Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2014 May 8;10(5):e1004317.
doi: 10.1371/journal.pgen.1004317. eCollection 2014 May.

Paternal Poly (ADP-ribose) Metabolism Modulates Retention of Inheritable Sperm Histones and Early Embryonic Gene Expression

Free PMC article

Paternal Poly (ADP-ribose) Metabolism Modulates Retention of Inheritable Sperm Histones and Early Embryonic Gene Expression

Motomasa Ihara et al. PLoS Genet. .
Free PMC article


To achieve the extreme nuclear condensation necessary for sperm function, most histones are replaced with protamines during spermiogenesis in mammals. Mature sperm retain only a small fraction of nucleosomes, which are, in part, enriched on gene regulatory sequences, and recent findings suggest that these retained histones provide epigenetic information that regulates expression of a subset of genes involved in embryo development after fertilization. We addressed this tantalizing hypothesis by analyzing two mouse models exhibiting abnormal histone positioning in mature sperm due to impaired poly(ADP-ribose) (PAR) metabolism during spermiogenesis and identified altered sperm histone retention in specific gene loci genome-wide using MNase digestion-based enrichment of mononucleosomal DNA. We then set out to determine the extent to which expression of these genes was altered in embryos generated with these sperm. For control sperm, most genes showed some degree of histone association, unexpectedly suggesting that histone retention in sperm genes is not an all-or-none phenomenon and that a small number of histones may remain associated with genes throughout the genome. The amount of retained histones, however, was altered in many loci when PAR metabolism was impaired. To ascertain whether sperm histone association and embryonic gene expression are linked, the transcriptome of individual 2-cell embryos derived from such sperm was determined using microarrays and RNA sequencing. Strikingly, a moderate but statistically significant portion of the genes that were differentially expressed in these embryos also showed different histone retention in the corresponding gene loci in sperm of their fathers. These findings provide new evidence for the existence of a linkage between sperm histone retention and gene expression in the embryo.

Conflict of interest statement

The authors have declared that no competing interests exist.


