Sperm mitochondria play an important role in spermatozoa because of the high ATP demand of these cells. Different mitochondrial DNA (mtDNA) mutations and haplogroups influence sperm function. The mtDNA dose also contributes to genetic variability and pathology in different tissues and organs, but nothing is known about its relevance in the performance of spermatozoa. We estimated the variability in mtDNA content within a population of men. Different mtDNA:nuclear DNA ratios were characteristic of progressive and nonprogressive spermatozoa, confirming the influence of mtDNA content on sperm functionality. We also estimated that the absolute content of mtDNA was 700 and 1200 mtDNA copies per cell in progressive and nonprogressive human spermatozoa, respectively. These results suggest that a marked increase of mtDNA copy number per cell volume takes place during spermatogenesis.
Sperm mitochondria play an important role in spermatozoa functionality; therefore, genetic alterations of mitochondrial DNA (mtDNA) may have consequences for normal fertilization. These genetic alterations could be pathological mutations or common mtDNA variants that only affect male fertility because mtDNA is maternally inherited [1, 2]. Thus, mutations responsible for encephalomyopathies decrease sperm motility [3, 4], and single  and multiple [6–8] mtDNA deletions have been associated with sperm dysfunction. However, results contradictory to the latter observations have been also reported [9, 10]. In addition, the cumulative effect of large scale deletions in total  or in a time-dependent manner associated with antiretroviral therapy  have also been associated with male infertility. Frequent mtDNA variants in Caucasians influence sperm function in such a way that the mtDNA haplogroup H accumulates significantly in nonasthenozoospermic men, and T mtDNA is associated with asthenozoospermic men . Specific alleles of the nucleus-encoded mtDNA polymerase also are associated with male infertility .
These observations lead to the question of how mtDNA population variability affects to sperm function within males [14, 15]. Oxidative phosphorylation activity is variable depending on the mitochondrial genetic background and is correlated with sperm motility [16, 17].
Mitochondrial DNA copy number contributes in great extent to mitochondrial genetic variability and pathology, and some changes occur within tissues and with aging [18–20]. Depletion of mtDNA is a severe pathological disorder characterized by extreme reduction of mtDNA content in a tissue-specific manner. This depletion has been repeatedly described in humans with mitochondrial disorders; it is present in different tissues and organs that have a high ATP demand, such as liver [21–25], skeletal muscle [26, 27], and nervous system [28, 29].
The copy number of mtDNA in human spermatozoa is not well established, and its variability in men is unknown. In addition, its effect on sperm function parameters has not been determined. The aim of this study was to estimate the human spermatozoon mtDNA content and to examine its population variability in relation to different parameters of seminal quality.
Materials and Methods
Patients and Control Participants
Semen samples from 440 donors from the areas of Madrid and Zaragoza (Spain) were collected by masturbation under hygienic conditions after a 3- to 5-day period of sexual abstinence. The samples were allowed to liquefy for 30 min at 37°C and were analyzed according to World Health Organization recommendations within a period of 2 h. The volume of the ejaculate was measured, and the number and percentage of motile spermatozoa was determined . A total of 73 semen samples were classified as normal, 255 as asthenozoospermic, 105 as astheno-oligozoospermic, and 7 as oligozoospermic.
Spermatozoa morphology was assessed by light microscopy, and the proportion of sperm cells with apparent head defects (large, small, or tapering heads), midpiece malformations, or tail abnormalities was estimated. Samples showing sperm cell agglutination, severe flagellar defects, or very high semen viscosity were discarded. All the samples utilized in this study were collected after consent of the donors and under the supervision of the Comité de Ética de Investigación Clínica (Universidad de Zaragoza) under approval number 575.
A human osteosarcoma-derived cell line 143B and the ρ° 206 cell line, which is without mtDNA , were respectively used as a control for a known amount of mtDNA and for the detection of cells with totally depleted mtDNA.
