Control of Juvenile Hormone Biosynthesis EVIDENCE FOR PHOSPHORYLATION OF THE 3-HYDROXY-3-METHYLGLUTARYL COENZYME A REDUCTASE OF INSECT CORPUS ALLATUM*

The activity of 3-hydroxy-3-methylglutaryl coenzyme A reductase, a key enzyme of juvenile hormone biosynthesis, has been measured in the supernatants of ho- mogenates (10,000 X g) prepared from the corpora al-lata of the adult tobacco hornworm moth, Munducu sextu Enzyme activity was inhibited 80% by 50 m~ NaF, a known phosphoprotein phosphatase inhibitor, if pres- ent during extirpation of the glands and all subsequent workup of the tissue. Reductase activity was also sig- nificantly decreased (20-30%) in homogenates preincubated with 4 m~ MgClz and 2 m~ ATP. These results provide evidence that reductase in the insect undergoes phosphorylation and dephosphoryla- tion similar to that occurring with reductase of mammalian liver. If so, this would provide a rapid method for modulating juvenile hormone synthesis.

The activity of 3-hydroxy-3-methylglutaryl coenzyme A reductase, a key enzyme of juvenile hormone biosynthesis, has been measured in the supernatants of homogenates (10,000 X g) prepared from the corpora allata of the adult tobacco hornworm moth, Munducu sextu Enzyme activity was inhibited 80% by 50 m~ NaF, a known phosphoprotein phosphatase inhibitor, if present during extirpation of the glands and a l l subsequent workup of the tissue. Reductase activity was also significantly decreased (20-30%) in homogenates preincubated with 4 m~ MgClz and 2 m~ ATP.
These results provide evidence that reductase in the insect undergoes phosphorylation and dephosphorylation similar to that occurring with reductase of mammalian liver. If so, this would provide a rapid method for modulating juvenile hormone synthesis.
Insect juvenile hormones are sesquiterpenoids produced in the corpora allata, a small pair of glands posterior to the brain. These compounds play important roles in the maintenance of the larval form and in maturation of the reproductive system in the adult female tobacco hornworm moth, Manduca sexta (1). Hormone levels fluctuate in both larvae and adults and often reach immeasurably low levels prior to metamorphosis (2). Production of juvenile hormone is believed to be controlled by feedback pathways (3) and by neural and humoral signals (3), and under certain circumstances, to be affected by the steroid molting hormone, 20-hydroxyecdysone (4).
The pathway for JH' biosynthesis shares common intermediates with sterol biosynthesis, and may share some aspects of the control mechanisms which operate on the sterol pathway. However, insects do not synthesize sterols (5) and juvenile hormone synthesis is a major isoprenoid pathway in insects. Our current understanding of the biosynthesis of cholesterol suggests that control is exerted principally upon the rate-limiting step catalyzed by the enzyme 3-hydroxy-3methylglutaryl-CoA reductase (EC 1.1.1.34) (6). Either cholesterol or cholesteryl esters contained in low density lipoproteins (7) or oxygenated derivatives of cholesterol (8) influence the activity and the amount of HMG-CoA reductase in cells.
* This work was supported by National Institute of General Medical Sciences Grants GM 13863 and 29238, and National Institute of Environmental Health Sciences Grants ES 05160. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Furthermore, the activity of HMG-CoA reductase in mammalian liver is markedly influenced by either Mg-ATP or F- (9). Inactivation of the reductase is effected through phosphorylation by ATP catalyzed by a protein kinase. A phosphoprotein phosphatase, which is inhibited by F-, can reactivate the reductase by removal of the covalently bound phosphate. Evidence to support this phosphorylation of the mammalian reductase has been provided by immunoprecipitation of the labeled protein after incubation of rat liver microsomes with [y-32P]ATP (10). In addition, it has been reported that in vitro incubation of purified HMG-CoA reductase with purified reductase kinase, MgC12, and [y-32P] ATP yielded a single labeled protein band on sodium dodecyl sulfate gel electrophoresis (11). The phosphorylation-dephosphorylation system has been implicated in short term feedback regulation of HMG-CoA reductase (12,13).
We have obtained evidence that the levels of HMG-CoA reductase in the CA of the tobacco hornworm, M. sexta, parallel, in most cases, the ability of the gland to synthesize J H (14). We have shown that hormone production by isolated glands is not influenced directly by J H or by 25-hydroxycholesterol, a potent inhibitor of cholesterol biosynthesis. We also reported that in CA homogenates, Mg-ATP and Fdid not influence the activity of HMG-CoA reductase. However, upon more careful examination, it has become clear that the activity of the CA enzyme is, in fact, altered by both Mg-ATP and Fin much the same way as the liver enzyme. We report here the results of these investigations.

