Testis-Specific Cytochrome c-Null Mice Produce Functional Sperm but Undergo Early Testicular Atrophy

Abstract

Differentiating male germ cells express a testis-specific form of cytochrome c (Cyt c_T) that is distinct from the cytochrome c expressed in somatic cells (Cyt c_S). To examine the role of Cyt c_T in germ cells, we generated mice null for Cyt c_T. Homozygous Cyt c_T^−/− pups were statistically underrepresented (21%) but developed normally and were fertile. However, spermatozoa isolated from the cauda epididymis of Cyt c_T-null animals were less effective in fertilizing oocytes in vitro and contain reduced levels of ATP compared to wild-type sperm. Sperm from Cyt c_T-null mice contained a greater number of immotile spermatozoa than did samples from control mice, i.e., 53.1% ± 13.7% versus 33.2% ± 10.3% (P < 0.0001) for vas deferens sperm and 40.1% ± 9.6% versus 33.2% ± 7.5% (P = 0.0104) for epididymal sperm. Cyt c_T-null mice often exhibit early atrophy of the testes after 4 months of age, losing germ cells as a result of increased apoptosis. However, no difference in the activation of caspase-3, -8, or -9 was detected between the Cyt c_T^−/− testes and controls. Our data indicate that the Cyt c_T-null testes undergo early atrophy equivalent to that which occurs during aging as a consequence of a reduction in oxidative phosphorylation.

Cytochrome c, a heme-containing protein, resides in the space between the outer and inner membranes of mitochondria and transports electrons from complex III to complex IV in the respiratory chain. Mammalian testes express two types of cytochrome c, somatic cytochrome c (Cyt c_S) and testis-specific cytochrome c (Cyt c_T) (6, 12). Within the epithelium of seminiferous tubules, differentiation of the germ cells progresses towards the luminal surface. Spermatogonia, diploid germ cells which reside along the basement membrane, express only Cyt c_S. From the preleptotene to the pachytene stages of meiotic prophase, expression of Cyt c_S decreases while levels of Cyt c_T protein increase (13). Expression of Cyt c_T becomes dominant by the pachytene stage, and Cyt c_T is the sole cytochrome c expressed in all postmeiotic male germ cells, including mature spermatozoa. Cyt c_T localizes in the mitochondria in a manner similar to that of Cyt c_S.

Cytochrome c also plays an important role during apoptosis (27). Activated mitochondria release cytochrome c from the intermembrane space into the cytoplasm, and the released cytochrome c binds to apoptosis protease-activating factor 1 (Apaf-1) (33). The resulting cytochrome c-Apaf-1 complex can then activate caspase-9 (20), followed by activation of downstream death effectors such as caspase-3, caspase-6, and caspase-7 (31). In the testis, apoptosis is considered an important physiological mechanism that regulates the number of sperm produced (26), and it has also been suggested that apoptosis removes germ cells with damaged DNA (2). Fas is expressed in germ cells and the Fas ligand (FasL) is expressed in Sertoli cells. The Fas system is known to regulate apoptosis in the testis of young rats (16). Ectopic expression as well as inactivation of apoptosis-related genes have been shown to cause abnormalities in murine spermatogenesis. For example, accumulation of spermatogonia and degeneration of germ cells were observed in transgenic mice misexpressing Bcl-2 (5). Likewise, the testis of the Bax knockout (KO) mouse showed accumulation of atypical premeiotic germ cells and an absence of elongated spermatids (15), suggesting that the mitochondria-mediated cascade of apoptosis is important in spermatogenesis.

Apoproteins of both Cyt c_T and Cyt c_S are encoded in the nuclear genome; in the mouse, the Cyt c_T gene localizes to chromosome 2 and the Cyt c_S gene localizes to chromosome 6 (10). While the cDNAs of the two genes are 74% homologous, their genomic organization is quite different. The Cyt c_T gene contains four exons and three introns and spans a length of about 6 kb, while the Cyt c_S gene contains only one intron and is only about 1.4 kb long. Multiple pseudogenes of Cyt c_S exist in the genome of humans and mice (21); however, no pseudogene of Cyt c_T has been reported. In the testis, both Cyt c_T and Cyt c_S show alternative splicing mRNA variants, although the biological significance of this variability is unclear (8, 9). Cyt c_T and Cyt c_S both encode apoproteins consisting of 104 amino acids, which are about 86% identical. The carboxyl-terminal regions show the least similarity. Both Cyt c_T and Cyt c_S sequences contain a mitochondria anchoring site; however, it is not known whether the retention of Cyt c_T to the mitochondrial inner membrane is different from that of Cyt c_S. It is not clear whether Cyt c_T has functions distinct from those of Cyt c_S.

We have generated and characterized KO mice deficient in Cyt c_T. Our analysis reveals that mature spermatozoa expressing no cytochrome c retain motility and are capable of fertilization. However, the testes of mice lacking Cyt c_T undergo apoptosis and display early onset atrophy, likely due to a dysfunction in oxidative phosphorylation.

MATERIALS AND METHODS

Targeting vector.

