Tyrosylprotein sulfotransferase-2 expression is required for sulfation of RNase 9 and Mfge8 in vivo

Abstract

Protein-tyrosine sulfation is mediated by two Golgi tyrosyl-protein sulfotransferases (TPST-1 and TPST-2) that are widely expressed in vivo. However, the full substrate repertoire of this enzyme system is unknown and thus, our understanding of the biological role(s) of tyrosine sulfation is limited. We reported that whereas Tpst1(-/-) male mice have normal fertility, Tpst2(-/-) males are infertile despite normal spermatogenesis. However, Tpst2(-/-) sperm are severely defective in their motility in viscous media and in their ability to fertilize eggs. These findings suggest that sulfation of unidentified substrate(s) is crucial for normal sperm function. We therefore sought to identify tyrosine-sulfated proteins in the male genital tract using affinity chromatography on PSG2, an anti-sulfotyrosine monoclonal antibody, followed by mass spectrometry. Among the several candidate tyrosine-sulfated proteins identified, RNase 9 and Mfge8 were examined in detail. RNase 9, a catalytically inactive RNase A family member of unknown function, is expressed only in the epididymis after onset of sexual maturity. Mfge8 is expressed on mouse sperm and Mfge8(-/-) male mice are subfertile. Metabolic labeling coupled with sulfoamino acid analysis confirmed that both proteins are tyrosine-sulfated and both proteins are expressed at comparable levels in wild type, Tpst1(-/-), and Tpst2(-/-) epididymides. However, we demonstrate that RNase 9 and Mfge8 are tyrosine-sulfated in wild type and Tpst1(-/-), but not in Tpst2(-/-) mice. These findings suggest that lack of sulfation of one or both of these proteins may contribute mechanistically to the infertility of Tpst2(-/-) males.

FIGURE 1.

PSG2 immunoblotting of epididymal/sperm proteins. Epididymides from wild type (WT), Tpst1^-/- (T1), and Tpst2^-/- (T2) mice were subjected to subcellular fractionation as described under “Experimental Procedures.” Samples of the soluble fraction and detergent extracts of the membrane fraction (15 μg of total protein) were electrophoresed on 4–15% SDS-polyacrylamide gels under either reducing (A) or non-reducing (B) conditions. Proteins were then electroblotted onto PVDF membranes and subjected to immunoblotting with PSG2.

FIGURE 2.

Silver staining and PSG2 immunoblot (IB) analysis of a ≈31-kDa protein. Proteins from the soluble fraction of wild type epididymal homogenates were subjected to affinity chromatography on a PSG2 column as described under “Experimental Procedures.” Fractions isocratically eluting from the column were concentrated and electrophoresed on a 4–15% Tris-HCl SDS-polyacrylamide gel under reducing (R) and non-reducing (NR) conditions in adjacent lanes and subjected to silver staining (left panel) or electroblotted onto PVDF membranes and subjected to immunoblotting with PSG2 (right panel).

FIGURE 3.

Summary of RNase 9 analysis. Mouse RNase 9 is predicted to be a 184-residue polypeptide with a 26-residue signal peptide. Peptides detected by MS/MS analysis of in-gel tryptic digests are underlined. The peptides underlined with dashed lines were detected only after treatment of RNase 9 with PNGaseF. This observation, coupled with the increased electrophoretic mobility of RNase 9 after PNGaseF treatment (see Fig. 4), suggests that RNase 9 is glycosylated at both Asn¹⁴⁷ and Asn¹⁷⁹ (boxed). N-terminal sequencing showed that the first five amino acids of native RNase 9 are NYWDF (boxed region), indicating that signal peptide cleavage occurs between Gly²⁶ and Asn²⁷. Asterisks are placed above the 3 tyrosine residues that meet threshold criteria for tyrosine sulfation as predicted by the Sulfinator program (29). The disulfide bonding pattern shown is based on homology with RNase A.

FIGURE 4.

Glycosidase treatment of RNase 9. Proteins in the RNase 9-containing fraction were electrophoresed under reducing conditions on 4–15% Tris-HCl SDS-polyacrylamide gels before (-) and after (+) digestion with PNGaseF. The gel was then silver stained. Treatment with PNGaseF caused a shift in mobility of the ≈31-kDa band. In-gel tryptic digestion and MS/MS analysis of the band after PNGaseF treatment (arrowhead) yielded two additional tryptic peptides that contain the two potential sites of N-glycosylation (Fig. 3). PNGaseF itself is ≈36 kDa and can be seen as a new band appearing in the treated sample (asterisk).

FIGURE 5.

Immunoprecipitation of RNase 9 from wild type and Tpst2^-/- epididymides. RNase 9 was immunoprecipitated from the soluble fraction of homogenized epididymides from wild type and Tpst2^-/- mice using anti-RNase 9 serum (R9) or preimmune serum (PI). Six percent of input material and supernatants, as well as the immunoprecipitates (IP), were electrophoresed on 4–15% Tris-HCl SDS-polyacrylamide gels under reducing conditions and electroblotted onto PVDF membranes. Membranes were immunoblotted (IB) with either anti-RNase 9 serum (top panel) or the PSG2 mAb (bottom panel). Asterisks indicate bands resulting from a PSG2-binding protein(s) present in the rabbit serum.

