The Biosynthesis and Processing of Vitellogenin in the Fat Bodies of Females and Males of the Cockroach Leucophaea maderae

Pergamon

PII: S0965-1748(97)00071-4

Insect Biochem. Molec. Biol. Vol. 27, No. 11, pp. 901–918, 1997  1998 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0965-1748/98 $19.00 + 0.00

The Biosynthesis and Processing of Vitellogenin in the Fat Bodies of Females and Males of the Cockroach Leucophaea maderae GRACE DON-WHEELER,† FRANZ ENGELMANN*† Received 22 April 1997; revised and accepted 29 August 1997

The juvenile hormone analog (JHA) methoprene was used to induce the synthesis of the yolk protein precursor vitellogenin (Vg) in adult females and males of the cockroach Leucophaea maderae. The female- and male-produced vitellogenin (VgF and VgM, respectively) contained polypeptides of 112, 95, 92, and 54 kDa. Also present in the secreted vitellogenins was a small quantity of a short-lived transitional 155 kDa Vg polypeptide, and a variable amount of an 85 kDa species. Quantitatively, the VgF and VgM were significantly different in the Vg112 and Vg95 units (VgF ⬎ VgM), and in the Vg85 polypeptide (VgF ⬍ VgM). In the present study, the biosynthesis of Vg precursors in the fat bodies of females and males was examined using a short radiopulse with 35S-methionine/cysteine and 32P-orthophosphate. The glycosylation of the Vg precursors was examined by digestion with endoglycosidase H and by the inhibition of N-linked glycosylation with tunicamycin. The data showed that in both females and males, the synthesis of the vitellogenin precursor occurred in a stepwise fashion: (1) the co-translational glycosylation of Vg203; (2) the post-translational phosphorylation of Vg203 to form Vg220; (3) the proteolytic processing of Vg220 to form the constituent Vg polypeptides. The 203 and 220 kDa Vg precursors of females and males appeared to be similarly glycosylated and phosphorylated. The additional processing of Vg112 to Vg85 was more pronounced in the fat bodies of males than in females, and appears to account for the quantitative difference in the distribution of these polypeptides in VgF and VgM. Finally, the major oligosaccharides of VgF and VgM appear to be those of N-linked mannose residues. The treatment of females and males with tunicamycin indicated that the co-translational glycosylation of Vg precursors was required for the phosphorylation of the Vg precursor, as well as the secretion of Vg from the fat body.  1998 Elsevier Science Ltd. All rights reserved Juvenile hormone Vitellogenin

Glycosylation Post-translational modification

INTRODUCTION

In most species of insects, the synthesis of the major yolk protein precursor vitellogenin (Vg) is induced by the juvenile hormones (JH) and is normally limited to the fat body of the adult female. The specificity of Vg synthesis is in part regulated by the JH titer, and in part, by stageand sex-related differences in the responsiveness of the fat body to JH. Although JH is present in both sexes during most stages except for the last instar, the vitellogenic response of the fat body is normally seen only in adult females. The vitellogenic responsiveness of the adult female is thought to reflect the sexual differen-

*Author for correspondence. †Department of Biology, University of California, Los Angeles, CA 90024, USA

tiation of the fat body which is completed at the metamorphic molt. In some hemimetabolous insects such as locusts and cockroaches, it is possible to induce some Vg synthesis in adult males and in last instar nymphs of both sexes with large doses of JH analogs (Mundall et al., 1979; Dhadialla and Wyatt, 1983; Don-Wheeler and Engelmann, 1991). These induced males and nymphs synthesize only a fraction of the Vg produced by comparably treated adult females. For Locusta (Dhadialla and Wyatt, 1983) and the cockroach Leucophaea maderae (Don-Wheeler and Engelmann, 1991), the development of vitellogenically competent fat body is seen in both sexes during the last nymphal instar. Furthermore, in both species, the sexual dimorphism of the fat body is established during the last nymphal instar as seen by a higher production of Vg in the fat body of the female than that of the male. Although the magnitude of vitello-

901

902

GRACE DON-WHEELER AND FRANZ ENGELMANN

genin synthesis in the fat bodies of induced females and males has been described, less is known about the biosynthesis and processing of vitellogenin in the sexually dimorphic fat bodies. In Leucophaea the Vg precursor of the female fat body is synthesized as a high molecular weight phosphorylated protein, which is cleaved within the fat body to yield the four major Vg polypeptides (della-Cioppa and Engelmann, 1987). These polypeptides are assembled into the protein backbone of the secreted Vg dimer. In the current investigation, the sequence of events leading to the synthesis of Vg precursors and mature Vg polypeptides was determined by examining the glycosylation, phosphorylation, and proteolytic processing of Vg in the fat body. This study was in part prompted by our earlier findings that vitellogenins produced by females and males were similar in the native molecular weights, but different in the composition of polypeptides (Don-Wheeler and Engelmann, 1991). It is unclear how the different polypeptide compositions of the female- and male-produced vitellogenins (VgF and VgM, respectively) affect the function of Vg. It has been shown that competent ovaries implanted into vitellogenic males failed to complete vitellogenesis, despite the presence of relatively high titers of Vg (10 ␮g/␮l) in the hemolymph. Because of the different polypeptide composition of the female- and maleproduced vitellogenins, it was relevant to ask if this reflects the sexual dimorphic biosynthesis of Vg. This question was addressed by examining the co- and posttranslational processing of the Vg precursors in the vitellogenic fat bodies of adult females and males. MATERIALS AND METHODS

Chemicals Most of the reagents used in these studies were obtained from Sigma (St Louis, MO). The chemicals used for electrophoresis were purchased from Gibco– BRL (Gaithersburg, MD). Endoglycosidase H was from Boehringer–Mannheim (Indianapolis, IN) and alkaline phosphatase from USB (Cleveland, OH). Tunicamycin and phenylmethylsulfonyl fluoride (PMSF) were obtained from Calbiochem (San Diego, CA). The substrate 4-chloro-1-naphthol was from Bio-Rad. The radioisotopes and Ecolite scintillation cocktail were purchased from ICN Radiochemicals (Costa Mesa, CA). Animals Animals were reared on a 12:12 h light/dark cycle at 26°C, 75% relative humidity. Adults were collected shortly after molt and kept in plastic cages (six to eight animals per cage). Animals were freely fed rat chow and water. Allatectomy, ovariectomy, and methoprene treatment Adults were allatectomized on the fourth day after emergence. Females and males were narcotized with car-

bon dioxide, and the corpora allata were removed through an incision made in the neck membrane with watchmaker forceps. A few grains of penicillin were applied to the wound and the incisions were sealed with warm wax. The JHA methoprene (a gift from Zoecon) was diluted in acetone and applied topically to the dorsal abdominal segments. Methoprene was applied in a single dose ranging from 0.5 to 400 ␮g. The control animals were treated with acetone only. Radiolabelled vitellogenins Radiolabelled vitellogenins were obtained by pulsing methoprene-treated vitellogenic females and males in vivo with either 1 ␮Ci of 14C-leucine (specific activity 184 mCi/mmol), 25–100 ␮Ci of Trans35S-label (35Smethionine/cysteine; specific activity 1082 mCi/mmol), 0.5–1 mCi of 32P-orthophosphate (285 Ci/mg, carrier free), or 10 ␮Ci of 3H-mannose (specific activity 28 Ci/mmol). For the radiopulse in vivo, the isotopes were injected into the hemocoel of the cockroaches with a Hamilton 10 ␮l syringe; the animals were pulsed from 0.5 to 24 h. After the pulse, the animals were neck-ligated and a puncture wound was made in the dorsal neck membrane posterior to the ligation. The animals were placed head down in a Kolmer centrifuge tube and the hemolymph collected by low speed centrifugation for 5 min. Homogenization of the fat body: isolation of the microsomal fraction The fat body was dissected and rinsed in ice cold cockroach saline (0.005 M KCl, 0.19 M NaCl, 0.001 M CaCl2, containing 0.001 M PMSF), cleaned of the adherent tissues, and stored at −90°C until homogenization. All of the following operations were performed at 4°C. The fat bodies were homogenized in about 3–4 volumes of sucrose-TKM (0.035 M sucrose, 0.05 M Tris, pH 7.6, 0.025 M KCl, 0.035 M KHCO3, 0.01 M Mg acetate, 0.001 M EDTA, and 0.001 M PMSF) using 20 strokes in Potter–Elvehjem hand-held homogenizer (della-Cioppa and Engelmann, 1984). The homogente was centrifuged at 1000g for 3 min to remove the nuclei and large debris. The supernatant was then spun at 9000g for 10 min in a Beckman L2-65B ultracentrifuge to pellet the mitochondria. Finally, the post-mitochondrial supernatant was centrifuged for 1 h at 105,000g to pellet the microsomal fraction. The microsomal proteins were solubilized in membrane protein solubilization buffer containing 0.05 M Tris (pH 7.4) 0.4 M NaCl, 0.1% sodium dodecyl sulfate (SDS), 1% (w/v) Nonidet P-40, 1% sodium deoxycholate, and 0.001 M PMSF. The solubilized proteins were aliquoted and stored at −90°C. The protein content was determined by the Lowry protein assay (Lowry et al., 1951). The solubilized microsomal proteins or immunoprecipitated microsomal vitellogenin were analysed by SDS–polyacrylamide gel electrophoresis (PAGE).

