INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The cAMP-dependent protein kinase (PKA)¹ is a major enzyme in cellular signal transduction and is thought to mediate most of the physiological responses to cAMP in eukaryotic cells (1). Below a cAMP threshold concentration, PKA exists as an inactive tetramer of two catalytic and two regulatory subunits (CR₂C). The two R subunits form a dimer with each protomer attaching to the substrate-binding site of a C subunit. Some isoforms of R also associate with binding proteins collectively termed PKA anchoring proteins; it is believed that through these interactions PKA is targeted to specific subcellular compartments (for reviews, see Refs. 2-4). Activation of adenylate cyclase by extracellular signals raises the intracellular concentration of cAMP, and at a certain threshold concentration cAMP binds to the R subunits of the PKA tetramer, releasing C to phosphorylate its substrates (for reviews, see Refs. 1 and 5-8).

Cyclic AMP-dependent signaling has an important role in the control of sperm movement. Mammalian sperm are nonmotile in the testis, but as they pass through the epididymis they acquire the capacity for motility. This process is known as "epididymal maturation" and is essential for the sperm to fertilize an egg (9). Several types of studies have shown that changes in sperm cAMP levels are involved in epididymal maturation (see Refs. 10-12 for reviews). The most direct evidence for a role of cAMP in the acquisition of the capacity for motility has come from studies of demembranated, reactivated sperm (13-15). When caudal epididymal or ejaculated sperm were demembranated by treatment with nonionic detergents, and then placed in an appropriate solution containing ATP, they were reactivated with a waveform very similar to that of intact ejaculated sperm. In contrast, under the same conditions demembranated testicular sperm exhibited very poor motility. However, if cAMP was added to the reactivation medium, the demembranated testicular sperm began to beat with a motility similar to that of the mature sperm models. Cyclic AMP-dependent motility also could be demonstrated in the demembranated ejaculated sperm if the sperm were metabolically inhibited to reduce their motility prior to demembranation (16, 17). The requirement for cAMP could be bypassed by addition of exogenous C to the reactivation solution, confirming that cAMP was acting via PKA (17). Therefore, cAMP-dependent phosphorylation is critical for sperm motility.

Sperm C (C_s) appears to have unusual solubility properties. It is generally accepted that, in the presence of cAMP, C is soluble in the cytoplasm (1, 18). However, in a previous study we observed that neither detergent nor cAMP alone released C activity from demembranated sperm, but that cAMP plus nonionic detergent did release the activity (17). These results suggested that C_s is bound to internal sperm structures by two types of bonds, one sensitive to detergent and one sensitive to cAMP.

There are at least three different isoforms of mammalian C, C, C, and C, which are the products of different genes. C occurs in a wide variety of tissues (19); C also is widely distributed but is expressed in lesser amounts than C in most tissues except the brain (19, 20). C appears to be expressed only in testis (21). C appears to be the predominant isoform in mammalian germ cells (22). Splice variants of both C and C also are known (23-25).

We now have further characterized C_s and investigated its relationship to the previously known isoforms of C. We found that C_s is the major, and perhaps the only, PKA catalytic subunit in sperm. We isolated C_s from ram sperm, which could be obtained in amounts sufficient for protein biochemistry. Purification also was aided by the solubility properties of the subunit, which made it possible to obtain a homogeneous preparation of C_s by a simple two-step protocol. Mass spectrometry (MS) revealed that the mass of C_s was ~890 Da less than that of C from ovine striated muscle. The amino terminus of C_s was blocked, but partial amino acid sequence from internal tryptic and CNBr fragments was an exact match to the sequence of bovine C. C_s clearly differed in sequence from the C and C isoforms. An amino-terminal endoproteinase lysine-C fragment was isolated from C_s and sequenced by a combination of tandem mass spectrometry (MS/MS) and Edman degradation of an endoproteinase aspartate-N subfragment. The results revealed that the amino-terminal 14 amino acids and amino-terminal myristate of C are replaced by six different amino acids and an amino-terminal acetate in C_s. The predicted mass difference due to this replacement is 899 Da, in excellent agreement with the MS data. The region of homology between C_s and C begins at exactly the exon 1/exon 2 boundary in C, suggesting that C_s results from the use of an alternate exon 1 in the C gene. The shorter amino terminus and lack of a myristate moiety may expose a hydrophobic portion of the catalytic core of C_s, allowing C_s to bind to a hydrophobic site within the flagellum and thus accounting for the unusual solubility properties of the subunit.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Materials used and their sources were: 4-aminobenzamidine, cAMP, CNBr, 3,5-dimethoxy-4-hydroxycinnamic acid, and NTCB from Aldrich; ¹⁴C-methylated protein molecular weight markers were from Amersham; Aquapore RP-300 C₈ column (0.1 × 25 cm), Aquapore OD-300 C₁₈ column (1 × 100 mm), and C₁₈ 300 A column (0.5 × 150 mm) from Applied Biosystems; Tween 20 and SDS from Bio-Rad; endoproteinase aspartate-N from Boehringer Mannheim; bovine serum albumin (fatty acid free), hydrogenated Triton X-100, -octylglucoside from Calbiochem; guanidine hydrochloride and urea from ICN; Immobilon P (PVDF) transfer membranes, 0.45 µm, from Millipore; [-³²P]ATP from NEN Life Science Products Inc; CM Fast Flow, Source 15S, from Pharmacia LKB; endoproteinase lysine-C and modified trypsin from Promega; nitrocellulose BA 83, 0.2 µm, from Schleicher and Schuell; alkaline phosphatase-labeled goat anti-rabbit IgG (A3937), Amido Black, ammonium bicarbonate, 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium-buffered tablets, -cyano-4-hydroxycinnamic acid, DTT, E-64, leupeptin, pepstatin, poly(Glu,Tyr) 4:1, PKA catalytic subunit (bovine or porcine heart), PKI(5-24), Ponceau S, TLCK, and TPCK were from Sigma; DE52 was from Whatman. All chemicals for automated Edman degradation sequencing and peptide synthesis were from Applied Biosystems/Perkin Elmer. Water and acetonitrile for HPLC separations were from EM Science. Polyclonal antibody to bovine C was a gift from Dr. Brian A. Hemmings (Friedrich Miescher-Institut, Switzerland).

