- Email: [email protected]
Cyclic nucleotide-dependent protein phosphorylation in vitellogenic follicles of Hyalophora cecropia
Insect Biochemistry and Molecular Biology 30 (2000) 29–34 www.elsevier.com/locate/ibmb
Cyclic nucleotide-dependent protein phosphorylation in vitellogenic follicles of Hyalophora cecropia Yuren Wang *, William H. Telfer Department of Biology, University of Pennsylvania, Philadelphia, PA 19104-6018, USA Received 27 April 1999; received in revised form 13 August 1999; accepted 17 August 1999
Abstract (1) In homogenates of vitellogenic follicles from Hyalophora cecropia, cyclic nucleotides promoted the transfer of label from [γ32P]-ATP to at least four polypeptides. PKI (6–20) amide, an inhibitor of PKA (cAMP-dependent protein kinase), prevented all four reactions. Quantitative tests using kemptide as a substrate indicated that 80% of the total follicular PKA activity was localized in the follicle cells; labeling at 45, 32, and 27 kDa was particle-associated and also restricted to the follicle cells, while a 58 kDa substrate was labeled only in homogenates of the oocyte. (2) When intact follicles were incubated in [32P]-phosphate and okadaic acid, a protein phosphatase inhibitor, the 32 kDa substrate again exhibited cAMP-dependent labeling. There was thus a physiological relationship between PKA activation and 32 kDa protein phosphorylation, while exposure of at least two of the other three substrates to appropriate kinases required homogenization. The latter was illustrated by phosphorylation of the 42 kDa small subunit of vitellogenin, which occurred only when homogenization mixed the proteins of the yolk bodies with cytoplasmic kinases. (3) PKA activation is known to promote the termination of vitellogenesis, even in the absence of detectable labeling of the 32 kDa substrate. The possibility remains that phosphorylation at 32 kDa concerns later aspects of postvellogenic development that were not tested by the assay system used here. 2000 Published by Elsevier Science Ltd. All rights reserved. Keywords: cAMP; cGMP; PKA; S6; Silk moths; Vitellogenin
1. Introduction Pertussis toxin and cyclic nucleotides both stimulate ovarian follicles of Hyalophora cecropia to terminate patency and, hence, the uptake of vitellogenin (Wang and Telfer, 1997, 1998a,b). Inhibitors of cAMP-dependent protein kinase (PKA) block the response, and activators of this enzyme promote it. The cAMP content of the follicle is depressed by a pertussis toxin-sensitive G protein during vitellogenesis, but is apparently released from this inhibition when the follicle terminates vitellogenin uptake.
Abbreviations: MB-cAMP: N6-monobutyryl-cAMP; MB-cGMP: N2monobutyryl-cGMP; IBMX: and 3-isobutyryl-1-methyl-xanthine; PKA: cAMP-dependent protein kinase; PKI: PKI (6–20) amide, an inhibitor of PKA; Rp- and Sp-cAMPS: isomers of adenosine-3⬘,5⬘monophosphorothioate; PSS: physiological saline solution. * Corresponding author. Present address: Department of Molecular and Cellular Biology, Wyeth-Ayerst Research, CN-8000 Princeton, NJ 08543, USA.
The ability of PKA activation to trigger the terminatation of patency is mediated by the phosphorylation of cytoplasmic protein substrates. This was supported by the demonstration of cyclic nucleotide-dependent phosphorylation of a 32 kDa polypeptide in homogenates of H. cecropia follicles (Wang and Telfer, 1996). The response could not be detected at that time in intact follicles, but this deficiency has now been overcome by including the protein phosphatase inhibitor okadaic acid in the incubation medium. Here, we report further observations on phosphorylation of the 32 kDa substrate and consider what they can reveal about the role of cyclic nucleotides in the postvitellogenic development of ovarian follicles. Several cyclic nucleotide-dependent protein kinases have been isolated and characterized from the ovaries of Bombyx mori and shown to phosphorylate in homogenates primarily serine hydroxyl groups of yolk proteins (Takahashi, 1984, 1985, 1987). Yolk protein phosphorylation also occurs in homogenates of H. cecropia follicles, and we therefore need to determine whether it
0965-1748/00/$ - see front matter 2000 Published by Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 5 - 1 7 4 8 ( 9 9 ) 0 0 0 9 4 - 6
30
Y. Wang, W.H. Telfer / Insect Biochemistry and Molecular Biology 30 (2000) 29–34
too is promoted by cyclic nucleotides and is concerned with the control of vitellogenesis.
2. Materials and methods 2.1. Preparation of follicles and homogenates Vitellogenic follicles were dissected from the eight ovarioles of day 18 pharate adult females. At this stage of metamorphosis each of the eight ovarioles contains previtellogenic and a chain of approximaately 40 vitellogenic follicles with intact nurse cell caps. The dissecting medium (PSS) was a physiological saline containing 1 mM phenylthiourea to prevent melanin formation, and 70 mM sucrose to make it isosmotic with day 18 hemolymph (Wang and Telfer, 1996). Hemolymph proteins were removed from the intercellular spaces by washing the follicles twice for 20 min each in PSS with constant shaking. To separate oocytes from the epithelium of follicle cells, washed follicles were treated for 3 min with 1 mg/ml pronase in PSS. Sheets of follicle cells could then be separated from the oocyte by gentle pipetting. The oocytes were removed from the dissociated mixture with a wide-mouth pipet and the sheets of follicle cells were collected by centrifugation at 200g for 2 min. Both oocytes and follicle cells were washed with protease inhibitors (50 mg/ml PMSF, 1 mg/ml leupeptin, 5 mg/ml aprotinin, 5 mg/ml chymostatain, 5 mg/ml antipain, and 0.5 mg/ml pepstatin in PSS), and then with PSS. Individual follicles, single oocytes, and the follicle cells from single follicles were each homogenized in 100 µl of cold 50 mM HEPES (pH 7.0), 1 mM EDTA, 1 mM dithiothreitol (DTT), and protease inhibitors as listed above. A portion of the homogenate was saved for protein determination (DC protein assay from Bio Rad), while the remainder was either stored at ⫺70° or processed immediately as described below. Particulate fractions were prepared for phosphorylation by centrifuging a whole follicle homogenate successively at 1000, 20,000, 80,000 and 150,000g. The 1000g pellet was discarded, since it contained any cells and yolk bodies that had survived homogenization intact. The succeeding three pellets were each resuspended in 0.1 ml homogenization buffer. A fourth fraction, the 150,000g supernatant, had a volume of 0.5 ml. 2.2. Protein phosphorylation Phosphorylation of unfractionated homogenates utilized 10 µl samples—the equivalent of one tenth of a follicle. Resuspended centrifugal pellets, by contrast, were normalized for protein content, each reaction mixture containing 50 µg of pellet protein. These amounts of homogenate or pellet were mixed with a phosphorylation