MATERIALS AND METHODS
Animals-Colonies of M. sexta were raised as previously described (15). Eggs were generously provided by Dr. J . P. Reinecke, U. S. Department of Agriculture, Fargo, ND. Unfed adult animals were used 0-4 days after time of eclosion.
HMG-CoA Reductase-CA homogenates were assayed for HMG-CoA reductase activity as described by Kramer and Law (14) with the following modifications. CC-CA complexes were rinsed free of dissecting saline (16) by three 100-pl-washes with a buffer containing 50 mM KHzPO,, 200 mM KC1, 5 m~ dithiothreitol, 5 m~ EDTA, and 0.25% Kyro EOB, pH 7.4 (17). MgC12 and CaC12 were omitted from this dissecting saline if 50 m~ NaF was present. In experiments involving the addition of 2 mM ATP and 6 m~ MgC12, the phosphate buffer contained 1 m~ EDTA, rather than 5 mM EDTA. The washed CC-CA glands (20-30) were homogenized in 70 pl or 100 p1 of the above buffer, blended with a Vortex mixer, and warmed at 37 "C for 10 min. Tissue homogenates were centrifuged at 10, OOO X g and 23 CA).
"C for 10 min in a water-cooled microcentrifuge (Misco, Berkeley, HMG-CoA reductase activity was assayed by a modification of the method of Shapiro et al. (18). For each incubation, all stock solutions were adjusted to pH 7.4 with KOH before addition to the homogenate. Assays were initiated by the addition of a cofactor-substrate solution containing 450 nmol of NADP, 4.5 pmol of glucose-6-phosphate, 0.67 IU of glucose-6-phosphate dehydrogenase, and 20 nmol of DL-[methyl-3H]HMG-CoA (specific activity, 11-12 dpm/pmol). Incubations were carried out at 37 "C for 90 min in sealed Reacti-Vials housed within a Temp-Blok module heater shaking at 170 rpm. Reactions were terminated by the addition of 25 pl (35 1.1 for incubating volumes greater than 250 pl) of 6 N HCl, followed with 1020 or 1160 dpm of DL-[2-'4C]mevalonolactone (specifk activity, 18 Ci/mol) as an internal standard and warmed at 40 "C for 30 min with shaking. Corrected recovery of the lactone typically ranged from 85-95%. Each sample was applied to two Silica Gel G thin layer chromatography plates and developed in benzene/acetone (LI, v/v). Mevalonolactone was recovered in the region RF 0.38-0.69 by scrapping the silica gel into 10 ml of toluene/ethanol/Omnifluor (21:12, v/v/w). In all cases, a sample of each CC-CA homogenate was treated with 6 N HCl before addition of the cofactor-substrate solution to serve as a denatured control. Enzyme activities are expressed as picomoles of mevalonic acid synthesized/h/CA gland pair equivalent. (11.7 Ci/mmol) was ob-

