Variation in the Length of Poly(A) Tails of Spam1 mRNA Populations in Testis and Epididymis
The Spam1 cDNA (from testis) reported in the database (GenBank accession number U33958) contains a poly(A) tail of 29 residues, similar to the finding in this study of an average of 32 residues for testicular transcripts. Our results also show that the testicular transcripts, which are spermatid-derived (Zheng and Martin-DeLeon, 1997), are on average 10–11 residues longer—a significant difference—than those in the epididymis (Figure 1). It should be noted that spermatidal mRNAs with poly(A) tracts of 30–150 bases long are translationally active (Kleene, 1989), and so the length of the poly(A) tract observed for testicular Spam1 transcripts would be consistent with translational activity.
Although a correlation between a specific number of residues in the poly(A) tract and translational activity has not been reported for epididymal mRNAs, it is known that androgens exert a positive effect on “short” (but not “long”) poly(A) tails of CD52, a glycoprotein secreted by the epididymis and found on the surface of epididymal sperm (Pera et al, 1997). Thus it is possible that the maintenance of epididymal Spam1 mRNAs with short tails might be regulated by androgens, which would not be the case for spermatidal mRNAs, because germ cells are without androgen receptors. The tissue difference in poly(A) tail length seen in this study may therefore well reflect regulation of expression (Curtis et al, 1995).
Oligosaccharides of Spam1 Show Tissue Differences in Epididymis and Testis, and Sperm and Epididymis Have Similar Glycan Structures
In the present study, Western blot analysis of two-dimensional SDS-PAGE was used to further identify biochemical characteristics of epididymal Spam1. Our results indicate that different patterns of isoforms of Spam1 exist in the testis and the epididymis. This correlates with the finding that Spam1 is synthesized independently in the testis and epididymis (Deng et al, 2000; Zhang and Martin-DeLeon, 2001), and suggests that its expression in the latter may not be redundant. There was more diversity in the testis which has four discrete isoelectric variants from 6.6 to 9.0 compared to the one-to-three isoforms in the epididymis spanning 7.3–9.0. The two isoforms with pI ranging from 7.6 to 9.0 observed in caudal sperm are within the range of those found in the testis and epididymis. This suggests that populations of Spam1 on sperm may be derived from both testis and epididymis. The pI variants are likely to be a result of different carbohydrate components in the glycan structure of Spam1 in the two tissues. It is well known that there is tissue-specificity of glycosyl-transferases and glycosidases (Kobata, 1992).
PNA, which preferentially binds to O-linked side chains that are present on sperm plasma membranes (Navaneetham et al, 1996) and apical cells (Calvo et al, 2000) recognized a 67 kd molecular mass in the epididymis, testis, and in caudal sperm. Evidence that this 67 kd membrane protein may be Spam1 was obtained by showing that a potential O-linked glycosylation site at Thr379 in mouse Spam1 that we identified from a computer analysis of Spam1 cDNA sequence (Deng et al, 1999) is functional (Figure 3). It is noteworthy that recently, Baker et al (2002) identified a 68 kd PNA-staining sperm membrane protein that localizes to the acrosomal crescent and the principal piece where we have localized Spam1 (Deng et al, 1999).
Our results demonstrate that the lectins PHA-E, LEL, and STL, which bind to N-linked side chains and which were also shown to stain PH-20 from macaque sperm (Li et al, 2002), identified a 67 kd protein molecular mass in epididymis and caudal sperm, but not in testis. Whereas LEL and STL have an affinity for N-acetyl glucosamine, PHA-E has a specificity for complex structures. Thus the findings in this study indicate that testicular and epididymal Spam1 glycoforms, which contain an identical peptide core, are differentially glycosylated: in the epididymal glycoform N-acetyl glucosamine is a major sugar within or at the terminal end of one or more linked glycans. Differential glycosylation has been reported for at least two other epididymal proteins: CD52 (SAGA-1), in which the difference is between lymphocytes and epididymis (Diekman et al, 1997); and clusterin (Ahuja et al, 1996), in which the difference is between testicular and epididymal glycoforms.