medium to yield a final volume of 50 µl containing 50 mM HEPES (pH 7.0), 10 mM MgCl2, 1 mM DTT, 20 µM ATP, and 10 µCi [γ 32P]-ATP (initial specific activity about 60,000 Ci/mmol). Dependence of phosphorylation on cyclic nucleotides was assayed by including in the reaction mixture predetermined concentrations of cAMP, cGMP, or analaogs of these, and/or 1 µM PKI (6–20) amide, a peptide inhibitor of PKA. The final mixture was incubated for 3 min at 30°. The reaction was terminated and the sample prepared for electrophoresis by adding 50 µl of 2X SDS-PAGE sample buffer (0.125 tris–HCl, pH 6.8, 4% SDS, 20% glycerol, 10% 2-mercaptoethanol, and 0.002% bromphenol blue) and heated for 4 min at 100°. To study phosphorylation in intact cells, 8–10 vitellogenic follicles were incubated in 200 µl PSS containing 10 µCi [32P]-phosphate and the indicated concentration of a cAMP analog or other reagent. Okadaic acid, a protein phosphatase inhibitor (Bialojan and Takai, 1988; Cohen et al., 1990), was also added to the incubation media. After a 2 h incubation at 23°, the follicles were homogenized in 500 µl SDS-PAGE buffer and boiled for 5 min. 2.3. Electrophoresis and autoradiography Phosphorylated proteins were separated by SDSPAGE and visualized by autoradiography. One-dimensional gel electrophoresis was performed with a 4% stacking gel and a 10% separating gel (Laemmle, 1969). For any one gel, equal sample volumes, standardized either for amount of protein or for follicle fraction (generally, about 1/20th of a follicle per lane), were added to each well. The gels were stained with Coomassie Blue R and then dried. Autoradiography utilized Du Pont Reflection film and a Reflection intensifying screen, with exposures of 4 h up to 2 days at ⫺72°. 2.4. Assay of PKA PKA was assayed with a kit from Gibco BRL (Gaithersberg, MD). Ten µl of homogenate, prepared as described above, were added to 29 µl of 50 mM tris– HCl (pH 7.5), 0.25 mg/ml bovine serum albumin, 50 µM Kemptide (peptide substrate for PKA) and 10 µM cAMP, with or without 1 µM PKI. After incubation at room temperature for 15 min, phosphorylation was started by adding 1 µCi [γ 32P]-ATP in 10 µl of the above buffer, and the tubes were incubated at 30° for an additional 5 min. Twenty µl of each reaction mixture were then spotted onto a phosphocellulose disc, and washed twice in 500 ml 1% phosphoric acid and once with distilled water. PKA activity was calculated by subtracting 32P incorporation in the presence of cAMP and PKI from that in the presence of cAMP alone.
Y. Wang, W.H. Telfer / Insect Biochemistry and Molecular Biology 30 (2000) 29–34
2.5. Chemicals cAMP, cGMP, Mb-cAMP, Deoxy-mb-cAMP, MbcGMP, and IBMX were from Sigma (St. Louis, MO). Forskolin, Dideoxy-forskolin, and PKI (Cheng and Kemp, 1986) were from Calbiochem (LaJolla, CA). Rpand Sp-isomers of cAMPS (Erneux et al., 1986) were from Biolog Life Science Institute (Bremen, Germany). [γ-32P]-ATP and [32P]O4 were from New England Nuclear (Wilmington, DE). All other chemicals were from Sigma. 3. Results 3.1. Cyclic AMP-dependent protein phosphorylation Detection of phosphorylation in both whole follicles and homogenates was more sensitive than reported earlier (Wang and Telfer, 1996). Okadaic acid, a cell-permeant protein phosphatase inhibitor, raised the level of 32 kDa protein labeling in living follicles to a clearly detectable level (Fig. 1). By itself, 10 µM okadaic acid
Fig. 1. SDS-PAGE autoradiogram of cyclic nucleotide-dependent phosphorylation in intact follicles. Vitellogenic follicles were incubated for 2 h in PSS containing [32P]-phosphate and the additions indicated below, before homogenization and electrophoresis. Additions to the labeling media included: lane 1, none; lane 2, 10 µM okadaic acid; lane 3, 1 mM MB-cAMP and 10 µM okadaic acid; lane 4, 250 µM forskolin and 10 µM okadaic acid. The asterisk indicates the position of the 32 kDa substrate. Molecular weight standards (×10⫺3 kDa) are indicated on the left. The label-free zones centered at 42 and 31 kDa in all lanes are the positions, respectively, of the small subunit of vitellogenin and microvitellogenin, both of which are so concentrated that they exclude other proteins from their positions (see Fig. 5A, lanes 4– 6). There is no indication that cyclic nucleotides promoted phosphorylation of the 45 and 27 kDa substrates that are demonstrated below in homogenates.
31
did not alter the protein phosphorylation patterns (Fig. 1, lane 2), but when combined with 1 mM Mb-cAMP (lane 3), forskolin (lane 4), Sp-cAMPS or IBMX (not shown), labeling of the 32 kDa polypeptide became evident. Neither deoxy-Mb-cAMP, dideoxy-forskolin, nor Rp-cAMPS promoted 32 kDa polypeptide phosphorylation, even when combined with 10 µM okadaic acid (not shown). In combination, these responses indicate that phosphorylation of the 32 kDa substrate can in fact be a physiological, in situ response to PKA activation. Improvement of detection in follicle homogenates was achieved by including a broad range mixture of protease inhibitors in the labeling medium. Cyclic nucleotidedependent phosphorylations were now detected not only at 32 kDa, but also at 58, 45, and 27 kDa (Fig. 2, lanes 2 and 5). Supplementing the incubation mixture with catalytic subunits of PKA yielded phosphorylation of additional polypeptides and a general intensification of label (lane 4). Revealed labeling patterns are influenced by the relative rates of kinase and phosphatase activities and the four bands indicated by asterisks in Fig. 2 are simply the most easily detected of a larger complement. In whole follicles (Fig. 1), labeling was not detectable at either 45 or 27 kDa; cAMP-independent labeling was so heavy above the 50 kDa level that detecting cyclic nucleotide-dependent labeling of a 58 kDa band was not
Fig. 2. SDS-PAGE autoradiogram showing 32P labeling of proteins in homogenates. Each lane on the gel contained labeled material from 0.02 follicle. Labeling had been for 5 min with [γ-32P]ATP in the following: lane 1, no additions; lane 2, 50 µM cAMP; lane 3, 50 µM cAMP and 1 mg/ml PKI; lane 4, 50 µM cAMP and 10 units catalytic subunits of PKA; lane 5, 50 µM cGMP; lane 6, 50 µM cGMP and 1 mg/ml PKI; lane 7, 1 mg/ml PKI. Molecular weight standards (×10⫺3 kDa) are indicated on the left. Asterisks mark the positions of four bands whose phosphorylations were promoted by cyclic nucleotides.