Seminal cells from semen samples donated by healthy men and showing ;lt10% nonsperm cells were fractionated into a discontinuous Percoll gradient. Most of the progressive spermatozoa were of normal morphology and high density and therefore sedimented preferentially into the 90% Percoll phase. Fewer progressive and most of the nonprogressive spermatozoa that showed abnormal morphology and were less dense sedimented into the 47% Percoll phase. Nonprogressive spermatozoa and part of the round cells sedimented into the 25% Percoll phase because of their lower density [5, 6]. Most round cells and residual bodies [32, 33] remained on top of the gradient within the seminal serum (this fraction referred to hereinafter as Rest). No more than 5% of round cells are expected to be present in the Percoll phases .
After the discontinuous Percoll gradient was divided into four fractions (90, 47, 25, and Rest), each one was diluted with the same volume of Dulbecco modified Eagle medium (DMEM) and centrifuged for 10 min at 600 × g at room temperature. To determine sperm motility, the pellet obtained from each phase was resuspended in 1 ml of DMEM.
Total DNA was extracted from 0.2 ml of total semen or cells fractionated by the Percoll gradient according to standard procedures  with some modifications. After centrifugation for 10 min at 600 × g at room temperature, seminal plasma was removed and the pellet was washed twice with isotonic saline solution (0.11 M sodium citrate and 0.11 M sodium acetate) and resuspended in 0.4 ml of spermatozoa medium (10 mM Tris-HCl pH 7.4, 10 mM EDTA, 2% SDS, 51 mM NaCl). Proteinase K to a final concentration of 150 μg/ml and 0.1 ml of 20 mM dithiothreitol in 10 mM sodium acetate were then added and incubated for 20 h at room temperature. After incubation, DNA was extracted by sequential addition of one volume of phenol-IAC (isoamilic-alcohol-chlorophorm), phenol/chloroform (1:1), and chloroform/diethyl ether (1:1). DNA in the aqueous phase was precipitated by addition of cold ethanol, recovered by centrifugation, and resuspended in sterile water.
DNA from 143B and ρ° 206 cell lines was prepared as previously reported .
To determine the mtDNA copy number, we amplified a 274-base pair (bp) fragment of the mitochondrial 16S rRNA gene and cloned it in the pGEM T easy vector. This fragment was amplified from preparations containing a known concentration of the plasmid under two different conditions: in the presence or the absence of a constant amount (0.5 μg) of total DNA extracted from mtDNA-depleted cells (ρ° cells). The primers used for polymerse chain reaction (PCR) amplification were localized at positions 1621–1643 (forward) and 1894–1874 (reverse) . Optimized PCR conditions were denaturation at 95°C for 45 sec, annealing at 64°C for 45 sec, and elongation at 72°C for 60 sec.
Quantitation of mtDNA was carried out by slot-blot hybridization (Micro-Sample Filtration Manifold; Schleicher and Schuell, Dassel, Germany) using 0.5 μg of total DNA from each semen sample, Percoll phase, or 143B cells. The filters were prehybridized in 6× saline-sodium citrate (SSC), 0.1% SDS, 10 mg/ml sperm salmon DNA, 1 μg/ml sodium pyrophosphate, and 5× Denhardt solution for 2 h and then hybridized in the same mix with a mitochondrial 32P-labeled probe for the mitochondrial 16S rRNA gene (nucleotides [nt]1621–1894 of the human mtDNA sequence). After autoradiography, the filter was stripped at 100°C for 15 min in 0.01 SSC, 1% SDS and hybridized again with a nuclear 32P-labeled probe for the nucleus-encoded 18S human rRNA gene (nt757–1188 of the gene sequence). Both probes were synthesized by PCR amplification. After hybridization, the filters were autoradiographed at −70°C using Hyperfilm (Amersham, Piscataway, NJ) and an intensifier screen. Filters were scanned with a laser LKB densitometer (ULTROSCAN XL) and Gelscan XL software (Pharmacia LKB Biotechnology, Uppsala, Sweden).
Calculation of mtDNA Copy Number
The mtDNA copy number was calculated for each healthy donor using a three-equation system:
where a, a′, and a″ indicate the decimal fraction of progressive spermatozoa, b, b′, and b″ indicate the decimal fraction of nonprogressive spermatozoa, and c, c′, and c″ indicate the decimal fraction of round cells from each 90%, 47%, and 25% Percoll fraction, respectively. Variables x, y, and z designate the mtDNA:nuclear DNA (nDNA) ratio of progressive spermatozoa, nonprogressive spermatozoa, and round cells, respectively.