BIOSYNTHESIS OF VITELLOGENIN

Immunological methods Antibodies against the Leucophaea Vg and vitellin were prepared by immunizing rabbits with purified vitellin. The immunoglobulins were partly purified and concentrated by the precipitation of the rabbit serum through 40% saturated ammonium sulfate (Mishell and Shiigi, 1980). The non-specific antibodies were removed from the rabbit serum by absorption with adult male hemolymph antigens. The radiolabelled vitellogenins from the hemolymph, fat body homogenates, or the solubilized microsomal fraction of the fat body were isolated by immunoprecipitation according to della-Cioppa and Engelmann (1984). Hemolymph samples were diluted with Tris buffered saline (0.05 M Tris, pH 7.5, 0.4 M NaCl) and then precipitated with anti-vitellogenin for 1–2 h at 4°C. Unlabelled Vg was added to samples containing low concentrations of Vg to facilitate antibody precipitation. The precipitates were pelleted by centrifugation in a Beckman microfuge B and washed four times with immunoprecipitation buffer (0.1 M Tris, pH 7.5, 0.4 M NaCl, 1% (w/v) Triton X-100, and 1% sodium deoxycholate). The pellet was solubilized in 25 ␮l of 1 N NaOH and added to 2 ml of Ecolite scintillation cocktail. The radioactivity was determined by liquid scintillation spectrometry in a Beckman LS 7000. Polyacrylamide electrophoresis The native molecular weights of the female- and maleproduced vitellogenins and vitellins were determined by non-denaturing PAGE (Davis, 1964; Sigma Chemical Co., 1986). Purified vitellogenins and vitellins were electrophoresed on different percentage (4–7%) polyacrylamide slab gels. The molecular mass was determined according to Hedrick and Smith (1968). The molecular weight standards used for this analysis were the bovine serum albumin (BSA) monomer (67 kDa) and dimer (134 kDa), and the urease trimer (270 kDa) and hexamer (540 kDa). The proteins were visualized by staining the gels with 0.25% Coomassie blue R-250 in 50% methanol–10% acetic acid, followed by destaining in 20% methanol–10% acetic acid. SDS–PAGE was performed according to Laemmli (1970). The proteins were solubilized in Laemmli sample buffer (0.0625 M Tris, pH 6.8, 2.3% SDS, 5% 2-mercaptoethanol, 10% (w/v) glycerol) and electrophoresed on 6 or 7.5% SDS–polyacrylamide gels. The polypeptides were visualized by staining with Coomassie blue R-250. High molecular weight protein standards ranging from 29 to 205 kDa were electrophoresed in a separate well and used to construct a calibration curve (molecular mass vs. relative mobility). Protease digestion and peptide map The fragment compositions of the vitellogenin polypeptides were determined by protease digestion with Staphylococcus aureus V-8 protease (Cleveland et al., 1977). 35S- or 32P-labelled Vg polypeptides were frac-

903

tionated by 5 or 6% SDS–PAGE and stained with Coomassie blue R-250. The polypeptides were excised from the gels and equilibrated for 1 h in Cleveland sample buffer (0.125 M Tris, pH 6.8, 0.1% SDS, 0.001 M EDTA, 0.3% 2-mercaptoethanol, 10% (w/v) glycerol). Individual gel slices were placed into a well of a 15% or 20% SDS–polyacrylamide gel (Giulian et al., 1985), and then overlayed with 10 ␮g of V-8 protease in Cleveland sample buffer. The digestion of the polypeptides was effected by electrophoresis at 5 mA in a 4 cm stacking gel. The fragment compositions of the Vg polypeptides were revealed by autoradiography. Low molecular weight markers ranging from 3 to 45 kDa (Gibco–BRL) were electrophoresed in a separate well and used to construct a calibration curve (molecular mass vs. mobility). Fluorography and autoradiography Polyacrylamide gels containing 35S-labelled proteins were fixed and stained in 0.25% Coomassie blue R-250 in 40% methanol-10% acetic acid. Excess water was removed by soaking the gels in two changes of dimethyl sulfoxide (DMSO) for a total of 40 min. The gels were fluorographed in 22% 2,5 diphenyloxazole in DMSO for 45 min followed by a 30 min wash with distilled water (Bonner and Laskey, 1974). The gels were laid onto Whatman 3 MM paper and then heat dried under vacuum on a Bio-Rad gel dryer. Gels containing 32P-labelled Vgs were fixed and stained and then dried directly onto Whatman 3 MM paper. Kodak X-AR film was used for both 35 S and 32P autoradiography. Endoglycosidase H digestion The presence of N-linked oligosaccharides in the Leucophaea Vg was determined by endoglycosidase H (endo H) digestion according to Trimble and Maley (1984). Thirty micrograms of either purified vitellogenin or microsomal vitellogenin (solubilized from the microsomal fraction of the fat body) were denatured by the addition of 36 ␮g of SDS and boiling for 1 min. Vitellogenin was then incubated with 30 mU of endo H in 0.05 M sodium phosphate (pH 6.0) and 1 × 10−4 M PMSF for 16 h at 30°C. The control treatment was the incubation of Vg in phosphate buffer alone. The reaction was terminated by adding an equal volume of 2 × Laemmli sample buffer followed by boiling for 1 min. The endo H- and non-digested vitellogenins were electrophoresed on 7.5% SDS–polyacrylamide gels. The Nlinked glycosylation of the Vg polypeptides was determined by comparing the molecular mass of polypeptides subjected to either the endo H- or control treatment. Concanavalin A binding The lectin concanavalin A was used to detect the carbohydrate moieties contained in the Vg polypeptides. The Vg polypeptides were separated by SDS–PAGE, and then transferred to nitrocellulose in a Bio-Rad Trans-Blot cell at 160–180 mA for 20 h (Towbin et al., 1979). The binding of concanavalin A was performed according to

904

GRACE DON-WHEELER AND FRANZ ENGELMANN

Clegg (1982). The nitrocellulose blots were blocked with phosphate-buffered saline (PBS)-5% BSA for 1 h, and then incubated for 30 min in 20 ml of PBS containing 10 ␮g/ml of horseradish peroxidase-labelled concanavalin A (Con A-HRP). PBS was composed of 0.025 M sodium phosphate (pH 6.5) 0.15 M NaCl, 0.5% (w/v) Triton X-100, and 1 × 10−5 M each of CaCl2, MnCl2, MgCl2. The blots were washed three times with Trisbuffered saline (TBS; 0.02 M Tris, pH 7.5, 0.4 M NaCl), and the binding of Con A-HRP was visualized by substrate development in a solution containing 50 ml of TBS, 3 mg/ml of 4-chloro-1-naphthol in 10 ml of methanol, and 30 ␮l of 30% (v/v) hydrogen peroxide. Alkaline phosphatase digestion The phosphorylation of the Vg precursors was assessed by alkaline phosphatase digestion of solubilized microsomal proteins of the fat body. Twenty micrograms of protein were incubated with 10 U calf alkaline phosphatase in 0.05 M Tris (pH 8.5) 1 × 10−4 M EDTA for 2 h at 37°C (Sherod et al., 1970). The control treatment was the incubation of the microsomal vitellogenins in Tris buffer alone. The digestion of the samples was terminated by adding an equal volume of 2 × Laemmli sample buffer and boiling. The samples were analyzed by 6% SDS–PAGE. Tunicamycin treatment The synthesis of Vg in allatectomized adult females was induced with 10 ␮g of methoprene and in allatectomized adult males with a 400 ␮g dose. On the fifth day of treatment, animals were injected with 12 ␮g of tunicamycin (dissolved in 3 ␮l of 2:1 DMSO:cockroach saline). The control animals were injected with DMSOcockroach saline only. After 4 h, animals were pulsed with the radioisotopes, 3H-mannose, Trans35S-label, or 32 P-orthophosphate. The radiolabelled Vgs of the hemolymph and the microsomal Vgs of the fat body were antibody-precipitated and analyzed by SDS–PAGE. RESULTS

