliers. The P-values were calculated using a two-sided t test. 188 pups born from 5 Trim66^gfp/gfp homozygous fathers and 199 born from 5 WT fathers were weighed at birth. Pups sired by Trim66^gfp/gfp homozygous males were on average 6.9% heavier than WT controls at birth. (B) Trim66’s loss of function in females does not impact the weight of their progeny. The violin plots show the weight distribution of pups at birth, mothered by Trim66^gfp/gfp homozygous females bred with WT males. The boundaries of the overlaid box plot show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, the rhombus indicates the mean, and the loose points show the data outliers. The P-values were calculated using a two-sided t test. 54 newborns from 3 Trim66^gfp/gfp mothers and 71 newborns from WT mothers were analyzed at birth. (C) Normal litter size sired by Trim66^gfp/gfp homozygous males when bred with WT females. The total number of litters with recorded pups number from 5 homozygous fathers was 20 and from 5 WT fathers was 24. Each litter size is shown as the single dot on the plot. The box plot boundaries show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, and the rhombus indicates the mean. The P-values were calculated using a two-sided t test. (D) Normal litter size sired by Trim66^gfp/gfp homozygous females when bred with WT males. The total number of litters with recorded pups from homozygous mothers was six (born from three mothers) and from WT mothers was 10 (born from five mothers). Each litter size is shown as the single dot on the plot. The boundaries of the box plot show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, and the rhombus indicates the mean. The P-values were calculated using a two-sided t test. (E) Trim66^phd/phd homozygous males sire progeny overweight at birth. The violin plots represent the weight distribution of pups at birth sired by Trim66^phd/phd homozygous males bred with WT females. The boundaries of the overlaid box plot show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, the rhombus indicates the mean, and the loose points show the data outliers. The P-values were calculated using a two-sided t test. Two hundred pups born from 11 Trim66^phd/phd homozygous fathers and 359 born from 13 WT fathers were weighed at birth. (F) Normal litter size sired by Trim66^phd/phd homozygous males when bred with WT females. The total number of litters with recorded pups from 11 homozygous fathers was 31 and from 13 WT fathers was 51. Each litter size is shown as the single dot on the plot. The boundaries of the box plot show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, and the rhombus indicates the mean. The P-values were calculated using a two-sided t test. (G) Average testicular weights in mg of adult mice of WT and gfp/gfp genotypes (n = 6). P-values were calculated using a two-sided t test. The boundaries of the box plot show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, and the rhombus indicates the mean. The P-values were calculated using a two-sided t test. Each dot represents the weight of one testis collected from the six animals. (H) Histological sections stained with periodic acid–Schiff stain and hematoxylin of paraffin-embedded testes dissected from Trim66^gfp/gfp animals. The representative cell type for each stage is labeled with an arrow accordingly: rSt, round spermatid; eSt, elongating spermatid; P, pachytene spermatocyte; Di, diplotene spermatocyte; L, leptotene spermatocyte; Z, zygotene spermatocyte; Z/P, zygotene/pachytene spermatocyte; M, meiosis I and II cells. The scale bar is 50 μm. (I) Super-resolution imaging of high-order chromatin organization of elongated spermatids after DNA staining by stimulated emission depletion microscopy. Representative raw, denoised, and MSSR processed images are shown. The chromocenter is delimited with a yellow line in the denoise micrographs. WT and Trim66^gfp/gfp elongated spermatids are shown. The scale bar is 1 μm. (J) Quantification of the chromatin arrangements imaged by stimulated emission depletion microscopy. The calculated radial-averaged distance autocorrelation function for the DNA folding pattern in the chromocenter area (bottom right panel) and spermatid nucleus with exclusion of the chromocenter (bottom left panel). A regression line was calculated using a gamma function. The horizontal bar in the box in the box plot shows the median, and the square shows the mean. The P-values were calculated using a one-sided Wilcoxon test. The violin plot represents the data distribution for all collected points. (K) Caudal sperm count of 12-wk-old males for WT (n = 4) and gfp/gfp (n = 10). P-values were calculated using a two-sided t test. The boundaries of the box in the box plot show the data within the first and third quantile, the whiskers show the maximum and minimum quantiles, and the horizontal bar indicates the median. (L) Sperm mobility was assessed on animals aged 12 wk. The mature sperm was sourced from the cauda from WT (n = 4) and gfp/gfp (n = 7). P-values were calculated using a two-sided t test. The boundaries of the box in the box plot show the data within the first and third quantile, and the single dot shows data outliers. The whiskers show the maximum and minimum quantiles, and the horizontal bar indicates the median. (M) In vitro fertilization with sperm sourced from the cauda for WT (n = 4) and gfp/gfp (n = 7). The fertilized eggs were grown in vitro until they reached the two-cell stage. P-values were calculated using a two-sided t test. The boundaries of the box in the box plot show the data within the first and third quantile, and the single dot shows data outliers. The whiskers show the maximum and minimum quantiles, and the horizontal bar indicates the median.