Figure 1
Figure 1. Experimental design to ascertain the impact of sperm chromatin structure on early embryonic gene expression.
(A) Efficiency of histone-to-protamine exchange in spermiogenesis depends in part on levels of poly(ADP-ribose) (PAR) formed transiently by the interplay of PAR polymerases (PARP1, PARP2) and PAR glycohydrolase (PARG). Inhibition of PAR synthesis by PJ34 or disruption of normal PARG activity in the Parg(110)−/− mouse leads to abnormal chromatin remodeling with retention of histones in sperm . (B) Natural mating of Parg(110)−/− males or males treated with PJ34 with wild-type control females was used to obtain 2-cell stage embryos (2CE) for genome-wide transcriptional profiling at the individual embryo level using microarrays and high throughput sequencing (HTS). Cauda epididymal sperm from the fathers were used to identify genes associated with nucleosomes rather than protamines using micrococcal nuclease digestion (MND). Aberrant histone association of gene loci with differential expression of genes in two-cell embryos was assessed and compared to embryo expression data.
Figure 2
Figure 2. Aberrant chromatin composition in mouse models of altered PAR metabolism.
Chromomycin A3 (CMA3) intercalation into the DNA indicates incomplete chromatin condensation in sperm from Parg(110)−/− (A) and PJ34-treated (C) males with histone retention. (B, D) Histogram of sperm CMA3-staining intensities reflects that severity of CM3A staining varied at the level of individual sperm and individual fathers (n>200 nuclei/sample, 3 males/group). (E) Immunoblot analyses of sperm protein lysates showing increase in histone retention in PJ34 treated males. TUBA1A: alpha tubulin loading control. (F) Overlaps of genes identified as differentially histone associated in sperm from 3 individual Parg(110)−/− males (“PargA”, “PargB”, “PargC”, the fathers of the embryos analyzed below) by micrococcal nuclease digests (MND) compared to the wild-type controls. The “PargAll” data set contains all genes commonly identified as differentially MNase-sensitive across 10 Parg(110)−/− males compared with 9 wild-type control animals. The red circle indicates common genes that were differentially histone associated in all groups (1604+216 = 1820, red circle) compared with wild-type. (G) PJ34: differentially MNase-sensitive genes in three different males (like in E) and overlap with a surrogate dataset (“PJ34All”) consisting of data from all 4 PJ34-treated males compared with 9 wild-type control males. The overlap of 2,489 genes that were commonly differentially histone associated in sperm samples is indicated (blue circle). (H) Overlap of genes commonly affected by differential histone association between the Parg(110)−/− and the PJ34 models compared to wild-type controls (red and blue circles in F and G). A Pearson correlation examining significance of this overlap using a genetic background of 19,472 genes was calculated with a resulting P<0.0001, dismissing the null hypothesis that the observed overlap is coincidental (predicted number). The list and GO-term analysis of the 583 genes is contained in Dataset S4 (MS Excel).
Figure 3
Figure 3. Perturbing PAR metabolism results in differential sperm histone association of gene loci with either excessive or reduced retention of nucleosomes.
A) Functional GO-term enrichment of genes affected by elevated histone association (MAT(+)) or local failure to retain histones in regulatory gene sequences (MAT(−)) in sperm from Parg(110)−/− (left panels) and PJ34-treated males (right hand panels). The y-axis shows GO terms and logarithmic scale indicates their p-values of GO-terms returned by DAVID. False discovery rates (FDR) are indicated above the graphs. The numbers of genes in a given GO-term are in parentheses. (B) Overlaps of relative histone enrichment or deficiency in Parg(110)−/− or PJ34-treated mouse models compared to wild-type controls. (C, D) Comparison of genes that are differentially histone associated in Parg(110)−/− or PJ34 sperm with known maternal transcripts or newly expressed embryonic transcripts or spermatogenesis-specific genes indicates the potential relevance of aberrant histone association on genes expressed in the 2-cell embryo (Embryo). Maternal: transcripts present in 1-cell embryos prior to the major wave of genome activation . The genes in the maternal, embryonic and spermatogenic groups are listed in Dataset S4 (MS Excel).
Figure 4
Figure 4. Differential sperm histone association of genes in Parg(110)−/− and PJ34-treated males is significantly associated with differential expression of these genes in offspring 2-cell embryos.
(A) Outline of the comparison procedure, shown here only for the Parg(110) gene disrupted mouse model. A similar regimen was used for the PJ34-treated males and their offspring. Differential histone association of genes in sperm from individual males (PargA, PargB, PargC) was determined by pair-wise comparison with all individual wild-type data sets obtained from 9 individually analyzed control males. Differential gene expression was determined by pair-wise comparison of individual offspring from the Parg(110)−/− males (PargA1-PargC3) with all 9 individual wild-type 2-cell embryo data sets by either microarrays or next generation sequencing. Differential gene expression was determined using ANOVA analyses and adjusted P-value calculation (with Padj<0.1 considered to be significant). (B) Pearson (uncorrected) and Yates (corrected) Chi-squared tests were used to determine the significance of overlaps of the lists of genes that were differentially histone associated in sperm samples of the sires (“Sperm samples”) compared to controls (Parg A–C and PJ34A–C), with the lists of genes that were differentially expressed in at least one of the 3 or 4 offspring 2-cell embryos from these sires (“2CE DE”, Parg(110)−/−: A 1–3, B 1–3, C 1–3, and for PJ34: A1–4, B1–4, C1–2). A genetic background of 19,472 genes interrogated by the microarrays and 20,018 genes interrogated by the tiling arrays and sequencing platforms was used for the calculations. Ranges of P-values resulting from Yates or Pearson are indicated in different shades of blue if P≤0.05, i.e., the overlaps were significant (see color legend in figure). The P-value denotes the confidence with which the null-hypothesis can be dismissed that the overlap between the list of genes with abnormal histone association in the sire with the list of genes that are DE in the offspring could be predicted by statistical probability, i.e. coincidence. Upper two panels: Parg(110)−/− group of fathers and offspring embryos, microarray expression analyses with either Yates' chi-squared test (upper triangular portion of each cell) or Pearson's (lower triangular portion of cells). HTS: Parg(110) group, high throughput sequencing of 2CE gene expression. Lower panel: overlaps of PJ34-treated group of fathers (sperm) and offspring embryos (2-cell embryo differential gene expression, “2CE DE”). Note that mainly overlaps between genes with lower histone retention in sperm and differentially expressed genes in offspring embryos are significant with P≤0.05 (left 1/3 portion of each panel).
Figure 5
Figure 5. Genes that are affected by altered sperm histone association and differential gene expression in 2-cell embryo progeny in the Parg(110)−/− (A, B, C) and PJ34-treated (D, E, F) mouse models share some common functional relevance.
Analyses of functional ontology of genes that are both affected by elevated (MAT(+)) or lower (MAT(−)) differential histone association in sperm and differential expression in resulting embryos (2CEDE/MND) were performed in the Parg(110)−/− (A) or PJ34 (D) model system. Genes are listed according to up- (ratio>1) or down (ratio<1) regulation compared to controls in offspring from Parg(110)−/− or PJ34 treated males and p-values of the GO terms (DAVID) associated with the corresponding group of genes are given on the logarithmic x-axis. The number of genes in each GO functional category is in parentheses. FDR: false discovery rates. (B, C, E, F) Comparing lists of genes that were either down-regulated (B, E) or up-regulated (C, F) in offspring of Parg(110)−/− or PJ34 treated males compared to controls (taken from (A) or (D)) to lists of genes of known embryonic , or maternal , origin demonstrated that maternal transcripts left over from the zygote were largely unchanged. Down-regulated transcripts (ratio<1) in 2-cell embryos of fathers with perturbed PAR metabolism were mostly of embryonic, but not maternal origin (B, E). Up-regulated genes are those that are not normally expressed during this phase of embryo development (C, F, underlined). These genes appear precociously expressed here and are normally only active after the blastocyst stage of development.
Figure 6
Figure 6. Shared differential gene expression in offspring from males of the two mouse models with perturbed PAR metabolism (Parg(110)−/−, PJ34 treated).
(A) The overlap of 150 differentially expressed 2-cell embryo genes (2CEDE) from Parg(110)−/− and PJ34 treated males is highly significant (Yates' and Pearson's Chi-squared tests using a genetic background of 19,472 genes, P<0.0001 in both analyses, the null hypothesis would be 70 genes in the overlap). Of the 150 genes commonly differentially expressed in embryos of the two mouse models of reduced PAR metabolism, 107 are commonly expressed at higher levels and 30 are commonly down-regulated (right panel). The shaded fields indicate genes with variable expression; these also have high coefficients of variation (Cv>5%) in the variance analyses. (B) There is also a significant overlap of 33 genes that were both differentially expressed in individual embryos and differentially histone associated in the corresponding sperm sample (2CEDE/MND genes) between the two models (Parg(110)−/− and PJ34-treated fathers (P<0.0001, the null hypothesis would have been 4.8 genes in the overlap)). The relationships of differential expression of these genes (ratio>1: R>1 or ratio<1: R<1) are again very similar for the genes in the overlap (box panel to the right). (C) Identity of the genes in the overlaps shown in (A). The 33 genes in the overlap shown in (B) are bold and in brackets. Variably (ambiguously) expressed genes are listed in column 5.
Figure 7
Figure 7. Chromatin remodeling events in spermiogenesis affect sperm histone-dependent regulation of gene expression during embryonic genome activation (working model).
Chromatin remodeling during spermiogenesis (left panel) leads to the exchange of nucleosomes (blue) with their specific paternal tail modifications by protamines that normally leads to regulated retention of histones in certain domains of a given gene locus (A). Modulation of this remodeling process, for instance by altering PAR metabolism, results in either insufficient exchange of histones (B) or excessive remodeling causing more intense depletion of histones from that locus (C). As a result, histone association of this locus can be variable in sperm (middle panel). After fertilization, the paternal chromatin becomes rapidly remodeled, again with the exchange of protamines, but presumably not paternal histones, for maternally provided histones (pink) with maternal tail modifications that are mostly activating or nondescript in nature (right hand panel). As a result, the ratio of maternal and paternal histones can vary at the time point of genome activation, leading to continued differential epigenetic remodeling of the locus and ensuing differential expression (DE) in the early embryo.

