Subjects and sample collection

The study was approved by the Ethics Review Board of the Nanjing Medical University. All the studies involving human subjects were conducted in full compliance with government policies and the Declaration of Helsinki. A total of 1,657 infertile patients, diagnosed with unexplained male factor infertility, were drawn from the Centre of Clinical Reproductive Medicine between April 2005 and March 2009 (NJMU Infertile Study). All participants completed an informed consent and a questionnaire, including detailed information, such as age, cigarette smoking, alcohol drinking, tea and vitamin consumption, and abstinence time. All patients underwent at least two semen analyses, and those with a history of orchitis, obstruction of the vas deferens, chromosomal abnormalities, or micro-deletions of the azoospermia factor region on the Y chromosome were excluded [22]. In the final analysis, 1,292 idiopathic infertility patients aged 24 to 42 years old were included, and were divided into three subgroups: 268 infertility patients with non-obstructive azoospermia, 256 infertility patients with oligozoospermia (sperm counts < 20 × 10⁶/ml) and 768 infertility patients with normal count (sperm counts ≥ 20 × 10⁶/ml).

The control group included 480 fertile men ranging from 25 to 40 years of age who had fathered at least one child without assisted reproductive technologies and had normal semen parameters. The semen analysis for sperm concentration, motility and morphology was performed following the World Health Organization criteria [23].

SNP selection and genotyping

We selected the tagging SNPs by using genotype data obtained from unrelated Han Chinese individuals from Beijing in the HapMap project (HapMap Data Rel 24/Phase II Nov08, on NCBI B36 assembly, dbSNP b126). To examine the gene extensively, we searched the MMR genes, including 2,000 bp of the flanking regions both upstream and downstream of the gene, using the pairwise option of the Haploview 4.0 software (Mark Daly's Lab, Broad Institute of MIT and Harvard, Cambridge, Massachusetts, USA). The tagging SNPs were selected on the basis of pairwise linkage disequilibrium with a r² threshold of 0.8 and minor allele frequency ≥ 0.05 to capture all the common SNPs. In total, 19 SNPs were chosen in these 5 genes. In addition, a non-synonymous SNP (rs1799977) in MLH1 and a non-synonymous SNP (rs2075789) in MSH5 that cause missense mutations were included.

Genotyping was performed using TaqMan 7900HT Sequence Detection System and GenomeLab SNPstream high-throughput 12-plex genotyping platform (Beckman Coulter, Fullerton, CA, USA). Sequences of forward, reverse and extension primers are listed in Additional file 1 (Table S1). For quality control, the genotyping was done without knowledge of case/control status of the subjects, and a random 5% of cases and controls were genotyped twice by different individuals, and the reproducibility was 100%. To confirm the genotyping results, selected PCR-amplified DNA samples (n = 2, for each genotype) were examined by DNA sequencing and the results were also consistent.

DNA fragmentation analysis

After a period of 48 to 72 h of sexual abstinence, semen samples were collected by masturbation into wide-mouthed sterile containers and were delivered to the laboratory within 1 h of ejaculation. The diluted samples were cooled gradually at 5°C for 2 h, frozen at -70°C for Tdt-mediated dUTP nick end labelling (TUNEL) evaluation. A detailed protocol of the TUNEL assay for human sperm has been described previously [24]. TUNEL labeling was carried out using a Cell Death Detection kit (APO-DIRECT kit; BD Biosciences PharMingen, San Diego, CA, USA) according to the manufacturer's instructions. Briefly, semen samples were thawed in a 37°C water bath and immediately diluted with buffer (0.15 M NaCl, 0.01 M Tris, 0.001 M EDTA, pH 7.4) to obtain a sperm concentration of 1 to 2 × 10⁶/ml. Washed sperm was resuspended in 2% paraformaldehyde for 30 minutes at room temperature. After rinsing in PBS, samples were resuspended in permeabilization solution (0.2% Triton X-100, 0.1% sodium citrate) for 10 minutes on wet ice. TUNEL reagent (50 μl) was added to each sample. For each batch, a negative control lacking the terminal deoxynucleotidyl transferase and a positive control treated with DNase I were included to ensure assay specificity. After incubation for 1 h at 37°C, samples were analyzed immediately by flow cytometry (FACSCalibur; BD Biosciences Pharmingen, San Diego, CA, USA). Flow during the analysis was controlled at approximately 500 spermatozoa/sec, and 10,000 cells were analyzed for each sample. The percentage of FITC-positive cells (FL1 channel) was calculated as the percentage of cells with a fluorescence intensity exceeding the threshold obtained with the negative control.

Plasmid construction

To evaluate the potential effects of PMS2 rs1059060 (Ser775Asn) polymorphisms on the interaction between MLH1 and PMS2, fluorescence resonance energy transfer (FRET) technology and immunoprecipitation were performed. The cDNA encoding MLH1 or PMS2 was generated by PCR from a human testis cDNA library.