32
Y. Wang, W.H. Telfer / Insect Biochemistry and Molecular Biology 30 (2000) 29–34
possible. Lack of a cyclic nucleotide response at 45 and 27 kDa could be explained in two ways. Either these products are particularly sensitive to hydrolytic degradation during homogenization, or compartmentation in intact follicles keeps them isolated from cytoplasmic kinases in the absence of homogenization. 3.2. Kinase activity promoted by cGMP Like Mb-cAMP, Mb-cGMP can trigger the end of vitellogenesis by intact follicles (Wang and Telfer, 1996), and also, shown here, the phosphorylation of the 58, 45, 32, and 27 kDa substrates in homogenates (Fig. 2, lane 5). Since PKI blocked the effects of cGMP (Fig. 2, lanes 6), these responses are mediated by an ability to activate PKA. Ten-fold higher concentrations of cGMP than cAMP were required to effect the same levels of phosphorylation, however (Fig. 3), so binding of the former to the regulatory subunit of PKA must be weaker than that of cAMP. 3.3. Substrate and PKA localization In a step-wise centrifugation procedure, three of the four cAMP-promoted bands indicated by asterisks in Fig. 2 proved to be particle-associated (Fig. 4). The 32 kDa band was labeled in all three resuspended pellets, but was not detected in the 150,000g supernatant. Supplementing the labeling medium with excess catalytic subunits of PKA intensified the labeling at 32 kDa in all three pellets, but did not make it detectable in the 150,000g supernatant or change the status of the 20– 80,000g pellet as the fraction in which labeling was the heaviest (Fig. 4, lanes c). The profile thus reflected the distribution of the 32 kDa substrate, rather than of PKA. The 27 kDa band, though weaker, was distributed similarly. The 45 kDa band was most concentrated in the 1–20,000g pellet. The 150,000g supernatant showed no evidence of the three particle-associated bands, yielding instead three bands that labeled in response to cAMP in the general region of 58 kDa. In homogenates of isolated follicle cells, cAMP-
Fig. 3. SDS-PAGE autoradiogram showing the effects of cyclic nucleotide concentration on 32 kDa protein phosphorylation. Homogenate preparation and labeling were as described for Fig. 2. Cyclic nucleotide concentrations were: lane 1, 0 µM; lanes 2–4, 1, 10, and 100 µM cAMP, respectively; lanes 5–7, 1, 10, and 100 µM cGMP, respectively. The asterisk indicates the position of the 32 kDa band.
Fig. 4. SDS-PAGE autoradiograms of phosphorylation in centrifugal fractions of a whole follicle homogenate. Lanes a, no additions; lanes b, 50 µM cAMP, lanes c, 1 unit of catalytic subunits of PKA. Asterisks on the left indicate the labeling corresponding to the 45, 32, and 27 kDa bands of cyclic nucleotide-dependent phosphorylation seen in unfractionated homogenates (Fig. 2). The 58 kDa substrate was not visible in the particulate fractions shown here, but may be represented by one of the bands near the asterisk on the right. Solubilized vitellogenin was concentrated in the 80–150,000g pellet, its small subunit being heavily labeled at the 40–44 kDa level.
dependent phosphorylation was detected at 45, 32, and 27 kDa (Fig. 5B, lanes 1 and 2); as in Fig. 2, addition of catalytic subunits of PKA (lane 3) resulted in intensification of all bands. Of the four cAMP-responsive phosphorylations indicated by asterisks in Fig. 2, only the 58 kDa band was visible in homogenates of isolated oocytes. That the method of separating follicle cells and oocytes was effective was indicated by the absence of yolk protein bands from follicle cell homogenates when this gel was stained with Coomassie blue (Fig. 5A, lanes 1–3). Also examined was the distribution of PKA between the two cell types. PKA activity was four times greater in homogenates of follicle cells than of oocytes. Measured as picomoles of phosphate transferred from ATP to kemptide/min/follicle, the activities were 57.4±2.3 for whole follicles, 44.5±0.74 for isolated follicle cells, and 10.9±1.3 for isolated oocytes (average±s.e., N=5). 3.4. Phosphorylation of yolk proteins The 43 kDa subunit of vitellogenin was heavily phosphorylated in oocyte homogenates (Fig. 5B, lanes 4–6),
Y. Wang, W.H. Telfer / Insect Biochemistry and Molecular Biology 30 (2000) 29–34
33
(Takahashi, 1987; Sato and Yamashita, 1989; Tsuchida et al., 1992). Cyclic nucleotide-independent labeling in this region of the autoradiogram in Fig. 1 was too complex and dark to yield this information for H. cecropia. Finally, the 31 kDa microvitellogenin band was not labeled by 32P, even in homogenates of isolated oocytes (Fig. 5B, lanes 4–6) and thus presumably lacks appropriate phosphorylation sites.
4. Discussion
Fig. 5. SDS-PAGE gels of homogenates of isolated oocytes and follicle cells. Homogenates were each divided into three aliquots; labeling was done with [γ32P]-ATP under standard phosphorylating conditions, but with the following additions: lanes 1 and 4, none; lanes 2 and 5, 10 µM cAMP; lanes 3 and 6, 1 unit catalytic subunit of PKA. Asterisks indicate the positions of bands whose phosphorylation was promoted by cAMP in Fig. 2.
even in the absence of cAMP and catalytic subunits of PKA (lane 4). It is unlikely that this was promoted by endogenous cyclic nucleotides, because the 58 kDa substrate in the same preparation failed to phosphorylate unless either cAMP or catalytic subunits of PKA were added (Fig. 5B, lanes 4–6, asterisk). When phosphorylation was performed in intact follicles, by contrast, the 43 kDa vitellogenin subunit was not phosphorylated (Fig. 1). The high concentration of yolk protein at this location excludes other proteins with similar mobility in SDS-PAGE, so that a blank space appears in the autoradiogram at 41–44 kDa. The absence of vitellogenin labeling in intact follicles can be ascribed to compartmentation—yolk proteins are enclosed by membranes in intact follicles (Telfer, 1961), while the kinases that phosphorylate them in homogenates are presumably cytoplasmic. Heavy labeling also bracketed the 58 kDa band in homogenates of isolated oocytes (Fig. 5B, lanes 4–6). As with the 43 kDa vitellogenin subunit, the bracketing bands of label occurred in the absence of added cAMP or catalytic subunits of PKA. This is the location of paravitellogenin, the follicle cell contribution to the yolk (Bast and Telfer, 1976; Rubenstein, 1979), whose phosphorylation might be expected in intact follicles, since yolk proteins synthesized by follicle cells in other lepidoptera have been shown to be phosphorylated