The female- and male-produced vitellogenins: the native molecular mass and polypeptide composition Vitellogenins from the hemolymph of females and males were purified by QAE ion-exchange chromatography as previously described (Don-Wheeler and Engelmann, 1991), and then resolved on non-denaturing polyacrylamide gels. The native molecular mass of both female- and male-produced vitellogenins (VgF and VgM) was approximately 540 kDa (Fig. 1(A)). The molecular mass of VgF and VgM was comparable to that previously reported for the Leucophaea vitellogenin (520 kDa) (Engelmann et al., 1976). Differences in the polypeptide structures of VgF and VgM were revealed by SDS–PAGE. VgF was composed of polypeptides of 112, 95, 92, 54 kDa, and quantitatively

small amounts of an 85 kDa species (Fig. 1(B)). The polypeptide composition of VgM was similar with respect to the quantities of the Vg92 and Vg54 units, but there was quantitatively less of the Vg112 and Vg95 polypeptides (Fig. 1(B)). The Vg85 polypeptide was more prominent in VgM than in VgF (Fig. 1(B)). The biosynthesis of phosphorylated Vg precursors and polypeptides in the fat bodies of vitellogenic females and males The biosynthesis of the Vg polypeptides in the fat bodies of induced females and males was determined during an in vivo radiopulse of methoprene-treated animals with 32 P-orthophosphate. The Vg polypeptides of the microsomal fraction of the fat bodies were fractionated by SDS– PAGE and visualized by autoradiography. During a brief pulse of females (15 min) the synthesis of a 220 kDa polypeptide (Vg220) and the smaller Vg polypeptides, Vg155, Vg112, and Vg54, was evident (Fig. 2(A)). In males pulsed for 30 min, the Vg220 species was apparent as well as the same small Vg polypeptides seen in the female fat body (Fig. 2(B)). For both sexes, a 1 h radiopulse resulted in an increased radiolabelling of the smaller Vg polypeptides, i.e., Vg155, Vg112, Vg92, Vg85, Vg54 (Fig. 2(A,B)). An 195 kDa polypeptide (Vg195) was present in the fat body of the female but not in that of the male (Fig. 2(A,B)); possibly the Vg195 is made only in small amounts or is rapidly processed in the male fat body. Despite this difference, the overall compositions of Vg polypeptides of the fat bodies of females and males were comparable. The time course of Vg synthesis and secretion in vivo The synthesis and secretion of VgF and VgM were analyzed in vitellogenic females and males by radiopulsing with Trans35S-label in vivo for 0.5–4 h. The 35S-labelled Vg polypeptides of the fat body (microsomal fraction) and the hemolymph were fractionated by SDS– PAGE (Fig. 3(A,B)). Because of the low rates of Vg synthesis in males, the 35S autoradiogram of the microsomal vitellogenins was exposed for a longer time (7 days) than that used for the females (2 days). For the male fat body, the radiolabelling of the Vg polypeptides and several additional polypeptides was seen; the latter polypeptides were presumably degradation products (Fig. 3(B)). We analyzed the putative Vg precursors and Vg polypeptides that were previously identified by 32P autoradiography (Fig. 2(A,B)). For the female, already during the first 30 min of radiopulse, some processing of the Vg precursor was apparent as evidenced by the radiolabelling of Vg195 and the smaller Vg polypeptides: Vg155, Vg112, Vg95, Vg92, and Vg85 (Fig. 3(A)). After 1 h of radiopulse, the synthesis of Vg220 and a 203 kDa species (Vg203) was seen (Fig. 3(A)). The 203 kDa polypeptide was poorly labelled and was obscured on the autoradiogram by the more heavily labelled Vg220 and Vg195. The low radiolabelling of Vg203 suggests that this polypeptide does

BIOSYNTHESIS OF VITELLOGENIN

905

FIGURE 1. The electrophoresis of the female- and male-produced vitellogenins (VgF and VgM) on (A) 5% native polyacrylamide gel and (B) 7.5% SDS–polyacrylamide gel. The vitellogenins were isolated by QAE-ion exchange column chromatography. VgF and VgM were visualized by staining with Coomassie blue R-250. The indicated molecular standards for native PAGE were the urease hexamer (540 kDa) and trimer (270 kDa), and the BSA dimer (134 kDa) and monomer (67 kDa). For SDS–PAGE, the molecular weights of the Vg polypeptides are given.

not accumulate in the fat body and may be rapidly modified to another form. During the 2–4 h of the pulse, there was a progressive decrease in the radiolabelling of the Vg220 and Vg195 units of the females which was concomitant with an increase in the radiolabelling of the smaller VgF polypeptides of the fat body, and the secreted VgF of the hemolymph (Fig. 3(A)). The large polypeptides Vg220, Vg203, and Vg195 were found only in the fat body, whereas the smaller Vg polypeptides of the fat body were obviously assembled into the secreted protein (Fig. 3(A)). The Vg85 unit was apparent intracellularly in the female fat body yet did not appear to be a major component of the secreted Vg (Fig. 3(A)). The synthesis of Vg in the male fat body was similar except that the radiolabelling of Vg polypeptides was less than that found for females (Fig. 3(B)). The fat body of the male contained the large polypeptides Vg220, Vg203, and Vg195 as well as the smaller units Vg155, Vg112, Vg92, and Vg85 (Fig. 3(B)). The radioactivity of the Vg polypeptides of the male fat body declined during the second to fourth hour of the pulse and corresponded to an increase in the radiolabelling of the secreted VgM after the second hour of the pulse (Fig. 3(B)). The synthesis of Vg112 was apparent in the male fat body, but this polypeptide was not always seen in the secreted Vg (Fig. 3(B) and Fig. 9(B)). Likewise, the Vg95 unit was quantitatively reduced in the fat bodies of males, and in

the secreted VgM of the hemolymph (Fig. 1(B) and Fig. 3(B)). The synthesis of the high molecular weight Vg polypeptides of the fat body The transient appearance of Vg220, Vg203, and Vg195 during the Trans35S-label radiopulse of vitellogenic females and males suggested that these polypeptides were modified to another form (Fig. 3(A,B)). The synthesis of these high molecular weight Vg polypeptides was analysed by pulsing females and males with Trans35S-label for brief periods ranging from 15 to 90 min. The microsomal Vgs of the fat body were separated by SDS–PAGE. During the first 15 min of the pulse, the synthesis of Vg195 and Vg203 was evident in the fat bodies of females and males (Fig. 4(A,B)). After 30 min of radiopulse, Vg220 was detected in the fat bodies of both sexes. The quantities of these three polypeptides appeared to decline after 90 min of radiopulse (Fig. 4(A,B)). The glycosylation and phosphorylation of Vg220, Vg203, and Vg195 were examined by the digestion of the microsomal vitellogenins with endoglycosidase H (endo H) or with alkaline phosphatase (AP), respectively. Under the conditions of the endo H digestion (16 h at 30°C), Vg195 was not apparent in the control samples (Fig. 5(A)); possibly this polypeptide was proteolyzed

906

GRACE DON-WHEELER AND FRANZ ENGELMANN

FIGURE 2. The 32P autoradiogram of the microsomal vitellogenins of the fat bodies of (A) females and (B) male. Methoprenetreated adult females and males were pulsed with 0.5 mCi of 32P-orthophosphate for the indicated times in vivo. The microsomal fraction of the fat body was isolated by the differential centrifugation of tissue homogenates. The microsomal proteins were solubilized in Tris buffer containing 0.1% SDS, 1% NP-40, and 1% sodium deoxycholate. The vitellogenins were isolated by the immunoprecipitation of solubilized microsomal proteins. The Vg polypeptides were separated by 7.5% SDS–PAGE.

during this long incubation. For females, both Vg220 and Vg203 were sensitive to endo H and appeared to be comparable in the degree of glycosylation (Fig. 5(A)). The Vg203 species was digested with endo H to yield an 190 kDa polypeptide (Fig. 5(A)). Because the synthesis of a 190 kDa Vg species was normally not seen in the fat bodies of vitellogenic females or males, Vg203 is presumably synthesized by the co-translational glycosylation of the primary translation product. The Vg220 of the female fat body was sensitive to AP digestion, whereas Vg203 did not appear to be digested with this enzyme. The absence of phosphorylation in Vg203 was further supported by the previous data showing that Vg203 was not radiolabelled with 32P-orthophosphate (Fig. 2(A,B)). Moreover, the AP digestion of Vg220 appeared to increase the quantity of the 203 kDa Vg species. The non-phosphorylated 203 kDa polypeptide produced by AP digestion was presumably equivalent to Vg203. These data suggest that Vg203 is normally phosphorylated to form Vg220. Finally, for the female Vg, we assume that the alkaline phosphatase digestion of Vg195

generated the small amount of a 178 kDa polypeptide (Fig. 5(B)). The degree of phosphorylation of Vg220 and Vg195 polypeptides appeared to be similar. The glycosylation of the Vg220 of males could not be determined by endo H digestion as this polypeptide and its digestion products were not stable under the assay conditions. However, the Vg203 of males was stable and was digested with endo H to yield an 190 kDa polypeptide (Fig. 5(A)). The Vg220 of males was digested with AP to yield a 203 kDa polypeptide (Fig. 5(B)). Based on these data, it appears that the Vg220 and Vg203 units of both sexes are similarly modified with respect to glycosylation and phosphorylation. The proteolytic processing of the Vg polypeptides The precursor–product relationships of the Vg polypeptides were assessed by the protease digestion and “peptide mapping” of the high molecular weight Vg polypeptides (Vg220 and Vg195), and the smaller Vg polypeptides (Vg155, Vg112, Vg95, Vg92, Vg54, Vg85). The radiolabelled (32P or 35S) Vg of adult females was

BIOSYNTHESIS OF VITELLOGENIN

FIGURE 3. The synthesis, processing, and secretion of vitellogenins in the fat bodies of (A) adult females and (B) adult males during a 4 h radiopulse in vivo. Vg synthesis in females and males was assessed on the sixth day after methoprene-treatment (400 ␮g) by radiopulsing with 50 ␮Ci of Trans35S-label from 0.5 to 4 h. The radiolabelled vitellogenins of the fat body were immunoprecipitated from 50 ␮g of solubilized fat body microsomal proteins. The secreted vitellogenins were immunoprecipitated from 1 ␮l of hemolymph. The microsomal and secreted Vgs were analysed by 7.5% SDS–PAGE and autoradiography. Arrows denote the Vg precursors and Vg polypeptides previously identified by 32P-autoradiography (Fig. 2(A,B)).

