I have compared the quantity and the length of the poly(A) tracts of five haploid-expressed mRNAs in the polysomal and nonpolysomal fractions of round and elongating spermatids in mice: transition proteins 1 and 2, protamines 1 and 2, and an unidentified mRNA of about 1050 bases. Postmitochondrial supernatants of highly enriched populations of round and elongating spermatids (early and late haploid spermatogenic cells) were sedimented on sucrose gradients, and the size and amount of each mRNA in gradient fractions were analyzed in Northern blots. In round spermatids, all five mRNAs are restricted to the postpolysomal fractions, but in elongating spermatids about 30–40% of each mRNA is associated with the polysomes. The distribution of these mRNAs in sucrose gradients suggests that all five mRNAs are stored in a translationally repressed state in round and early elongating spermatids, and that they become translationally active in middle and late elongating spermatids. The translationally repressed forms of all five mRNAs are long and homogenous in size, whereas the polysomal forms are shorter and more heterogenous due to shortening of their poly (A) tracts. The relationship between translational activity and poly(A) size exemplified by these five mRNAs may be typical of mRNAs which are translationally repressed in round spermatids and translationally active in elongating spermatids.

The specialized organelles of the mammalian spermatozoon are formed during the haploid (spermatid) phase of spermatogenesis which is known as spermiogenesis (reviewed in Bellvé & O’Brien, 1982). Spermiogenesis in mice lasts approximately two weeks and is divided roughly in half into round spermatids (steps 1–8) and elongating spermatids (steps 9–15). In round spermatids, the acrosome and flagellum start to differentiate, and the rate of RNA synthesis declines ceasing entirely by the beginning of the elongating spermatid stage (Monesi et al. 1978; Kierszenbaum & Tres, 1975). In elongating spermatids, the mitochondria become arranged in a spiral around the base of the flagellum, a fibrous sheath is deposited around the flagellar axoneme, and the nuclei become elongate, dense, compact and resistant to disruption by sonication (Kierszenbaum & Tres, 1975; Meistrich et al. 1976). The differentiation of the spermatid nucleus is accompanied by the replacement of the histones in two stages, first by a group of spermatid-specific basic nuclear proteins, called transition proteins, which are replaced in turn by protamines, the sole basic chromosomal protein in the sperm of most mammals (Kistler et al. 1973; Bellvé et al. 1975; Grimes et al. 1977; Balhorn et al. 1984; Mayer & Zirkin, 1979; Mayer et al. 1981).

The absence of detectable transcription in elongating spermatids in mammals necessitates post-transcriptional regulation of mRNA translation and degradation. Iatrou & Dixon (1977) found that protamine mRNAs in the trout are synthesized and stored in a translationally inactive state in meiotic cells and are translated in spermatids. Subsequently, the mRNAs for mouse protamine 1 (Kleene et al. 1984), rat transition protein 1 (Heidaran & Kistler, 1987; Heidaran et al. 1988) and rat protamine 1 (Mali et al. 1988) were demonstrated to be translationally regulated but according to a different schedule than the trout: the mRNAs are synthesized and translationally repressed in round spermatids and are translationally active in elongating spermatids. The protamine and transition protein mRNAs must be very stable because they persist at high levels for about 6–7 days after the cessation of transcription (Kleene et al. 1983, 1984; Heidaran & Kistler, 1987; Heidaran et al. 1988; Mali et al. 1988). In addition, there is reason to suspect that some spermatidal mRNAs may be degraded after the requisite amount of protein has been synthesized because rat transition protein 1 mRNA disappears from spermatids as the transition proteins are replaced by protamines (Heidaran et al. 1988).

It is also known that the translation of trout protamine, mouse protamine 1 and rat transition protein 1 mRNAs is paralleled by poly(A) shortening (Iatrou & Dixon, 1977; Kleene et al. 1984; Heidaran & Kistler, 1987). The poly (A) tracts on translationally inactive forms of mouse protamine 1 and rat transition protein 1 mRNA are about 160 bases long and remarkably homogenous in size, whereas the poly(A) tracts on the translationally active mRNAs are heterogenous in size varying from about 30 to 160 bases (Kleene et al. 1984; Heidaran & Kistler, 1987). The shortening of poly(A) tracts on spermatidal mRNAs resembles the poly(A) shortening that normally starts when newly synthesized mRNAs enter the cytoplasm (Sheiness & Darnell, 1973). Poly (A) shortening may have a role in controlling post-transcriptional gene expression during spermiogenesis because mRNAs with poly(A) tracts less than about 30 residues long are thought to be degraded rapidly (Marbaix et al. 1975; Nudel et al. 1976) and/or translated inefficiently (Jacobson & Favreau, 1983; Palatnik et al. 1984).

In this paper, I have compared the relationship between poly(A) length and translational activity of five mRNAs which are synthesized after meiosis and are present at high levels in round and elongating spermatids in mice: transition protein 1, transition protein 2, protamine 1, the precursor for protamine 2 and a new unidentified haploid-expressed mRNA of about 1050 bases, HEM1050 (Kleene et al. 1983, 1985, 1988, unpublished; Yelick et al. 1987; Kleene & Flynn, 1987). The experiments were designed to answer two questions: First, does the distribution of these five mRNAs in the polysomal and nonpolysomal fractions of sucrose gradients change between round and elongating spermatids? Second, are there differences in the lengths of poly(A) tracts on the polysomal and nonpolysomal forms of each mRNA? The results demonstrate striking similarities in these mRNAs: all five mRNAs are present as nontranslated messenger ribonucleoprotein particles (mRNPs) with homogenous poly(A) tracts about 150 bases long in round spermatids, and 30–40 % of all five mRNAs are associated with the polysomes in elongating spermatids with heterogenous, shortened poly(A) tracts.