Similar articles

See all similar articles

Cited by 22 articles

See all "Cited by" articles


    1. Rivera RM (2010) Epigenetic aspects of fertilization and preimplantation development in mammals: lessons from the mouse. Syst Biol Reprod Med 56: 388–404 10.3109/19396368.2010.482726 - DOI - PubMed
    1. Braun RE (2001) Packaging paternal chromosomes with protamine. Nat Genet 28: 10–12 10.1038/88194 - DOI - PubMed
    1. Balhorn R (2007) The protamine family of sperm nuclear proteins. Genome Biol 8: 227 10.1186/gb-2007-8-9-227 - DOI - PMC - PubMed
    1. Arpanahi A, Brinkworth M, Iles D, Krawetz SA, Paradowska A, et al. (2009) Endonuclease-sensitive regions of human spermatozoal chromatin are highly enriched in promoter and CTCF binding sequences. Genome Res 19: 1338–1349 10.1101/gr.094953.109 - DOI - PMC - PubMed
    1. Brykczynska U, Hisano M, Erkek S, Ramos L, Oakeley EJ, et al. (2010) Repressive and active histone methylation mark distinct promoters in human and mouse spermatozoa. Nat Struct Mol Biol 17: 679–687 10.1038/nsmb.1821 - DOI - PubMed

Publication types

LinkOut - more resources