FIGURE 4. The time course of the synthesis of the Vg precursors of (A) adult females and (B) adult males. Methoprenetreated females and males were pulsed with 50 ␮Ci of Trans35S-label in vivo from 15 to 90 min. The radiolabelled vitellogenins of the fat bodies were immunoprecipitated from 50 ␮g of solubilized fat body microsomal protein. The Vg precursors were identified by 6% SDS–PAGE and autoradiography.

907

908

GRACE DON-WHEELER AND FRANZ ENGELMANN

FIGURE 5. The digestion of the Vg precursors of females and males with (A) endoglycosidase H (endo H) and (B) alkaline phosphatase. The Vg precursors were radiolabelled in vivo with 50 ␮Ci Trans35S-label. The Vg contained in 20 ␮g of solubilized fat body microsomal proteins was digested with 10 mU of endo H or with 10 U of AP. The control treatment was the incubation of the microsomal proteins in buffer only: 0.05 M Tris (pH 8.0) (alkaline phosphatase), or 0.05 M sodium phosphate (pH 6.0) (endo H). The microsomal vitellogenins were isolated by immunoprecipitation. The Vg precursors were resolved by 6% SDS–PAGE. The molecular mass of the deglycosylated or dephosphorylated Vg precursors was compared with that of the undigested Vg precursors of the respective control treatment. (+) enzymatic digestion; (−) undigested control.

of a high specific activity which facilitated the detection of labelled fragments on autoradiograms. Although the 32 P-VgM was not as highly labelled as that of 32P-VgF, the pattern of 32P-fragments of the smaller VgM polypeptides (Vg112, Vg92, Vg85, Vg54) appeared to be similar to those of the corresponding VgF polypeptides (Fig. 6(B,C)). In the following analysis of 32P-VgF, we assume that Vg220 is the single phosphorylated precursor from which all of the phosphorylated Vg polypeptides are ultimately derived. The fragment compositions of Vg220 and Vg195 were nearly identical suggesting that Vg220 is cleaved to form Vg195 (Fig. 6(A)). The different fragment compositions of Vg155 and Vg112 indicated that these two polypeptides were unrelated and derived from different portions of the Vg220 (Fig. 6(A,B)). Because of the large molecular sizes of Vg195, Vg155, and Vg112, it appears that the cleavage of these polypeptides from Vg220 are mutually exclusive. A limitation of the peptide map of the 32P-labelled Vg was that this analysis only detected phosphorylated fragments. A more complete representation of the fragment compositions of the Vg polypeptides was obtained by the peptide map of 35Slabelled Vg. The 35S-fragment compositions of Vg220 and Vg195 were nearly the same except for a 43 kDa fragment that was only present in Vg220 (Fig. 7). Because Vg220 and Vg112 contained several of the same fragments including the 43 kDa species (Fig. 7), Vg112 is apparently derived from Vg220 and not Vg195. The Vg195 may instead be processed to Vg155 because these two polypeptides have several small fragments in common (6–18 kDa; Fig. 6(A,B) and Fig. 7). The commonality of fragments between Vg155 and those of Vg95,

Vg92, and Vg54 suggest that these smaller units are derived from Vg155 (Fig. 6(A,B)). Also, Vg95 and Vg92 have nearly identical fragment compositions. Although Vg85 was present in the fat bodies of females and males (Fig. 2(A,B) and Fig. 3(A,B)), it was a major polypeptide of VgM, whereas it was only a minor component of VgF (Fig. 1(B) and Fig. 3(A,B)). The fragment compositions of the Vg85 of females and males appear to be comparable suggesting that this polypeptide is derived through a common pathway in the respective fat bodies (Fig. 6(C) and Fig. 7). The Vg85 unit could be derived from the cleavage of either the Vg220, Vg155, or Vg112 polypeptides because some of the 32P-labelled fragments of Vg85 (14–20 kDa) were also contained in the larger polypeptides (Fig. 6(B,C)). The peptide maps of 35S-labelled Vg polypeptides showed that Vg112 and Vg85 had similar fragment patterns (Fig. 7); Vg112 is possibly cleaved to Vg85. It is noteworthy that a 25 kDa fragment in Vg112 was phosphorylated, whereas the same fragment in Vg85 was not radiolabelled with 32Porthophosphate (Fig. 6(C) and Fig. 7). The glycosylation of the secreted vitellogenins of females and males The glycosylation of the polypeptides of the secreted VgF and VgM was ascertained by endo H digestion and by the detection of carbohydrates with peroxidase-labelled concanavalin A (Con A-HRP). Vitellogenins that were subjected to either endo H or control treatment were fractionated by 7.5% SDS–PAGE, and then transferred to nitrocellulose filters. The polypeptides were visualized by staining with Coomassie blue R-250, or by the bind-

BIOSYNTHESIS OF VITELLOGENIN

909

FIGURE 6. The 32P autoradiogram of the “peptide map” of the Vg polypeptides. The vitellogenins of females and males were radiolabelled in vivo during a 2 h pulse with 32P-orthophosphate. The 32P-Vgs were isolated from the microsomal fraction of the fat body or from the hemolymph and subjected to 5% SDS–PAGE. The PAGE-fractionated Vg polypeptides were digested with 10 ␮g of Staphylococcous aureus V-8 protease during the electrophoresis through a 5% stacking gel. The resulting fragments were separated by a 20% SDS–PAGE and visualized by autoradiography. The molecular weight of the various Vg fragments are given. The proposed processing schemes for the Vg220 precursor (A,B); the Vg155 polypeptide (B); and Vg112 polypeptide (C) are given below the autoradiogram. The grouping of the polypeptides was based on the commonalities in the fragment compositions as shown in (A–C).

ing of Con A-HRP to the carbohydrate moieties. For all of the polypeptides of VgF and VgM, endo H effected a reduction in the molecular weights (Fig. 8(A)) and the loss of Con A-HRP binding (Fig. 8(B)). This loss of lectin binding showed that endo H treatment causes the quantitative release of oligosaccharides from VgF and VgM. Endo H effected a similar reduction in the molecular mass of the corresponding polypeptides of VgF and VgM (112, 92, 54 kDa) (Fig. 8(A)). There appears to be no quantitative differences in the glycosylation of corresponding polypeptides of VgF and VgM. Glycosylation and the biosynthesis of vitellogenin The function of N-linked glycosylation in Vg synthesis, processing, and secretion was determined with tunicamycin treatment of vitellogenic females and males. The dose of tunicamycin used in the present study was based on preliminary experiments in which methoprenetreated females (10 or 25 ␮g of JHA) were injected with increasing doses of tunicamycin (1–10 ␮g). The degree to which Vg secretion in vivo was inhibited in these treated females was directly correlated to the tunicamycin dose (1 ␮g: 5% inhibition; 2 ␮g: 16% inhibition; 4 ␮g: 45% inhibition; 10 ␮g: 89% inhibition). In the present investigation, a 12 ␮g dose of tunicamycin was applied to vitellogenic animals, and the short-term effect of this drug on the glycosylation and synthesis of Vg was

determined by radiopulsing the treated animals with 3Hmannose or Trans35S-label. Tunicamycin treatment resulted in an 80% inhibition of 3H-mannose incorporation into the microsomal vitellogenin of the fat bodies of females and males (Table 1). Although the synthesis of 35S-labelled Vg in the fat bodies of tunicamycintreated females and males was inhibited by 21–49%, the secretion of 35S-Vg was inhibited by 78–85% (Table 1). The low quantities of radiolabelled Vg in the hemolymph of treated animals reflected both the inhibitory effects of tunicamycin on the rates of Vg synthesis, and the more dramatic inhibition of its secretion into the circulation. Following tunicamycin treatment, vitellogenic females and males produced an 190 kDa Vg polypeptide that was found only in the fat body (Fig. 9(A)). In females, an additional 185 kDa polypeptide was seen as well as smaller Vg polypeptides that were of lower molecular weights than those produced by normal females. These smaller polypeptides are presumably artifacts derived from the proteolysis of the 190 kDa Vg polypeptide. During the radiopulse, the secretion of newly synthesized Vg could not be detected in the hemolymph of tunicamycintreated females and males (Fig. 9(B)). The digestion of the 190 kDa Vg polypeptide with endo H and alkaline phosphatase showed that this polypeptide was neither glycosylated nor phosphorylated (Fig. 10(A,B)). The 190 kDa polypeptide is an apo-proVg polypeptide and