A genomic fragment containing intron 2 and intron 3 was amplified by PCR from genomic DNA from the 129sv^+/+ mouse strain by using a 5′ primer from exon 1 (5′-GGC CCA GGG CAC GGC TGC TGT GAT TGT G-3′) and a 3′ primer from exon 3 (5′-CAA AGA TCT TCT TGC CTG CTT CAG CAT CTC CC-3′). The PCR was carried out with a combination of AmpliTaq (PE Applied Biosystems, Foster City, Calif.) and TaqExtender (Stratagene, La Jolla, Calif.) following the manufacturers' instructions. After 40 rounds of annealing and extension at 72°C for 12 min and denaturing at 94°C for 1.5 min, a 4.2-kb fragment was amplified. The 4.2-kb DNA was subcloned into pCR2.1 (Invitrogen, Carlsbad, Calif.) and sequenced. An intermediate KS plasmid vector (Stratagene) containing the pMCNeo poly(A) and the diphtheria toxin A cassette was cut with SpeI and NotI, and the 4.2-kb fragment was ligated to the 5′ end of the Neo cassette. To delete the second half of exon 3, which contains the heme binding site of Cyt c_T, a 2.7-kb KpnI-HindIII fragment was isolated from the 3.5-kb genomic DNA fragment and used as the right arm for homologous recombination, as shown in Fig. 1c.

FIG. 1.

Disruption of the Cyt c_T gene. (a) Amino acid sequences of Cyt c_S and Cyt c_T. Vertical lines indicate unmatched amino acids. Underlined amino acid sequences shown in this panel were used for the production of antibodies (anti-c_S antibody, carboxyl-terminal 16 amino acids; anti-c_T antibody, carboxyl-terminal 16 amino acids). (b) Genomic organization of the Cyt c_T gene. White boxes are exons and shadowed regions are introns. Translation starts at ATG in exon III. Probes used for Southern blot analysis are shown beneath the genomic DNA. (c) Targeting vector. (d) PCR screening performed for ES cell clones obtained from the vector. PCR using the 5′ primer shown in panel c and the 3′ primer shown in panel b amplified a predicted 3-kb fragment (left panel). EcoRI digestion of this fragment showed the predicted 1.8- and 1.2-kb fragments (right panel). D3, parental ES cell. 6 and 64 are ES clones used for generation of mouse colonies. (e) Southern blot analysis of Cyt c_T-targeted mice. The targeted allele in Cyt KO mice showed a 6.6-kb BamHI fragment with probe 360.

Production of KO mice.

Thirty micrograms of linearized targeting vectors was electroporated, and the transfected R1 embryonic stem (ES) cells (23) were processed using standard protocols (25). Fourteen out of 266 G418-resistant clones were confirmed as homologous recombinants by Southern blotting using five restriction enzymes (BamHI, EcoRV, EcoRI, HindIII, and PstI) and four probes (900, 530, 360, and a Neo probe [Fig. 1b]). Homologous recombination was also confirmed by PCR with a 3′ primer (5′-CTC CCA GCC TGA CGC TCC TTC CTC CTC-3′) from the 3′ region of the genomic gene (Fig. 1b) and a 5′ primer (5′-CGT GAC CCA TGG CGA TGC CTG CTT GCC G-3′) from the Neo cassette (Fig. 1c). The PCR conditions were 40 rounds of a cycle consisting of annealing and extension at 72°C for 15 min and denaturing at 94°C for 1.5 min. A combination of AmpliTaq (PE Applied Biosystems) and TaqExtender (Stratagene) was used to amplify the 3-kb fragment (Fig. 1d).

Homologous recombinant clones carrying normal karyotypes were injected into blastocysts of C57BL/6 mice (Charles River, Wilmington, Mass.) to produce chimeric animals. Chimeric males with high contributions of agouti coat color were bred to C57BL/6 females (Charles River). Germ line heterozygote mice, G₁, were bred to homozygosity. Most of the mating pairs were siblings, and homozygous animals were analyzed together with littermate controls. Results of the PCR of the ES clones used for production of germ line mice are shown in Fig. 1d. The restriction pattern of DNA derived from the tail of the Cyt c_T KO is shown in Fig. 1d (right panel). Two independent mouse colonies, derived from two independent ES clones, were maintained and their genotypes were determined by Southern blot hybridization (Fig. 1e). All male mice used in this study had been individually separated in single cages for at least 10 days to avoid influence from cage-mate males, such as suppression of testosterone production. All mouse experiments complied with all relevant federal guidelines and institutional policies.

Southern and Northern blot hybridizations.