For the FRET assay, the primers used for amplifying PMS2 (amino acids 655-856) were 5'-CGTTAAGCTTGGAGAAAATCAAGCAGCCGAAG-3'/5'-ATACGGATCC CAGGTTGGCGATGTGTCTCAT -3', including HindIII and BamHI restriction sites (underlined sequences). Point mutations for PMS2 were performed using QuikChange Site-Directed Mutagenesis Kit (Stratagene, La. Jolla, CA, USA). The amplified fragment of PMS2 and its genetic variants were cloned into the pEYFP-C1 vector (Clonetech, Palo Alto, CA, USA). Similarly, the cDNA sequence encoding MLH1 (amino acids 506-756) was amplified by PCR using the following primers: 5'-CGTTGAATTCGTGTTTTGAGTCTCCAGGAAGAAA-3'/5'-ATACGGATCCACACCTCTCAAAGACTTTGTAT-3', which contain EcoRI and BamHI restriction sites (underlined sequences). This amplified fragment was ligated into pECYP-C1 vector (Clonetech, USA). For immunoprecipitation, the cloning of the full-length PMS2 and MLH1 cDNA constructs into pcDNA3.1 (Invitrogen, Carlsbad, CA, USA), between NheI and BamHI, has already been described [25]. The integrity of the inserts was confirmed by sequence analysis.

Cell culture and transfection

MutLα-deficient HEK293T cells were cultured in DMEM: F12 (1:1) (Gibco, Carlsbad, CA, USA), supplemented with 10% foetal bovine serum and 0.1% streptomycin/penicillin (Gibco, USA) in a humidified atmosphere with 5% CO₂ at 37°C. Cells were seeded onto 30 mm dishes with poly-L-lysine-coated glass coverslips and co-transfected with YFP recombinant plasmid (YFP-PMS2 or variants of YFP-PMS2) and CFP recombinant plasmid (CFP-MLH1) using Lipofectamine 2000 (Invitrogen) until the cells were at 50 to 60% confluence, according to the manufacturer's protocols. The transfection efficiency was compared by Western blotting at 72 hours after transfection using anti-PMS2 (A16-4) (1:100; BD Biosciences), anti-MLH1 (G168-728) (1:100; BD Biosciences), and anti-β actin (1:5000; Santa Cruz Biotechnology, CA, USA) antibodies.

Image analysis and calculation of fluorescence resonance energy transfer ratios

We used a Zeiss LSM710 confocal microscope (Carl Zeiss, Jena, Germany) operating with a 40 mW argon laser. Filter-cube specifications for the fluorescent channels were as follows for excitation and emission, respectively: enhanced cyan fluorescent protein (ECFP), 430 ± 25 and 470 ± 30 nm; enhanced yellow fluorescent protein (EYFP), 500 ± 20 and 535 ± 30 nm; and fluorescence resonance energy transfer (FRET), 430 ± 25 and 535 ± 30 nm.

Image analysis involved three basic operations: subtraction of background autofluorescence and blurred light, quantification of fluorescence intensity, and calculation of a corrected FRET (FRETc) by the following equation:

FRETc = (I _DA - a I _AA - d I _DD )/I _AA , where I _DA is the fluorescence intensity from the FRET filter set and I _DD and I _AA are the fluorescent intensities from ECFP (the donor) and EYFP (the acceptor), respectively. The cross-talk coefficients a and d were considered constant. The corrected FRET ratio was defined as FRETc/I_DD.

Co-Immunoprecipitation and Western blotting

Proteins were extracted from co-transfected HEK293T cells by the M-PER^® Mammalian Protein Extraction Reagent (Pierce Bio, Thermo, Rockford, IL, USA) according to the manufacturer's instruction. Approximately 200 μg total cell protein was transferred to a 1.5 ml microcentrifuge tube, and 20 μl of Protein A/G PLUS-Agarose (Santa Cruz Biotechnology, CA, USA) was added to the supernatant and the mixture was incubated at 4°C on a rocker platform for one hour. After this incubation, 2 μg anti-MLH1 N-20 (Santa Cruz Biotechnology, CA, USA) was added and incubated with shaking at 4°C overnight. The immunoprecipitates were collected by centrifugation at 1,000 × g for 5 minutes at 4°C, washed 4 times with lysis buffer and then the precipitates were collected for the Western blotting detection with the anti-PMS2 (A16-4) (1:100; BD Biosciences) antibody. Proteins were then detected with a Phototope-HRP Western Blot Detection kit (Cell Signalling Technology, Inc., Beverly, MA, USA).