910

GRACE DON-WHEELER AND FRANZ ENGELMANN

FIGURE 7. The 35S autoradiogram of the fragments generated by protease digestion of the Vg polypeptides. The Vg of adult females and males (VgF and VgM) were radiolabelled with Trans35S-label during a 1–2 h pulse in vivo. The 35S-Vgs were isolated from the microsomal fraction of the fat body or from the hemolymph. The Vg polypeptides were separated by 6% SDS–PAGE and digested with 10 ␮g of Staphylococcus aureus V-8 protease as previously described. The resulting fragments were separated on a 20% SDS–PAGE and visualized by autoradiography. The molecular weights of the fragments are given. 220M, 195M and 85M denote microsomal Vgs; 155S, 112S and 85S are from secreted vitellogenins.

is presumably equivalent to the primary Vg translation product. Although these data showed that glycosylation is not essential for the synthesis of Vg, the apparent retention of aglycosylated Vg polypeptides in the fat bodies of tunicamycin-treated animals suggests that the glycosylation of Vg is important for its secretion. The absence of phosphorylation of the 190 kDa polypeptide suggested that either this apo-proVg polypeptide could not be phosphorylated, or that tunicamycin inhibits phosphorylation. We addressed these possibilities by examining the phosphorylation of Vg in tunicamycintreated females during a radiopulse with 32P-orthophosphate. The incorporation of 32P-orthophosphate into the 220 kDa pro-Vg and the smaller Vg polypeptides (155, 112, 95 kDa) was seen in these females (Fig. 11). It appears that the formation of Vg220 in the tunicamycintreated females occurs through the phosphorylation of Vg203, albeit at a lower rate than that seen in control females (Fig. 11).

DISCUSSION

The biosynthesis and processing of Vg precursors In the present study, the secreted vitellogenins produced by vitellogenic females and males of Leucophaea were comparable in the native molecular mass (Fig. 1(A)), but differed in composition of polypeptides (Fig. 1(B)). The Vg precursors in the female and male fat body appeared to be similarly glycosylated and phosphorylated. Differences in the proteolytic processing of some of the Vg polypeptides apparently account for the different polypeptide compositions of VgF and VgM. The biosynthesis and processing of Vg precursors in females and males was determined by the isolation of the newly synthesized vitellogenin of the fat body. Vitellogenin precursors of high specific activity were obtained by the radiopulsing of animals in vivo. This in vivo approach was chosen because of the low rates of Vg synthesis in the fat bodies of males cultured in vitro; conse-

BIOSYNTHESIS OF VITELLOGENIN

911

FIGURE 8. The endo H digestion of the polypeptides of VgF and VgM. Twenty micrograms of the respective vitellogenins were digested with 10 mU of endo H (+). The vitellogenins of the control treatment (−) were incubated in 0.05 M phosphate buffer (pH 6.0) containing 1 × 10−4 M PMSF. The VgF and VgM polypeptides were resolved by 7.5% SDS–PAGE and then transferred to nitrocellulose. The Vg polypeptides were visualized by (A) Coomassie blue R-250 staining and (B) the binding of Con A-HRP to carbohydrates. The bound Con A-HRP was visualized by development with the substrate 4-chloro-1-naphthol. The protein standards are given on the left and used to construct a calibration curve (molecular mass vs. relative mobility). The degree of glycosylation of the Vg polypeptides in (A) was determined by the difference in the molecular mass between the endo H-digested (+) and undigested (−) Vg polypeptides.

quently, these low rates made it difficult to determine the exact pathway of precursor processing. Radiopulse-chase studies in Locusta, Leucophaea, Blattella, and Aedes showed that during the biosynthesis of Vg, the radioactivity of the large Vg precursor is eventually transferred to the mature Vg polypeptides (Chen et al., 1978; Wojchowski et al., 1986; della-Cioppa and Engelmann, 1987; Dhadialla and Raikhel, 1990). In the current study on females and males of Leucophaea, the data showed that a decrease in the radioactivity of Vg220 was concomitant with an increase in the labelling of the smaller Vg polypeptides (Fig. 2(A) and Fig. 3(A,B)). Additional information on the processing of the Vg precursor was obtained by: (1) endoglycosidase H and alkaline phos-

phatase digestion; (2) peptide map analysis; and (3) the effect of tunicamycin on Vg precursor synthesis. For Leucophaea females and males, two species of Vg precursors, i.e., the 220 and 203 kDa provitellogenins (proVgs), were apparent (Fig. 2(A,B) and Fig. 4(A,B)). The 220 kDa proVg was phosphorylated and glycosylated whereas the 203 kDa proVg was glycosylated only (Fig. 5(A,B)). The 220 and 203 kDa Vg precursors reported here correspond to the respective 215 kDa phosphorylated and 200 kDa non-phosphorylated Vg precursors previously described by della-Cioppa and Engelmann (1987) for the adult female of Leucophaea. The digestion of these proVgs with endo H and alkaline phosphatase indicated that the glycosylation and phosphoryl-

912

GRACE DON-WHEELER AND FRANZ ENGELMANN

TABLE 1. The effect of tunicamycin on the incorporation of 3H-mannose and Trans 35S-label into the Vg of the fat body, and on the secretion of Vg into the hemolymph 3

Control Female Incorporation of radiolabel into Vg (Vg dpm/100 ␮g of microsomal protein) Vg secretion (Vg dpm/5 ␮l hemolymph) Male Incorporation of radiolabel into Vg (Vg dpm/100 ␮g of microsomal protein) Vg secretion (Vg dpm/5 ␮l hemolymph)

Trans35S-label

H-mannose Tunicamycin

%I

Control

Tunicamycin

%I

1387 ± 479 *(n = 3)

295 (avg.) *(n = 2)

79

37,600 ± 1400 *(n = 3)

29,700 ± 3000 *(n = 3)

21

2036 ± 59 (n = 4)

86 ± 1 (n = 3)

96

675,000 ± 60,000 (n = 6)

81,745 ± 1931 (n = 7)

78

1500 *(n = 1)

312 *(n = 1)

80

20,200 ± 2900 *(n = 3)

10,400 ± 2800 *(n = 3)

49

1081 ± 123 (n = 3)

82 ± 14 (n = 3)

92

53,100 ± 1075 (n = 9)

8105 ± 1500 (n = 10)

85

*For the incorporation of radiolabel into microsomal Vgs, 1n, pool of fat body from three animals. Adult females were treated with 10 ␮g of methoprene and males were treated with a 400 ␮g dose. On the fifth day after treatment, 12 ␮g of tunicamycin (in DMSO–PBS) was injected into the hemocoel. Control animals were injected with the DMSO–PBS vehicle only. After 4 h, females and males were pulsed in vivo for 3 h with either 5 ␮Ci 3H-mannose or 50 ␮Ci of Trans35S-label. The incorporation of radiolabel into the Vg of the fat body was determined by the immunoprecipitation of the Vg contained in solubilized fat body microsomal protein. Vg secretion was assessed by the immunoprecipitation of radiolabelled Vg in 5 ␮l of hemolymph. The data presented are the average ± SE for the number of animals given in parenthesis. % I, (control–experimental/control) × 100.

ation of the respective precursors in the fat bodies of females and males were quantitatively the same (Fig. 5(A,B)). The biosynthesis of vitellogenin in the fat bodies of females and males appears to occur in the same stepwise fashion: (1) the co-translational glycosylation of the Vg nascent chains to form the 203 kDa proVg; (2) the post-translational phosphorylation of the Vg203 to the 220 kDa proVg; (3) the proteolysis of the 220 kDa proVg to the transient Vg polypeptides (Vg195, Vg155) and the mature Vg polypeptides (Vg112, Vg95, Vg92, Vg54, Vg85). Although the details for the biosynthesis of insect Vgs are lacking for most species, the existing data for Locusta, Aedes, and Blattella indicate that the co- and post-translational modifications of the Vg precursors of these species may also occur in a stepwise manner (Wojchowski et al., 1986; Dhadialla and Raikhel, 1990). In Leucophaea, the glycosylated Vg203 was the smallest proVg that could be identified in the fat bodies of vitellogenic females and males (Fig. 4(A)) and was presumably synthesized by the co-translational glycosylation of nascent Vg polypeptides. The synthesis of a non-glycosylated Vg precursor, i.e., the 190 kDa apo-proVg polypeptide, was seen only when vitellogenic animals were treated with tunicamycin which inhibits N-linked glycosylation (Fig. 9(A)). It is noteworthy that the endo H digestion of Vg203 also yielded an 190 kDa polypeptide (Fig. 5(A)). For Aedes, a 240 kDa Vg precursor appeared to be synthesized by the co-translational glycosylation of