Collection of Ovine Sperm

Ejaculated sperm were collected and washed as described earlier (26). Epididymis and testis were obtained from a freshly killed ram. Epididymal sperm was collected by retrograde extrusion of the epididymal contents with wash buffer (27). The extruded sperm were then washed as described for the ejaculated sperm. To collect testicular sperm, a testis was minced, suspended in wash buffer, and the suspension filtered twice through cheesecloth. The filtrate containing testicular sperm was transferred to 15-ml polypropylene tubes and centrifuged at 900 rpm for 20 min (Model CL Clinical Centrifuge, IEC). The supernatant was removed and the sperm pellet was washed and centrifuged again as above. The sperm concentrations were determined using a hemacytometer.

Isolation of Ovine Sperm Flagella and Heads

The flagella were isolated using a modified version of an earlier procedure (28). About 6 ml of ejaculated ram semen (sperm density 3-4 × 10⁹/ml) were collected and washed. The sperm density was adjusted to 1.5 × 10⁹/ml with the wash buffer to yield about 12 ml of sperm suspension. Three-ml aliquots of the washed sperm were transferred to 50-ml polypropylene tubes, each containing 9 ml of chilled phosphate-buffered saline with protease inhibitors (150 mM NaCl, 10 mM 4-aminobenzamidine, 10 µM E-64, 4 mM EDTA, 15 µM pepstatin, 100 µM TLCK, 200 µM TPCK, 4 mM DTT, 20 mM potassium phosphate, pH 6.5) (PBSI). Subsequent steps were done at 4 °C. The sperm suspension in each tube was sonicated on ice (Branson Sonifier Model 200 fitted with a 3-mm double-stepped microtip) using four 10-s continuous pulses with a power setting of 25% (37.5 watts). The above sonication conditions produced about 70% sperm decapitation.

The flagella were separated from heads and residual whole sperm by centrifugation at 4 °C in a discontinuous sucrose gradient as described earlier (28) except that the bottom 2.2 M sucrose layer was omitted. The sonicated sperm (48 ml) were stirred into 108 ml of 2.2 M sucrose and 13-ml aliquots were pipetted into 32-ml polycarbonate centrifuge tubes each containing 19 ml of 2.05 M sucrose. The tubes were centrifuged at 72,000 × g for 1.5 h (Beckman SW 28 rotor). Heads and whole sperm pelleted at the bottom of the tube while the flagella collected at the 1.5-2.05 M sucrose boundary.

The isolated flagella were transferred to 12-ml thick-walled polypropylene tubes, diluted 1:2 with PBSI, and centrifuged at 18,000 × g for 15 min (Sorvall SS 34 rotor). The pellets were resuspended in PBSI, and the concentration of the flagellar suspension determined using a hemacytometer. Usually about 10 × 10⁹ flagella were obtained from 6 ml of ram semen.

Isolation of sperm heads was as described (28). Two passes through the sucrose gradient were done to make sure that the preparation was free of flagella and undecapitated sperm.

Preparation of Demembranated Ovine Ejaculated Sperm, Sperm Heads, and Sperm Flagella

Ovine ejaculated sperm were demembranated as described earlier (26). Isolated sperm heads and flagella were treated the same way except that flagella were separated from the demembranation medium by centrifugation through 40% Percoll, and heads were separated from the demembranation medium by centrifugation through 70% Percoll.

Renaturation of Protein Kinases Blotted on PVDF Membrane

Samples were dissolved in SDS-PAGE sample buffer (10% glycerol, 3% SDS, 0.03% bromphenol blue, 50 mM DTT, 62.5 mM Tris-HCl, pH 6.8) and then electrophoresed in a 1.5-mm thick gradient gel (5-15% acrylamide, 12 × 14 cm). Protein kinases were renatured on the blot and detected using a protocol adapted from Ferrell and Martin (29) and described earlier (30). In some experiments, 1% poly(Glu,Tyr) 4:1, a tyrosine kinase substrate, was used as blocking solution in place of 5% bovine serum albumin.

Western Blotting

Samples were dissolved in SDS-PAGE sample buffer, electrophoresed in 0.75-mm thick 10% minigels (4.5 × 8 cm), and blotted to PVDF membranes (TE 22 transfer apparatus, Hofer Scientific, 28 V for 5 min followed by 84 V for 20 min). The transfer buffer composition was 50 mM Tris base, 192 mM glycine, 20% methanol, and 0.01% SDS. After transfer, the membrane was blocked with TBS-Tween 20 (30 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.05% Tween 20) for 20 min, incubated with anti-bovine C (in TBS-Tween 20, 1:200 dilution) for 1 h, then washed four times (5 min each wash) with 200 ml of TBS-Tween 20. Incubation with the secondary antibody (alkaline phosphatase-labeled goat anti-rabbit IgG in TBS-Tween 20, 1:800 dilution) was for 1 h, followed by washing twice, 10 min each wash, with 200 ml of TBS-Tween 20. Final wash was 200 ml of 30 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.5% Triton X-100 for 20 min. The blot was then exposed to 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium (1 tablet dissolved in 10 ml water) to reveal cross-reacting proteins.

Isolation of the PKA Catalytic Subunit (C_s) from Ovine Sperm Flagella

Extraction of C_s from Sperm Flagella by cAMP-- Extraction was done at 4 °C. Sperm flagella (in PBSI) were centrifuged (1750 × g, 15 min) and resuspended for 30 min in a Triton X-100/NaCl buffer (5 mM potassium phosphate, pH 6.5, 0.5%, v/v, Triton X-100, 150 mM NaCl, 1 mM EDTA, 25 µM leupeptin, 1 mM DTT), at a concentration of 3 × 10⁸ flagella/ml. This treatment removed the plasma membrane and most of the soluble flagellar proteins that otherwise would coextract with C_s upon subsequent treatment with cAMP; C_s itself remained bound to the demembranated flagella.² The suspension then was centrifuged (1750 × g, 30 min), the supernatant discarded, and the pellet dispersed in KPNELD wash buffer (5 mM potassium phosphate, pH 6.5, 50 mM NaCl, 1 mM EDTA, 25 µM leupeptin, 1 mM DTT, 0.22 ml of buffer/10⁸ flagella) and centrifuged (1750 × g, 10 min). This was repeated with 0.167 ml of buffer/10⁸ flagella. The washed pellet was then extracted for 20 min with KPNELD + 10 µM cAMP, 0.167 ml/10⁸ flagella. The cAMP extract, which contained C_s, was transferred to a polypropylene tube that had been treated with Triton X-100 to minimize nonspecific binding of C_s, and then centrifuged at 27,000 × g for 15 min to remove residual flagella.