FIGURE 6.

Immunoblot analysis of RNase 9 in wild type, Tpst1^-/-, and Tpst2^-/- epididymides. Partially purified native mouse RNase 9 (70 ng of protein) and 15 μg of total protein from the soluble fraction of epididymal homogenates from wild type (WT), Tpst1^-/- (T1), and Tpst2^-/- (T2) mice were electrophoresed on duplicate 4–15% Tris-HCl SDS-polyacrylamide gels under reducing conditions and electroblotted onto PVDF membranes. One membrane was immunoblotted (IB) with RNase 9 antiserum (left panel) and the other was immunoblotted with preimmune serum (right panel).

FIGURE 7.

Metabolic labeling and sulfotyrosine analysis of native RNase 9. A, wild type epididymis was metabolically labeled with Na ³⁵₂SO₄ and RNase 9 was immunoprecipitated from conditioned medium using preimmune or specific antisera as described under “Experimental Procedures.” Immunoprecipitates (IP) were electrophoresed on a 4–12% BisTris SDS-polyacrylamide gels and electroblotted onto PVDF membrane and the membrane was subjected to autoradiography. B, the band indicated by the arrow was excised from the membrane and subjected to alkaline hydrolysis. The alkaline hydrolysate was spiked with internal sulfoamino acid standards and then subjected to thin layer electrophoresis. The plates were sprayed with ninhydrin to reveal the internal standards (right lane) and then exposed to radiographic film to reveal radiolabeled amino acids (left lane).

FIGURE 8.

Enrichment of tyrosine-sulfated proteins from the membrane fraction of epididymis. A sample of detergent extract (15 μg) of the membrane fraction of wild type epididymal homogenate (Input) and 20 μl of the peak elution fraction (Eluate) from the PSG2 column were electrophoresed on the same 4–15% Tris-HCl SDS-polyacrylamide gel under nonreducing conditions, electroblotted onto a PVDF membrane, and immunoblotted (IB) with PSG2. The elution fractions from the column were pooled, concentrated, and electrophoresed on a single lane of another gel. This gel was stained with colloidal Coomassie Blue and the bands indicated were excised and subjected to in-gel tryptic digestion for identification by MS/MS analysis. The proteins identified from these bands are listed in Table 1.

FIGURE 9.

Schematic representation of the long isoform of mouse Mfge8. Mouse Mfge8 cDNA predicts a polypeptide with an N-terminal signal peptide (SP) that targets it to the secretory pathway. The long isoform contains two N-terminal EGF domains linked to two C-terminal discoidin or F5/8 type C domains by a 38-residue Pro/Thr-rich mucin-like domain that is absent in the short isoform. This segment contains three tyrosine residues (asterisk) and two potential O-glycosylation sites (arrowheads) according to the NetOGlyc 3.1 prediction server (40).

FIGURE 10.

Immunoblot analysis of mouse epididymal proteins and recombinant Mfge8 isoforms. Detergent extracts from the membrane fraction of epididymal homogenates (15 μg of total protein) from wild type, Tpst1^-/-, and Tpst2^-/- mice were electrophoresed on the same non-reducing 4–12% BisTris SDS-polyacrylamide gel along with HPC4-tagged recombinant short and long isoforms of Mfge8. The gel was transferred to PVDF and immunoblotted with a mAb to mouse Mfge8.

FIGURE 11.

Immunoprecipitation of Mfge8 from wild type, Tpst1^-/-, and Tpst2^-/- epididymides. Mfge8 was immunoprecipitated (IP) from detergent extracts of membrane fractions of wild type (WT), Tpst1^-/- (T1), and Tpst2^-/- (T2) epididymides using anti-Mfge8 mAb (clone 2422) or isotype control IgG (Con). Immunoprecipitates were electrophoresed under non-reducing conditions in duplicate on the same 4–15% Tris-HCl SDS-polyacrylamide gel and electroblotted onto a PVDF membrane. Half of the membrane was immunoblotted with an anti-Mfge8 mAb (clone 18A2-G10), whereas the duplicate half was immunoblotted (IB) with PSG2. Arrowheads indicate the long and short forms of Mfge8.

FIGURE 12.

Metabolic labeling and sulfotyrosine analysis of recombinant Mfge8. HEK293T cells were transfected with vectors encoding HPC4-tagged versions of the long or short isoforms of mouse Mfge8. The cultures were metabolically labeled with Na ³⁵₂SO₄ and recombinant Mfge8 was immunoprecipitated from conditioned medium harvested 48 h after transfection using HPC4 mAb as described under “Experimental Procedures.” Immunoprecipitates were electrophoresed on duplicate 4–12% BisTris SDS-polyacrylamide gels and electroblotted onto PVDF membranes. One membrane was subjected to autoradiography (A), whereas the other membrane was immunoblotted (IB) with HPC4 (B). The bands indicated by an asterisk were excised from the membrane and subjected to alkaline hydrolysis. The alkaline hydrolysate was spiked with internal sulfoamino acid standards and then subjected to thin layer electrophoresis. The plates were sprayed with ninhydrin to reveal the internal standards and then exposed to radiographic film to reveal radiolabeled amino acids (C). The locations of the internal standards are indicated.