the Vg primary translation product (Dhadialla and Raikhel, 1990) as determined by the in vitro translation of Vg mRNA in the presence of canine microsomal membranes. However, this 240 kDa glycosylated precursor could not be identified in the vitellogenic fat body, whereas the fully glycosylated and phosphorylated 250 kDa Vg precursor was apparent. Presumably, the glycosylated 240 kDa Vg precursor of Aedes is rapidly phosphorylated, and therefore, does not persist in the fat body long enough to be detected. In Leucophaea, the synthesis of Vg203 in the fat bodies of females and males appears to precede that of Vg220 (Fig. 4(A,B)). The posttranslational phosphorylation of Vg203 to the Vg220 proVg was supported by two lines of evidence: (1) Vg203 and Vg220 were similarly glycosylated which suggested that the synthesis of Vg220 does not involve additional glycosylation; and (2) the alkaline phosphatase digestion of Vg220 yielded a polypeptide equivalent in size to Vg203. The stepwise synthesis of the Leucophaea Vg precursors appears to be similar to that described for the synthesis of the “egg specific protein” (ESP) in the ovarian follicle cells of Bombyx (Sato and Okitsugu, 1989). For Bombyx, the co-translational glycosylation of the nascent ESP polypeptide leads to the production of a discrete glycosylated precursor which is then posttranslationally phosphorylated. In the current study, Vg195 was a phosphorylated Vg polypeptide that was synthesized early during the radiopulse of vitellogenic fat bodies (Fig. 3(A,B) and Fig.

BIOSYNTHESIS OF VITELLOGENIN

FIGURE 9. The synthesis and secretion of vitellogenin (Vg) in tunicamycin-treated and control females and males. Adult females were treated with 10 ␮g of methoprene and adult males with 400 ␮g. On the fifth day, animals were injected with a 12 ␮g dose of tunicamycin (Tn) in DMSO–PBS. After 4 h of tunicamycin exposure, animals were pulsed in vivo with 50 ␮Ci of Trans35S-label for 3 h. (A) The microsomal vitellogenin of the fat bodies was immunoprecipitated from the solubilized microsomal proteins. (B) The secreted Vg was immununoprecipitated from 5 ␮l of hemolymph. The vitellogenins were analysed by 6% SDS–PAGE. (+) Tunicamycin; (−) control.

FIGURE 10. The digestion of the apo-provitellogenin of tunicamycin-treated females (A) and males (B) with endo H (EH) and alkaline phosphatase (AP). Methoprene-treated females and males were treated with 12 ␮g of tunicamycin as previously described, and then radiopulsed with 50 ␮Ci of Trans35S-label for 3 h pulse in vivo. The Vg contained in 20 ␮g of solubilized microsomal protein of the fat body was digested with either 10 mU of endo H or 10 mU of alkaline phosphate. The control (C) apo-vitellogenins were incubated in 0.05 M Tris (pH 8.0) for 2 h. The enzyme-digested and control apo-vitellogenins were subjected to 6% SDS–PAGE.

913

914

GRACE DON-WHEELER AND FRANZ ENGELMANN

FIGURE 11. The phosphorylation of the Vg of control and tunicamycin-treated adult females. Females were treated with 10 ␮g of methoprene for 5 days. On the fifth day, females were injected with 12 ␮g of tunicamycin (Tn) in DMSO–PBS. After 4 h, females were pulsed for 3 h in vivo with 0.5 mCi of 32P-orthophosphate. The vitellogenins were immunoprecipitated from solubilized fat body microsomal proteins. The Vg polypeptides were resolved by 6% SDS–PAGE. (+) Tunicamycin-treated (−) control.

4(A,B)). The peptide map analysis of Vg220 and Vg195 shows a high degree of similarity between these two polypeptides (Fig. 6(A) and Fig. 7). We assume that there is a single phosphorylated Vg precursor for Leucophaea and that Vg195 is derived from the proteolytic processing of Vg220. Proteolytic processing of the 220 kDa proVg For several species of insects, the high molecular weight Vg precursors are cleaved into a number of large and small polypeptides which are eventually assembled into mature secreted vitellogenin (Harnish and White, 1982). Previously, the proteolytic processing of the proVg and transient Vg polypeptides of Leucophaea was determined by the peptide map of radiolabelled Vg polypeptides (della-Cioppa and Engelmann, 1987). For the present investigation, a similar approach was used to determine the proteolytic processing of Vg polypeptides in the fat bodies of females and males. Because males have lower rates of Vg synthesis than females, the autoradiographic intensities of labelled fragments were weaker in the autoradiogram of VgM polypeptides than in that of VgF polypeptides (Fig. 6(B,C) and Fig. 7). Despite the difference in the intensities of the fragments, the respective fragment patterns of Vg220 and Vg155 of females and males were comparable (data not shown). Similarly, there appeared to be only minor differences in

the fragment compositions of corresponding VgF and VgM polypeptides of 112, 92, 85 and 54 kDa (Fig. 6(B,C) and Fig. 7). These data suggest the proteolytic processing of Vg precursors and transient Vg polypeptides occur in a likewise fashion in the fat bodies of females and males. Earlier studies in Leucophaea showed that the phosphorylated proVg was unevenly cleaved to form either the mature Vg112 polypeptide or the large transient Vg155 polypeptide (della-Cioppa and Engelmann, 1987). It was postulated that Vg155 is a short-lived transitional polypeptide that was processed to the Vg95-92 and Vg54. In the current study, the data suggest that Vg220 is cleaved to form either Vg112, Vg195, or Vg155. The processing fate of Vg195 is unclear at the present, but this large polypeptide is possibly the precursor of Vg155. The data confirm the earlier report that Vg155 is cleaved to Vg95-92 and Vg54 as all of the fragments of these smaller Vg polypeptides were found in Vg155 (Fig. 6(B)). The peptide map of Vg95 and Vg92 revealed that these polypeptides were nearly identical in the composition of fragments (Fig. 6(B)). The virtual absence of Vg95 in the microsomal and secreted Vg of males (Fig. 1(A) and Fig. 3(B)) suggests the rapid processing of Vg95 to Vg92 in the male fat body. It is noteworthy that the processing of Vg95 to Vg92 is also seen in vitellogenic oocytes where the progressive loss of the Vg95-92 units is concomitant with the appearance of a 90 kDa vitellin polypeptide (Don-Wheeler, 1996). The similarity in the fragment composition of Vg112 and Vg85 (Fig. 6(C) and Fig. 7) suggested that the Vg85 unit may arise from the additional processing of Vg112 in the fat body. This precursor–product relationship between Vg112 and Vg85 is supported by the polypeptide composition of VgM, i.e., the quantity of Vg112 is reduced whereas Vg85 is a prominent polypeptide (Fig. 1(B) and Fig. 3(B)). For the female- and male-produced vitellogenins, the quantitative difference in Vg112 (VgF ⬎ VgM) and Vg85 (VgF ⬍ VgM) suggests that the proteolytic processing of Vg112 occurs at a lower rate in the fat body of the female than in the male. Although we did not detect any quantitative difference in the glycosylation and phosphorylation of the Vg precursors (Vg203 and Vg220) in the fat bodies of females and males (Fig. 5(A)), the differential phosphorylation of polypeptides cannot be ruled out. Analysis of the peptide maps of 35Sand 32P-labelled vitellogenins revealed that Vg112 and Vg85 both contained a 25 kDa fragment that was phosphorylated in Vg112 but not in Vg85 (Fig. 6(C) and Fig. 7). Possibly the strategic dephosphorylation of sites within the Vg112 unit exposes a protease sensitive site. Such a mechanism has been implicated in the proteolytic digestion of vitellin during embryogenesis in Blattella (Purcell et al., 1988a). It is noteworthy that for the proVgs of Bombyx and Aedes, polyserine domains flank the putative cleavage sites in these precursors, which further suggests that the phosphorylation and/or dephosphorylation of these sites may play a role in the specific