Fast Protein Liquid Chromatography of cAMP Extract of Sperm Flagella-- The cAMP extract was made 0.2% (w/v) in -octylglucoside and applied at 0.3 ml/min flow rate to a CM Fast Flow column (0.5 × 5 cm) previously equilibrated with chilled buffer A (20 mM potassium phosphate, pH 6.5, 1 mM EDTA, 50 mM NaCl, 0.1%, w/v, -octylglucoside, 1 mM DTT). The column was washed with buffer A until the absorbance, which rose due to cAMP in the extract, returned to baseline. The column was eluted with 7 ml of a cold linear NaCl gradient of buffer A plus buffer B (20 mM potassium phosphate, pH 6.5, 1 mM EDTA, 300 mM NaCl, 0.1%, w/v, -octylglucoside, 1 mM DTT) at 0.2 ml/min flow rate. C_s was detected in fractions between 180 and 215 mM NaCl. Yield was typically 20-25 µg/6 ml of semen. -Octylglucoside was included in the buffers to prevent nonspecific binding of C_s to the glass column, tubings, and tubes, as well as to stabilize the native conformation of C_s (31).

Isolation of the PKA Catalytic Subunit (C_sm) from Ovine Skeletal Muscle

The procedure was adapted from Okuno and Fujisawa (31) for the isolation of bovine heart C. Working at 4 °C, about 600 to 800 g of skeletal muscle tissue from the hind legs and back of a ram were stripped of fat and connective tissue and then passed through a meat grinder (coarse setting). The ground tissue was mixed with 1.5 liters of homogenization buffer (10 mM potassium phosphate, pH 6.8, 1 mM EDTA, 0.1 mM DTT) and further processed to a smooth consistency in a Waring blender. The homogenized mass was then centrifuged at 18,000 × g for 30 min and the supernatant was collected, filtered through glass wool, and applied to a packed DE52 column (5 × 15 cm) equilibrated with the homogenization buffer. About 8 liters of wash buffer (55 mM potassium phosphate, pH 6.8, 1 mM EDTA, 0.1 mM DTT) were then passed through the column, followed by about 2 liters of the DE52 equilibration buffer (45 mM potassium phosphate, pH 6.8, 0.1 mM EDTA, 1 mM DTT, 0.1%, v/v, Tween 20). The PKA holoenzyme was bound to the DE52 resin; C_sm was released by washing with DE52 equilibration buffer containing 100 µM cAMP. DE52 sequestered cAMP, so that C_sm was eluted only after about twice the bed volume of the elution buffer was applied. The fractions containing C_sm were identified by SDS-PAGE, and then applied to a CM Fast Flow column (1 × 8.5 cm, equilibrated with DE52 equilibration buffer). The column was washed with 20 mM potassium phosphate, pH 6.8, 1 mM EDTA, 50 mM NaCl, 0.1% (w/v) -octylglucoside, 1 mM DTT and then subjected to a linear salt gradient of 50 mM to 300 mM NaCl (total volume of gradient, 50 ml; flow rate 0.4 ml/min). C_sm eluted between 160 and 280 mM NaCl. The resulting C_sm preparation was about 95% pure, with a yield of about 400 µg of C_sm/600 g of skeletal muscle. It was made 40% in glycerol and stored at 20 °C. Homogeneous C_sm was obtained by applying an aliquot of the glycerol-stabilized preparation (2-3 ml, 40-60 µg of C_sm) to a Source 15S column (0.5 × 5 cm, equilibrated with buffer A) and eluting with a linear NaCl gradient (buffer A plus buffer B, total volume of 7 ml, flow rate 0.4 ml/min). C_sm eluted as a sharp peak between 250 and 265 mM.

HPLC and Protein Sequencing

All HPLC separations were carried out on a modified Hewlett-Packard HP1090 M system equipped with a UV photodiode array detector with a 1.7-µl microflow cell. For 20 µl/min flow rates an LC packing Accurate Microflow splitter was used. Automated Edman degradations were performed on an Applied Biosystems/Perkin-Elmer 494 Procise sequencing system.

Mass Spectrometry

MALDI TOF MS was performed on a Perseptive Biosystems linear Voyager BioSpectrometry Workstation. Electrospray MS/MS was performed using a Perkin-Elmer Sciex API 365 benchtop triple quadrupole mass spectrometer equipped with MicroIonSpray and the BioToolBox^TM software package. Product ion MS/MS experiments were carried out using nitrogen as the dissociating gas at a collision cell pressure of 2.2 × 10³ torr and a collision energy of 40 eV. Scans were obtained in the positive-ion mode from m/z 30 to 1500 with a step size of 0.25 atomic mass units and a dwell time of 0.75 ms. In order to increase sensitivity, the first MS was operated using low resolution (full width, half-height ~ 3 atomic mass units). Additionally, samples were infused at a flow rate of 200 nl/min using the MicroIonSpray source, thus allowing 140 scans to be signal averaged to improve signal-to-noise ratio.

Trypsin Digestion of C_s

C_s was blotted onto PVDF membrane as described above (see "Western Blotting"). The membrane was cut into small pieces (1 × 1 mm) and submerged under 50 µl of Digest Buffer (10% acetonitrile, 1% hydrogenated Triton X-100, 100 mM ammonium bicarbonate, pH 8.2). One µg of trypsin in 4 µl of 50 mM acetic acid was added to the sample, which was then incubated overnight at 37 °C followed by direct injection of the supernatant onto the HPLC. Tryptic peptides were separated on a 1-mm × 25-cm Applied Biosystems (Aquapore RP-300) C₈ column using a linear gradient from 100% solvent A (0.1% trifluoroacetic acid) to 55% solvent B (0.08% trifluoroacetic acid in acetonitrile/water: 70/30) in 30 min, then from 55% solvent B to 85% solvent B in 10 min at a flow rate of 150 µl/min. The eluent was monitored at 210 nm and fractions were collected manually.