Labeling of 32 kDa substrate in intact follicles raises the likelihood that this reaction is a physiological response to a rise in the level of cAMP. In homogenates, additional cyclic nucleotide-dependent phosphorylations occurred, but since labeling was not evoked at these locations in whole follicles, even when incubated in okadaic acid, homogenates may well generate artifactual products whose precursors are not normally available to PKA or to PKA-activated kinases. A clear example for this model was provided by vitellogenin, whose 43 kDa subunit was heavily phosphorylated in homogenates of oocytes (Fig. 5B) but remained unlabeled in whole follicles (Fig. 1), where it is confined to membrane-limited yolk spheres. It is unlikely that phosphorylation of the 32 kDa substrate, at least to a level that is autoradiographically detectable, is required for the termination of vitellogenin uptake, for concentrations of cell permeant cAMP analogs that stimulate vitellogenic follicles to stop forming yolk have consistently failed to induce detectable labeling at 32 kDa (Wang and Telfer, 1996). In the most fully characterized PKA-based signal transduction system in lepidoptera, cAMP treatment of intact prothoracic glands promotes the labeling of a 32/34 kDa polypeptide (Rountree et al., 1992) that has been implicated as ribosomal protein S6 (Song and Gilbert, 1994). In a variety of mammalian cells and tissues, S6 is also selectively phosphorylated in response to cAMP, as well as to other developmental and mitogenic stimuli (reviewed by Stewart and Thomas (1994)). In another insect system, P-element insertions into the S6 gene of Drosophila melanogaster resulted in lethal tumorigenesis of hematopoetic tissues. The 32 kDa phosphorylation substrate of H. cecropia follicle cells has not been identified, but its size in SDS-PAGE, its distribution in centrifugal fractions, and its status as the most prominent cAMP-dependent phosphorylation product of intact follicles are characteristics suggestive of S6. PKA-activated phosphorylation at 32 kDa also resembles many examples of S6 phosphorylation in its timing relative to a complex of developmental and secretory changes. Loss of patency and termination of vitellogenin uptake are followed in situ by a set of devel-
34
Y. Wang, W.H. Telfer / Insect Biochemistry and Molecular Biology 30 (2000) 29–34
opmental changes that include cessation of synthesis of the yolk protein produced by the follicle cells (Rubenstein, 1979), secretion of a hydrophobic layer around the oocyte (Telfer and Smith, 1970), and a complex sequence of chorionic protein secretions (Paul and Kafatos, 1975). The latter is not initiated until 20 h after the cessation of vitellogenin uptake (Telfer and Anderson, 1968), and was thus not tested by our assay system. Cyclic nucleotide-induced phosphorylation at 32 kDa, while not involved in blocking patency and the termination of vitellogenin uptake, could still turn out to be involved in the ensuing developmental events. Concerning yolk proteins, both the large and small subunits of B. mori vitellogenin have been shown to be post-translationally phosphorylated at their site of synthesis in the fat body (Takahashi, 1987), and there was in the same report some evidence that phosphates are also added in vitellogenic follicles. One of several forms of PKA that were isolated from B. mori ovaries was shown in that study to be able to phosphorylate isolated vitellogenin; but as in the present study (Fig. 5) the heavy labeling of vitellogenin that can occur in follicle homogenates was not promoted by cAMP. If, despite this counter-indication, phosphorylation of vitellogenin by PKA should in fact prove to be an ongoing feature of yolk deposition in silk moths, two requirements would have to be met. PKA would have to be transferred to the endosomal system to which vitellogenin is restricted (reviewed for insects by Raikhel and Dhadialla (1992)); and cAMP would have to be prevented from crossing gap junctions from the oocyte to the follicle cells (Woodruff, 1979) in sufficient quantity to trigger the termination of patency. These restrictions, combined with the inability of Mb-cAMP to promote yolk labeling in intact follicles (Fig. 1), make the cyclic nucleotidedependence of vitellogenin phosphorylation seem unlikely at the present time.
Acknowledgements Supported by GM-32909, a grant to WT from the National Institutes of Health.
References Bast, R., Telfer, W., 1976. Follicle cell protein synthesis and its contribution to the yolk of the cecropia moth oocyte. Devel. Biol. 52, 83–97. Bialojan, C., Takai, A., 1988. Inhibitory effect of a marine-sponge toxin, okadaic acid, on protein phosphatases. Biochem. J. 256, 283–290. Cheng, H.-C., Kemp, B.C., 1986. A potent synthetic peptide inhibitor of the cAMP-dependent protein kinase. J. Biol. Chem. 261, 989– 992. Cohen, P., Holmes, C.F.B., Tsukitani, Y., 1990. Okadaic acid: a new
probe for the study of cellular regulation. Trends Biochem. Sci. 15, 98–102. Erneux, C., Sande, J.V., Jastorff, B., Dumont, J.E., 1986. Modulation of cAMP action in the dog thyroid by its agonist and antagonist Sp- and Rp-adenosine 3⬘, 5⬘-monophosphothioate. Biochem. J. 234, 193–197. Laemmle, U., 1969. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227, 680–685. Paul, M., Kafatos, F.C., 1975. Specific protein synthesis in cellular differentiation. II. The program of protein synthetic changes during chorion formation by silkmoth follicles, and its implementation in organ culture. Devel. Biol. 42, 141–159. Raikhel, A.S., Dhadialla, T.S., 1992. Accumulation of yolk proteins in insect oocytes. Ann. Rev. Entomol. 37, 217–251. Rountree, D.B., Combest, W.L., Gilbert, L.I., 1992. Prothoracicotropic hormone regulates the phosphorylation of specific protein in the prothoracic glands of the tobacco hornworm, Manduca sexta. Insect Biochem. Molec. Biol. 22, 353–362. Rubenstein, E.C., 1979. The role of an epithelial occlusion zone in the termination of vitellogenesis in Hyalophora cecropia ovarian follcles. Devel. Biol. 71, 115–127. Sato, Y., Yamashita, O., 1989. Post-translational processing in the synthesis of egg-specific protein in the silkworm Bombyx mori. Insect Biochem. 19, 293–300. Song, Q., Gilbert, L.I., 1994. S6 phosphorylation results from prothoracicotropic hormone stimulation of insect prothoracic glands: a role for S6 kinase. Devel. Genetics 15, 332–338. Stewart, M.J., Thomas, G., 1994. Mitogenesis and protein synthesis: a role for ribosomal protein S6 phosphorylation. BioEsays 16, 809–815. Takahashi, S.Y., 1984. Ovarian protein kinases and endogenous substrates from the silkworm, Bombyx mori. Insect Biochem. 14, 219–229. Takahashi, S.Y., 1985. Characterization of guanosine 3⬘:5⬘-monophosphate-dependent protein kinase from silkworm eggs and analysis of the endogenous protein substrates. J. Comp. Physiol. B 155, 693–701. Takahashi, S.Y., 1987. Studies on the phosphorylation of ovarian proteins from the silkworm, Bombyx mori. Insect Biochem. 17, 141–152. Telfer, W., 1961. The route of entry and localization of blood proteins in the oocytes of saturniid moths. J. Biophys. Biochem. Cytol. 9, 747–759. Telfer, W., Anderson, L., 1968. Functional transformations accompanying the initiation of a terminal growth phase in the cecropia moth oocyte. Devel. Biol. 17, 512–535. Telfer, W., Smith, D., 1970. Aspects of egg formation. In: Neville, C. (Ed.), Symposium of Royal Entomology Society 5. Blackwell, Oxford, pp. 117–134. Tsuchida, K., Kawooya, J., Law, J., Wells, M., 1992. Isolation and characterization of two follicle-specific proteins from eggs of Manduca sexta. Insect Biochem. Molec. Biol. 22, 89–89. Wang, Y., Telfer, W., 1996. Cyclic nucleotide-induced termination of vitellogenin uptake by Hyalophora cecropia follicles. Insect Biochem. Molec. Biol. 26, 85–94. Wang, Y., Telfer, W., 1997. cAMP-stimulated termination of vitellogenesis in Hyalophora cecropia: formation of a diffusion barrier and the loss of patency. J. Insect Physiol. 43, 675–684. Wang, Y., Telfer, W., 1998a. cAMP-induced water uptake in a moth ovary: inhibition by bafilomycin and anthracene-9-carboxylic acid. J. Exper. Biol. 201, 1627–1635. Wang, Y., Telfer, W., 1998b. Pertussis-toxin sensitive G protein that supports vitellogenin uptake by promoting patency. Arch. Insect Biochem. Physiol. 39, 36–45. Woodruff, R., 1979. Electrotonic junctions in Cecropia moth ovaries. Devel. Biol. 61, 281–295.