Isolation of polysomal and nonpolysomal RNA from purified spermatogenic cells

CD-I mice at least 60 days old were obtained from Charles River Laboratories. Single cell suspensions from the testes of 10 mice were fractionated by sedimentation at unit gravity on linear 2–4 % bovine serum albumin gradients in a Staput Chamber (Meistrich et al. 1973; Romrell et al. 1976). The round spermatids and nuclear and cytoplasmic fragments of elongating spermatids from a cell separation were resuspended in 3 ml RPMI 1640 medium containing 10 % fetal bovine serum, 6 mm-L-lactate, 1 mm-sodium pyruvate (O’Brien, 1987) and 0·2 μg ml−1 cycloheximide (Walden & Thach, 1986) and cultured 15 min at 33 °C in an atmosphere of 5 % CO2 in air to facilitate the recovery of polysomes after cell separation (Kleene et al. 1984). The following operations were carried out at 0–4 °C. The cells were collected by centrifugation at 300 g for 5 min. The cells were lysed in a motor-driven Teflon-glass homogenizer in 300 μl 0·l m-NaCl, l·5 mm-MgC12, 20 mm-Hepes (pH 7·2), 0·5 % Triton N-101 and 0·1 % diethylpyrocarbonate. The nuclei and mitochondria were pelleted by centrifugation at 13750g for 2 min. The supernatants were adjusted to 0·5 m-NaCl, 30 mm-MgC12 and centrifuged again. 250 μ1 of the supernatant was layered over a 3·7 ml linear 15–40 % sucrose gradients in 0·5 m-NaCl, 30 mm-MgCl2, 20 mm-Hepes, pH 7·2 (Weber et al. 1979). In some experiments, the postmitochondrial supernatants were prepared and centrifuged on gradients in which 20mm-EDTA was substituted for MgCl2. The gradients were centrifuged 100 min at 50 000 revs min−1 in the Beckman SW60 rotor, and deaccelerated to 8000 revs min−1 with the brake on. These centifugation conditions are designed to separate nonpolysomal mRNAs from polysomes containing <5 ribosomes, but polysomes containing ≤5 ribosomes pellet. The gradients were collected from the bottom in a cold room, analyzed at 254 nm in a Beckman Model 153 flow cell (2mm pathlength), SDS was added to 0·5 % to each fraction, and the fractions were extracted once with 2 vol. of cold phenol: chloroform (1:1), and twice with chloroform. The pellets in the bottom of the ultracentrifuge tubes were dissolved in 0·5 ml sucrose gradient buffer containing 0·5 % SDS and extracted identically. The RNA samples were precipitated with ethanol, digested with proteinase K, extracted with phenol and chloro-form again and precipitated with ethanol.

Deadenylation with RNase H

Total testis polysomal and nonpolysomal RNAs were prepared by centrifugation in sucrose gradients, extracted with phenol and chloroform, hybridized to oligo(dT) and digested with RNase H as described by Kleene et al. (1984).

RNA blots

RNA samples were denatured with glyoxal, electrophoresed through 25 cm 1·5–2% agarose gels for 24–26 h at 35 volts, and blotted to nitrocellulose (Thomas, 1980). About 60 ng of cDNA inserts was labeled by random primer extension (Feinberg & Vogelstein, 1986), hybridized as described in Kleene et al. (1984) except that the blots were washed in 0·125 × SSPE 0·2 % SDS at 65 °C. For RNA dot blots, the RNAs were bound to nitrocellulose (Thomas, 1980), hybridized to probes, and small pieces of nitrocellulose were dissolved in 1ml 2-methoxyethanol and counted in 10 ml Aquasol (New England Nuclear).

Distribution of mRNAs in sucrose gradients

Since the rate of mRNA translation is usually regulated at the level of initiation, changes in the rate of translation of an mRNA are generally accompanied by a shift in the distribution of the mRNA between the polysomal and nonpolysomal fractions. It follows that changes in the rate of translation of individual mRNAs can be detected by comparing the distribution of mRNAs in sucrose gradients in Northern blots with cDNA probes (Rosenthal et al. 1983; Kleene et al. 1984; Heidaran & Kistler, 1987).

Here, I have compared the size and translationally active and inactive fractions of five small mRNAs in round and elongating spermatids from mice: transition proteins 1 and 2, protamines 1 and 2, and an unidentified mRNA of about 1050 bases (HEM1050). Round and elongating spermatids were purified on bovine serum albumin gradients. Postmitochondrial extracts were sedimented on sucrose gradients, and the size and levels of mRNAs in gradient fractions were determined in Northern blots. The sequence-of each cDNA probe is known and each probe hybridizes to a single mRNA (Kleene et al. 1983, 1984, 1985, 1988, unpublished; Kleene & Flynn, 1987; Yelick et al. 1987).