CNBr Cleavage of C_sm and C_s at Methionyl Residues

For electrophoretic analysis of fragments, about 2 µg each of C_sm and C_s in 50 µl of buffer A were lyophilized. To each lyophilized sample was added 12.5 µl of 100 mM DTT, and the mixture then was allowed to stand at room temperature for 1 h. To start the cleavage reaction, 37.5 µl of 47 mM CNBr (in 88% formic acid) were added. The reaction mixture was incubated overnight (about 19 h) at room temperature in the dark. To stop the reaction, 100 µl of water was added and the resulting solution was lyophilized to remove the formic acid and unreacted CNBr. The pellet was resuspended in 100 µl of water and lyophilized again, and afterward 15 µl of SDS sample buffer (50 mM Tris-HCl, pH 6.8, 15% glycerol, 5% SDS, 0.003% bromphenol blue) was added to the lyophilized sample. The sample was electrophoresed in a 0.75-mm thick 13% Tris-Tricine/SDS-polyacrylamide gel (32); the gel was silver stained to reveal the fragments.

For sequencing of fragments, the starting sample contained about 10-15 µg of C. The cleavage products were separated by electrophoresis in a 1.5-mm gel and transferred to a PVDF membrane according to the protocol of Otter et al. (33), except that the transfer time was shortened from 17 to 12 h. The blot was stained with Amido Black and the fragments excised and sequenced.

NTCB Cleavage of C_sm and C_s at Cysteinyl Residues

The procedure was adapted from Jacobson et al. (34). The concentration of C_sm and C_s (in fresh buffer A) was adjusted to 40 µg/ml and 50 µl of each was then lyophilized. The dried samples were redissolved in 125 µl of denaturation buffer (100 mM HEPES, pH 8.5, 8.1 M urea) and allowed to stand for 30 min at room temperature. A 15-µl aliquot of freshly prepared NTCB (66.5 mM) was added and allowed to react for 20 min. The amount of NTCB added gave about a 10-fold excess of NTCB over the combined sulfhydryl groups of DTT (present in buffer A) and C_sm or C_s (presumed to be two, as with bovine C). NTCB (15 mg) was first dissolved in 0.333 ml of ethanol and then made up to 1 ml with the denaturation buffer. The pH of the reaction mixture after the addition of NTCB was about 8.4. The cleavage reaction was initiated by the addition of 4.8 µl of 1 N NaOH, which brought the pH up to about 11.6. The reaction was allowed to proceed for about 16 h. The mixture was then transferred to a Centriplus 10 concentrator (Amicon) and washed with buffer A until the urea and excess NTCB were reduced about 3000-fold and the final volume was reduced to about 200 µl. The concentrate was then mixed with an equal volume of 2 × Schägger sample buffer and electrophoresed in a 10% Tris-Tricine/SDS-polyacrylamide gel (32). The fragments were revealed by silver staining.

Preparation of NTCB fragments for mass spectrometry was the same, except that 30 µg each of C_sm and C_s were used as starting material. The Tris-Tricine gel was transferred to nitrocellulose (33) at 28 V for 5 min and then at 84 V for 20 min. The bands were revealed by staining with Ponceau S and then cut out for MALDI TOF MS.

Endoproteinase Lysine-C Digestion of C_sm and C_s

C_sm and C_s were blotted on nitrocellulose which was then cut into 1 × 1-mm pieces and submerged under 50 µl of Digest Buffer. An aliquot of endoproteinase lysine-C (0.5 µg in 0.5 µl of 25 mM sodium phosphate, pH 7.5, 1 mM EDTA) was then added and the samples were incubated overnight at 37 °C followed by direct injection of the supernatant onto the HPLC. Endoproteinase lysine-C peptides were separated on a 0.5 × 150-mm Applied Biosystems column (C₁₈, 300 A) using a linear gradient from 100% solvent A to 46% solvent B in 35 min, then from 46% solvent B to 60% solvent B in 10 min at a flow rate of 20 µl/min. The eluent was monitored at 210 nm and fractions collected manually.

Endoproteinase Aspartate-N Digestion of Blocked Amino-terminal Peptide of C_s

The blocked amino-terminal peptide isolated from the endoproteinase lysine-C digest of C_s was dissolved in 25 µl of 100 mM ammonium bicarbonate, and 0.12 µg of endoproteinase aspartate-N in 3 µl of 10 mM Tris-HCl, pH 7.5, was added. Digestion proceeded overnight at 37 °C. The digested peptide was desalted in a C₁₈ microcartridge (0.8 mm × 5 mm, LC packing, San Francisco, CA) prior to direct application to Edman sequence analysis.

Peptide Synthesis

Four peptides were synthesized in order to compare their MS/MS product ion spectra to the MS/MS product ion spectrum of the amino-terminally blocked endoproteinase lysine-C peptide derived from C_s: 1) acetyl-SANPNDVQEFLAK; 2) acetyl-ASNPNDVKEFLAK; 3) acetyl-ASGGPNDVKEFLAK; and 4) acetyl-ASNPGGDVKEFLAK. Peptides were synthesized on a Perkin-Elmer 432A Synergy Peptide Synthesizer using HBTU activation and the 9-fluorenylmethoxycarbonyl protecting strategy. Crude peptide mixtures were purified by reversed phase HPLC using an Aquapore OD-300 C₁₈ column (1 × 100 mm) and a water/acetonitrile/trifluoroacetic acid gradient at 40 µl/min and 37 °C.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

C_s Has an Unusual Mobility in SDS-PAGE-- Our previous results (17) showed that demembranated ram sperm models retained >90% of the sperm PKA activity; moreover, preparations of demembranated sperm with initially poor motility in the presence of ATP became highly motile when cAMP was added, indicating that the cAMP-dependent pathway that initiates motility was still intact. To begin characterization of C_s and other protein kinases in these cytosol-free models, the models were analyzed using protein kinase blots. Proteins of the demembranated sperm were separated by SDS-PAGE, blotted to PVDF membranes, renatured by incubation in a buffer containing the non-ionic detergent Tween 20, and then incubated in the presence of [-³²P]ATP to allow the bound kinases to phosphorylate themselves or the blocking reagent. A typical autoradiogram of such a blot is shown in Fig. 1A. Six to eight putative kinase bands were observed in the demembranated sperm. Two prominent bands at ~M_r 40,000 and 39,000 migrated slightly faster than porcine C (~M_r 41,000). To determine if either of these was C_s, the incubation with [-³²P]ATP was carried out in the presence of PKI(5-24), a potent inhibitor of C (35), including ram C_s (17) activity. The inhibitor completely and specifically blocked the labeling of porcine C and the ~M_r 40,000 band from sperm (Fig. 1B), indicating that the latter is likely to represent C_s. Labeling of the other bands was not affected by PKI(5-24). Interestingly, labeling of the prominent band at ~M_r 69,000 was enhanced when the blocking reagent was the tyrosine kinase substrate poly(Glu,Tyr) instead of bovine serum albumin (results not shown), suggesting that this band represents a tyrosine kinase. Tyrosine kinases also have been implicated in the control of sperm motility (36, 37).