Cyclic nucleotide-dependent protein phosphorylation in vitellogenic follicles of Hyalophora cecropia Yuren Wang *, William H. Telfer Department of Biology, University of Pennsylvania, Philadelphia, PA 19104-6018, USA Received 27 April 1999; received in revised form 13 August 1999; accepted 17 August 1999
Abstract (1) In homogenates of vitellogenic follicles from Hyalophora cecropia, cyclic nucleotides promoted the transfer of label from [γ32P]-ATP to at least four polypeptides. PKI (6–20) amide, an inhibitor of PKA (cAMP-dependent protein kinase), prevented all four reactions. Quantitative tests using kemptide as a substrate indicated that 80% of the total follicular PKA activity was localized in the follicle cells; labeling at 45, 32, and 27 kDa was particle-associated and also restricted to the follicle cells, while a 58 kDa substrate was labeled only in homogenates of the oocyte. (2) When intact follicles were incubated in [32P]-phosphate and okadaic acid, a protein phosphatase inhibitor, the 32 kDa substrate again exhibited cAMP-dependent labeling. There was thus a physiological relationship between PKA activation and 32 kDa protein phosphorylation, while exposure of at least two of the other three substrates to appropriate kinases required homogenization. The latter was illustrated by phosphorylation of the 42 kDa small subunit of vitellogenin, which occurred only when homogenization mixed the proteins of the yolk bodies with cytoplasmic kinases. (3) PKA activation is known to promote the termination of vitellogenesis, even in the absence of detectable labeling of the 32 kDa substrate. The possibility remains that phosphorylation at 32 kDa concerns later aspects of postvellogenic development that were not tested by the assay system used here. 2000 Published by Elsevier Science Ltd. All rights reserved. Keywords: cAMP; cGMP; PKA; S6; Silk moths; Vitellogenin
1. Introduction Pertussis toxin and cyclic nucleotides both stimulate ovarian follicles of Hyalophora cecropia to terminate patency and, hence, the uptake of vitellogenin (Wang and Telfer, 1997, 1998a,b). Inhibitors of cAMP-dependent protein kinase (PKA) block the response, and activators of this enzyme promote it. The cAMP content of the follicle is depressed by a pertussis toxin-sensitive G protein during vitellogenesis, but is apparently released from this inhibition when the follicle terminates vitellogenin uptake.
Abbreviations: MB-cAMP: N6-monobutyryl-cAMP; MB-cGMP: N2monobutyryl-cGMP; IBMX: and 3-isobutyryl-1-methyl-xanthine; PKA: cAMP-dependent protein kinase; PKI: PKI (6–20) amide, an inhibitor of PKA; Rp- and Sp-cAMPS: isomers of adenosine-3⬘,5⬘monophosphorothioate; PSS: physiological saline solution. * Corresponding author. Present address: Department of Molecular and Cellular Biology, Wyeth-Ayerst Research, CN-8000 Princeton, NJ 08543, USA.
The ability of PKA activation to trigger the terminatation of patency is mediated by the phosphorylation of cytoplasmic protein substrates. This was supported by the demonstration of cyclic nucleotide-dependent phosphorylation of a 32 kDa polypeptide in homogenates of H. cecropia follicles (Wang and Telfer, 1996). The response could not be detected at that time in intact follicles, but this deficiency has now been overcome by including the protein phosphatase inhibitor okadaic acid in the incubation medium. Here, we report further observations on phosphorylation of the 32 kDa substrate and consider what they can reveal about the role of cyclic nucleotides in the postvitellogenic development of ovarian follicles. Several cyclic nucleotide-dependent protein kinases have been isolated and characterized from the ovaries of Bombyx mori and shown to phosphorylate in homogenates primarily serine hydroxyl groups of yolk proteins (Takahashi, 1984, 1985, 1987). Yolk protein phosphorylation also occurs in homogenates of H. cecropia follicles, and we therefore need to determine whether it
0965-1748/00/$ - see front matter 2000 Published by Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 5 - 1 7 4 8 ( 9 9 ) 0 0 0 9 4 - 6
30
Y. Wang, W.H. Telfer / Insect Biochemistry and Molecular Biology 30 (2000) 29–34
too is promoted by cyclic nucleotides and is concerned with the control of vitellogenesis.
2. Materials and methods 2.1. Preparation of follicles and homogenates Vitellogenic follicles were dissected from the eight ovarioles of day 18 pharate adult females. At this stage of metamorphosis each of the eight ovarioles contains previtellogenic and a chain of approximaately 40 vitellogenic follicles with intact nurse cell caps. The dissecting medium (PSS) was a physiological saline containing 1 mM phenylthiourea to prevent melanin formation, and 70 mM sucrose to make it isosmotic with day 18 hemolymph (Wang and Telfer, 1996). Hemolymph proteins were removed from the intercellular spaces by washing the follicles twice for 20 min each in PSS with constant shaking. To separate oocytes from the epithelium of follicle cells, washed follicles were treated for 3 min with 1 mg/ml pronase in PSS. Sheets of follicle cells could then be separated from the oocyte by gentle pipetting. The oocytes were removed from the dissociated mixture with a wide-mouth pipet and the sheets of follicle cells were collected by centrifugation at 200g for 2 min. Both oocytes and follicle cells were washed with protease inhibitors (50 mg/ml PMSF, 1 mg/ml leupeptin, 5 mg/ml aprotinin, 5 mg/ml chymostatain, 5 mg/ml antipain, and 0.5 mg/ml pepstatin in PSS), and then with PSS. Individual follicles, single oocytes, and the follicle cells from single follicles were each homogenized in 100 µl of cold 50 mM HEPES (pH 7.0), 1 mM EDTA, 1 mM dithiothreitol (DTT), and protease inhibitors as listed above. A portion of the homogenate was saved for protein determination (DC protein assay from Bio Rad), while the remainder was either stored at ⫺70° or processed immediately as described below. Particulate fractions were prepared for phosphorylation by centrifuging a whole follicle homogenate successively at 1000, 20,000, 80,000 and 150,000g. The 1000g pellet was discarded, since it contained any cells and yolk bodies that had survived homogenization intact. The succeeding three pellets were each resuspended in 0.1 ml homogenization buffer. A fourth fraction, the 150,000g supernatant, had a volume of 0.5 ml. 2.2. Protein phosphorylation Phosphorylation of unfractionated homogenates utilized 10 µl samples—the equivalent of one tenth of a follicle. Resuspended centrifugal pellets, by contrast, were normalized for protein content, each reaction mixture containing 50 µg of pellet protein. These amounts of homogenate or pellet were mixed with a phosphorylation