The populations of round spermatids used here contain about 80 % round spermatids, 10 % spermatogonia and 10 % elongating spermatids and the elongating spermatid fraction contained >95 % intact elongating spermatids and nuclear and cytoplasmic fragments of elongating spermatids. Fig. 1 (A and B) shows the distribution of absorbance at 254 nm in sucrose gradients prepared from purified populations of round and elongating spermatids. Note the large absorbances of ribosomal subunits and polysomes that are present in postmitochondrial extracts of purified round and elongating spermatids. These sucrose gradients contain a high concentration of NaCl (0·5 M) to minimize contamination of the polysomal regions of the sucrose gradients by nonpolysomal mRNPs (Weber et al. 1979). The peaks for the 40S (fractions 2 and 3) and 60S (fractions 3 and 4) ribosomal subunits are much larger than the 80S single ribosomes (fraction 4) in Fig. 1A and B, because the high salt dissociates the subunits of translationally inactive single ribosomes (Weber et al. 1979). The peaks of polysomes (fractions 5–8) sedimenting more rapidly than single ribosomes from elongating spermatids were dissociated to ribosomal subunits by EDTA as expected (Fig. 1C). (EDTA was equally effective in dissociating polysomes from round spermatids – data not shown.) In seven experiments, the amounts of ribosomal subunits were always much larger than the polysomes in both round and elongating spermatids, and the proportion of polysomes to ribosomal subunits in round spermatids was greater than in elongating spermatids as shown in Fig. 1A and B. The preponderance of translationally inactive ribosomes in spermatids resembles a variety of cells in which the overall rate of protein synthesis is limited at the level of initiation (reviewed by Jackson, 1982). These data also suggest that the fraction of translationally active ribosomes decreases between round and elongating spermatids.

Fig. 1.

Distribution of polysomes in sucrose gradients. Postmitochondrial supernatants were prepared from purified round and elongating spermatids, centrifuged on sucrose gradients, and the absorbance at 254 nm was monitored during collection of gradient fractions. The absorbances were monitored using a flow cell with a 2 mm pathlength and have been converted to the absorbance expected for a 1cm pathlength. (A) Round spermatids, gradient with Mg2+; (B) elongating spermatids, gradient with Mg2+; (C) elongating spermatids, gradient with 20 mm-EDTA instead of Mg2+.

Fig. 1.

Distribution of polysomes in sucrose gradients. Postmitochondrial supernatants were prepared from purified round and elongating spermatids, centrifuged on sucrose gradients, and the absorbance at 254 nm was monitored during collection of gradient fractions. The absorbances were monitored using a flow cell with a 2 mm pathlength and have been converted to the absorbance expected for a 1cm pathlength. (A) Round spermatids, gradient with Mg2+; (B) elongating spermatids, gradient with Mg2+; (C) elongating spermatids, gradient with 20 mm-EDTA instead of Mg2+.

The Northern blots demonstrate that in round spermatids (Fig. 2, RS), the vast majority of all five mRNAs sediments slower than single ribosomes in fractions 2 and 3, so these mRNAs are translationally inactive. Furthermore, each nonpolysomal mRNA migrates as a narrow band in agarose gels implying that there is little variation in the size of the mRNA. In elongating spermatids, a large fraction of the five mRNAs is also found in the nonpolysomal region of the gradient, and these mRNAs are indistinguishable in size from the nonpolysomal mRNAs in round spermatids (Fig. 2, ES). Fig. 3 (lane 2) demonstrates the homogeneity of each nonpolysomal mRNA more clearly because the autoradiogram was exposed for a shorter period than in Fig. 2.

Fig. 2.

Distribution of mRNAs for protamine 1, protamine 2, transition protein 2, and HEM1050 mRNAs in polysome gradients from purified round and elongating spermatids. Postmitochondrial supernatants from round and elongating spermatids were sedimented on sucrose gradients, and the gradients were collected in 8 fractions as shown in Fig. 1, and the pellet. RNA was extracted from each fraction, denatured with glyoxal, electrophoresed through 1·5% agarose (HEM1050) or 2 % agarose (protamine and transition protein mRNAs), blotted to nitrocellulose and hybridized to the indicated 32P-labeled cDNA probes. For protamine and transition protein mRNAs, each lane contained 10 % of RNA from a sucrose gradient fraction, and for HEM1050 mRNA each lane contained 40% of the RNA from a single cell separation. Each probe was hybridized to three sets of sucrose gradient fraction RNAs: RS, round spermatid, Mg2+ gradient; ES, elongating spermatids, Mg2+ gradient; ES-EDTA, elongating spermatid, EDTA gradient. The exposures varied from 3h to 3 days. The apparent differences in rate of sedimentation of the various nontranslated mRNAs are due to slight differences in fractionating the gradients, since there is no difference in the distribution of the various mRNPs in the same gradients.

Fig. 2.

Distribution of mRNAs for protamine 1, protamine 2, transition protein 2, and HEM1050 mRNAs in polysome gradients from purified round and elongating spermatids. Postmitochondrial supernatants from round and elongating spermatids were sedimented on sucrose gradients, and the gradients were collected in 8 fractions as shown in Fig. 1, and the pellet. RNA was extracted from each fraction, denatured with glyoxal, electrophoresed through 1·5% agarose (HEM1050) or 2 % agarose (protamine and transition protein mRNAs), blotted to nitrocellulose and hybridized to the indicated 32P-labeled cDNA probes. For protamine and transition protein mRNAs, each lane contained 10 % of RNA from a sucrose gradient fraction, and for HEM1050 mRNA each lane contained 40% of the RNA from a single cell separation. Each probe was hybridized to three sets of sucrose gradient fraction RNAs: RS, round spermatid, Mg2+ gradient; ES, elongating spermatids, Mg2+ gradient; ES-EDTA, elongating spermatid, EDTA gradient. The exposures varied from 3h to 3 days. The apparent differences in rate of sedimentation of the various nontranslated mRNAs are due to slight differences in fractionating the gradients, since there is no difference in the distribution of the various mRNPs in the same gradients.

