Introduction

Reproduction is a highly regulated process that requires coordination of the functions of numerous genes. Infertility affects 10–15 % of human couples, and a male factor is estimated to be involved in nearly half of these cases (Visser and Repping 2010). Among domestic animals, reproductive performance has been declining for a long time in dairy cows (Lucy 2001; Royal et al. 2008; Maas et al. 2009).

Approximately, 600 testis-specific protein-coding genes have been identified. Null mutations have been introduced in nearly 400 genes associated with spermatogenesis using knockout mouse technology (Matzuk and Lamb 2008; Jamsai and O’Bryan 2011; Massart et al. 2012). Information regarding genetic abnormalities in spermatogenesis obtained using reverse genetics approaches is expected to further help understand male infertility. Recent advances in genetics have paved the way for the development of effective methods to study male infertility and subfertility. Accordingly, the number of “repro” mouse strains produced by the JAX Reproductive Mutagenesis Program (Handel et al. 2006) has reached more than a hundred, and a number of genes responsible for male infertility and subfertility have been identified (http://reprogenomics.jax.org). The bidirectional approaches described above are intended to study infertility caused by a single gene. However, reproduction is temporally regulated by the coordinated action of a number of genes.

Our approach was also genetic, and the phenotype was polygenic. Previously, 17 strains of mice were surveyed to determine the frequency of abnormal sperm-head morphology, and B10.M/Sgn (B10.M) and B10.MOL-TEN1 (TEN1) strains were ranked first and second, with respect to the frequency of sperm-head morphological abnormalities (Gotoh 2010). Both strains of mice showed teratozoospermia, producing sperm cells displaying a wide variety of morphological abnormalities. Segregation analysis revealed that the sperm phenotype was heritable in both strains. The frequency of occurrence of each type of morphological abnormality varied among individuals (Gotoh et al. 2012; Hirawatari et al. 2015). Further quantitative trait loci (QTL) analysis and fine mapping of B10.M strain identified two causative loci, Sperm-head morphology 1 (Shm1) on chromosome (Chr) 1 and Shm2 on Chr 4 (Gotoh et al. 2012). As to male fertility capacities of the two strains, the B10.M strain showed low fecundity while the TEN1 strain was fully fertile (Hirawatari et al. 2015). The relationship between sperm-head morphology and male subfertility is yet to be established.

Sperm phenotype of TEN1 strain has been reported to be controlled by at least three responsible loci (Hirawatari et al. 2015). In this study, we analysed the genetic control of sperm-head morphological abnormalities in the TEN1 strain by QTL analysis using backcross males between TEN1 and C3H/HeN strains, and gene mapping using F2 intercross males.

Materials and methods

Mice

All the experimental procedures were approved by the Institutional Animal Care and Use Committee of the National Institute of Agrobiological Sciences. The identification code for the animal experiments in this study at the institute is H20-009. Animals were housed and cared for according to the guidelines established by the Committee. B10.MOL-TEN1 (TEN1) mice were provided by the RIKEN BioResource Center (Tsukuba, Ibaraki, Japan). C3H/HeNCrlCrlj (C3H) mice were purchased from Charles River Japan (Yokohama, Japan). B10.M/Sgn (B10.M) mice are maintained at our facility. The animals were maintained on a cycle of 12 h of light and 12 h of darkness under specific pathogen-free conditions. A commercial mouse diet and water were provided to them.

TEN1 strain

This strain was one of the congenic lines established at the National Institute of Genetics, Japan (Mishima, Shizuoka, Japan) to define H2 alleles on Chr 17 in the Japanese wild mouse population (Shiroishi et al. 1982). The original donor was a wild non-inbred animal (Mus musculus molossinus). The wild-derived H2 allele was transmitted to B10 over twelve generations. Subsequently, this strain has been maintained via sibling breeding.