Southern blot hybridization for genotyping mice and Northern blot hybridization for analyses of gene expression were performed using standard protocols (25). The following probes used for Northern blot hybridization were obtained by reverse transcription-PCR by using mouse testis RNA: Pgk2 (a 300-bp fragment amplified with primers provided by John McCarrey [Southwest Foundation for Biomedical Research, San Antonio, Tex.]), cyt b (a 945-bp reverse transcription-PCR fragment amplified with the 5′ primer 5′-CCC ATC CAA CAT TTC ATC ATG ATG A-3′ and 3′ primer 5′-CCA ATT CAG GTT AAG ATA AGT AGG T-3′), cytochrome c oxidase subunit 1 (COX1; a 374-bp fragment amplified with the 5′ primer 5′-GAG GGC ACC CAT GAA GTC ATT C-3′ and 3′ primer 5′-CCA TTT ACA CTA TGT TCC ATC A-3′), Apaf-1 (a 555-bp cDNA fragment amplified with the 5′ primer 5′-CTG CGC GTT CTC CTC AGA CGA CAG-3′ and 3′ primer 5′-GTC ACA GTA CTG GAT GGT GCT GTG ATG-3′), testicular androgen binding protein (a 552-bp fragment amplified with the 5′ primer 5′-GGC GAC TGC TTC TGC TGT TGC TAC TA-3′ and 3′ primer 5′-ATA CAG CCG TCC AGG GCA GGC ACG-3′). The COX3 cDNA probe (250 bp) was prepared from an IMAGE Consortium clone (no. 520010) purchased from the American Type Culture Collection (Manassas, Va.). The protamine 2 probe is a 150-bp fragment from the 3′ untranslated region of the mouse protamine 2 gene. The ldhc probe is a 267-bp partial fragment of the mouse lactate dehydrogenase (LDH)-C₄ cDNA.

Antibodies against Cyt c_T and Cyt c_S.

To generate Cyt c_T-specific and Cyt c_S-specific antibodies, rabbits were immunized with keyhole limpet hemocyanin-conjugated synthetic peptides of 16 amino acids derived from the carboxy terminal of Cyt c_T and Cyt c_S (Cyt c_T, SEREDLIKYLKQATSS; Cyt c_S, GERADLIAYLKKATNE). All antisera were affinity purified with each peptide attached to a solid phase. Peptide synthesis, immunization, and affinity purification were performed by Quality Controlled Biochemicals, Inc. (Hopkinton, Mass.).

Immunostaining, terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) staining, and electron microscopy.

Mouse tissues were fixed in 10% formalin in phosphate-buffered saline for 72 h and processed as previously described (24). The affinity-purified rabbit anti-c_S antibody (0.5 μg/ml) and anti-c_T antibody (4 μg/ml) were incubated for 60 min. For detection of caspase-3, 5 μg of goat anti-caspase-3 antibody per ml (recognizing the p20 subunit of caspase-3) (Santa Cruz Biotechnology, Santa Cruz, Calif.) was used. To confirm specificity of staining, 10 μg of blocking peptide for the anti-caspase-3 antibody (Santa Cruz Biotechnology) per ml was preincubated for 60 min on sections and added to the primary antibody. To detect caspase-9, 5 μg of rabbit antibody recognizing the p10 subunit of caspase-9 (Santa Cruz Biotechnology) per ml was used. Localization of these primary antibodies was detected by using a Vectastain ABC kit (Vector Laboratories, Burlingame, Calif.). Mayer's hematoxylin was used for the counterstain. Apoptotic fragmentation of DNA was detected using an ApopTag Peroxidase In Situ Apoptosis Detection kit (Intergen Company, Purchase, N.Y.). Testis samples were fixed for electron microscopy using 3.2% paraformaldehyde-2% glutaraldehyde in 0.2 M cacodylate buffer (pH 7.4) for 2 h and processed using standard methods.

Western blotting.

Tissue samples were homogenized in lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% NP-40, 0.1% sodium dodecyl sulfate (SDS), 1 mM phenylmethylsulfonyl fluoride, and protein inhibitor cocktail (50 μl/0.1 g of tissue; Sigma, St. Louis, Mo.). The protein concentration was determined with bicinchoninic reagent (Pierce, Rockford, Ill.). Sperm samples were dissolved in buffer containing 125 mM Tris-HCl (pH 6.8), 4% SDS, and 0.05% β-mercaptoethanol. SDS-polyacrylamide gel electrophoresis was performed under denaturing and reducing conditions. Peroxidase-labeled secondary antibodies were purchased from Santa Cruz Biotechnology. Bound peroxidase was detected using the ECL Plus system (Amersham Pharmacia Biotech, Little Chalfont, Buckinghamshire, England).

Sperm collection, in vitro fertilization, and motility assay.

Cauda epididymides and vasa deferentia were minced in 1.5 ml of TYH medium, a modified Krebs-Ringer bicarbonate solution (11) and incubated at 33°C in a 7% CO₂ incubator for 15 min. Tissue debris was removed with a 40-μm-pore-size nylon filter (BD Falcon, Franklin Lakes, N.J.), and samples were incubated at 33°C. The number of spermatozoa was determined using a Neubauer hemacytometer. One set of samples was diluted with TYH medium to count immotile spermatozoa, and another was diluted with 10% formalin-phosphate-buffered saline to determine total sperm number. The number of sperm acrosomes per 10⁻⁴ ml was counted six times to calculate the median value.

For in vitro fertilization, spermatozoa were isolated from 6.5-month-old male mice. Donor eggs were collected from female mice of C57/B6 × 129sv^+/+ F₁ hybrids purchased from Taconic (Germantown, N.Y.) and superovulated with progesterone and human chorionic gonadotropin. The numbers of fertilized eggs reaching the two-cell stage were determined by using a microscope.