Fig. 3.

Effects of deadenylation with RNase H on the sizes of polysomal and nonpolysomal mRNAs for protamine 1 (Pl), protamine 2 (P2), transition protein 1 (TP1), transition protein 2 (TP2), and HEM1050 mRNAs. Postmitochondrial supernatants from total testes were sedimented on sucrose gradients and separated into fractions which sediment faster and slower than single ribosomes. (The region of the gradient containing single ribosomes was discarded.) A sample of polysomal and nonpolysomal RNA was hybridized to oligo(dT) and digested with RNase H to remove poly(A) tracts. Intact and RNase H-treated RNAs were denatured with glyoxal, electrophoresed through 1·5% agarose (HEM1050 mRNA) or 2 % agarose (protamine and transition protein mRNAs), blotted to nitrocellulose, and hybridized to the indicated 32P-labeled DNA probes. The RNA size markers were obtained from Bethesda Research Laboratories. Lane 1, 1·2 μg intact polysomal RNA; lane 2, 0·2 μg intact postpolysomal RNA; lane 3, 1·2 μg RNase H-treated polysomal RNA; lane 4, 0·2 μg RNase H-treated postpolysomal RNA.

Fig. 3.

Effects of deadenylation with RNase H on the sizes of polysomal and nonpolysomal mRNAs for protamine 1 (Pl), protamine 2 (P2), transition protein 1 (TP1), transition protein 2 (TP2), and HEM1050 mRNAs. Postmitochondrial supernatants from total testes were sedimented on sucrose gradients and separated into fractions which sediment faster and slower than single ribosomes. (The region of the gradient containing single ribosomes was discarded.) A sample of polysomal and nonpolysomal RNA was hybridized to oligo(dT) and digested with RNase H to remove poly(A) tracts. Intact and RNase H-treated RNAs were denatured with glyoxal, electrophoresed through 1·5% agarose (HEM1050 mRNA) or 2 % agarose (protamine and transition protein mRNAs), blotted to nitrocellulose, and hybridized to the indicated 32P-labeled DNA probes. The RNA size markers were obtained from Bethesda Research Laboratories. Lane 1, 1·2 μg intact polysomal RNA; lane 2, 0·2 μg intact postpolysomal RNA; lane 3, 1·2 μg RNase H-treated polysomal RNA; lane 4, 0·2 μg RNase H-treated postpolysomal RNA.

Fig. 2 (ES) also demonstrates that in elongating spermatids substantial quantities of all five mRNAs sediment in the polysomal region, and that these polysomal mRNAs migrate as a broad band of heterogenous-sized RNAs ranging from the size of nonpolysomal RNAs to shorter sizes. There are two additional reasons for believing that the heterogenous-sized mRNAs sedimenting in the polysomal regions are translationally active: First, mRNAs encoding larger primary translation products – protamine 2 precursor, 109 amino acids and transition protein 2, 117 amino acids (Yelick et al. 1987; Kleene & Flynn, 1987), sediment with larger polysomes than mRNAs encoding smaller primary translation products – protamine 1, 51 amino acids and transition protein 1, 55 amino acids (Kleene et al. 1985, 1988). Although the size of the primary translation product of the HEM1050 mRNA is not known, it is reasonable that this mRNA sediments with the largest polysomes as shown in Fig. 2E (ES), because the mRNA is the largest studied here (Fig. 3). Second, to verify that the heterogenous forms of the mRNAs are associated with polysomes, the postmito-chondrial supernatants from elongating spermatids were sedimented on sucrose gradients in which the Mg2+ was replaced by EDTA. EDTA dissociates mRNA and ribosomes causing polysomal mRNA to sediment slower than single ribosomes (Penman et al. 1968). Most of the polysomes are dissociated to sub-units by EDTA (cf. Fig. 1B and C), and most of the heterogenous forms of all five mRNAs are shifted to the postpolysomal region of the gradient (Fig. 2, ES-EDTA).

It should be pointed out that the exposure range of the autoradiograms in Fig. 2 is so great that a single exposure cannot illustrate all of the information in the Northern blots. Longer exposures reveal that low levels of the heterogenous polysomal forms of all five mRNAs are also detectable in the nonpolysomal fractions of elongating spermatids (not shown). I do not know whether the heterogenous mRNAs in the nonpolysomal fractions are an artifact of cell separation (Kleene et al. 1984) or a normal part of the translation cycle.

In addition, RNA dot blots were used to estimate the proportion of protamine 1 and transition protein 2 mRNA in each sucrose gradient fraction from round and elongating spermatids (data not shown). In round spermatids, about 93–95 % of the protamine 1 and transition protein 2 mRNAs sediments in the nonpolysomal region (fractions 1–4). The small amounts of these mRNAs sedimenting in the polysomal regions (fractions 5–8) from round spermatids (5–7%) are presumably derived from contamination of the round spermatids by elongating spermatids and contamination of the polysomal fractions of the gradient by nonpolysomal mRNPs. In four experiments, the average fraction of protamine 1 and transition protein 2 mRNAs sedimenting in the polysomal fractions of elongating spermatids was 28 % and 37 %, respectively.