medium to yield a final volume of 50 µl containing 50 mM HEPES (pH 7.0), 10 mM MgCl2, 1 mM DTT, 20 µM ATP, and 10 µCi [γ 32P]-ATP (initial specific activity about 60,000 Ci/mmol). Dependence of phosphorylation on cyclic nucleotides was assayed by including in the reaction mixture predetermined concentrations of cAMP, cGMP, or analaogs of these, and/or 1 µM PKI (6–20) amide, a peptide inhibitor of PKA. The final mixture was incubated for 3 min at 30°. The reaction was terminated and the sample prepared for electrophoresis by adding 50 µl of 2X SDS-PAGE sample buffer (0.125 tris–HCl, pH 6.8, 4% SDS, 20% glycerol, 10% 2-mercaptoethanol, and 0.002% bromphenol blue) and heated for 4 min at 100°. To study phosphorylation in intact cells, 8–10 vitellogenic follicles were incubated in 200 µl PSS containing 10 µCi [32P]-phosphate and the indicated concentration of a cAMP analog or other reagent. Okadaic acid, a protein phosphatase inhibitor (Bialojan and Takai, 1988; Cohen et al., 1990), was also added to the incubation media. After a 2 h incubation at 23°, the follicles were homogenized in 500 µl SDS-PAGE buffer and boiled for 5 min. 2.3. Electrophoresis and autoradiography Phosphorylated proteins were separated by SDSPAGE and visualized by autoradiography. One-dimensional gel electrophoresis was performed with a 4% stacking gel and a 10% separating gel (Laemmle, 1969). For any one gel, equal sample volumes, standardized either for amount of protein or for follicle fraction (generally, about 1/20th of a follicle per lane), were added to each well. The gels were stained with Coomassie Blue R and then dried. Autoradiography utilized Du Pont Reflection film and a Reflection intensifying screen, with exposures of 4 h up to 2 days at ⫺72°. 2.4. Assay of PKA PKA was assayed with a kit from Gibco BRL (Gaithersberg, MD). Ten µl of homogenate, prepared as described above, were added to 29 µl of 50 mM tris– HCl (pH 7.5), 0.25 mg/ml bovine serum albumin, 50 µM Kemptide (peptide substrate for PKA) and 10 µM cAMP, with or without 1 µM PKI. After incubation at room temperature for 15 min, phosphorylation was started by adding 1 µCi [γ 32P]-ATP in 10 µl of the above buffer, and the tubes were incubated at 30° for an additional 5 min. Twenty µl of each reaction mixture were then spotted onto a phosphocellulose disc, and washed twice in 500 ml 1% phosphoric acid and once with distilled water. PKA activity was calculated by subtracting 32P incorporation in the presence of cAMP and PKI from that in the presence of cAMP alone.
Y. Wang, W.H. Telfer / Insect Biochemistry and Molecular Biology 30 (2000) 29–34
2.5. Chemicals cAMP, cGMP, Mb-cAMP, Deoxy-mb-cAMP, MbcGMP, and IBMX were from Sigma (St. Louis, MO). Forskolin, Dideoxy-forskolin, and PKI (Cheng and Kemp, 1986) were from Calbiochem (LaJolla, CA). Rpand Sp-isomers of cAMPS (Erneux et al., 1986) were from Biolog Life Science Institute (Bremen, Germany). [γ-32P]-ATP and [32P]O4 were from New England Nuclear (Wilmington, DE). All other chemicals were from Sigma. 3. Results 3.1. Cyclic AMP-dependent protein phosphorylation Detection of phosphorylation in both whole follicles and homogenates was more sensitive than reported earlier (Wang and Telfer, 1996). Okadaic acid, a cell-permeant protein phosphatase inhibitor, raised the level of 32 kDa protein labeling in living follicles to a clearly detectable level (Fig. 1). By itself, 10 µM okadaic acid
Fig. 1. SDS-PAGE autoradiogram of cyclic nucleotide-dependent phosphorylation in intact follicles. Vitellogenic follicles were incubated for 2 h in PSS containing [32P]-phosphate and the additions indicated below, before homogenization and electrophoresis. Additions to the labeling media included: lane 1, none; lane 2, 10 µM okadaic acid; lane 3, 1 mM MB-cAMP and 10 µM okadaic acid; lane 4, 250 µM forskolin and 10 µM okadaic acid. The asterisk indicates the position of the 32 kDa substrate. Molecular weight standards (×10⫺3 kDa) are indicated on the left. The label-free zones centered at 42 and 31 kDa in all lanes are the positions, respectively, of the small subunit of vitellogenin and microvitellogenin, both of which are so concentrated that they exclude other proteins from their positions (see Fig. 5A, lanes 4– 6). There is no indication that cyclic nucleotides promoted phosphorylation of the 45 and 27 kDa substrates that are demonstrated below in homogenates.
31
did not alter the protein phosphorylation patterns (Fig. 1, lane 2), but when combined with 1 mM Mb-cAMP (lane 3), forskolin (lane 4), Sp-cAMPS or IBMX (not shown), labeling of the 32 kDa polypeptide became evident. Neither deoxy-Mb-cAMP, dideoxy-forskolin, nor Rp-cAMPS promoted 32 kDa polypeptide phosphorylation, even when combined with 10 µM okadaic acid (not shown). In combination, these responses indicate that phosphorylation of the 32 kDa substrate can in fact be a physiological, in situ response to PKA activation. Improvement of detection in follicle homogenates was achieved by including a broad range mixture of protease inhibitors in the labeling medium. Cyclic nucleotidedependent phosphorylations were now detected not only at 32 kDa, but also at 58, 45, and 27 kDa (Fig. 2, lanes 2 and 5). Supplementing the incubation mixture with catalytic subunits of PKA yielded phosphorylation of additional polypeptides and a general intensification of label (lane 4). Revealed labeling patterns are influenced by the relative rates of kinase and phosphatase activities and the four bands indicated by asterisks in Fig. 2 are simply the most easily detected of a larger complement. In whole follicles (Fig. 1), labeling was not detectable at either 45 or 27 kDa; cAMP-independent labeling was so heavy above the 50 kDa level that detecting cyclic nucleotide-dependent labeling of a 58 kDa band was not
Fig. 2. SDS-PAGE autoradiogram showing 32P labeling of proteins in homogenates. Each lane on the gel contained labeled material from 0.02 follicle. Labeling had been for 5 min with [γ-32P]ATP in the following: lane 1, no additions; lane 2, 50 µM cAMP; lane 3, 50 µM cAMP and 1 mg/ml PKI; lane 4, 50 µM cAMP and 10 units catalytic subunits of PKA; lane 5, 50 µM cGMP; lane 6, 50 µM cGMP and 1 mg/ml PKI; lane 7, 1 mg/ml PKI. Molecular weight standards (×10⫺3 kDa) are indicated on the left. Asterisks mark the positions of four bands whose phosphorylations were promoted by cyclic nucleotides.