Sperm morphology tests

Sperm collection from the cauda epididymis and sperm counts were performed according to methods in the report on the mouse sperm morphology test (Wyrobek et al. 1983). Sperm samples were collected from 3- to 5-month-old male mice. The animals were deeply anesthetised using 8.0 % isoflurane for inhalation until death, following which the epididymides were dissected. To obtain sperm for assessing its morphology, the epididymis from one side was removed and sperm were transferred into 1 mL of phosphate-buffered saline (pH 7.0) with 0.1 % glucose. The sperm-head morphology test was performed by the method described previously (Gotoh et al. 2012). Briefly, 1–5 µL of the sperm suspension was spread on a glass slide, air dried, fixed using ethanol, and sperm morphology was observed by differential interference contrast microscopy (DMRXA2; Leica Microsystems, Cambridge, UK) (×400 magnification). Two independent samples, each containing at least 200 sperm cells, were analysed. The classification of sperm-head morphological abnormalities is shown in Supplementary Fig. S1.

Preparation of genomic DNA

Mouse tails were resuspended in 300 μL Tris buffer (50 mM, pH 7.8) containing 100 mM ethylenediaminetetraacetic acid (EDTA), 100 mM NaCl, and 1 % sodium dodecyl sulphate (SDS) (w/v). Proteinase K (Wako, Osaka, Japan) was then added to a final concentration of 500 μg/mL. The samples were incubated on a shaker (Thermomixer comfort; Eppendorf, Hamburg, Germany), at 1000 rotation per minute, overnight, at 56 °C. RNase A (Wako, Osaka, Japan) was then added to a final concentration of 10 μg/mL, and the samples were incubated for 1 h at 37 °C. Following this, 300 μL DNA-binding reagent containing 4.5 M guanidine hydrochloride (Wako, Osaka, Japan), 0.5 M potassium acetate (pH 5.0), and 40 mg/mL silica gel (Sigma-Aldrich #288519, St. Louis, MO, USA) was added to the samples. The samples were then mixed by shaking for 1 min to allow the DNA to bind to the silica gel. Silica gel particles were retrieved and washed three times with a solution containing 80 % ethanol (v/v), 10 mM potassium acetate (pH 5.0), and 20 μM EDTA. After air drying, the bound DNA was eluted from the silica gel using TE buffer (pH 8.0). DNA concentration was calculated indirectly on the basis of the absorbance at 260 nm measured using a UV spectrophotometer (GeneQuant; Pharmacia Biotech, Cambridge, UK).

Genotyping

A genome-wide scan was conducted using single nucleotide polymorphism (SNP) markers spaced at 10 cM for backcross (BC) males (n = 176) produced from a genetic cross of ♀TEN1 × ♂(♀C3H × ♂TEN1)F1 mice. The 103 SNP markers used for the QTL analysis are listed in Supplementary Table S1. The markers were chosen from the Mouse SNA database (http://www.broad.mit.edu/snp/mouse/) as described previously (Furuse et al. 2012). We also used microsatellite markers (Supplementary Table S2) for a detailed mapping analysis for F2 males (n = 592) produced from intercross of (♀C3H × ♂TEN1)F1 mice. Procedures were same as described earlier (Gotoh et al. 2012).

QTL analysis

QTL analysis was performed to identify loci controlling frequencies of sperm-head abnormalities using R software (ver. 3.1.3 available at http://www.rqtl.org/) and qtl package (ver. 1.35-6) (Browman et al. 2003; Browman and Sen 2009). Genetic model was estimated using 176 bc males. We carried out single locus scans to identify independent effect QTLs. Because distribution of the phenotype was not normal, sperm abnormality data were transformed using Johnson SU function of the JMP statistical software (ver. 12.0.0, SAS Institute Inc., Cary, NC, USA). However, truly normalized distribution was not obtained (Supplementary Fig. S2). LOD scores were computed at 2-cM intervals across the genome using the “EM” method in R/qtl. Independent effect QTLs were considered significant (suggestive) if they exceeded the 95 % (63 %) genome-wide adjusted threshold based on 1000 permutations for autosomes and on 22,849 permutations for X chromosome. The significant (suggestive) thresholds were 2.71 (1.74) and 2.81 (1.86) for autosomes and X chromosome, respectively. For significant QTLs, a location of a 95 % confidence interval (CI) was estimated by a decline of 1.5 LOD. Epistatic effects were investigated using the genome-wide all-pairs scans. The significant thresholds for the pairwise scans were obtained from 1000 permutations. Two significant pairs of QTLs, 24.3 cM on chromosome 1 and 29.0 cM on X chromosome, and 61.8 cM on chromosome 6 and 29.0 cM on X chromosome, were suggested. Because one of QTLs in both pairs was positioned on X chromosome, multilocus scans were not applicable in R/qtl package. We then used 592 F2 males for further analysis. Microsatellite markers which cover the 95 % CI of the significant QTLs on chromosome 1, 6, and X were picked up, and then F2 samples were genotyped. Discrimination between epistatic effects and additive effects was judged by the results obtained from “effectplot” function of R/qtl.