View larger version (80K): [in this window] [in a new window]   Fig. 1.   Protein kinase blots of demembranated ram sperm. Panel A, autoradiogram of blotted and renatured protein kinases after labeling with [-32P]ATP. Lane S, demembranated ram sperm (3.3 × 106); lane C, porcine C (400 ng). Panel B, same as panel A except labeling with [-32P]ATP was done in the presence of 600 nM of the PKA inhibitor PKI(5-24). Arrows indicate PKI(5-24)-sensitive protein kinase bands. Asterisks indicate a possible tyrosine kinase (see text). Numbers indicate molecular weights of 14C-methylated marker proteins.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Immunodetection of Cs and Csm by Western blot using polyclonal antibodies to bovine C. Panel A, ram sperm (lane S, 1 × 106 washed sperm) contain a cross-reacting protein with a slightly faster mobility than porcine C (lane C, 62 ng). This protein is released from sperm flagella by extraction with Triton X-100 in the presence of cAMP (lane Ex, extract from 2 × 106 flagella). Panel B, comparison of electrophoretic mobilities of porcine C (lane C, 20 ng), ram skeletal muscle C (lane Csm, 20 ng), ram sperm head C (lane H, 3 × 106 heads), and ram sperm flagellar C (lane F, 1 × 106 flagella). Lane F + Csm shows a mixture of ram skeletal muscle C (20 ng) and ram sperm flagella (1 × 106 flagella). Panel C, the C subunits from ram testicular, epididymal, and ejaculated sperm have identical electrophoretic mobilities. Lane 1, porcine C (31 ng); lanes 2 and 7, ejaculated ram sperm flagella (6 × 105 flagella); lanes 3-5, ram sperm from the cauda, corpus, and caput regions, respectively, of the epididymis (1 × 106 sperm each); lane 6, ram sperm from rete testis (1 × 106 sperm); Numbers indicate molecular weight of marker proteins (not shown).

C_s Is a Tissue Variant-- The unusual mobility of ram C_s could have been due to species (porcine versus ovine) or tissue (somatic versus sperm) variation. To clarify this, ram skeletal muscle C (C_sm) was partially purified and compared directly to the sperm isoform (Fig. 2B). Because the predominant isoform of C in skeletal muscle is C (19), the isoform that we purified from muscle is presumed to be predominantly or entirely C. The skeletal muscle subunit had the same mobility as porcine C, and both migrated slightly slower than the sperm subunit. When purified ram C_sm and ram sperm flagella were mixed together, two distinct, nonmerging bands were observed. These results clearly indicated that the unusual mobility of C_s is tissue specific.

C_s Is Contained Primarily in Flagella-- The relative distribution of C_s in demembranated, isolated sperm heads and tails also was investigated. The vast majority of C_s was located in the flagella (Fig. 2B, lanes H and F; note that lane H was loaded with 3 × 10⁶ sperm heads versus 1 × 10⁶ flagella in lane F); however, some C_s was detectable in the sperm heads. Similar results were obtained with intact sperm heads and intact sperm tails (results not shown). These results are consistent with previous reports that PKA is located primarily in the sperm flagella (38-43).

The Unusual Mobility of C_s Is Not the Result of Epididymal Processing-- Some sperm proteins undergo processing during epididymal maturation (44-46). To determine if such processing was responsible for the unusual mobility of C_s, sperm were isolated from the testis and regions of the epididymis, and the relative mobilities of their PKA catalytic subunits compared in Western blots. Fig. 2C shows a Western blot of C_s from demembranated ram testicular sperm, demembranated epididymal sperm (cauda, corpus, and caput), and demembranated ejaculated sperm flagella. The mobility of C_s was identical in sperm from all stages, and in all cases was slightly faster than that of somatic C. These results indicate that the apparently smaller size of C_s is not due to processing during sperm maturation.

Purification of C_s-- To understand the difference between C_s and C_sm, it was necessary to purify C_s for structural characterization. We previously reported that demembranation of ram sperm with Triton X-100 (0.2%, 0.2 ml per 1.5 × 10⁹ sperm, 35 s) did not release significant PKI(5-24)-inhibitable protein kinase activity, and that subsequent extraction of the demembranated sperm with cAMP (100 µM, 5 min) did not release the activity, but that demembranation with Triton X-100 in the presence of cAMP solubilized ~50% of the activity (17). Unfortunately, the resulting extract also contained a large number of other proteins that complicated purification of C_s. More recently, we have found that C_s can be removed from purified sperm flagella in a near homogeneous state by a two-step procedure consisting of longer extraction with Triton X-100 (0.5%, 5 ml per 1.5 × 10⁹ sperm, 30 min) in the presence of 150 mM NaCl to remove the detergent- and salt-soluble proteins, followed by extraction with cAMP (10 µM, 20 min) in the absence of detergent to remove C_s. Apparently, the longer extraction with Triton X-100 disrupted bonds that impeded the release of C_s under our previous extraction conditions. Inclusion of 150 mM NaCl in the Triton X-100 extraction caused release at this step of some proteins that otherwise would be removed by the cAMP buffer and thus contaminate the C_s. Fig. 3, TX-100, shows the flagellar proteins released by the Triton X-100/NaCl extraction. After two washes with buffer (washes), the flagella were exposed to 10 µM cAMP, which removed nearly pure C_s from the flagella (cA ex). In some preparations, this cAMP extract also contained small amounts of a 20-kDa protein (results not shown). This apparently was not a proteolytic fragment of C_s, as it was not recognized by polyclonal antibodies to C (results not shown). C_s was further purified by passing the cAMP extract through a CM Fast Flow column (Fig. 3, lanes marked 10-21). This removed any trace of the 20-kDa protein, which eluted later than C_s. Using this procedure, 20-25 µg of purified C_s could be obtained from ~6 ml of semen (~10¹⁰ sperm).