32
Y. Wang, W.H. Telfer / Insect Biochemistry and Molecular Biology 30 (2000) 29–34
possible. Lack of a cyclic nucleotide response at 45 and 27 kDa could be explained in two ways. Either these products are particularly sensitive to hydrolytic degradation during homogenization, or compartmentation in intact follicles keeps them isolated from cytoplasmic kinases in the absence of homogenization. 3.2. Kinase activity promoted by cGMP Like Mb-cAMP, Mb-cGMP can trigger the end of vitellogenesis by intact follicles (Wang and Telfer, 1996), and also, shown here, the phosphorylation of the 58, 45, 32, and 27 kDa substrates in homogenates (Fig. 2, lane 5). Since PKI blocked the effects of cGMP (Fig. 2, lanes 6), these responses are mediated by an ability to activate PKA. Ten-fold higher concentrations of cGMP than cAMP were required to effect the same levels of phosphorylation, however (Fig. 3), so binding of the former to the regulatory subunit of PKA must be weaker than that of cAMP. 3.3. Substrate and PKA localization In a step-wise centrifugation procedure, three of the four cAMP-promoted bands indicated by asterisks in Fig. 2 proved to be particle-associated (Fig. 4). The 32 kDa band was labeled in all three resuspended pellets, but was not detected in the 150,000g supernatant. Supplementing the labeling medium with excess catalytic subunits of PKA intensified the labeling at 32 kDa in all three pellets, but did not make it detectable in the 150,000g supernatant or change the status of the 20– 80,000g pellet as the fraction in which labeling was the heaviest (Fig. 4, lanes c). The profile thus reflected the distribution of the 32 kDa substrate, rather than of PKA. The 27 kDa band, though weaker, was distributed similarly. The 45 kDa band was most concentrated in the 1–20,000g pellet. The 150,000g supernatant showed no evidence of the three particle-associated bands, yielding instead three bands that labeled in response to cAMP in the general region of 58 kDa. In homogenates of isolated follicle cells, cAMP-
Fig. 3. SDS-PAGE autoradiogram showing the effects of cyclic nucleotide concentration on 32 kDa protein phosphorylation. Homogenate preparation and labeling were as described for Fig. 2. Cyclic nucleotide concentrations were: lane 1, 0 µM; lanes 2–4, 1, 10, and 100 µM cAMP, respectively; lanes 5–7, 1, 10, and 100 µM cGMP, respectively. The asterisk indicates the position of the 32 kDa band.
Fig. 4. SDS-PAGE autoradiograms of phosphorylation in centrifugal fractions of a whole follicle homogenate. Lanes a, no additions; lanes b, 50 µM cAMP, lanes c, 1 unit of catalytic subunits of PKA. Asterisks on the left indicate the labeling corresponding to the 45, 32, and 27 kDa bands of cyclic nucleotide-dependent phosphorylation seen in unfractionated homogenates (Fig. 2). The 58 kDa substrate was not visible in the particulate fractions shown here, but may be represented by one of the bands near the asterisk on the right. Solubilized vitellogenin was concentrated in the 80–150,000g pellet, its small subunit being heavily labeled at the 40–44 kDa level.
dependent phosphorylation was detected at 45, 32, and 27 kDa (Fig. 5B, lanes 1 and 2); as in Fig. 2, addition of catalytic subunits of PKA (lane 3) resulted in intensification of all bands. Of the four cAMP-responsive phosphorylations indicated by asterisks in Fig. 2, only the 58 kDa band was visible in homogenates of isolated oocytes. That the method of separating follicle cells and oocytes was effective was indicated by the absence of yolk protein bands from follicle cell homogenates when this gel was stained with Coomassie blue (Fig. 5A, lanes 1–3). Also examined was the distribution of PKA between the two cell types. PKA activity was four times greater in homogenates of follicle cells than of oocytes. Measured as picomoles of phosphate transferred from ATP to kemptide/min/follicle, the activities were 57.4±2.3 for whole follicles, 44.5±0.74 for isolated follicle cells, and 10.9±1.3 for isolated oocytes (average±s.e., N=5). 3.4. Phosphorylation of yolk proteins The 43 kDa subunit of vitellogenin was heavily phosphorylated in oocyte homogenates (Fig. 5B, lanes 4–6),
Y. Wang, W.H. Telfer / Insect Biochemistry and Molecular Biology 30 (2000) 29–34
33
(Takahashi, 1987; Sato and Yamashita, 1989; Tsuchida et al., 1992). Cyclic nucleotide-independent labeling in this region of the autoradiogram in Fig. 1 was too complex and dark to yield this information for H. cecropia. Finally, the 31 kDa microvitellogenin band was not labeled by 32P, even in homogenates of isolated oocytes (Fig. 5B, lanes 4–6) and thus presumably lacks appropriate phosphorylation sites.
4. Discussion
Fig. 5. SDS-PAGE gels of homogenates of isolated oocytes and follicle cells. Homogenates were each divided into three aliquots; labeling was done with [γ32P]-ATP under standard phosphorylating conditions, but with the following additions: lanes 1 and 4, none; lanes 2 and 5, 10 µM cAMP; lanes 3 and 6, 1 unit catalytic subunit of PKA. Asterisks indicate the positions of bands whose phosphorylation was promoted by cAMP in Fig. 2.
even in the absence of cAMP and catalytic subunits of PKA (lane 4). It is unlikely that this was promoted by endogenous cyclic nucleotides, because the 58 kDa substrate in the same preparation failed to phosphorylate unless either cAMP or catalytic subunits of PKA were added (Fig. 5B, lanes 4–6, asterisk). When phosphorylation was performed in intact follicles, by contrast, the 43 kDa vitellogenin subunit was not phosphorylated (Fig. 1). The high concentration of yolk protein at this location excludes other proteins with similar mobility in SDS-PAGE, so that a blank space appears in the autoradiogram at 41–44 kDa. The absence of vitellogenin labeling in intact follicles can be ascribed to compartmentation—yolk proteins are enclosed by membranes in intact follicles (Telfer, 1961), while the kinases that phosphorylate them in homogenates are presumably cytoplasmic. Heavy labeling also bracketed the 58 kDa band in homogenates of isolated oocytes (Fig. 5B, lanes 4–6). As with the 43 kDa vitellogenin subunit, the bracketing bands of label occurred in the absence of added cAMP or catalytic subunits of PKA. This is the location of paravitellogenin, the follicle cell contribution to the yolk (Bast and Telfer, 1976; Rubenstein, 1979), whose phosphorylation might be expected in intact follicles, since yolk proteins synthesized by follicle cells in other lepidoptera have been shown to be phosphorylated