For motility assays, plastic cell culture inserts with 3-μm pores (BD Falcon) were placed in wells of 12-well plates (Costar, Corning, N.J.). One milliliter of TYH medium with or without 0.05 mM iodoacetamide (Sigma) and/or 0.33 mM oxamic acid (Sigma) was added to each outer well. The 3-μm-pore-size inserts were filled with equal numbers of motile spermatozoa (4.5 × 10⁶ sperm/insert) isolated from epididymides and suspended in 250 μl of TYH medium containing the same reagents as the outer well. After 20 h of incubation at 33°C in a 7% CO₂ incubator, the number of spermatozoa that moved to the outer well was determined with a Neubauer hemacytometer. The number of sperm acrosomes per 10⁻³ ml was determined three times to calculate the mean.

Caspase activity assay.

Protein extracts from testes were measured for caspase activity using as substrates carbobenzoxy-Asp-Glu-Val-Asp-7-amino-4-trifluoromethyl coumarin (Z-DEVD-AFC) and carbobenzoxy-Ile-Glu-Thr-Asp-amino-4-trifluoromethyl coumarin (Z-IETD-AFC) (30). Testes extracts, containing 10 μg of total protein/μl, from eight pairs of homozygotes and their littermate controls were measured by release of AFC as emission at 505 nm upon excitation with 400 nm. Average V_max values from duplicated wells for each animal were compared.

ATP measurements.

Each testicle was homogenized in 1 ml of 3% trichloroacetic acid (TCA)-2 mM EDTA, and spermatozoa collected from epididymides and vas deferentia from each animal were resuspended in 70 μl of 3% TCA-2 mM EDTA (22). After incubation for 15 min at room temperature and centrifugation at 16,000 × g, the ATP content in the supernatants was measured by using a Bioluminescent Somatic Cell Assay kit (Sigma) (28).

RESULTS AND DISCUSSION

Inactivation of the cytochrome c_T gene.

The construction of the targeting vector, production of targeted mice, and results of the Southern blot analysis of the targeted allele are presented in Fig. 1 and were described in Materials and Methods. Northern blot analysis of testis RNA revealed that no mRNA of Cyt c_T was detected in the testis of Cyt c_T^−/− mice, and Cyt c_T^+/− mice expressed Cyt c_T mRNA at half the level of wild-type controls (Fig. 2A ). The expression of Cyt c_S was unchanged in the testes of Cyt c_T KO mice (Fig. 2A).

FIG. 2.

Confirmation of Cytc_T inactivation. (A) Northern blot analysis of the following fragments: c_T, partial cDNA probe corresponding to a DNA fragment between BglII in exon III and BglII in exon IV of the Cyt c_T shown in Fig. 1b; c_T entire cDNA, c_T probe using entire Cyt c_T cDNA; c_S, c_S probe using entire Cyt c_S cDNA; Neo, Neo gene probe; L32, ribosomal protein L32-A4, used as a standard. In the Cyt c_T^−/− mice, the Cyt c_T probe gave no signal (arrowhead); however, the entire Cyt c_T cDNA probe detected a faint shorter signal (thin arrowhead), suggesting that there was transcription until the stop codon in exon III. The expression pattern of Cyt c_S in the Cyt c_T^−/− testis was normal and not distinguishable from Cyt c_S expression in heterozygotes and wild-type mice. Samples are 10 μg of total RNA from testes. (B) Western blotting using anti-Cyt c_S and Cyt c_T antibodies. Sperm cells were isolated from wild-type and Cyt c_T^−/− mice (2 × 10⁵ sperm/lane); liver extract was from wild-type mice (100 μg/lane); testis extracts were from mice (30 μg of protein/lane). Sperm cells were purified on a Percoll gradient (Amersham Pharmacia, Uppsala, Sweden) and washed in 0.45% NaCl hypotonic solution to remove contaminating somatic cells. (a) Black arrowheads indicate the predicted 12-kDa Cyt c_T or Cyt c_S protein. (b) Enriched population of mitochondria extracted from testes of wild-type, heterozygous, and Cyt c_T KO mice (12.5 μg of protein/lane). (C) Fertility and weight of testes. (a) Breeding record of Cyt c_T KO homozygous males. Litter size is the mean number of pups in each litter. (b) Average weight of testes from each individual mouse. Wild-type (+/+), n = 50, mean = 110.8 ± 17.1 mg; heterozygote (+/−), n = 21, mean = 113.0 ± 15.7 mg; homozygote (−/−), n = 60, mean = 102.9 ± 22.2 mg. Homozygous animals and their littermate controls, either wild type or heterozygotes, were analyzed at the same time. The ages of the mice subjected to analyses were 3 to 19 months. (D) Immunohistochemistry and morphology of the Cyt c_T^−/− testis. Panels on the right (b, d, f, g, i, and j) show Cyt c_T^−/− sections, and panels on the left (a, c, e, and h) show wild-type control sections. (a to d) Immunohistochemistry of testes fixed with Bouin's and embedded in paraffin. Panels a and b were stained with anti-c_T antibody; no Cyt c_T protein was seen in spermatocytes and spermatids of the Cyt c_T^−/− testes. Panels c and d were stained with anti-c_S antibody showing expression of Cyt c_S in spermatogonia, Sertoli cells, and Leydig cells in both +/+ and −/− mice. The mice are 3 months old. Bar, 100 μm. (e to j) Hematoxylin-eosin staining of testis and epididymis (ep; square insets in panels e and f) fixed with formalin and embedded in OCT compound. (e to g) Three-month-old mice. Testis weight of Cyt c_T^−/− mice (f) was the same as control (e) weights. At this age, homozygote testes appeared to be similar to controls except for the unusual basophilic structures in the luminal region of the seminiferous tubules and epididymis (arrowheads). Panel g is an enlarged view of the basophilic structure in homozygote testis (arrowheads). (h to j) Seven-month-old mice. At this age, testis weight of Cyt c_T^−/− mice (i) was approximately 27% of controls (h). Seminiferous tubules showed a dramatic loss of spermatocytes, spermatids, and spermatozoa, although some seminiferous tubules in the same testis still contain mature spermatozoa (i). Spermatogonia, Sertoli cells, and Leydig cells appeared unchanged in Cyt c_T^−/− and wild-type mice (i and j). Bars: e, f, h, and i, 100 μm; j, 10 μm. L, Leydig cell; St, Sertoli cell; Sg, spermatogonia. (k to m) Hematoxylin-eosin staining of testes of 17- to 19-month-old wild-type mice. (k) Testis of a 17-month-old mouse, showing marked loss of germ cells. (l) Testis of a 19-month-old mouse with reduced germ cells and basophilic structures (arrowheads). (m) High magnification of the basophilic structure from the same testis as shown in panel l. Bars: k and l, 100 μm; m, 10 μm.