Gene mapping

In attempt to narrow the region of QTL loci, we observed phenotype/genotype relationship using F2 males. One reason to perform this approach was that multiple QTL function of R/qtl package does not usable for a locus on Chr X. The other reason was that the sperm abnormality phenotype is clearly judged “high” or “low” with small ambiguity in this case.

Results

Sperm-head morphological abnormalities

The TEN1 strain displayed teratozoospermia. A wide variety of sperm-head morphologies, including a shortened apical hook, a bent hook, spermatozoa with ectopic attachment of the flagella, and several types of amorphous heads, were observed (Supplementary Fig. S1). The mean frequencies of abnormal sperm were 50.2 % for TEN1 and 2.4 % for C3H.

QTL analysis: main effects

Three statistically significant LOD score peaks appeared: Chr 1 at 24.3 cM with a LOD score of 29.25, Chr 6 at 63.8 cM with a LOD score of 3.58, and Chr X at 32.0 cM with a LOD score of 6.80 (Fig. 1; Table 1). The locus on Chr 1 was considered to be the major locus for the trait, while the other two loci on Chr X and Chr 6 were considered to be the minor loci. We named these loci on Chrs 1, X, and 6 as Sperm-head morphology 3 (Shm3), Shm4, and Shm5, respectively.

Table 1 QTLs identified for a single gene genome-wide scan
Full size table

QTL analysis: epistasis

Pairwise genome scans estimated two candidates for significant interactions: Chr1 at 24.1 cM with Chr X at 31.5 cM, and Chr 6 at 61.8 cM with Chr X at 31.5 cM. LOD scores by R/qtl package (LODf, LODfv1, LODint, LODa, LODav1) for these interactions were (35.7, 8.81, 0.0462, 35.6, 8.76) for Chr 1:Chr X interaction, and (12.2, 4.61, 0.0166, 12.2, 4.59) for Chr 6:Chr X interaction. Graphical draws representing LOD scores by two-dimensional, two-QTL genome scan were also shown in Supplementary Fig. S3. We then used frequencies of sperm-head abnormalities of F2 males, and analysed estimated average frequency of abnormal sperm as a function of genotype of a marker at responsible QTLs (Fig. 2). Epistatic interaction between Shm3 on Chr 1 (represented by 1@24.3) and Shm4 on Chr X (represented by X@32.0) was evident because Shm4 C3H/− suppressed sperm abnormality only when Shm3 genotype was homozygous for B10.M allele (Fig. 2a). Effect of Shm5 on Chr 6 (represented by 663.8) differed depending on the genotype of Shm4 (Fig. 2b).

Fig. 1
figure 1

Identification of the three loci responsible for the TEN1 sperm abnormalities phenotype by QTL analysis. Three statistically significant logarithms of the odds (LOD) score peaks appeared on Chrs 1, 6, and X. These peaks were found on Chr 1 at 24.6 cM with a LOD score of 29.25, on Chr 6 at 63.8 cM with a LOD score of 3.58, and on Chr X at 32.0 cM with a LOD score of 6.80. The statistically significant LOD threshold (P = 0.05) for the phenotype was 2.77 for autosomes and 2.81 for X chromosome (a dotted horizontal line), as determined by the permutation test (n = 1000) for autosomes and by the permutation test (n = 22,849) for X chromosome, respectively

Full size image
Fig. 2
figure 2

Allele effects for frequency of abnormal sperm as a function of genotype at two putative QTL. a Shm3 (1@24.3) and Shm4 (X@32.0). b Shm4 and Shm5 (6@63.8). B10.M allele is represented by M, and C3H allele is represented by C. y-axes show mean values of sperm abnormality. Error bars represent SE