Statistical analyses

Differences in select demographic variables, as well as smoking and alcohol status, between the cases and the controls, were evaluated using the χ² test. The Student's t test was used to evaluate continuous variables, including age and pack-years of cigarette smoking. The Hardy-Weinberg equilibrium was tested using a goodness-of-fit χ² test. We used unconditional multivariate logistic regression analysis to examine associations between genetic polymorphisms and male infertility risk by estimating ORs and 95% confidence intervals (95% CI). To reduce the potential for spurious findings due to multiple testing, we applied the False Discovery Rate (FDR) method to the P-values for the differences of genotype distributions among cases and controls. False Discovery Rate (FDR) is a new approach to the multiple comparisons problem. Instead of controlling the chance of any false positives (as Bonferroni methods do), FDR controls the expected proportion of false positives among suprathreshold voxels [26].

Sperm DNA fragmentation was normalized by natural logarithm (ln) transformation. Linear regression models were used to estimate the association with ln-transformed sperm fragmentation values for each SNP independently. Models were adjusted for age, smoking status, drinking status and abstinence time. All P-values presented are two-sided and all analyses were carried by the Statistical Analysis Software, version 9.1.3 (SAS Institute, Cary, NC, USA).

Subject characteristics

Allelic frequencies and genotype distributions of MMR polymorphisms

Association between MMR polymorphisms and sperm DNA fragmentation

Considering the importance of the MMR pathway in maintenance of DNA integrity, we further evaluated the effects of these three SNPs on sperm DNA fragmentation. In the present study, semen samples were pre-treated with cryopreservation prior to TUNEL analyses. However, it has been demonstrated that the process of cryopreservation can lead to an increase in oxidative stress and percentage DNA fragmentation [27]. To determine whether the results of the TUNEL analyses were profoundly influenced by cryopreservation in our study, 10 semen samples were pre-treated with or without cryopreservation prior to TUNEL analyses. As shown in Additional file 3 (Table S3), modest but significant elevated levels of sperm DNA fragmentation were induced by cryopreservation (P = 0.001). However, all the semen samples undergo the same cryopreservation process, thus we believe that the effect of cryopreservation, if any, is unlikely to be substantial. The non-normal distribution of sperm DNA damage levels and sperm concentration were natural log (ln) transformed for further association studies (skewness-kurtosis tests P > 0.05). After adjustment for age, smoking, alcohol use and length of abstinence, we found that subjects who carried the rs4647269 CT/TT genotypes displayed markedly higher levels of sperm DNA fragmentation compared with the CC homozygotes (mean ± S.D., 11.82% ± 2.66% vs. 26.58% ± 1.97%; P < 0.001) (Figure 1A). Moreover, a gradual increase in sperm DNA fragmentation was found among the three PMS2 rs1059060 subgroups (mean ± S.D., 12.30% ± 2.72%, 17.99% ± 2.27%, and 24.78% ± 1.70% for GG, GA and AA, respectively; P _trend < 0.001) (Figure 1B). However, no significant difference was observed for the MSH5 rs2075789 (Figure 1C).

Effects of the PMS2 Ser775Asn polymorphism on MLH1 and PMS2 interaction

The PMS2 Ser775Asn polymorphism (rs1059060) was potentially located within the MLH1-PMS2 interacting domain. Therefore, we examined whether PMS2 Ser775Asn polymorphisms influence binding between MLH1 and PMS2. HEK293T cells were transiently co-transfected with plasmids encoding the MLH1 (amino acids 506-756) and wild-type or genetic variants of PMS2 (amino acids 675-850). The schematic diagram of the FRET assay is summarized in Additional file 4 (Figure S1). By confocal fluorescence detection, we found that there was a weak interaction between MLH1 and PMS2-775Asn proteins, for little FRETc was detected in cells co-expressing CFP-MLH1 and YFP-PMS2-775Asn plasmid (Figure 2B). Cells co-expressing CFP-MLH1 + YFP-PMS2-775Ser had a four-fold increase in FRETc values (0.031 ± 0.013, n = 12; 0.008 ± 0.005, n = 12; P < 0.001) (Figure 2A). This result suggested that the PMS2 Ser775Asn polymorphism could significantly influence the interaction between MLH1 and PMS2.

We also used a co-immunoprecipitation assay to detect the effects of PMS2 variants on the MLH1 and PMS2 interaction. Full-length MLH1 and PMS2 775Ser or PMS2 775Asn were constructed and co-transfected into HEK293T cells. Western blot analysis of whole cell lysates showed satisfied transfection efficiency (Figure 3A). The co-immunoprecipitated result also suggested that binding between MLH1 and PMS2-775Ser was more robust compared with binding between MLH1 and the PMS2-775Asn variant (Figure 3B, lane 4 vs. 3) which suggested that mutation of PMS2 significantly attenuated the protein-protein interaction of MLH1 and PMS2.