View larger version (40K): [in this window] [in a new window]   Fig. 3.   Purification of Cs. Silver-stained SDS-polyacrylamide gel illustrating, from left to right, sequential steps in Cs purification. Starting material was 8.2 × 109 ram sperm flagella. Lanes are as follows: TX-100, proteins extracted by Triton X-100/NaCl (from 5.6 × 105 flagella); washes, supernatants recovered from washes of the flagella after extraction with Triton X-100/NaCl; cA ex, cAMP extract from washed flagella (4.5 × 106); FT, flow-through during introduction of cAMP extract to the CM Fast Flow column; 10, 12, 14, 15, 16, 17, 18, 19, 20, 21, fractions collected from the CM Fast Flow column during elution with the NaCl gradient; Csm, ram Csm (140 ng); M, molecular weight markers. Fractions 12-20 were pooled to yield about 25 µg Cs.

C_s and C_sm Differ in Mass-- The availability of pure C_s and C_sm allowed us to use MALDI TOF MS to determine if the difference in their electrophoretic mobilities was due to their masses. The MALDI TOF mass spectra for C_s had a single peak corresponding to a mass of ~39.9 kDa, whereas those for C_sm had a single peak of ~40.8 kDa. Fig. 4 shows the spectrum for an equimolar mixture of C_s and C_sm; two well separated peaks with masses of 39,832 and 40,722 Da were observed. Although some error is to be expected for the estimated masses of proteins of this size, the mass difference obtained by comparing two proteins in the same spectrum is quite accurate. Therefore, C_s is ~890 Da smaller than C_sm, confirming the apparent difference in mass observed by SDS-PAGE. The results also confirm the purity of the C_s and C_sm preparations. The observed mass for ovine C_sm was reasonably close to that predicted for bovine somatic C with a myristylated glycine at the amino terminus (40,858 Da) (see below).

View larger version (14K): [in this window] [in a new window]   Fig. 4.   MALDI TOF mass spectrum of Cs and Csm mixture. Equimolar amounts of ram Cs and Csm were mixed together and analyzed. Two discrete peaks were observed, differing in mass by 889.9 Da.

The Partial Amino Acid Sequence of C_s Is Identical to That of C-- Beebe et al. (21) reported the isolation and sequencing of human cDNA clones encoding an unusual tissue-specific isoform of C, which they termed C. C mRNA was found in detectable levels only in testis. C has not yet been isolated from testis, but when it was expressed and purified from transfected cells it migrated in SDS-PAGE at 39-40 kDa versus 41-42 kDa for C (49). Although C reportedly is not sensitive to PKI, its expression in testis and the similarity in its apparent mass to that of C_s raised the possibility that C_s might be C. Human C and C differ at 74 amino acid residues, suggesting that the ovine homologs of these two isoforms should be distinguished readily by even partial amino acid sequences. Attempts to obtain amino-terminal sequence directly from the intact ram C_s were unsuccessful, suggesting that its amino terminus is blocked. Consequently, purified C_s was digested with trypsin to generate tryptic fragments, and the fragments purified by HPLC. Similarly, purified ram C_s or C_sm was cleaved with CNBr, and the larger fragments separated by SDS-PAGE and transferred to PVDF membrane. Selected tryptic and CNBr fragments were then sequenced by automated Edman degradation. Fig. 5 shows the resulting amino acid sequences compared with similar sequences predicted for bovine C and C, and human C and C. There is no published sequence for ovine C or C. However, the ovine C_s sequence exactly matched that of bovine C (78 out of 78 residues). Moreover, in 17 out of 18 positions where human C differed from human C, ovine C_s was identical to human C. Therefore, C_s is not C. Similarly, ovine C_s was identical to bovine C at 5 out of 5 positions where bovine C differed from bovine C, indicating that C_s is not C. These results strongly suggested that C_s is a short variant of C.

View larger version (61K): [in this window] [in a new window]   Fig. 5.   Partial sequence of Cs and Csm from CNBr and tryptic fragments. Sequences obtained for ram Cs and Csm fragments are aligned with the corresponding fragments from bovine C and C, and human C and C. Residues different from those of Cs and Csm are highlighted.

Localization of the Region of Mass Difference-- To delimit the regions that are different between C_s and C_sm, the purified subunits were treated with NTCB, which cleaves at cysteinyl residues. There are only 2 cysteinyl residues (Cys-199 and Cys-343) out of a total of 350 amino acids in either bovine or human C. The three fragments resulting from a complete cleavage of C by NTCB are predicted to have masses of 0.9 kDa (residues 343-350), 16.6 kDa (residues 199-342), and 23.0 kDa (residues 1-198). When the NTCB fragments of C_s and C_sm were electrophoresed in a Tris-Tricine/SDS-polyacrylamide gel, three major bands were seen (Fig. 6A). The largest band corresponded to the intact polypeptide. Based on the predicted sizes of the fragments, fragment 1 must be the amino-terminal fragment, and fragment 2 must correspond to residues 199-342. Fragment 2 from C_s and fragment 2 from C_sm had identical mobilities, whereas fragment 1 of C_s migrated more rapidly than fragment 1 of C_sm. The fragments were then transferred to nitrocellulose and analyzed by MALDI TOF MS (Fig. 6B). Fragment 2 of C_s and fragment 2 of C_sm had nearly identical masses of 17,970 and 17,967 Da, respectively. In contrast, fragment 1 of C_s had a mass of ~23,620 Da, whereas fragment 1 of C_sm had a mass of ~24,444 Da, a difference of ~824 Da. Because this difference is similar to the difference in masses between the intact polypeptides (~890 Da), most of the difference in mass must be due to structural differences in the amino-terminal halves of the proteins. Substantial sequence in this part of C_s (residues 30-45, 72-91 and 129-133; see Fig. 5) already had been found to match the sequence of C, so these regions could be ruled out as being the source of the difference.