To confirm the inactivation of the Cyt c_T gene at the protein level, we produced, characterized, and used peptide-specific antibodies in Western blotting and immunohistochemistry. No Cyt c_T protein (12 kDa) was detected in protein extracts from either whole testes or enriched testicular mitochondrial fractions of Cyt c_T^−/− mice (Fig. 2B). Expression of endogenous Cyt c_T in Cyt c_T-heterozygous mice was lower than that seen in wild-type controls (Fig. 2B, panel b). The expression levels of Cyt c_S in the Cyt c_T KO testes were not different from those in wild-type controls. Figure 2B, panel b, indicates that no Cyt c_T protein was expressed in either the testis or the sperm of Cyt c_T^−/− mice and that Cyt c_S levels were unchanged in Cyt c_T KO samples (Fig. 2B).

Ratio, fertility, and testicular weight of the homozygous mice.

While both male and female Cyt c_T KO mice were fertile, the ratio of homozygous animals was less than 25%, suggesting that the targeted allele does not transmit in a Mendelian fashion. Thus, of 434 offspring from Cyt c_T heterozygote parents, the genotype frequencies were 133 Cyt c_T^+/+ (30.6%), 209 Cyt c_T^+/− (48.2%), and 92 Cyt c_T^−/− (21.1%). Nevertheless, all homozygous males tested were able to produce offspring (Fig. 2C, panel a).

We examined whether the testes from Cyt c_T^−/− mice were of normal size by plotting the average wet tissue weight of each testicle for each individual mouse (Fig. 2C, panel b). The mean values were as follows: Cyt c_T^+/+, 110.8 ± 17.1 mg (n = 50); Cyt c_T^+/−, 113.0 ± 15.7 mg (n = 21); Cyt c_T^−/−, 102.9 ± 22.2 mg (n = 60). Several Cyt c_T KO testes were small in size and low in weight, as reflected in the larger standard deviation for the Cyt c_T^−/− group. We considered a Cyt c_T KO mouse testis weighing as much as a wild-type mouse testis to be a normal-sized testis, and a testis weighing less than 70 mg or less than 80% of the littermate control weight was considered a small-sized testis. Nine of 60 testes from Cyt c_T KO mice were small (15%) (Fig. 2C, panel b).

Lack of Cyt c_T and abnormal morphology in the testes.

No Cyt c_T protein was detected in the seminiferous tubules of Cyt c_T KO mice by immunohistochemistry using the anti-c_T antibody, while germ cells after the pachytene spermatocyte stage in control testes stained positive for Cyt c_T (Fig. 2D, panel b). In contrast, the anti-Cyt c_S antibody detected Cyt c_S in spermatogonia, Sertoli cells, and Leydig cells of KO and control testes, and the expression pattern of Cyt c_S remained unchanged in the Cyt c_T KO testis (Fig. 2D, panels c and d).

Histological examination of the testes of 2- to 3-month-old Cyt c_T KO mice showed the presence of abnormal basophilic structures near the luminal region of the seminiferous tubules (Fig. 2D, panels f and g). Older (4- to 10-month-old) Cyt c_T KO mice tended to have small testes. Examination of these small testes revealed that they were highly atrophic and had a reduced number of spermatocytes, spermatids, and spermatozoa (Fig. 2D, panel i). The spermatogonia, Sertoli cells, and Leydig cells appeared normal, and some seminiferous tubules still contained mature spermatozoa (Fig. 2D, panels i and j).