Differences between means were evaluated by ANOVA and a Fisher protected least significant difference post hoc test. Correlations and linear regression analysis were performed with the Stat View 5.0 software (SAS Institute, Cary, NC).
Estimation of the mtDNA copy number in different cell types and tissues is generally achieved by slot-blot or Southern hybridization and the relative values normalized against the signal given by a nuclear gene probe. For sperm, this estimation is hampered by the presence of both haploid and diploid cells and by the very high variability in the concentration of the different cell types within samples. In addition, the low number of mitochondria and mtDNA molecules per sperm cell introduces additional complications.
The available data for mammalian species indicate very divergent values. In addition to natural variability, the use of different methodologies prevents any meaningful conclusions [38–40]. Thus, quantitative real-time PCR or conventional slot- or Southern-blot hybridization have been used by different researchers. In both cases, a plasmid containing a discrete mtDNA insert of known concentration was used to establish a quantitative reference for copy number estimation [39, 40].
Before starting our study, we tested the appropriateness of both experimental approaches. We amplified a 274-bp fragment of the mitochondrial 16S rRNA gene, cloned in pGEM T easy, from preparations containing a known concentration of the plasmid in either the presence or the absence of a constant amount of total cell DNA extracted from mtDNA-depleted cells (ρ° cells). Because the amount of amplification product was greatly influenced by the presence of nuclear DNA (Fig. 1), the PCR-based approaches were not considered suitable for our study. Therefore, the estimation of mtDNA content and the mtDNA:nDNA ratio of sperm cells were calculated by slot-blot hybridization using specific mitochondrial and nuclear rDNA probes. A representative experiment is shown in Figure 2.
The relative content of mtDNA (mtDNA:nDNA) was analyzed in semen samples of 440 individuals. To understand the basis of the high variability found among the samples, we examined the potential correlation of this ratio with the sperm cellular parameters (Fig. 3). A significant negative correlation was found with both percentage of progressive spermatozoa and sperm concentration. The percentage of round cells was positively correlated with the relative amount of mtDNA.
Given the complexity and the high variability of the cell contents in semen, accurate determination of the spermatozoa mtDNA amount is not possible. Therefore, to quantify the relative mtDNA content of the different major types of seminal cells, fresh sperm samples were fractionated in a 90%–47%–25% discontinuous Percoll gradient. We obtained four fractions (90%, 47%, 25%, and Rest) from seven healthy donors. The proportion of each different seminal cell type was different among phases (Fig. 4). The highest progressive spermatozoa content was found in the 90% fraction, and the nonprogressive spermatozoa and some round cells were found in the 47% and 25% fractions. The majority of round cells remained in the supernatant (Rest). The relative mtDNA content also differed among the Percoll fractions. The lower relative mtDNA content corresponded to the progressive sperm-enriched fraction, whereas the higher relative mtDNA content was found in the Rest fraction, where most of the round cells were concentrated (Fig. 4).
The mtDNA:nDNA ratio corresponding to different types of seminal cells was calculated. These values, expressed as mean ± SD, were 0.507 ± 0.279 in progressive spermatozoa, 0.903 ± 0.337 in nonprogressive spermatozoa, and 1.815 ± 0.970 in round cells (Fig. 5A). To transform these relative values into absolute mtDNA copy number, the human 143B cell line, whose mtDNA copy number has been established as 9100 , was used as a reference. Taking into account the different genomic nuclear content in each cellular type, the mtDNA copy number of the different seminal cells was determined. The values obtained were 717 ± 394 for progressive spermatozoa (haploid), 1278 ± 477 for nonprogressive spermatozoa (assuming that they are all haploid), and 3850 ± 2119 for round cells (haploid and diploid) (Fig. 5B). The average mtDNA copy number per cell in each semen sample was then recalculated taking into consideration its cell composition. As expected, the values thus calculated were strongly correlated (P = 0.03) with the experimental mtDNA:nDNA ratio found in the whole population.