Measurement of poly (A) tracts on polysomal and nonpolysomal mRNAs

The lengths of the poly(A) tracts on the nonpolysomal and polysomal mRNAs for the transition proteins and protamines were analyzed in Northern blots before and after selective degradation of the poly(A) tracts by hybridization to oligo(dT) and digestion with RNase H. Fig. 3 shows the relative sizes in a 2 % agarose gel of intact and deadenylated polysomal and nonpolysomal mRNAs encoding transition proteins 1 and 2 and protamines 1 and 2. The final panel shows the same data for the HEM1050 mRNA in a 1·5 % agarose gel. Note that the sizes of the polysomal mRNAs (lane 1) are much more heterogenous than the nonpolysomal mRNAs (lane 2). After deadenylation with RNase H, the mRNAs from the polysomal and postpolysomal fractions are reduced to virtually the same length (lanes 3 and 4), demonstrating that the differences in size of the nonpolysomal and polysomal mRNAs are primarily due to differences in the length of poly(A) tracts. The approximate sizes of the intact and deadenylated non-polysomal mRNAs based on three experiments are, respectively: protamine 1, 550 and 400 bases; transition protein 1, 590 and 440 bases; transition protein 2, 690 and 540 bases; and protamine 2, 760 and 610 bases; HEM1050 mRNA, 1050 and 900 bases. Estimates of the lengths of the poly(A) tracts on the five mRNAs varied between about 120 and 180 bases. The difference in size between the shortest polysomal mRNAs and deadenylated mRNAs indicates that the majority of polysomal mRNAs have poly(A) tracts at least 30 bases long.

Despite the similarity in the sizes of the poly(A) tracts on the five spermatid-specific mRNAs, there are obvious differences in the size distribution of the polysomal forms of these mRNAs. The polysomal protamine 1 mRNAs with the shortest poly(A) tracts are consistently much more abundant than the mRNAs with longer poly(A) tracts (Kleene et al. 1984). By comparison, the polysomal transition protein 1, transition protein 2 and protamine 2 mRNAs usually show two distinct size classes, one corresponding to mRNAs with little or no poly(A) shortening and the other to mRNAs with shortened poly(A) tracts. The HEM1050 mRNA differs from the transition protein and protamine mRNAs because the majority of its polysomal mRNAs shows no obvious poly(A) shortening (Fig. 2E, ES).

Five haploid-specific mRNAs including the mRNAs encoding the four major basic nuclear proteins that replace the histones in elongating spermatids in the mouse – transition proteins 1 and 2 and protamines 1 and 2 – have been demonstrated here to undergo similar changes in translational activity and poly(A) length during spermiogenesis in the mouse. In round spermatids, all five mRNAs are present as translationally inactive, nonpolysomal mRNPs bearing homogenoussized poly(A) tracts about 150 bases long. In elongating spermatids, approximately 30– 40% of each mRNA is associated with the polysomes bearing heterogenoussized poly (A) tracts ranging from about 30 to 150 bases. These observations confirm and extend previous studies of the mRNAs for trout protamines (Iatrou & Dixon, 1977), mouse protamine 1 (Kleene et al. 1984) and rat transition protein 1 (Heidaran & Kistler, 1987). Since the mRNAs for transition proteins 1 and 2 and protamines 1 and 2 account for all of the abundant basic nuclear proteins in sonication-resistant spermatids in mice (Kleene, unpublished), the HEM1050 mRNA does not encode a spermatidal basic nuclear protein. Therefore, the correlation between active translation and poly(A) shortening in these five mRNAs may extend to all mRNAs that are translationally repressed in round spermatids and translationally active in elongating spermatids.

The change in the distribution of mouse transition protein 1 and 2 and protamine 1 and 2 mRNAs between the nonpolysomal and polysomal fractions in round and elongating spermatids parallels a lag between the synthesis of these mRNAs in round spermatids and the synthesis of the transition proteins and protamines during steps 12–15 of elongating spermatids (Grimes et al. 1977; Mayer & Zirkin, 1979; Mayer et al. 1981; Kleene et al. 1983; Kleene & Flynn, 1987; Balhom et al. 1984; Heidaran & Kistler, 1987; Heidaran et al. 1988). The large fraction (approximately 2/3) of the transition protein and protamine mRNAs that are translationally inactive in elongating spermatids is at least partly explained by the repression of translation of the transition protein and protamine mRNAs in early elongating spermatids (Bellvé et al. 1975; Mayer & Zirkin, 1979; Balhom et al. 1984). However, it is also possible that protamine and transition protein mRNAs are totally active in translation for a very brief period in elongating spermatids.

The shortening of poly(A) on spermatidal mRNAs is analogous to a general phenomenon of poly(A) shortening in eukaryotic cells (reviewed in Brawerman, 1981). Typically, long homogenous poly(A) tracts are added post-transcriptionally to newly synthesized mRNAs in the nucleus, and the poly (A) tracts gradually shorten and become more heterogenous in the cytoplasm. It seems more likely that poly(A) shortening in spermatids is a consequence of, rather than a prerequisite for, translation. Poly (A) shortening is not sufficient for translation because there is no obvious poly(A) shortening on a portion of each of the polysomal forms of the mRNAs studied here. In fact, most of the polysomal HEM1050 mRNAs have undergone no detectable poly(A) shortening. Furthermore, nonhistone mRNAs with longer poly(A) tracts are translated more efficiently than mRNAs with shorter or no poly(A) tracts (Jacobson & Favreau, 1983; Palatnik et al. 1984; Drummond et al. 1985; Galili et al. 1988). The alternative that poly(A) shortening is a consequence of translation explains why poly(A) tracts on polysomal mRNAs are shortened more rapidly than on nonpolysomal mRNAs in spermatids, mammalian tissue culture cells (Sheiness & Darnell, 1973; Merkel et al. 1976), and sea urchin embryos (Dworkin et al. 1977). Several lines of evidence also suggest that mRNA degradation and poly(A) shortening are coupled to the interaction of mRNAs and ribosomes (Sheiness et al. 1975; Shaw & Kamen, 1986; Graves et al. 1987; Pachter et al. 1987; Brewer & Ross, 1988; Wilson & Treisman, 1988).