The 67 kd band for the different N-linked lectins was verified to be Spam1 by two methods. First, lectin blot analysis of two-dimensional SDS-PAGE of protein in epididymal tissue and luminal fluid identified isoforms at 67 kd molecular mass with pIs ranging from 7.6 to 8.3. This correlates with the results from Western blot analysis of two-dimensional SDS-PAGE in epididymal tissue and luminal fluid. Second, the LEL bands at 67 kd in epididymis and sperm were confirmed to be Spam1-specific with the use of anti-Spam1 serum. After preincubating protein extracts from epididymis and caudal sperm with preimmune serum or anti-Spam1 serum, the lectin bands disappeared in the latter indicating Spam1 specificity. It must be pointed out that Spam1 N-linked sites (Asn46, 165, 293, and 401) are evenly spaced (Deng et al, 1999) and that the antipeptide Spam1 that was used is generated from a 15-mer at positions 381–395, which would be between Asn293 and Asn401. It is possible that steric hindrance from the antibody binding could be responsible for occluding the N-linked chains at Asn293 and Asn401.
N-glycans have been shown to mediate apical sorting of GPI-anchored proteins (Benting et al, 1999), a finding consistent with Spam1 location on sperm and on the apical surface of the epididymis from where it was shown to be released in vivo and in vitro (Deng et al, 2000; Zhang and Martin-DeLeon, 2001). Our finding that three lectins specific for N-linked glycans stain a protein identified as Spam1 from the epididymis and caudal sperm, but not testis, suggests the following: 1) TS on sperm may undergo deglycosylation during epididymal transit to generate N-linked sites not found on the protein in the testis (Deng et al, 1999), and 2) ES may be directly involved in the N-linked sites on sperm by either functioning as the glycosidase (in the luminal fluid) involved in the deglycosylation process or by binding to sperm. It is note-worthy that Baker et al (2002) identified a 68 kd mouse sperm membrane protein that stained with LEA (LEL) and showed it to be located on intact sperm in a pattern distinctly different from that of PNA. It was seen on the anterior crescent and the posterior head where we have localized Spam1 (Deng et al, 1999). The similarities in molecular weights, staining, and localization of Spam1 and that of the protein described by Baker et al (2002), strongly argue that the latter is Spam1. The staining distribution reported by Baker et al (2002) was seen on some sperm but not on others, suggesting that it may be associated with epididymal maturational states and implicating the involvement of ES.
Epididymal Spam1 Is Released in Epididymosomes (Exosomes) With its Lipid Anchor
After ultracentrifugation of the luminal fluid, Spam1 was found to be present in both the supernatant and the pellet although preferentially in the latter. This indicates that there are two populations of this membrane protein, as is the case for epididymal protein DE in rats (Cohen et al, 2000) and for prostasome-like particles in rat epididymal fluid (Fornes et al, 1995). However, it is possible that there may be two sources of Spam1 in the luminal fluid: molecules that come from sperm and those that are released from the epididymis. Because ES was observed in vesicles that were seen in the process of being released in the lumen (Deng et al, 2000), it is likely to reside in the insoluble pellet. This insoluble membranous material, which is similar to prostasomes, has been referred to as epididymosomes and serves as a means of transferring proteins to the sperm surface (Rooney et al, 1996; Yeung et al, 1997).
We have demonstrated that ∼60% of the Spam1 from the insoluble particles obtained after the initial ultracentrifugation of the luminal fluid was released into the medium following 15 minutes of exposure to PI-PLC (Figure 5, E and F). This confirms that ES is secreted into the lumen via lipid vesicles or epididymosomes, probably as a complex of GPI-linked proteins and signaling molecules. Our inability to release the lipid anchor in 100% of Spam1 molecules might result from the inaccessibility of some of the molecules to the enzyme, or it could indicate the presence of a population of Spam1 that is either not PI-anchored or has a GPI anchor that is insensitive to the particular PI-PLC (Cooper, 1998).
GPI-anchored proteins are often known to be resistant to detergent extraction (Triton X-100) and are referred to as detergent-insoluble glycolipids, which float on top of sucrose gradients (Primakoff et al, 1988; Cherr et al, 2001). However, 100% of the Spam1 from the insoluble particles was extracted by exposure to 1% Triton X-100 in our study (Figure 5, G and H), supporting the evidence that ES is released with its lipid anchor, a form in which it can bind to sperm.
Our studies, which provide further evidence that Spam1 is produced independently in the testis and epididymis, reveal that TS and ES are different glycoforms and that the latter shares identical glycans with caudal sperm. There is therefore evidence to suggest that ES may not be a redundant protein. The findings of this study implicate an interaction between sperm and ES, similar to other GPI-anchored proteins such as CD52 (Rooney et al, 1996) and P26h, P34H, and P25b (Frenette and Sullivan, 2001).