BIOSYNTHESIS OF VITELLOGENIN

cleavage of the proVg (Chen et al., 1994; Izumi et al., 1994). Finally, it is relevant to consider that the fat bodies of females and males may differ with respect to inherent protease activity. The quantitative reduction of Vg112 and Vg95 in the male-produced vitellogenin (Fig. 1(B) and Fig. 3(B)) is perhaps due to a high protease activity of the male fat body, which in turn promotes the additional cleavage of Vg112 and Vg95. We considered the possibility that the hemolymph of vitellogenic males contained factors that promoted the additional proteolysis of the Vg polypeptides within the secreted vitellogenin. However, the incubation of VgF in male hemolymph did not change the polypeptide composition of this vitellogenin (data not shown). The processing of Vg112 to Vg85 thus appears to occur within the fat body. The similarity in the molecular mass of the native VgF and VgM suggests that the proteolytic processing of Vg112 to yield the Vg85 unit and other small peptides occurs without the loss of these fragments from the native protein. In this instance, the Vg85 and the smaller peptides of the proteolysed Vg112 are non-covalently associated and are not apparent under non-denaturing conditions. The different polypeptide compositions of VgF and VgM only become obvious under the denaturing conditions of SDS–PAGE. This proteolytic processing of Vg polypeptides without the loss of peptides is also seen in the vitellogenic oocytes of Blattella. In these oocytes, the large vitellogenin polypeptide (160 kDA) is cleaved to two smaller vitellin polypeptides (95 and 55 kDa) (Purcell et al., 1988b). For Blattella, the vitellogenin and vitellin dimers are of the same molecular mass despite the difference in the composition of polypeptides. The role of glycosylation in Vg synthesis, processing, and secretion To date, all of the insect Vgs that have been characterized appear to contain only asparagine N-linked high mannose oligosaccharides, i.e., GlcMan9GlcNac21 to Man5GlcNac2 (Kunkel and Nordin, 1985). For Leucophaea, we showed that all of the mature Vg polypeptides were glycosylated and that the carbohydrates were removed by endo H (Fig. 8(A,B)). These data confirm ¨ the findings of Konig et al. (1988), who reported that the carbohydrates of the Leucophaea Vg were exclusively composed of N-linked mannose oligosaccharides. We did not find any difference in the glycosylation of the corresponding polypeptides of the VgF and VgM (Fig. 8(A)). These data support our previous findings that the Vg precursors of females and males were similarly glycosylated. The 12 ␮g dose of tunicamycin used in the present experiments was shown to profoundly inhibit the glycosylation and secretion of Vg in treated animals, while

1

(glucose)(mannose)9 (N-acetylglucosamine)2 to (mannose)5 (Nacetylglucosamine)2.

915

causing the moderate inhibition of Vg synthesis (Table 1). Because methoprene-treated females have higher than normal rates of Vg synthesis (Don-Wheeler and Engelmann, 1991; Don-Wheeler, 1996), and the large size of the adult female (2 g), it was prudent to use high doses of tunicamycin to block the glycosylation of Vg effectively. Tunicamycin-treated vitellogenic females and males produced an apo-proVg of 190 kDa which accumulated in the fat body (Fig. 9(A) and Fig. 10(A,B)). Previous studies showed the in vitro translation of the Vg mRNA of Leucophaea in rabbit reticulocyte lysate yielded polypeptides of 180–190 kDa (della-Cioppa and Engelmann, 1987). More recently, Harmon (1993) reported that the size of the Vg mRNA of Leucophaea is 5400 nucleotides and theoretically encodes a protein of approximately 180 kDa. Taken together, these data argue that the 190 kDa apo-proVg is equivalent to the primary Vg translation product. Similar findings have been reported for Locusta and Aedes, in which the aglycosylated Vg precursors produced by tunicamycin-treatment of vitellogenic females were comparable in size to the primary Vg translation products (Wyatt et al., 1984; Dhadialla and Raikhel, 1990). The effect of tunicamycin on Vg synthesis in vitellogenic females has been reported for Locusta, Aedes, and Blattella (Wyatt et al., 1984, Wojchowski et al., 1986; Dhadialla and Raikhel, 1990). In all of these cases, including Leucophaea, the secretion of Vg was inhibited and concomitant with the intracellular accumulation of high molecular weight apo-Vg polypeptides. These studies indicate that although the co-translational glycosylation of Vg was not obligatory for the synthesis of Vg in Leucophaea, it is required for vitellogenin secretion (Table 1; Fig. 9(A,B)). In a number of secretory and membrane proteins, the inhibition of N-linked glycosylation was shown to induce the misfolding of polypeptides; these misfolded polypeptides are retained in the endoplasmic reticulum (ER) and are eventually degraded (reviewed by Pfeffer and Rothman, 1987; Hurtley and Helenius, 1989). For Aedes and Blattella (Wojchowski et al., 1986; Dhadialla and Raikhel, 1990), and for Leucophaea (Fig. 9(A,B)) the absence of Vg secretion coupled with the accumulation of the apo-Vg polypeptides in the fat body would be consistent with the notion that the aglycosylated Vg polypeptides are retained in the ER. In females of Leucophaea, the proteolysis of the 190 kDA apoVg precursor to smaller apo-Vg polypeptides was seen (Fig. 9(A)). However, it is unclear if this processing of the apoVg is related to the normal proteolytic processing of the Vg precursor, or represents the degradation of the Vg precursor. In Aedes, the proteolysis of the apo-proVg to apoVg subunits was also seen (Dhadialla and Raikhel, 1990). For both Aedes and Leucophaea, these aglycosylated subunits were not assembled into the secreted Vg. Some fully processed Vg polypeptides were still produced in tunicamycin-treated females as evidenced by the production of a small quantity of the 220 kDa phos-

916

GRACE DON-WHEELER AND FRANZ ENGELMANN

phorylated Vg precursor and mature Vg polypeptides (Fig. 11). Similar results were reported for Aedes in which the synthesis of glycosylated and phosphorylated Vg precursors occurred in tunicamycin-treated vitellogenic fat bodies, although the quantity was reduced in comparison to normal fat body (Dhadialla and Raikhel, 1990). For Aedes and Leucophaea, the low quantities of phosphorylated proVg and mature Vg polypeptides produced by the treated females is possibly due to the direct inhibition of phosphorylation and proteolytic processing by tunicamycin, or may reflect the small number of glycosylated Vg precursors available for modification. In the latter case, the co-translational glycosylation of the Vg precursor may be necessary for the transport of this polypeptide to the site of phosphorylation, and/or the glycosylated Vg precursor is the preferred substrate for phosphorylation and proteolytic processing.

CONCLUSIONS

The present study examined JH-induced Vg biosynthesis in the sexually dimorphic fat bodies of females and males of Leucophaea. Whereas the glycosylation and phosphorylation of Vg precursors in the fat bodies of females and males appear to be comparable, quantitative differences in the processing of Vg112 to Vg85 and Vg95 to Vg92 seem to exist. The synthesis of Vg in the fat bodies of females and males can be summarized as shown in Scheme 1. The data of the current study confirm the previous findings on the biosynthesis and modification of Vg in the fat body of Leucophaea (della-Cioppa and Engelmann, 1987). Although the earlier report by della-Cioppa and Engelmann (1987) did not comment on the two polypeptides Vg195 and Vg85, these polypeptides were

SCHEME 1. Summary of Vg synthesis and processing

BIOSYNTHESIS OF VITELLOGENIN

present in their autoradiograms. It should be noted that the Vg isolation procedure used by della-Cioppa and Engelmann (1987) included the use of the protease inhibitors PMSF, EDTA, and aprotinin. The identification of Vg195 and Vg85 in the present study was in part facilitated by the isolation of the microsomal fraction of the fat body which is enriched in newly synthesized Vg polypeptides. Also, methoprene treatment induced high rates of Vg synthesis in the female, and thus the synthesis of large amounts of precursors. Although the Vg polypeptides of the fat bodies of females and males appear to be comparable, there was a greater degree of proteolytic processing of the Vg112 and Vg95 in the fat body of the male (Fig. 3(A,B)). The difference in the proteolysis of these polypeptides may be related to stage- and sex-related differences in the quantity and quality of processing enzymes in the fat body, inherent protease activity of the fat body, or in the structure of Vg112 and Vg95. Previously we showed that the quantity of Vg85 unit present in the Vg of induced last instar females was greater than that of adult females, but less than that of the Vg of induced males (Don-Wheeler and Engelmann, 1991). Possibly stage- and sex-related differences in the modification of Vg112 affect the proteolysis of this polypeptide in the fat bodies of females and males. At present, little is known about how the proteolytic processing of the Vg polypeptides is effected in Leucophaea. The occurrence of phosphoserine domains at the putative cleavage sites in the proVgs of Bombyx and Aedes (Chen et al., 1994; Izumi et al., 1994) suggest that phosphorylation may play a role in the proteolytic processing of Vg in the fat body. This idea is consistent with the observation that the insect proVgs are phosphorylated before their proteolytic cleavage (dellaCioppa and Engelmann, 1987; Dhadialla and Raikhel, 1990). Possibly the differential phosphorylation of Vg112 in the fat bodies of females and males affects the proteolysis of this polypeptide in the respective fat bodies. Thus far, sexual differences in Vg structure have been described only for Leucophaea. In the few species where male-produced vitellogenins have been characterized, i.e., Drosophila, Athalia (sawfly), Apis (bee), and the cockroach Diploptera, the polypeptide compositions of the female- and male-produced Vgs were comparable (Mundall et al., 1983; Shirk et al., 1983; Trenczek et al., 1989; Hatakeyama and Oishi, 1990). It is also noteworthy that ovaries implanted into males of Diploptera incorporated Vg at similar rates as in the female (Mundall et al., 1979). Leucophaea may be a useful model in which to examine how the differentiation of a tissue affects the structure and processing of a specific protein. Such a phenomenom has been described for the rat in which thyrotropin is differentially glycosylated during prenatal and perinatal development (Gyves et al., 1989). This differential glycosylation of thyrotropin is thought to reflect developmental changes in the expression and activity of the various glycosylation