Labeling of 32 kDa substrate in intact follicles raises the likelihood that this reaction is a physiological response to a rise in the level of cAMP. In homogenates, additional cyclic nucleotide-dependent phosphorylations occurred, but since labeling was not evoked at these locations in whole follicles, even when incubated in okadaic acid, homogenates may well generate artifactual products whose precursors are not normally available to PKA or to PKA-activated kinases. A clear example for this model was provided by vitellogenin, whose 43 kDa subunit was heavily phosphorylated in homogenates of oocytes (Fig. 5B) but remained unlabeled in whole follicles (Fig. 1), where it is confined to membrane-limited yolk spheres. It is unlikely that phosphorylation of the 32 kDa substrate, at least to a level that is autoradiographically detectable, is required for the termination of vitellogenin uptake, for concentrations of cell permeant cAMP analogs that stimulate vitellogenic follicles to stop forming yolk have consistently failed to induce detectable labeling at 32 kDa (Wang and Telfer, 1996). In the most fully characterized PKA-based signal transduction system in lepidoptera, cAMP treatment of intact prothoracic glands promotes the labeling of a 32/34 kDa polypeptide (Rountree et al., 1992) that has been implicated as ribosomal protein S6 (Song and Gilbert, 1994). In a variety of mammalian cells and tissues, S6 is also selectively phosphorylated in response to cAMP, as well as to other developmental and mitogenic stimuli (reviewed by Stewart and Thomas (1994)). In another insect system, P-element insertions into the S6 gene of Drosophila melanogaster resulted in lethal tumorigenesis of hematopoetic tissues. The 32 kDa phosphorylation substrate of H. cecropia follicle cells has not been identified, but its size in SDS-PAGE, its distribution in centrifugal fractions, and its status as the most prominent cAMP-dependent phosphorylation product of intact follicles are characteristics suggestive of S6. PKA-activated phosphorylation at 32 kDa also resembles many examples of S6 phosphorylation in its timing relative to a complex of developmental and secretory changes. Loss of patency and termination of vitellogenin uptake are followed in situ by a set of devel-
34
Y. Wang, W.H. Telfer / Insect Biochemistry and Molecular Biology 30 (2000) 29–34
opmental changes that include cessation of synthesis of the yolk protein produced by the follicle cells (Rubenstein, 1979), secretion of a hydrophobic layer around the oocyte (Telfer and Smith, 1970), and a complex sequence of chorionic protein secretions (Paul and Kafatos, 1975). The latter is not initiated until 20 h after the cessation of vitellogenin uptake (Telfer and Anderson, 1968), and was thus not tested by our assay system. Cyclic nucleotide-induced phosphorylation at 32 kDa, while not involved in blocking patency and the termination of vitellogenin uptake, could still turn out to be involved in the ensuing developmental events. Concerning yolk proteins, both the large and small subunits of B. mori vitellogenin have been shown to be post-translationally phosphorylated at their site of synthesis in the fat body (Takahashi, 1987), and there was in the same report some evidence that phosphates are also added in vitellogenic follicles. One of several forms of PKA that were isolated from B. mori ovaries was shown in that study to be able to phosphorylate isolated vitellogenin; but as in the present study (Fig. 5) the heavy labeling of vitellogenin that can occur in follicle homogenates was not promoted by cAMP. If, despite this counter-indication, phosphorylation of vitellogenin by PKA should in fact prove to be an ongoing feature of yolk deposition in silk moths, two requirements would have to be met. PKA would have to be transferred to the endosomal system to which vitellogenin is restricted (reviewed for insects by Raikhel and Dhadialla (1992)); and cAMP would have to be prevented from crossing gap junctions from the oocyte to the follicle cells (Woodruff, 1979) in sufficient quantity to trigger the termination of patency. These restrictions, combined with the inability of Mb-cAMP to promote yolk labeling in intact follicles (Fig. 1), make the cyclic nucleotidedependence of vitellogenin phosphorylation seem unlikely at the present time.
Acknowledgements Supported by GM-32909, a grant to WT from the National Institutes of Health.
References Bast, R., Telfer, W., 1976. Follicle cell protein synthesis and its contribution to the yolk of the cecropia moth oocyte. Devel. Biol. 52, 83–97. Bialojan, C., Takai, A., 1988. Inhibitory effect of a marine-sponge toxin, okadaic acid, on protein phosphatases. Biochem. J. 256, 283–290. Cheng, H.-C., Kemp, B.C., 1986. A potent synthetic peptide inhibitor of the cAMP-dependent protein kinase. J. Biol. Chem. 261, 989– 992. Cohen, P., Holmes, C.F.B., Tsukitani, Y., 1990. Okadaic acid: a new
probe for the study of cellular regulation. Trends Biochem. Sci. 15, 98–102. Erneux, C., Sande, J.V., Jastorff, B., Dumont, J.E., 1986. Modulation of cAMP action in the dog thyroid by its agonist and antagonist Sp- and Rp-adenosine 3⬘, 5⬘-monophosphothioate. Biochem. J. 234, 193–197. Laemmle, U., 1969. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227, 680–685. Paul, M., Kafatos, F.C., 1975. Specific protein synthesis in cellular differentiation. II. The program of protein synthetic changes during chorion formation by silkmoth follicles, and its implementation in organ culture. Devel. Biol. 42, 141–159. Raikhel, A.S., Dhadialla, T.S., 1992. Accumulation of yolk proteins in insect oocytes. Ann. Rev. Entomol. 37, 217–251. Rountree, D.B., Combest, W.L., Gilbert, L.I., 1992. Prothoracicotropic hormone regulates the phosphorylation of specific protein in the prothoracic glands of the tobacco hornworm, Manduca sexta. Insect Biochem. Molec. Biol. 22, 353–362. Rubenstein, E.C., 1979. The role of an epithelial occlusion zone in the termination of vitellogenesis in Hyalophora cecropia ovarian follcles. Devel. Biol. 71, 115–127. Sato, Y., Yamashita, O., 1989. Post-translational processing in the synthesis of egg-specific protein in the silkworm Bombyx mori. Insect Biochem. 19, 293–300. Song, Q., Gilbert, L.I., 1994. S6 phosphorylation results from prothoracicotropic hormone stimulation of insect prothoracic glands: a role for S6 kinase. Devel. Genetics 15, 332–338. Stewart, M.J., Thomas, G., 1994. Mitogenesis and protein synthesis: a role for ribosomal protein S6 phosphorylation. BioEsays 16, 809–815. Takahashi, S.Y., 1984. Ovarian protein kinases and endogenous substrates from the silkworm, Bombyx mori. Insect Biochem. 14, 219–229. Takahashi, S.Y., 1985. Characterization of guanosine 3⬘:5⬘-monophosphate-dependent protein kinase from silkworm eggs and analysis of the endogenous protein substrates. J. Comp. Physiol. B 155, 693–701. Takahashi, S.Y., 1987. Studies on the phosphorylation of ovarian proteins from the silkworm, Bombyx mori. Insect Biochem. 17, 141–152. Telfer, W., 1961. The route of entry and localization of blood proteins in the oocytes of saturniid moths. J. Biophys. Biochem. Cytol. 9, 747–759. Telfer, W., Anderson, L., 1968. Functional transformations accompanying the initiation of a terminal growth phase in the cecropia moth oocyte. Devel. Biol. 17, 512–535. Telfer, W., Smith, D., 1970. Aspects of egg formation. In: Neville, C. (Ed.), Symposium of Royal Entomology Society 5. Blackwell, Oxford, pp. 117–134. Tsuchida, K., Kawooya, J., Law, J., Wells, M., 1992. Isolation and characterization of two follicle-specific proteins from eggs of Manduca sexta. Insect Biochem. Molec. Biol. 22, 89–89. Wang, Y., Telfer, W., 1996. Cyclic nucleotide-induced termination of vitellogenin uptake by Hyalophora cecropia follicles. Insect Biochem. Molec. Biol. 26, 85–94. Wang, Y., Telfer, W., 1997. cAMP-stimulated termination of vitellogenesis in Hyalophora cecropia: formation of a diffusion barrier and the loss of patency. J. Insect Physiol. 43, 675–684. Wang, Y., Telfer, W., 1998a. cAMP-induced water uptake in a moth ovary: inhibition by bafilomycin and anthracene-9-carboxylic acid. J. Exper. Biol. 201, 1627–1635. Wang, Y., Telfer, W., 1998b. Pertussis-toxin sensitive G protein that supports vitellogenin uptake by promoting patency. Arch. Insect Biochem. Physiol. 39, 36–45. Woodruff, R., 1979. Electrotonic junctions in Cecropia moth ovaries. Devel. Biol. 61, 281–295.