There are two implications of the idea that poly(A) shortening of spermatidal mRNAs is a consequence of translation. First, it raises the question whether the marked differences in the size distribution of poly(A) tracts on the polysomal forms of the five mRNAs studied here are due to differences in the rate of shortening of poly(A) tracts of different lengths or changes in the rate of translation at different stages of spermiogenesis. Second, it predicts that mRNAs that are synthesized and translated efficiently in round spermatids will be exclusively associated with polysomes with shortened poly(A) tracts. Similarly, mRNAs that are synthesized and translated inefficiently in round spermatids will be found in both polysomal and nonpolysomal fractions with shortened poly(A) tracts because the mRNAs interchange between the polysomal and nonpolysomal compartments. The mRNA for the testis-specific isozyme of lactate dehydrogenase (LDH-X) is synthesized and translated inefficiently in meiotic and haploid spermatogenic cells (Meistrich et al. 1977; Wieben, 1981; Fujimoto et al. 1988). In accordance with these predictions, the LDH-X mRNA is found in the polysomal and nonpolysomal fractions of meiotic and haploid stages with shortened, heterogenous poly(A) tracts (Fujimoto et al. 1988).

Although the function of poly(A) shortening in mRNA metabolism is disputed, most workers agree that the loss of poly(A) tracts from nonhistone mRNAs results in loss of mRNA function. In some cases, nonadenylated mRNAs are degraded rapidly (Marbaix et al. 1975; Brewer & Ross, 1988; Wilson & Treisman, 1988; Bernstein et al. 1989) and, in other cases, non-adenylated mRNAs are translated inefficiently (Jacobson & Favreau, 1983; Drummond et al. 1985; Galili et al. 1988). The conclusion that spermatidal mRNAs with poly(A) tracts 30–150 bases long are translationally active is consistent with this work because the loss of mRNA function is observed only when shortening reduces the poly(A) tracts to less than about 30 bases (Nudel et al. 1976; Jacobson & Favreau, 1983). It may be possible to elucidate whether poly(A) shortening on spermatidal mRNAs is associated with degradation or translational inhibition by microinjecting nonadenylated and polyadenylated mRNAs or by examining the size and distribution of nonadenylated mRNAs in sucrose gradients.

This work was supported by Grants DCB-8510350 and DCB-8710485 from the National Science Foundation.