917

enzymes. Whether the sexual differentiation of the fat body of the insect also involves the expression of a specific complement of processing enzymes has yet to be determined. REFERENCES Bonner W. M. and Laskey R. A. (1974) A film detection method for tritium-labelled proteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem. 46, 83–88. Chen J.-S., Cho W.-L. and Raikhel A. (1994) Analysis of mosquito vitellogenin cDNA. Similarity with vertebrate phosvitins and arthropod serum proteins. J. Mol. Biol. 237, 641–647. Chen T. T., Strahlendorf P. W. and Wyatt G. R. (1978) Vitellin and vitellogenin from locusts (Locusta migratoria). J. Biol. Chem. 253, 5325–5331. Clegg J. C. S. (1982) Glycoprotein detection in nitrocellulose transfers of electrophoretically separated protein mixtures using concanavalin A and peroxidase: application to arenavirus and flavivirus proteins. Analyt. Biochem. 127, 389–394. Cleveland D. W., Fischer S. G., Kirschner M. W. and Laemmli U. K. (1977) Peptide mapping by limited proteolysis in sodium dodecyl sulfate and analysis by gel electrophoresis. J. Biol. Chem. 252, 1102–1106. Davis B. J. (1964) Disc electrophoresis. II. Method and application to human serum proteins. Ann. N.Y. Acad. Sci. 121, 404–427. della-Cioppa G. and Engelmann F. (1984) Phospholipid synthesis in the fat body endoplasmic reticulum during primary and secondary juvenile hormone stimulation of vitellogenesis in Leucophaea maderae. Roux Arch. Dev. Biol. 193, 78–85. della-Cioppa G. and Engelmann F. (1987) The vitellogenin of Leucophaea maderae: synthesis of a large phosphorylated precursor. Insect Biochem. 17, 401–415. Dhadialla T. S. and Raikhel A. S. (1990) Biosynthesis of mosquito vitellogenin. J. Biol. Chem. 265, 9924–9933. Dhadialla T. S. and Wyatt G. R. (1983) Juvenile hormone-dependent vitellogenin synthesis in Locusta migratoria fat body: inducibility related to sex and stage. Dev. Biol. 96, 436–444. Don-Wheeler G. and Engelmann F. (1991) The female- and male-produced vitellogenins of Leucophaea maderae. J. Insect Physiol. 37, 869–882. Don-Wheeler, G. (1996) The development of the vitellogenic competence of the fat body and ovary to respond to juvenile hormone in Leucophaea maderae. Ph.D. dissertation. University of California, Los Angeles. Engelmann F., Friedel T. and Ladduwahetty M. (1976) The native vitellogenin of the cockroach Leucophaea maderae. Insect Biochem. 6, 211–230. Giulian G. G., Shanahan M. F., Graham J. M. and Moss R. L. (1985) Resolution of low molecular weight polypeptides in a non-urea sodium dodecyl sulfate slab polyacrylamide gel system. Fed. Proc. 44, 686. Gyves P. W., Gesundheit N., Stannard B. S., DeCherney G. S. and Weintraub B. D. (1989) Alterations in the glycosylation of secreted thyrotropin during ontogenesis. J. Biol. Chem. 264, 6104–6110. Harmon, M. A. (1993) Cloning and characterization of a cDNA encoding the C-terminal of Leucophaea maderae vitellogenin. Ph.D. dissertation. University of California, Los Angeles. Harnish D. G. and White B. N. (1982) Insect vitellin: identification, purification and characterization from eight orders. J. Exp. Zool. 220, 1–10. Hatakeyama M. and Oishi K. (1990) Induction of vitellogenin synthesis and maturation of transplanted previtellogenic eggs by juvenile hormone III in males of the sawfly, Athalia rosae. J. Insect Physiol. 36, 791–797. Hedrick J. L. and Smith A. J. (1968) Size and charge isomer separation and estimation of molecular weights of proteins by disc gel electrophoresis. Arch. Biochem. Biophys. 126, 155–164.

918

GRACE DON-WHEELER AND FRANZ ENGELMANN

Hurtley S. M. and Helenius A. (1989) Protein oligomerization in the endoplasmic reticulum. Annu. Rev. Cell Biol. 5, 277–307. Izumi S., Yano K., Yamamoto Y. and Takahashi S. Y. (1994) Yolk proteins from insect eggs: structure, biosynthesis and programmed degradation during embryogenesis. J. Insect Physiol. 40, 735–746. ¨ Konig R., Nordin J. H., Gochoco C. H. and Kunkel J. G. (1988) Studies on ligand recognition by vitellogenin receptors in follicle membrane prepartions of the German cockroach, Blattella germanica. Insect Biochem. 18, 395–404. Kunkel, J. G. and Nordin, J. H. (1985) Yolk proteins. In Comprehensive Insect Physiology, Biochemistry, and Pharmocology, Vol. 1, eds L. I. Gilbert and G. A. Kerkut, pp. 83–111. Pergamon Press, New York. Laemmli U. K. (1970) Cleavage of structural proteins during assembly of the head of the bacteriophage T4. Nature, Lond. 227, 680–685. Lowry O. H., Rosebrough N. J., Farr A. L. and Randall R. J. (1951) Protein measurement with the folin phenol reagent. J. Biol. Chem. 193, 265–275. Mishell, B. B. and Shiigi, S. M. (1980) Purification of immunoglobulins and their fragments. In Selected Methods in Cellular Immunology, eds B. B. Mishell and S. M. Shiigi, pp. 278–286. W. H. Freeman, New York. Mundall E. C., Tobe S. S. and Stay B. (1979) Induction of vitellogenin and growth of implanted oocytes in male cockroaches. Nature, Lond. 282, 97–98. Mundall E. C., Szibbo C. M. and Tobe S. S. (1983) Vitellogenin induced in adult male Diploptera punctata by juvenile hormone and juvenile hormone analogue: identification and quantitative aspects. J. Insect Physiol. 29, 201–207. Pfeffer S. R. and Rothman J. E. (1987) Biosynthetic protein transport by the endoplasmic reticulum and Golgi. Annu. Rev. Biochem. 56, 829–852. Purcell J. P., Kunkel J. G. and Nordin J. H. (1988a) Yolk hydrolase activity associated with polypeptide and oligosaccharide processing of Blattella germanica vitellin. Arch. Insect Biochem. Physiol. 8, 39–58. Purcell J. P., Quinn T. M., Kunkel J. G. and Nordin J. H. (1988b) Correlation of yolk phosphatase expression with the programmed proteolysis of vitellin in Blattella germanica during embryonic development. Arch. Insect Biochem. Physiol. 40, 237–251.

Sato Y. and Okitsugu Y. (1989) Post-translational processing in the synthesis of egg specific protein in the silkworm, Bombyx mori. Insect Biochem. 19, 293–300. Sigma Chemical Co. (1986) Technical Bulletin No. MKR-137. St. Louis, MO. Sherod G., Johnson G. and Chalkley R. (1970) Phosphorylation of mouse ascites tumor cell lysine-rich histone. Biochemistry 9, 4611–4615. Shirk P. D., Minoo P. and Postlethwait J. H. (1983) 20-Hydroxyecdysone stimulates the accumulation of translatable yolk polypeptide gene transcript in adult male Drosophila melanogaster. Proc. Natl. Acad. Sci. U.S.A. 80, 186–190. Towbin H., Staehelin T. and Gordon J. (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. U.S.A. 76, 4350–4354. Trenczek T., Zillikens A. and Engels W. (1989) Developmental patterns of vitellogenin haemolymph titre and rate of synthesis in adult drone honey bees (Apis mellifera). J. Insect Physiol. 35, 475–481. Trimble R. B. and Maley F. (1984) Optimizing hydrolysis of N-linked high mannose oligosaccharides by endo-␤-N-acetylglucosaminisidase H. Analyt. Biochem. 141, 515–522. Wojchowski D. M., Parsons P., Nordin J. H. and Kunkel J. G. (1986) Processing of pro-vitellogenin in insect fat body: a role for high mannose oligosaccharides. Dev. Biol. 116, 422–430. Wyatt, G. R., Locke, J. and Bradfield, J. Y. (1984) The vitellogenin genes of Locusta migratoria and other insects. In Advances in Invertebrate Reproduction, Vol. 3, eds W. Engels, W. H. Clark, A. Fischer, P. J. W. Olive and D. F. Went, pp. 73–80. Elsevier Science Publisher B.V. Amsterdam, Netherlands.

Acknowledgements—The research results reported here was supported by a research grant from the National Science Foundation (IBN9004609), and a University Research Grant (F.E.), and by a Sigma Xi Grants in Aid of Research award (G.D.W.). Methoprene was a generous gift from Dr D. Cerf (Sandoz, Palo Alto, CA, U.S.A.).