RESULTS
The CA is a minute organ (one CC-CA pair weighs -20 pg, wet weight) and the amount of HMG-CoA reductase present is correspondingly small. Hence, activity measurements must be conducted over 1.5 h in order to obtain meaningful values for a small number of glands (3-6 CA pair equivalents/incubation). Mevalonate production has been shown to be linear for at least two h under the assay conditions. Extirpation of CA is practical only when the closely adhering CC and a portion of the aorta are also included. Measurement of reductase activity in CC and in aorta showed values representing only 6 and 2% of that measured in the CA, respectively. Therefore, the three tissues were homogenized together.
Nearly a 2-fold (70-95%) enhancement in reductase activity was observed when 0.25% (v/v) of the detergent, Kyro EOB, was added to the homogenates. This was determined by homogenizing only the right CA of each gland pair in 0.25% Kyro EOB, and using the left CA to serve as a nondetergent control, thereby avoiding animal-animal variation. This detergent has been used by other investigators to enhance reductase activity in preparations of mammalian tissue. EDTA (5 nm) had been routinely included in homogenization buffers for the measurement of HMG-CoA reductase. It was found that 1-5 mM EDTA enhanced CA reductase activity by 25-35%. It was not clear whether this elevation in activity was due to the removal of divalent metal ions which directly inhibited the reductase, due to inhibition of HMG-CoA lyase or mevalonate kinase, which could decrease substrate or products of the reductase reaction, respectively, or due to the removal of ions (specifically Mg") essential to a kinase involved in modulation of reductase activity through phosphorylation. Therefore, we investigated the effect of ATP and Mg+2 on CA reductase activity. It was reasoned that if a kinase were involved in inactivating HMG-CoA reductase, then a phosphatase would be involved in reactivating the reductase. NaF, a known inhibitor of phosphoprotein phosphatases, was, therefore, included when preincubating homogenates with Mg-ATP to avoid this reactivation and to enhance the inhibition of reductase activity observed. Table I shows that preincubation of CA homogenates with MgC12, ATP, and NaF, followed by assay in the presence of 30 mM EDTA, caused a significant decrease in enzymatic activity in both sexes when compared to controls. These results are consistent with a protein kinase activity which decreases reductase activity through Mg-ATP-dependent phosphorylation of the reductase. This inhibition was not due to a change in the ionic strength of the reaction mixture upon addition of Mg-ATP, since Mg-CTP controls did not differ from buffer controls. If an excess of EDTA (30 m~) was not added, a further apparent decrease in reductase activity was observed in male homogenates preincubated with Mg-ATP. There was no difference in the level of reductase activity measured in female homogenates preincubated with Mg-ATP and assayed with or without excess EDTA present. This apparent further decrease is most likely due to removal of mevalonate by mevalonate kinase, and this suggests that mevalonate kinase is active in homogenates of glands of males, but not in those of females.
Table I1 (A) shows that addition of 50 m~ NaF to CA immediately following homogenization and just prior to measurement of reductase activity resulted in only small decreases in activity with CA of either sex. We reasoned that the phosphatase, if present, might be active enough to dephosphorylate most of the reductase during the interim between extirpation of the glands and addition of F-. Therefore, we in the supernatant fraction (l0,OOO X g ) of male and female CA or buffer (Mg-ATP test with 1 m~ EDTA) immediately prior to the homogenate (50 pl) prepared from male and female CA. These were addition of the cofactor-substrate solution. The preincubation conpreincubated for 1.0 h at 37 "C with buffer only (buffer control; Mg-centrations of MgC12 and ATP were 6 m~ and 2 m~, respectively. CTP control) or with MgC12, ATP, and NaF. Preincubation volumes The fiial concentrations of NaF and EDTA were 50 mM and 30 m~, were 125 pl. Incubations (275-p1 final volumes) were carried out as respectively. n = 6 for each incubation except for Experiment I,  and fiquots (25 pl) were taken in duplicate. These were diluted with either 25 pl of 100 m~ NaF or additional buffer to serve as controls. In 2, male or female adults were decapitated, a section of the exoskeleton was removed from the heads to expose the CA, and the heads were placed in either control dissection saline or in saline supplemented with 50 m~ NaF. NaF was also present in the homogenization buffer and in the cofactor-substrate solution to maintain test tissues at 50 m~ NaF throughout the experiment.

DISCUSSION
The results reported here show that HMG-CoA reductase of M. sexta CA responds to Mg-ATP and Fin a manner quite analogous to the reductase of mammalian liver. This implies that the enzyme of CA exists in two forms, an active, nonphosphorylated form, and an inactive, phosphorylated form which are interconvertible by a protein kinase and a phosphoprotein phosphatase. Of course, confirmation of this hypothesis must await more definitive experiments, e.g. demonstration of the transfer of labeled phosphate from ATP to the enzyme as has been done with liver reductase. The difficulty in obtaining the required amount of purified enzyme from the minute CA will delay such work, however. If this scheme is correct for the insect CA, then it would seem that both reductase kinase and phosphoprotein phosphatase levels are about equal in CA homogenates prepared from adults of either sex. It is impossible at this stage to decide what the actual ratios of active and inactive forms of reductase are in the intact CA of the animals. It is of interest that we were able to show earlier (14) that CA of adult males, unlike those from adult females, produced no JH in maintenance culture, while CA homogenates had reasonable HMG-CoA reductase levels. It may be that extirpation and subst" quent maintenance of intact male CA resulted in rapid inactivation of the reductase through activation of the kinase, while in homogenates that contained EDTA the phosphatase rapidly reactivated the phosphorylated reductase. Another possible explanation is that denervation of the CA in males stimulates some alternate isoprenoid synthetic pathway beyond phosphomevalonate and this shunts intermediates away from juvenile hormone biosynthesis. Further experimentation is necessary to determine the nature of these sex differences. Whatever the reason, it appears that phosphorylation-dephosphorylation may provide rapid modulation of J H biosynthetic capacity in insects of both sexes.