Balhorn
,
R.
,
Weston
,
S.
,
Thomas
,
C.
&
Wyrobek
,
A. J.
(
1984
).
DNA packaging in mouse spermatids: synthesis of protamine variants and four transition proteins
.
Expl Cell Res
.
150
,
298
308
.
Bellvé
,
A. R.
,
Anderson
,
E.
&
Hanley-Bowdoin
,
L.
(
1975
).
Synthesis and amino acid composition of basic proteins in mammalian sperm nuclei
.
Devl Biol
.
47
,
349
365
.
Bellvé
,
A. R.
&
O’Brien
,
D. A.
(
1983
).
The mammalian spermatozoon: Structure and Temporal Assembly
.
In Mechanism and Control of Animal Fertilization
(ed.
J. F
Hartman
), pp.
55
137
.
New York
:
Academic Press
.
Bernstein
,
P.
,
Peltz
,
S. W.
&
Ross
,
J.
(
1989
).
The poly(A)-Poly(A)-binding protein complex is a major determinant of mRNA stability in vitro
.
Molec. cell. Biol
.
9
,
659
670
.
Brawerman
,
G.
(
1981
).
The role of the poly(A) sequence in mammalian messenger RNA
.
CRC Cm. Rev. Biochem
.
10
,
1
38
.
Brewer
,
G.
&
Ross
,
J.
(
1988
).
Poly(A) shortening and degradation of the 3’ A+U rich sequences of human c-myc mRNA in a cell-free system
.
Molec. cell. Biol
.
8
,
1697
1708
.
Drummond
,
D. R.
,
Armstrong
,
J.
&
Colman
,
A.
(
1985
).
The effect of capping and polyadenylation on the stability, movement and translation of synthetic messenger RNAs in Xenopus oocytes
.
Nuc. Acid. Res
.
13
,
7375
7394
.
Dworkin
,
M. B.
,
Rudensey
,
L. M.
&
Infante
,
A. A.
(
1977
).
Cytoplasmic nonpolysomal ribonucleoprotein particles in sea urchin embryos and their relationship to protein synthesis
.
Proc, natn. Acad. Sci. U.S.A
.
74
,
2231
2235
.
Feinberg
,
A. P.
&
Vogelstein
,
B.
(
1986
).
A technique for radiolabeling DNA restriction enzyme fragments to high specific activity
.
Analyt. Biochem
.
132
,
6
13
.
Fujimoto
,
H.
,
Erickson
,
R. P.
&
Tone
,
S.
(
1988
).
Changes in polyadenylation of lactate dehydrogenase-X mRNA during spermatogenesis in mice
.
Mol. Reprod. Dev
.
1
,
27
34
.
Galili
,
G.
,
Kawata
,
E. E.
,
Smith
,
L. D.
&
Larkins
,
B. A.
(
1988
).
Role of the 3’-poly(A) sequences in translational regulation of mRNAs in Xenopus laevis oocytes
.
J. biol. Chem
.
263
,
5764
5770
.
Graves
,
R. A.
,
Prandey
,
N. B.
,
Chodchoy
,
N.
&
Marzluff
,
W. F.
(
1987
).
Translation is required for regulation of histone mRNA degradation
.
Cell
48
,
615
626
.
Grimes
,
S. R.
, Jr
,
Meistrich
,
M. L.
,
Platz
,
R. D.
&
Hnilica
,
L. S.
(
1977
).
Nuclear protein transitions in rat testis spermatids
.
Expl Cell Res
.
110
,
31
39
.
Heidaran
,
M. A.
&
Kistler
,
W. S.
(
1987
).
Transcriptional and translational regulation of the mRNA for transition protein 1 in the rat
.
J. biol. Chem
.
262
,
13309
13315
.
Heidaran
,
M. A.
,
Showman
,
R. M.
&
Kistler
,
W. S.
(
1988
).
A cytochemical study of the transcriptional and translational regulation of nuclear transition protein 1 (TP1), a major chromosomal protein of mammalian spermatids
.
J. Cell Biol
.
106
,
1427
1433
.
Iatrou
,
K.
&
Dixon
,
G. H.
(
1977
).
The distribution of poly(A) + and poly(A)-protamine messenger RNA sequences in the developing trout testis as determined by protamine cDNA-RNA hybridization
.
Cell
10
,
433
441
.
Jackson
,
R. J.
(
1982
).
The cytoplasmic control of protein synthesis
.
In Protein Synthesis in Eukaryotes
(ed.
Perez-Bercaff
). pp.
363
418
.
Plenum Publishing Corp
.
New York
.
Jacobson
,
A.
&
Favreau
,
M.
(
1983
).
Possible involvement of poly(A) in protein synthesis
.
Nucleic Acid. Res
.
11
,
6353
6358
.
Kierszenbaum
,
A. L.
&
Tres
,
L. L.
(
1975
).
Structural and transcriptional features of the mouse spermatid genome
.
J. Cell Biol
.
65
,
258
270
.
Kistler
,
W. S.
,
Geroch
,
M. E.
&
Williams-Ashman
,
H. G.
(
1973
).
Specific basic proteins from mammalian testes: isolation and properties of small basic proteins from rat testes and epidiymal spermatozoa
.
J. biol. Chem
.
248
,
4532
4543
.
Kleene
,
K. C.
,
Borzorgzadeh
,
A.
,
Flynn
,
J. F.
,
Yelick
,
P. C.
&
Hecht
,
N. B.
(
1988
).
Nucleotide sequence of a cDNA clone encoding mouse transition protein 1
.
Biochim. biophys. Acta
950
,
215
220
.
Kleene
,
K. C.
,
Distel
,
R. J.
&
Hecht
,
N. B.
(
1983
).
cDNA clones encoding cytoplasmic poly(A)+ RNAs which are first present at detectable quantities in haploid phases of spermatogenesis in the mouse
.
Devl Biol
.
98
,
455
464
.
Kleene
,
K. C.
,
Distel
,
R. J.
&
Hecht
,
N. B.
(
1984
).
Translational regulation and deadenylation of a protamine mRNA during spermatogenesis in the mouse
.
Devl Biol
.
105
,
71
79
.
Kleene
,
K. C.
,
Distel
,
R. J.
&
Hecht
,
N. B.
(
1985
).
Nucleotide sequence of a cDNA clone encoding mouse protamine 1
.
Biochemistry
24
,
719
722
.
Kleene
,
K. C.
&
Flynn
,
J. F.
(
1987
).
Characterization of a cDNA clone encoding a basic protein, TP2, involved in chromatin condensation during spermiogenesis in the mouse
.
J. biol. Chem
.
262
,
17272
17277
.
Mali
,
P.
,
Sandberg
,
M.
,
Vuorio
,
E.
,
Yelick
,
P. C.
,
Hecht
,
N. B.