Figure S5.. Intercrossing Trim66^gfp/gfp animals produces overweight pups along with smaller litter sizes at birth.

(A) Violin plot of weaning weight distribution for pups sired by Trim66^gfp/gfp males. The number of pups sired by Trim66^gfp/gfp males was 169, and for WT, the number of pups was 193; the P-value was calculated by a one-sided (greater) t test. The overlaid box plot boundaries show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, and the rhombus indicates the mean. The P-values were calculated using a two-sided t test. (B) Violin plot of weaning weight distribution for pups sired by Trim66^phd/phd males. The number of pups sired by Trim66^phd/phd males was 135, and for WT, the number of pups was 274; the P-value was calculated by a one-sided (greater) t test. The overlaid box plot boundaries show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, and the rhombus indicates the mean. The P-values were calculated using a two-sided t test. (C) Violin plot of birth weight distribution of pups at birth sired by Trim66^gfp/gfp homozygous males bred with homozygous Trim66^gfp/gfp females. P-values were calculated using a two-sided t test. The homozygous pups born from homozygous cross show overweight phenotype with P-value = 0.00000494. The number of homozygous Trim66^gfp/gfp pups was 85, and for WT, the number of pups was 134. The boundaries of the overlaid box plot show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, the rhombus indicates the mean, and the loose points show the data outliers. The P-values were calculated using a two-sided t test. (D) Distribution of the number of pups born from Trim66^gfp/gfp homozygous males and females cross. The total number of litters from homozygous Trim66^gfp/gfp males and females was 10 and 13 from the WT. Each litter size is shown as the single dot on the plot. The box plot boundaries show the data within the first and third quantile, whiskers indicate minimum and maximum quartiles, the horizontal bar in the box plot shows the median, and the rhombus indicates the mean. The P-values were calculated using a two-sided t test.

Figure S6.. Morphology of spermatozoa produced by Trim66^gfp/gfp males and sperm parameters of Trim66^phd/phd.

(A) Representative sperm observed by light microscopy. Left panel: WT sperm; right panel: sperm produced by homozygous Trim66^gfp/gfp males. Scale bars: 10 μm. (B) Average testicular weights in mg of adult mice from 9 Trim66^phd/phd males and 10 WT males with total n = 18 for Trim66^phd/phd, and n = 20 for the WT. P-values were calculated using a two-sided t test. The whiskers show the maximum and minimum quantiles, and the horizontal bar indicates the median. Each dot represents the weight of one testis collected from the 20 animals. (C) Caudal sperm count of 12-wk-old males for phd/phd (n= 6) and WT (n = 6). P-values were calculated using a two-sided t test. The boundaries of the box in the box plot show the data within the first and third quantile. (D) Sperm motility measurements from 12-wk-old males. P-values were calculated using a two-sided t test. The whiskers show the maximum and minimum quantiles, and the horizontal bar indicates the median. The boundaries of the box in the box plot show the data within the first and third quantile, and the single dot shows data outliers. (E) In vitro fertilization with sperm sourced from the cauda. The number of technical replicates was three for each biological replicate (n = 6). The fertilized eggs were grown in vitro until they reached the two-cell stage. P-values were calculated using a two-sided t test. The whiskers show the maximum and minimum quantiles, and the horizontal bar indicates the median. The boundaries of the box in the box plot show the data within the first and third quantile, and the single dot shows data outliers.