View larger version (29K): [in this window] [in a new window]   Fig. 6.   NTCB fragments from Cs and Csm. Panel A, Tris-Tricine gel of products resulting from NTCB cleavage of ram Cs (Cs) and Csm (Csm). The three bands are intact C (C) and fragments 1 (1) and 2 (2). Starting material was 2 µg each of Cs and Csm. Positions of molecular weight markers (M) are indicated. Panel B, MALDI TOF mass spectra of fragments 1 and 2 of Cs and Csm. Singly, doubly, and triply charged ion peaks are evident.

Identification of a Unique Endoproteinase Lysine-C Fragment from C_s-- There are 34 lysyl residues in bovine C, of which 8 occur in the first 59 residues. It therefore seemed likely that digestion of C_s and C_sm with endoproteinase lysine-C would allow the detection of any dissimilar fragments. Fig. 7A shows HPLC chromatograms of endoproteinase lysine-C fragments from C_s and C_sm. A prominent peak eluting at 26 min was observed in the C_s digest but not in the C_sm digest. MALDI TOF MS analysis of this peak indicated that it contained a single peptide with a mass of 1474 Da (results not shown). Similarly, a peak at 1475 Da was observed in a MALDI TOF mass spectrum of the endoproteinase lysine-C digest of C_s, but not in the C_sm digest (Fig. 7B). An attempt to determine the amino-terminal sequence of the 1474-Da peptide obtained by HPLC was unsuccessful, suggesting that its amino terminus was blocked. These results indicated that this fragment probably represented the amino terminus of C_s.

View larger version (26K): [in this window] [in a new window]   Fig. 7.   Endoproteinase lysine-C fragments of Cs and Csm. Panel A, HPLC chromatograms of endoproteinase lysine-C digests of Cs and Csm. The arrow indicates a unique peptide in the Cs map. Panel B, MALDI TOF mass spectra of the digests. The arrow indicates the unique 1,475 Da peptide in the Cs digest.

Amino Acid Sequence of a Unique Amino-terminal Fragment from C_s-- The structure of the 1474-Da peptide was solved by a combination of MS/MS on a triple quadrupole mass spectrometer and Edman sequence analysis of an endoproteinase aspartate-N cleavage product.

View larger version (25K): [in this window] [in a new window]   Fig. 8.   Analysis of a unique 1474-Da peptide by tandem mass spectrometry. MS/MS product ion spectra for the doubly charged ion with m/z 738.5 from the synthetic peptide acetyl-ASNPNDVKEFLAK (Panel A), and the precursor-ion with m/z 738.5 of the 1474-Da amino-terminally blocked peptide derived from the endoproteinase lysine-C digest of Cs (Panel B). The two spectra are in excellent agreement with regard to fragment ions observed and their relative intensities. As would be expected from a peptide with a carboxyl-terminal lysine, a strong y-ion series is observed; doubly charged fragment ions are prominent due to the internal lysine. Peaks due to cleavage at the amino-terminal amide bond of the proline residue are particularly prominent, as has been reported by others (59). Ions marked with an asterisk (*) correspond to fragment ions originating from a contaminant Triton X-100 precursor ion with a nominal mass of 740 Da. Because Q1 was operated in a low resolution mode, some ions with m/z 740 were admitted into the collision cell for fragmentation. In a subsequent set of experiments the resolution on Q1 was increased in order to prohibit the contaminant Triton X-100 precursor ions from entering the collision cell. Peptide MS/MS product ion spectra obtained under these conditions did not contain the "*" fragment ions, whereas high resolution MS/MS product ion scans for m/z 740 (i.e. prohibiting fragmentation of m/z 738.5 peptide ions) did produce the "*" contaminant ions.

View larger version (11K): [in this window] [in a new window]   Fig. 9.   Amino acid sequences of amino-terminal regions of ovine Cs and bovine C. Residues 1-6 of Cs and 1-14 of bovine C are highlighted. The position of the exon 1/exon 2 boundary in mouse C is indicated by brackets.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The PKA catalytic subunit of ram sperm (C_s) is unusual in its solubility properties (17) and its electrophoretic mobility in SDS-PAGE (see Figs. 1 and 2). The solubility properties were utilized as a critical first step in a simple purification protocol that permitted the isolation of ram sperm C_s in amounts sufficient for characterization by peptide mapping, protein sequencing, and mass spectrometry. The results showed that ovine C_s is identical to bovine C in 85/85 amino acids sequenced between C residues 15 and 317, but that the amino-terminal 14 amino acids of C have been replaced with 6 amino acids of completely different sequence in C_s. Moreover, the amino terminus of C_s is acetylated, rather than myristylated as in C. As a result, C_s is predicted to be 899 Da smaller than C, in excellent agreement with the mass difference (~890 Da) determined by MS. This difference undoubtedly accounts for the more rapid mobility of C_s in SDS-PAGE.

A closer inspection of the amino-terminal sequences of ovine C_s and bovine C reveals that the homology between the two subunits begins precisely at the site (Val-15 in C) of the exon 1/exon 2 junction in the mouse gene (51) (Fig. 9). Inasmuch as the residues carboxyl-terminal to this site that have been sequenced in C_s are an exact match to those of bovine C, it seems likely that C_s is a splice variant of C resulting from use of an alternative 5' exon. If C_s and C were different genes, some differences in amino acid residues carboxyl-terminal to the exon 1/exon 2 boundary probably would have been detected in the tryptic and CNBr fragments that we sequenced. Øyen et al. (22) reported that a probe specific for C mRNA detected both 2.3- and 2.4-kilobase mRNAs in rat pachytene spermatocytes and round spermatids; the smaller band may have been C_s mRNA.

The position of the first intron is conserved between mouse C and C subunits (51), and two isoforms of bovine C have been identified that similarly appear to use alternative versions of exon 1 (24). Different amino termini resulting from alternative promoter use associated with two different 5' exons also have been reported for the neuronal PKA catalytic subunit of Aplysia (52). It has been suggested that the different amino termini may be important for targeting specific isoforms of C to specific sites in the cell (Refs. 24 and 52, and see below).