Loss of germ cells from the seminiferous tubules was also observed in normal testes of aged wild-type animals (Fig. 2D, panels k and l). In these older control mice, we observed basophilic round structures that were indistinguishable from those structures observed in the Cyt c_T^−/− testis (Fig. 2D, panels l and m). However, in the case of wild-type mice, such basophilic structures were recognized only in animals older than 17 months, while these structures were clearly evident in the testes of 6-week-old Cyt c_T^−/− mice.

Evidence of accelerated apoptosis in homozygous Cyt c_T KO testes.

To examine if the loss of germ cells in Cyt c_T^−/− testes occurs as a result of apoptosis, we used TUNEL staining to analyze DNA fragmentation. The seminiferous tubules of small testes from Cyt c_T KO mice contained a substantial number of pachytene stage spermatocytes undergoing apoptosis (Fig. 3b and c). Electron microscopy confirmed the presence of apoptotic chromatin condensation in the smaller testes of Cyt c_T KO animals (Fig. 3d). In contrast, no differences were seen in 2- to 3-month-old Cyt c_T KO mouse testes compared to littermate controls (data not shown).

FIG. 3.

Apoptosis in the Cyt c_T^−/− testis. (a to c) TUNEL staining of testes from 4.5-month-old mice; (b) Cyt c_T^−/− mouse; (a) Cyt c_T^+/+ littermate control of animal shown in panel b (weight of the −/− testis was approximately 71% of control); (c) high magnification of image shown in panel b. Arrowheads in panel c indicate positive cells. (d) Electron microscopic view of apoptotic cells in testis of 5.5-month-old Cyt c_T^−/− mouse (weight of −/− testis was approximately 53% of control). (e to g) Immunostaining with anti-caspase-3 antibody on the same testis as shown in panel b. (f) Control inhibition of immunoreaction by a specific blocking peptide for the epitope; (g) high magnification of panel e. (h and i) Immunostaining with anti-caspase-9 antibody on testes from the same animals as shown in panels a and b. (j) Staining with normal rabbit immunoglobulin G (4 μg/ml) using same testis samples as shown in panels b, e, and i. Ap, apoptotic cell; P, pachytene spermatocyte; L, Leydig cell. Bars: a, b, e, and f, 100 μm; c, g, h, i, and j, 20 μm; d, 2.5 μm.

Since release of cytochrome c from mitochondria is required for the activation of downstream caspases in the mitochondria-mediated apoptotic cascade, we analyzed the expression of two key proteases, caspase-3 and caspase-9. Using an antibody against the p20 subunit of caspase-3, we found positive staining in Leydig cells and in apoptotic cells in the Cyt c_T KO testes (Fig. 3e, f, and g). This reaction was inhibited in the presence of the caspase-3 peptide used to raise the antibody, indicating the specificity of the signal (Fig. 3e and f). TUNEL staining and caspase-3 were detected in the same seminiferous tubule in the same testis (Fig. 3c and g). Immunostaining with an antibody against the p10 subunit of caspase-9 showed that pachytene spermatocytes in wild-type testes stained positive, as did apoptotic cells in the Cyt c_T KO testes (Fig. 3h and i). Western blotting experiments showed that there is activation of caspase-3 and caspase-9 in wild-type testis, but that their activation is not accelerated in the Cyt c_T-null testes (data not shown).

The main extrinsic apoptotic pathway operating in the testis is the Fas-FasL system (16). Fas is expressed in spermatogenic cells, and FasL is expressed in Sertoli cells of 4-week-old rats. Interactions between the germ cells and Sertoli cells through the Fas-FasL system serve as an important regulator of apoptosis in the testis (17) via caspase-8 activation. However, we observed no induction of Fas protein or caspase-8 in our apoptotic Cyt c_T-null testis. Caspase activity assays using Z-DEVD-AFC, a substrate preferred by caspase-3 and -7, and Z-IETD-AFC, a substrate preferred by caspase-8, yielded no significant differences in activity between wild-type and cyt c_T^−/− testes. We have also examined mRNA levels of apoptosis-related molecules using the GEArray system (SuperArray, Bethesda, Md.). Neither the membrane containing the cDNAs associated with mouse apoptosis nor the signal transduction pathways showed significant differences when hybridized with probes from homozygote testes and littermate controls. These cDNA arrays include the tumor necrosis factor (TNF) ligand family, TNF receptor family, and caspase family. These additional data are consistent with the Western blotting and caspase activity assays that showed no evidence of enhanced activation of caspase-8 in the Cyt c_T^−/− testes. We found no significant difference in the expression levels of other apoptosis-related molecules, including apoptosis-inducing factor, Apaf-1, Nip3 (a cell death-inducing protein) (3), and Fas, as shown in Fig. 4B, panel a.

FIG. 4.