Figure S7.. Overview and validation of the FACS gating strategy for the isolation of round and elongated spermatids.

(A, B, C, D, E, F) FACS gating strategy for sorting of WT round and elongated spermatids. (G, H, I, J, K, L) FACS gating strategy for sorting of round and elongated spermatids from homozygous gfp/gfp animals. The FACS plots are the output from BD FACSDiva software. (A, G) Gating for the propidium iodide (PI) and Hoechst 33342. PI is the dye for selection for live/dead cells. The Hoechst allows for selection of the cells according to their cell cycle. The selected peak for a gate P1 corresponds to haploid cells. (B, H) Forward scatter for single-cell selection gate. (C, I) Side scatter for single-cell selection gate. (D, E, J, K) Final selection gates for elongating and round spermatids. (F, L) Spermatid population across all cell populations sorted from the testis labeled in red. (M) Representative micrographs of FACS-sorted round and elongated spermatids stained by Hoechst 33342. Left panel: sorted round spermatids; right panel: sorted elongated spermatids. The scale bar is 10 μm. (N) Principal component analysis of the transcriptome (total RNA-seq) of the FACS-sorted round and elongated spermatids. (O) Heat map showing the expression (Z-score) of marker genes known to be specifically expressed in round and elongated spermatids. Marker genes were taken from the single-cell transcriptomic analysis in Green et al (2018).

Figure 4.. Gene expression and H3K4 methylation changes in post-meiotic germ cells disrupted for Trim66.

(A) Normal silencing of retrotransposons in Trim66-deficient round spermatids and elongated spermatids. Scatter plot comparing the expression of retrotransposons in WT and Trim66^gfp/gfp round spermatids (left panel) and elongated spermatids (right panel). No significant difference was observed. The family of retrotransposons previously reported to be reactivated in Trim66 knockout mouse embryonic stem cells is shown (Zuo et al, 2022), and no change was detected in round spermatids and elongated spermatids lacking Trim66. (B) MA plots showing log₂ fold changes in gene expression (total RNA-seq) in Trim66^gfp/gfp round spermatids (left panel, n = 6) and Trim66^gfp/gfp elongated spermatids (right panel, n = 4). Significant gene expression changes are shown in red (P < 0.05, Wald’s t test). Significantly up-regulated histone-modifying enzymes are shown: H3K4-specific histone methyltransferases Set1b, Kmt2e, Kmt2d, and the histone acetyltransferase Ep400. Trim66 is significantly down-regulated and is labeled in blue. (C) Gene-concept network representing the results of the gene ontology functional enrichment analysis on significantly up-regulated genes. The graph shows the top five most enriched GO terms as blue nodes. The node size represents the number of significantly up-regulated genes annotated to a given term. Genes annotated to those terms are represented in shades of red. The color intensity represents the log₂ fold change. (D) Epigenomic profiling of H3K4me3 in spermatozoa isolated from WT and Trim66^gfp/gfp epidermis. Gene stack plots displaying ChIP-seq signal for all H3K4me3 peaks and average peak profiles in WT and Trim66^gfp/gfp sperm as indicated on the top. (E) Scatter plot comparing the average peak intensity of sperm H3K4me3 between Trim66^gfp/gfp and WT. Black line: true diagonal; blue line: linear regression fitted on the data. (F) Dot plot representing the number of H3K4me3 peaks in spermatozoa for the six analyzed samples. We detected on average 62 H3K4me3 peaks in Trim66-mutant sperm and 58 in WT. The slight increase in peak number in sperm chromatin of Trim66-mutant animal was not statistically significant (P = 0.44, t test).

Figure S8.. Up-regulation of genes encoding histone modification writers and readers in Trim66^gfp/gfp round spermatids.

expression level of differentially expressed genes identified in the network analysis as histone modifiers and readers of histone marks. Trim66^gfp/gfp round spermatids (n = 6) and WT control (n = 6). Several factors involved in the homeostasis of H3K4me3 and histone acetylation are unregulated in round spermatids lacking Trim66.