Full size image
Fig. 3
figure 3

Effects of Shm3, Shm4, and Shm5 on sperm-head abnormalities. The frequency of sperm-head abnormalities of F2 animals (n = 592) was plotted for each genotype concerning the responsible loci Shm3 on Chr 1, Shm4 on Chr X, and Shm5 on Chr 6. The genotypes of D1Mit236, DXMit114, and D6Mit59 were used for the genotypes of the Shm3, Shm4, and Shm5 loci, respectively. Each genotype is represented by three capital letters or stars. The first capital represents the genotype at the Shm3 locus, the second represents the genotype at the Shm4 locus, and the third represents the genotype at the Shm5 locus. The homozygous state of autosomal Shm3 and Shm5 loci is shown as “T” from the TEN1 allele and as “C” from the C3H allele. The hemizygous state of the Shm4 locus linked on Chr X is represented as “T” from the TEN1 allele and as “C” from the C3H allele. Heterozygotes are represented as “H”. The star “*” represents “T” or “H” or “C”. For example, the “TTT” represents an animal with the genotype Shm3 TEN1/TEN1, Shm4 TEN1/−, or Shm5 TEN1/TEN1. The genotype “TCT” was divided into two groups at 20 % of sperm-head abnormalities. The horizontal bar for each genotype represents the mean frequency of sperm-head abnormalities

Full size image

Genetic model of inheritance

The frequencies of sperm-head abnormalities of a total of 592 F2 animals were rearranged according to their Shm3, Shm4, and Shm5 genotypes, and analysed to estimate the roles of the three loci in inheritance. The results depicted in Fig. 3 and Table 2 summarise our understanding of the results. The TEN1 allele at the Shm3 locus (Shm3 TEN1) was considered to be the major locus for the high frequency of sperm-head abnormalities. It acts recessively because 97 % of heterozygotes and C3H homozygotes showed low frequencies of abnormalities. Shm3 TEN1/TEN1 homozygotes showed high and low sperm-head abnormality phenotypes. Both Shm4 and Shm5 were considered conditional loci because both loci act in specific genotypes. The C3H allele of the Shm4 locus on Chr X was considered to act in suppressing the action of Shm3 TEN1/TEN1 because 96 % of Shm3 TEN1/TEN1 and Shm4 TEN1/− animals showed a high frequency of sperm-head abnormalities, and 96 % of Shm3 TEN1/TEN1, Shm4 C3H/−, Shm5 C3H/TEN1 and 96 % of Shm3 TEN1/TEN1, Shm4 C3H/−, Shm5 C3H/C3H animals showed low frequency of sperm-head abnormalities (Fig. 3, Table 2). Animals with Shm3 TEN1/TEN1, Shm4 C3H/−, Shm5 TEN1/TEN1 showed high to low frequencies of abnormalities. Fifty-two percent of animals showed high frequency of abnormalities (≥20 %), and 48 % of animals showed low frequency of abnormalities (<20 %). We considered that this genotype of animals could be divided into “High” and “Low” groups. Our understanding is that the Shm5 TEN1/TEN1 allele acts to suppress the action of Shm4 C3H/− recessively, only when the genotype of Shm3 is homozygous for the TEN1 allele and the genotype of Shm4 is C3H. From the information that 48 % of Shm3 TEN1/TEN1, Shm4 C3H, and Shm5 TEN1/TEN1 animals showed low frequency of abnormalities, another locus is expected to act in suppressing the action of Shm5 TEN1/TEN1.

Table 2 Expected function of QTL loci Shm3, Shm4, and Shm5 in sperm-head morphological abnormalities
Full size table
Table 3 Estimated function of alleles at sperm-head morphology loci
Full size table

Genetic mapping

Figure 4 shows the results of mapping the Shm3, Shm4, and Shm5 loci. The Shm3 locus was mapped between the Ercc marker (Chr 1: 23.55 cM) and the D1Mit303 (Chr 1: 31.79 cM) marker. The Shm4 locus was mapped between the DXMit1 marker (Chr X: 37.29 cM) and the DXMit170 marker (Chr X: 45.87 cM). The Shm5 locus was mapped between the D6Mit219 marker (Chr 6: 64.03 cM) and the D6Mit14 marker (77.64 cM). Because the mapping results presented above were based on one recombinant in most cases, overinterpretation should be avoided.