Expression of testis-specific proteins and apoptosis-related proteins. (A) Northern blot hybridization of testis RNA. (a) Analysis of testis-specific genes. The first pair of samples are from normal-sized testes of Cyt c_T^−/− and control (+/+) mice. The second pair of samples are from small testes of Cyt c_T^−/− and control (+/+) mice. Weight of the small testis was approximately 27% of its control littermate. (b) Analysis of genes related to oxidative phosphorylation or apoptosis. The first pair of samples are testes from a Cyt c_T^−/− mouse and his littermate control (both testes were approximately the same weight). For the second pair of samples the weight of the Cyt c_T^−/− testis was approximately 60% of the littermate control value. D3 is RNA from D3 ES cells and was included as a positive control for Apaf-1. (B) Western blotting of protein extracts from testes and sperm. (a and b) Testis extracts from a Cyt c_T^−/− mouse (weight of −/− testis was approximately 75% of the +/+ control), with antibodies against apoptosis-related proteins (AIF, Apaf-1, Nip3, and Fas) (a) or antibodies to cytochrome b, transferrin, and vimentin (b); 100 μg of protein/lane. (c) Western blotting with anti-COX1 antibody on testis extracts from aged wild-type mice (14, 17, 18, or 19 months old) and Cyt c_T^−/− mice (weight of −/− testis was approximately 75% of the +/+ control), and control and Cyt c_T^−/− mice with testes of comparable weight; 100 μg of protein/lane. (d) Sperm cells isolated from epididymides of Cyt c_T^−/− and control mice with anti-c_T and anti-COX1 antibody. Extracts are of 3.3 × 10⁵ sperm/lane.

Are there changes in the glycolytic and/or oxidative phosphorylation pathways?

Normal expression of two meiotic and one postmeiotic marker, pgk2 (phosphoglycerate kinase 2), ldhc (lactate dehydrogenase c), and prot2 (protamine 2), was seen in the RNA from a normal-sized testis of a 6-month-old Cyt c_T mouse. PGK-2 (EC 2.7.2.3) and LDH-C₄ (EC 1.1.1.27) are glycolytic enzymes, and the fact that their mRNA levels remained unchanged indicates that the disruption of oxidative phosphorylation resulting from inactivation of the Cyt c_T gene does not accelerate the synthesis of glycolytic enzymes in normal-sized Cyt c_T KO testes. A small testis of a 7-month-old Cyt c_T KO mouse, weighing only 24% of the control, showed reduced expression of pgk2, ldhc, and prot2, while levels of Cyt c_S mRNA were increased. However, these changes are due to the reduction in the number of germ cells and the altered ratio of germ cells to Sertoli cells in these atrophic testes (Fig. 2D, panel i), since markers of Sertoli cells, i.e., testicular androgen binding protein and vimentin, also appeared to increase (Fig. 4A and B, panels b). The level of transferrin, another marker of Sertoli cells, was not increased (Fig. 4B, panel b), but this most likely reflects a down-regulation of transferrin expression caused by the loss of germ cells (29).

To determine whether the deletion of Cyt c_T alters the expression levels of molecules that participate in oxidative phosphorylation, we analyzed the expression of three mitochondrial genes. Cytochrome b (cyt b) is a component of complex III in the electron transport chain, while COX1 and COX3 are components of complex IV of the electron transport chain. Northern blot analysis revealed no changes in the mRNA levels of either cyt b, COX1, or COX3 in small testes (Fig. 4A, panel b). However, Western blotting with protein extracts from small (4-month-old) or normal-sized (3-month-old) testes of Cyt c_T KO mice and from spermatozoa isolated from Cyt c_T KO mice revealed a significant reduction of COX1 protein (Fig. 4B, panels c and d). This suggests that the synthesis or stability of the COX1 protein is reduced in the Cyt c_T^−/− germ cells and spermatozoa. Similar changes in COX1 expression were also observed in the testes of aged wild-type mice (14 to 19 months old), as shown in Fig. 4B, panel c.

It has been reported that aging damages complex III in oxidative phosphorylation in the human heart and decreases the heart's tolerance for ischemia and reperfusion (18). Aging mammals often show testicular atrophy with loss of germ cells (7). We found corresponding basophilic structures in the luminal surface of the seminiferous tubules and detected a reduction in COX1 mRNA levels in both aging normal testis and the atrophic Cyt c_T-null testis from young mice. Thus, the Cyt c_T^−/− testis appears to mimic a dysfunction in oxidative phosphorylation which normally occurs with aging but which in the Cyt c_T-null testis results in early testicular atrophy.

Cyt c_T^−/− sperm show reduced motility, ATP concentrations, and fertility.

To examine if the production of spermatozoa by the Cyt c_T^−/− testis is normal, we studied the number and motility of Cyt c_T^−/− spermatozoa. Isolating sperm from the vas deferens, we found an average of (0.71 ± 0.4) × 10⁷ Cyt c_T^−/− sperm, compared to (1.1 ± 0.36) × 10⁷ for control samples (P = 0.0014) (Fig. 5A, left). Furthermore, the Cyt c_T^−/− samples contained 53.1% ± 13.7% immotile sperm, while the control samples contained 33.2% ± 10.3% immotile sperm (P < 0.001). Similar sperm counts were obtained from the epididymis (Fig. 5A, right), in agreement with the results obtained for the vas deferens sperm; i.e., we found (3.6 ± 1.2) × 10⁷ sperm for the Cyt c_T^−/− samples, compared to (5.2 ±2.4) × 10⁷ sperm for the controls (P = 0.0104). The epididymal Cyt c_T^−/− samples had 40.1% ± 9.6% immotile sperm, while only 33.2% ± 7.5% immotile sperm were seen in the control samples (P = 0.0146). A total of 23 Cyt c_T^−/− mice and a total of 20 control mice were examined for these experiments (age range, 6 to 15 months; average, 10.5 months).