A mouse C pseudogene, termed Cx, has been cloned (53). It does not possess intron sequences and, based on Northern blot analysis and RNase protection assays of testis and brain mRNA, is not expressed, indicating that it is a retroposon. It was mapped to the mouse X chromosome, whereas C, which it most closely resembles, maps to chromosome 8. Compared with the C sequence, the Cx sequence contains frameshift deletions that give rise to numerous termination codons in the reading frame. The nucleotide sequence of Cx is homologous to that of C downstream from the exon 1/exon 2 junction, but is completely different upstream of this boundary. Cummings et al. (53) reasoned that this apparent truncation at the splice site for the first intron in C occurred because the mRNA intermediate for Cx was incompletely spliced. However, the predicted amino-terminal sequence for Cx is MPSSSNDV-, which is very similar to the amino-terminal sequence (M)ASNPNDV- of ovine C_s. Thus, it seems more likely that the Cx pseudogene arose by reverse transcription of an intact mouse C_s mRNA.

Beebe et al. (21) reported the cloning of cDNAs encoding a unique testis-specific isoform of C, which they termed C. C has not been isolated from testis, but when expressed in transformed adrenal cells, it migrated faster than C in SDS-PAGE (49). It also differed from C in substrate specificity and sensitivity to PKI(5-24). Comparison of the partial sequence of ovine C_s with that predicted for human C showed that C_s is not C. However, in our protein kinase blots, we detected a sperm protein kinase migrating slightly faster than C_s that was not inhibited by PKI(5-24). C cross-reacts weakly with an anti-bovine C polyclonal antibody (49), and although we did not observe any cross-reactivity with the faster, PKI-insensitive protein kinase band using the same antibody, the possibility remains that this band represents ovine C. If so, it may be possible to isolate C directly from ram sperm and conduct studies of its structure and properties.

In the course of cloning C from the human testis cDNA library, Beebe and colleagues (21) also isolated C clones, but they did not obtain a clone encoding the amino-terminal part of the molecule. Apparently, this was because EcoRI was used to excise the C cDNA inserts, and an internal EcoRI site was present at the codons for Glu-17 and Phe-18 of the human C gene (54). Because there was complete identity between their truncated cDNA sequence and the sequence of a C cDNA from a HeLa cell library (54), Beebe et al. (21) did not screen further for the 5' end of the testis C cDNA. It is possible that C_s is present in human testis, but went undetected because all clones characterized were missing exon 1 and a portion of exon 2.

The substitution of the short acetylated amino-terminal segment in C_s for the amino-terminal myristyl moiety and first 14 residues of somatic C is likely to have important consequences for the structure of the subunit. In C, this segment, which corresponds to exon 1 and is termed the "myristylated motif," forms the first two turns of a long amphipathic  helix, the A-helix, that extends across both the small and large lobes of the catalytic core (residues 42-297), then turns sharply inward to terminate with the myristyl group occupying a deep hydrophobic pocket (55). The myristate is tightly anchored by interactions with hydrophobic groups from four parts of the molecule that are widely separated in the linear sequence. The myristate in turn anchors one end of the A-helix, which covers a large and mostly hydrophobic portion of the catalytic core. Bacterially expressed recombinant C lacking the myristyl group is more heat labile than the myristylated subunit (56), and a deletion mutant lacking the myristyl group and residues 1-14 also has decreased thermal stability (57). These results indicate that the myristylated motif is important for stabilizing the protein. Interestingly, deletion of the myristylated motif did not affect the catalytic properties of the recombinant C. Therefore, C_s might be expected to be less thermodynamically stable than C, but to be similar with regard to catalytic activity. In crystals of the non-myristylated recombinant C, the amino-terminal 14 residues were not visible, indicating that they are unstructured in the absence of the myristyl group (58). It would be of interest to determine if the short amino-terminal domain of C_s is similarly unstructured.

In C the amino acid residues of the myristylation motif and the rest of the A-helix shield the myristate and hydrophobic surface of the catalytic core from the aqueous environment, with the result that free C is soluble. It is possible that replacement of the myristylation motif with a shorter stretch of residues, coupled with replacement of the myristate with an acetate, would leave a portion of the hydrophobic core exposed in C_s. Additionally, because the shorter A-helix would be more loosely held to the hydrophobic core as a consequence of the absence of the myristyl anchor, even more of the hydrophobic surface beneath it might be exposed. In this case, the surface of C_s would have a hydrophobic patch that might cause it to bind to other hydrophobic surfaces in the flagellum.

Indeed, such a hydrophobic site may exist on C_s. In the course of developing our purification protocol for C_s, we observed that C_s eluted from a Source 15S column as a broad peak (results not shown). This was in contrast to C_sm, which eluted as a sharp peak. Source 15S is an anion exchanger with a polystyrene/divinylbenzene matrix. Two possible explanations for the different elution profiles are that either the C_s sample was heterogeneous, or there was a substantial nonspecific interaction of the matrix with C_s but not with C_sm. The former is unlikely, because only one band was observed in silver-stained SDS-PAGE and only one peak was detected in mass spectra of purified C_s. The broad peak observed with C_s was more likely due to hydrophobic interaction with the column matrix. The corollary of this is that C_s has an exposed hydrophobic domain not present in C_sm (C).

The presence of a hydrophobic patch on C_s could explain why cAMP alone could not release C_s from demembranated sperm (17), but C_s was released from sperm in the presence of cAMP plus Triton X-100 (17) or by cAMP following extended extraction with Triton X-100 (this report). The Triton X-100 may disrupt hydrophobic interactions between C_s and a flagellar protein other than R. Consistent with this, when unmyristylated recombinant C was co-crystallized with a detergent, the hydrophobic pocket was found to be occupied by a molecule of the detergent (55). This raises the possibility that the unusual amino terminus of C_s may be related to a unique requirement for localization of the "free" subunit to a specific compartment within the highly ordered sperm tail.

Many proteins are potential substrates for phosphorylation by C. In the sperm, phosphorylation of specific proteins by C may be achieved by controlling access of C to its potential substrates. For example, if C were tethered to a flagellar structure by its unique amino terminus or a hydrophobic patch, it would be unable to diffuse freely upon cAMP-induced release from R, and could phosphorylate only those substrates within its immediate vicinity. Such anchoring would prevent promiscuous phosphorylation of incorrect substrates, with its potentially deleterious effects on flagellar motility. Alternatively, anchoring of C_s may keep the "free" subunit from sterically interfering with the motile machinery of the axoneme.