Fig. 4
figure 4

Fine mapping of Shm3, Shm4, and Shm5 loci. The genotypes of the informative recombinants whose break points were mapped within the responsible genetic regions on Chr 1, X, and 6 are shown. a Fine mapping of the Shm3 locus on Chr 1. b Fine mapping of the Shm4 locus on Chr X. The genotype at the Shm3 locus on Chr 1 had to be homozygous for the TEN1 allele. c Fine mapping of the Shm5 locus on Chr 6. The genotype of the Shm3 locus on Chr 1 had to be homozygous for the TEN1 allele, and the genotype of the Shm4 locus on Chr X had to be hemizygous for the C3H allele. Homozygotes for the TEN1 allele, heterozygotes, and homozygotes for the C3H allele are represented as T, H, and C, respectively. The dark shadowed regions represent the presence of the responsible loci on the TEN1 chromosomes for the Shm3 and Shm5 loci, and represent the presence of the responsible locus on the C3H chromosome for the Shm4 locus. The light shadowed chromosomal regions represent the absence of the responsible loci. The frequency of sperm-head abnormalities and the animal identification number are depicted to the right

Full size image

Discussion

Several types of sperm tests, including sperm morphology, sperm count, sperm mobility, progressive motility, and others are used to assess male fertility (Aitken 2006; Lewis 2007; Oehninger et al. 2014). However, no reliable correlation between the results of sperm tests and observed fertility has so far been established (Guzick et al. 2001). Our study was intended to focus on genetic regulation of sperm morphology in mice as a model system to analyse the complicated reproductive regulation in mammals, including humans and domestic animals. In this study, we analysed the genetic regulation of the TEN1 strain for sperm-head morphological abnormalities.

In the previous study, the sperm phenotype of the TEN1 strain was found to be heritable, and was expected to be controlled by at least three loci (Hirawatari et al. 2015). Supporting the results of the segregation analysis, three statistically significant loci, Shm3, Shm4, and Shm5, were obtained from the QTL analysis. Locus mapping and further analysis determined the complicated genetic regulation by these three loci and another predicted locus on teratozoospermia. Together with our previous study on genetic teratozoospermia of B10.M strain (Gotoh et al. 2012), Fig. 5 shows our current understanding of the model representing the polygenic regulations of sperm-head abnormalities in TEN1 and B10.M. In both cases, complicated cascades of epistatic interactions were predicted. In the case of the TEN1 strain, the homozygosity of Shm3 TEN1 allele was confirmed to act as the major effect on teratozoospermia. It resembles the action of Shm1 B10.M allele of B10.M. Sequential up- and down-regulations by Shm4, Shm5, and then by another predicted locus yet to be determined were estimated to constitute a cascade of interactions for teratozoospermia in TEN1 strain. Table 3 summarises the expected function of Shm loci obtained from both this study on TEN1 and our previous study on B10.M.

Fig. 5
figure 5

Schematic models of polygenic regulation of sperm-head abnormalities in B10.M and TEN1 strains of mouse. Italic letters represent the genotype of the each locus. Thin arrows next to the genotypes show the actions of the genotypes regarding to the frequencies of sperm-head abnormalities. Thick arrows show the sequential polygenic interactions of the loci involved. As C3H strain was used for genetic analysis of the B10.M strain (Gotoh et al. 2012), the Shm4 C3H/− has been confirmed to have no effect to the Shm1 B10.M/B10.M genotype. Interaction between the Shm3 TEN1/TEN1 and the Shm2 B10.M/B10.M has not been tested

Full size image

Coincidentally, the Shm3 locus in TEN1 and the Shm1 locus in B10.M were found to be located in the overlapping region on Chr 1. It may be that Shm1 and Shm3 encode the same gene. However, even if Shm1 B10.M and Shm3 TEN1 are allelic, they certainly differ in their functional properties; Shm4 C3H on Chr X interacts with Shm3 TEN1/TEN1 on Chr 1 suppressively, but it does not interact with Shm1 B10.M/B10.M (Gotoh et al. 2012). Determination of genes encoded by both Shm1 locus and Shm3 locus would expected to clarify the issue.