FIG. 5.

Reduced motility of Cyt c_T^−/− spermatozoa. (A) Numbers of spermatozoa collected from vas deferens (left graph) and cauda epididymis (right graph) from a single mouse. White bars indicate total spermatozoa. Control samples (cntrl) consist of 16 wild-type and four heterozygous mice (n = 20; mean = [1.1 ± 1] × 10⁷ per mouse) while homozygote samples were from 23 mice (mean = [0.71 ± 0.40] × 10⁷ per mouse). Bars with hatch marks show immotile spermatozoa in each sample, and percent numbers within the hatched bar show the ratio of immotile spermatozoa in the total sperm count. (B) Ratio of ATP levels in testes and isolated spermatozoa of Cyt c_T^−/− mice relative to values of littermate wild-type controls. The average ATP level in sperm from Cyt c_T^−/− mice was approximately 60% of control, while the average ATP level in testis was approximately the same as in controls (about 100%). (C) Motility assay of spermatozoa isolated from epididymis. Total numbers of spermatozoa passing through 3-μm pores in media with or without 0.33 mM oxamic acid (OX) were determined. The same assay was conducted with or without 0.05 mM iodoacetamide (IA).

We then measured and compared the ATP content in the testes and sperm samples from seven pairs of Cyt c_T^−/− mice and matched controls (3 to 19 months old). As shown in Fig. 5B, ATP content per single sperm in homozygotes was 60% ± 14.8% of the controls, while ATP content per testis weight (in milligrams) in homozygotes was 103% ± 42.7% of the controls.

Next, sperm motility was quantified in the presence or absence of agents that are known to inhibit glycolysis, such as oxamic acid, an inhibitor of LDHs, and iodoacetamide, an inhibitor of glyceraldehyde-3-phosphate dehydrogenase (EC 1.2.1.12). The Cyt c_T KO samples contained reduced numbers of highly motile sperm with and without oxamic acid (approximately 73%) (Fig. 5C). In the presence of 0.05 mM iodoacetamide but without oxamic acid, both homozygote and control sperm were less motile (approximately 56%); however, in the presence of both iodoacetamide and oxamic acid, wild-type sperm still contained 48.8% ± 1.2% motile sperm, while Cyt c_T KO sperm contained only 39.9% ± 0.6% motile sperm (P = 0.012).

To determine whether Cyt c_T KO spermatozoa are capable of fertilizing normal eggs in vitro as efficiently as wild-type spermatozoa, we used an in vitro fertilization assay. We found that out of 180 oocytes incubated with spermatozoa isolated from a 6.5-month-old Cyt c_T KO mouse, 14.5% ± 6.1% developed to the two-cell stage. In contrast, 60.5% ± 15.6% (n = 2; 157 out of 263 eggs) of the oocytes that were incubated with spermatozoa from a littermate wild-type mouse developed to the two-cell stage.

It has been suggested that deficiencies in the respiratory chain could affect spermatogenesis, producing nonfunctional sperm (4). However, a causal relationship between a dysfunction of the mitochondria in germ cells and male infertility is still unclear (1). Our data indicate that oxidative phosphorylation is not essential to support sperm motility and fertility under conditions where fructose, lactic acid, and/or glucose are available to maintain glycolysis.

Concluding remarks.

Cytochrome c is an essential protein in the electron transport chain functioning during oxidative phosphorylation. Since Cyt c_T is highly expressed in meiotic prophase during spermatogenesis and is the only cytochrome c expressed in mature spermatozoa, we predicted severe consequences to the function of the spermatozoa upon inactivation of Cyt c_T. Contrary to our expectations, Cyt c_T-null mice are fertile, even though a high percentage of their spermatozoa are immotile. Other investigators have shown that mice lacking Cyt c_S die in utero, and cells derived from these embryos are viable only under conditions that compensate for the resulting dysfunction in oxidative phosphorylation (19). The authors reported that Cyt c_S^−/− cells showed reduced caspase-3 activation, were resistant to proapoptotic stimuli, and became more susceptible to the extrinsic TNF-α-mediated pathway of apoptosis. Spermatogenesis in our Cyt c_T-null mice was ostensibly unaffected in young mice, although their germ cells underwent apoptosis in an age-related fashion. The fact that removal of cytochrome c accelerates apoptosis seems paradoxical. Cyt c_S in germ cells declines as the cells enter meiosis, with pachytene spermatocytes containing low levels of Cyt c_S at an approximate Cyt c_S to Cyt c_T ratio of 1:4 (13). One would have expected that a compensatory Cyt c_S induction in the Cyt c_T^−/− pachytene cells might explain the accelerated apoptosis. However, our data show no evidence of such a compensatory induction of Cyt c_S. Thus, it seems that Cyt c_T may play a role distinct from that of Cyt c_S in germ cell mitochondria. It has been reported that cells lacking mitochondrial DNA expression and lacking a functional respiratory chain undergo apoptosis (14, 32). The Cyt c_T^−/− testis appears to mimic a dysfunction in oxidative phosphorylation which normally occurs with aging.