The apparent correlations between the percentage of progressive spermatozoa and the spermatozoa concentration with the mtDNA copy number (Fig. 3) were reevaluated. A negative and significant correlation between motility and the mtDNA copy number was again observed even if total seminal cell or only spermatozoa DNA was considered (Fig. 6). However, the apparent correlation between sperm concentration and mtDNA copy number was entirely due to the contribution of round cell DNA to the total semen DNA (Fig. 7).
The results of this study allowed us to establish the mtDNA:nDNA ratio of progressive and nonprogressive spermatozoa. Consequently, by comparison with the absolute mtDNA copy number present in the 143B cell line  we could estimate the absolute mtDNA content of progressive (700 copies/cell) and nonprogressive (1200 copies/cell) spermatozoa.
Others have reported figures of 1300 , 75 , and 10  mtDNA copies/cell for human, bull, and mouse spermatozoa, respectively. For the mouse mtDNA, the amount of real-time PCR amplification product obtained with a cloned mtDNA fragment in the absence of nuclear genome was used to determine the absolute mtDNA copy number. Because the presence of the nuclear genome significantly reduces the yield of the PCR product (Fig. 1), these values could have been an underestimate of the absolute level of mtDNA. Another research group  used hybridization to detect a cloned mtDNA fragment as a reference to estimate the total mtDNA content in bull spermatozoa. The difference between their results and our results could reflect interspecies variability or technical differences. Nevertheless, our estimation of the mtDNA content in human spermatozoa (on average, 1000 mtDNA copies in both progressive and nonprogressive spermatozoa) is very close to the 1300 mtDNA copies/spermatozoon obtained by Manfredi et al.  using a strategy similar to ours. However, Manfredi et al. did not take into account the different contributions of each of the cellular types in the semen sample.
The estimation of seminal mtDNA content in a given sample could be interpreted as an index of the spermatogenic quality. Under normal circumstances, motile spermatozoa result from a correct spermatogenic maturation and generate semen with a high number of cells with a low mtDNA:nDNA ratio. In contrast, a disturbed process could generate high levels of nonprogressive cells or a low number of spermatozoa. Semen is enriched in cells with intermediate or high mtDNA:nDNA ratio. Therefore, an abnormally high level of mtDNA in a semen sample could reflect an impairment in spermatogenesis. Furthermore, a sample diagnosed as asthenozoospermic with a low relative mtDNA content could indicate a specific motility disorder not associated with general spermatogenic dysfunction.
Our results are in accord with the previously described downregulation of mtDNA copy number during rat  and mouse and human  spermatogenesis. During the process of spermatogenesis, the mtDNA copy number is reduced a 5- to 6-fold (3850 to 700). This decrease in mtDNA copy number is likely regulated by a mitochondrial transcription factor A, which is also downregulated in mice and humans [41–43]. This phenomenon contrasts with and can appear paradoxical to the fact that sperm motility is a highly ATP-demanding process that clearly depends on mitochondrial oxidative phosphorylation activity. Other tissues or cell types increase their mitochondrial number and mtDNA content with higher ATP demands [21, 22, 26, 27]. Closer examination of the profound effects of spermatogenesis is needed to clarify the apparent paradox.
We have estimated from data published elsewhere [44, 45] that there is a 40-fold reduction in cell volume (from 1500 μm3 to 40 μm3) during spermatogenesis in parallel with a 5- to 6-fold reduction in mtDNA copy number. As a consequence of the cellular volume reduction, a 7-fold increase in mtDNA copy number per cubic micrometer of cell is concurrent with the higher ATP requirement during spermatozoon maturation. Thus, despite the decline of mtDNA per cell, mtDNA copy number per cell volume increases as the energy requirement for the spermatozoon increases.
Despite the large number of samples analyzed, we have not found any individuals showing mtDNA depletion in their sperm. Although sperm mtDNA depletion probably would result in infertility, it is highly possible that an individual with such a depletion would suffer other, much more severe clinical signs associated with general mtDNA depletion. Such a patient probably would not be encountered at a reproduction clinic.
We thank Dr. Giuseppe Attardi for his generous gift of the 143B and ρ°-206 cells. We also thank S. Morales for his technical assistance. C.D. is the recipient of a License for Study from the Ministerio de Educación y Cultura of Spain.