Shm4 C3H and Shm4 TEN1 were also found to differ in their functional properties (Table 3). Shm4 C3H/− suppresses the action of Shm3 TEN1/TEN, while Shm4 TEN1/− does not. Because TEN1 is an H2 congenic strain on the background of B10 genome, sequence variation should be present between the genes encoded by Shm4 C3H and Shm4 B10. Within the region between the DXMit1 and DXMit170 markers, polymorphic non-synonymous SNPs within the coding region between C3H/HeJ and C57BL/10J strains were searched for using the Mouse SNP query of the public database at the Jackson Laboratory (http://www.informatics.jax.org/javawi2/servlet/WIFetch?page=snpQF, Accessed Dec. 1, 2014). A total of 13 polymorphic genes and one synthetic QTL were found; Tktl1, Llna, Gdi1, 4930468A15Rik, 4930595M18Rik, Heph, Stard8, Atrx, Atp7a, Hmgn5, Tex16, 2010106E10Rik, Cpxcr1 genes, and Hst3 QTL. These comprise the candidate genes encoded by Shm4. Interestingly, two male hybrid sterility QTL loci, Hstx1 (Storchová et al. 2004) and Mhysq2 (Vyskocilová et al. 2009), and one sperm-head anomaly locus, Spha2 (Oka et al. 2004), have been mapped within the same region. The relationship between these loci and Shm4 remains unknown.

No equivalent allele of Shm2 B10.M on Chr 4, which enhance the action of Shm1 B10.M/B10.M, was found in the TEN1 genome. This may be because the Shm2 B10.M locus contains a mutation that the TEN1 allele does not. Another possibility is that Shm2 B10.M and Shm2 TEN1 are one and the same, and interact with Shm1 B10.M but not with Shm3 TEN1.

Similarly, no equivalent allele of Shm5 TEN1 on Chr 6 was found in the B10.M genome. Shm5 TEN1 suppresses the action of Shm4 C3H/− recessively, if the genotype is Shm3 TEN1/TEN1 and Shm4 C3H/−. From our research so far, it is not possible to predict the function of Shm5 B10.M.

At least one additional interacting locus involved in sperm-head abnormalities was predicted to exist in the genetic system analysed in this study. We expected the function of Shm5 TEN1 to suppress the action of Shm4 C3H/− conditionally (Table 3). However, 48 % of the animals genotyped as Shm3 TEN1/TEN1, Shm4 C3H/−, and Shm5 TEN1/TEN1 showed low frequency of sperm-head abnormalities, contrary to our expectation (Fig. 3; Table 2). Another locus that suppresses the action of Shm5 TEN1/TEN1 should be present. The interacting genetic element may not be solitary. Further interpretation became difficult because the sample size of this study (n = 592) was not sufficient.

Because TEN1 strain is the H2 congenic strain, we predicted that the gene responsible for the abnormal sperm-head phenotype was located on Chr 17. However, the responsible loci were mapped outside of the H2 complex on Chr 17. The origin of Shm3 TEN1 and Shm5 TEN1 mutations is speculated to have either transferred from the progenitor wild mouse, the donor of the H2 complex of TEN1, or to have emerged de novo during the establishment of the TEN1 strain. Although both Shm1 B10.M and Shm3 TEN1 were mapped within the identical region on Chr 1, it was considered to be a coincidence because both the TEN1 and B10.M strains were developed independently at the National Institute of Genetics, Japan (Shiroishi et al. 1982) and at the Jackson Laboratory (Snell and Jackson, 1958), respectively.

From our studies using mutant strains showing teratozoospermia, epistatic interactions of Shm loci became apparent. Both the TEN1 and the B10.M strains are the useful tools for determining responsible genes for teratozoospermia. However, epistasis on teratozoospermia has been shown to be a common feature in the laboratory mice. In the study to observe sperm-head abnormalities in F2 animals produced between C3H/HeN and C57BL/6 J strains, involvement of numerous loci for teratoazoospermia has been reported (Gotoh and Aoyama, 2012). The observation that ~4 % of animals showed unexpected values for the frequency of sperm-head abnormalities in this study (Fig. 3; Table 2) can be interpreted in the same context. The involvement of epistasis in mouse reproductive performance has been reported in other studies (Peripato et al. 2004; Flachs et al. 2012). Our findings further confirm the notion that the reproductive system is controlled by the coordinated action of numerous genes.

Although TEN1 strain shows both teratozoospermia and considerable histological changes in testis, its male reproductive performance was observed normal. On the other hand, B10.M strain shows teratozoospermia and male subfertility (Hirawatari et al. 2015). Relationship between abnormal sperm morphology and male subfertility in B10.M strain has not been established. Identification of responsible gene(s) for B10.M male subfertility is necessary to clarify the important issue of reproduction.