&
Parvinen
,
M.
(
1988
).
Localization of protamine 1 mRNA in different stages of the cycle of the rat seminiferous epithelium
.
J. Cell Biol
.
107
,
407
412
.
Marbaix
,
G.
,
Huez
,
G.
,
Burny
,
A.
,
Cleuter
,
Y.
,
Hubert
,
E.
,
Leclerq
,
M.
,
Chantrenne
,
H.
,
Soreq
,
H.
,
Nudel
,
U.
&
Littauer
,
U. Z.
(
1975
).
Absence of polyadenylate segment in globin messenger RNA accelerates its degradation in Xenopus oocytes
.
Proc. natn. Acad. Sci. U.S.A
.
72
,
3065
3067
.
Mayer
,
J. F.
,
Chang
,
T. S. K.
&
Zirkin
,
B. R.
(
1981
).
Spermatogenesis in the mouse. 2. Amino acid incorporation into basic nuclear proteins of mouse spermatids and spermatozoa
.
Biol. Reprod
.
25
,
1041
1051
.
Mayer
,
J. F.
&
Zirkin
,
B. R.
(
1979
).
Spermatogenesis in the mouse 1. Autoradiographic studies of nuclear incorporation and loss of 3H-amino acids
.
J. Cell Biol
.
81
,
403
410
.
Meistrich
,
M.
,
Bruce
,
W. R.
&
Clermont
,
Y.
(
1973
).
Cellular composition of fractions of mouse testis cells following velocity sedimentation separation
.
Expl Cell Res
.
79
,
213
227
.
Meistrich
,
M. L.
,
Reid
,
B. O.
&
Barcellona
,
W. J.
(
1976
).
Changes in sperm nuclei during spermiogenesis and epididymal maturation
.
Expl Cell Res
.
99
,
72
78
.
Meistrich
,
M. L.
,
Trostle
,
P. K.
,
Frapart
,
M.
&
Erickson
,
R. P.
(
1977
).
Biosynthesis and localization of lactate dehydrogenase X in pachytene spermatocytes and spermatids of mouse testes
.
Devl Biol
.
60
,
428
441
.
Merkel
,
C. G.
,
Wood
,
T. G.
&
Lingrel
,
J. B.
(
1976
).
Shortening of the poly(A) region of mouse globin messenger RNA
.
J. biol. Chem
.
251
,
5512
5515
.
Monesi
,
V.
,
Geremia
,
R.
,
D’Agostino
,
A.
&
Boitani
,
C.
(
1978
).
Biochemistry of male germ cell differentiation in mammals: RNA synthesis in meiotic and post-meiotic cells
.
Curr. Topics devl Biol
.
12
,
11
36
.
Nudel
,
LL
,
Soreq
,
H.
,
Littauer
,
U. Z.
,
Marbaix
,
G.
,
Huez
,
G.
,
Leclercq
,
M.
,
Hubert
,
E.
&
Chantrenne
,
H.
(
1976
).
Globin mRNA species containing poly(A) segments of different lengths: their functional stability in Xenopus oocytes
.
Eur. J. Biochem
.
64
,
115
121
.
O’Brien
,
D. A.
(
1987
).
Stage-specific protein synthesis by isolated spermatogeneic cells throughout meiosis and early spermiogenesis in the mouse
.
Biol. Reprod
.
37
,
147
157
.
Pachter
,
J. S.
,
Yen
,
T. J.
&
Clevelond
,
D. W.
(
1987
).
Autoregulation of tubulin expression is achieved through specific degradation of polysomal tubulin mRNAs
.
Cell
23
,
283
292
.
Palatnik
,
C. M.
,
Wilkins
,
C.
&
Jacobson
,
A.
(
1984
).
Translational Control during early Dictyostelium development: possible involvement of poly(A) sequences
.
Cell
36
,
1017
1025
.
Penman
,
S.
,
Vesco
,
C.
&
Penman
,
M.
(
1968
).
Localization and kinetics of formation of nuclear heterodisperse RNA, cytoplasmic heterodisperse RNA and polyribosome-associated messenger RNA in HeLa cells
.
J. molec. Biol
.
34
,
149
169
.
Romrell
,
L. J.
,
Bellve
,
A. R.
&
Fawcett
,
D. W.
(
1976
).
Separation of mouse spermatogenic cells by velocity sedimentation
.
Devl Biol
.
49
,
119
131
.
Rosenthal
,
E. T.
,
Tansey
,
T. R.
&
Ruderman
,
J. V.
(
1983
).
Sequence specific adenylations and deadenylations accompany changes in translation of maternal messenger RNAs after fertilization of Spisula oocytes
.
J. molec. Biol
.
166
,
309
327
.
Shaw
,
G.
&
Kamen
,
R.
(
1986
).
A conserved AU sequence from the 3’ untranslated region of GM-CSF mRNA mediates selective mRNA degradation
.
Cell
46
,
659
667
.
Sheiness
,
D.
&
Darnell
,
J. E.
(
1973
).
Poly(A) segment in RNA becomes shorter with age
.
Nature New Biol
.
241
,
265
268
.
Sheiness
,
D.
,
Puckett
,
L.
&
Darnell
,
J. E.
(
1975
).
Possible relationship of poly(A) shortening to mRNA turnover
.
Proc, natn. Acad. Sci. U.S.A
.
72
,
1077
1081
.
Thomas
,
P. C.
(
1980
).
Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose
.
Proc. natn. Acad. Sci. U.S.A
.
77
,
5201
5205
.
Walden
,
W. E.
&
Thach
,
R. E.
(
1986
).
Translational control of gene expression in a normal fibroblast. Characterization of a subclass of mRNAs with unusual kinetic properties
.
Biochemistry
25
,
2033
2041
.
Weber
,
L. A.
,
Simili
,
M.
&
Baglioni
,
C.
(
1979
).
Binding of viral and cellular messenger RNAs to ribosomes in eukaryotic cell extracts
.
Methods in Enzymology LX
, pp.
351
360
.
Wieben
,
E.
(
1981
).
Regulation of the synthesis of lactate dehydrogenase-X during spermatogenesis in the mouse
.
J. Cell Biol
.
88
,
492
498
.
Wilson
,
T.
&
Treisman
,
R.
(
1988
).
Removal of poly(A) and consequent degradation of c-fos mRNA facilitated by 3’ AU-rich sequences
.
Nature, Lond
.
336
,
396
399
.
Yelick
,
P. C.
,
Balhorn
,
R.
,
Johnson
,
P. A.
,
Corzett
,
M.
,
Mazrimas
,
J.
,
Kleene
,
K. C.
&
Hecht
,
N. B.
(
1987
).
Mouse protamine 2 is synthesized as a precursor whereas mouse protamine 1 is not
.
Molec. cell Biol
.
7
,
2173
2179
.