Glycoprotein synthesis in Drosophila Kc cells. Biosynthesis of dolichol-linked saccharides.

The biosynthesis of dolichol and dolichol-linked saccharide intermediates in glycoprotein synthesis was studied in an embryonic Drosophila cell line (Kc) that lacks the squalene-cholesterol branch of the polyisoprenoid biosynthetic pathway. Kc cells were labeled with [5-3H]mevalonic acid and the radioactive lipids formed were analyzed. Although the major labeled product was coenzyme Q, dolichol and a variety of dolichol derivatives could be readily detected. On the basis of their chromatographic and chemical properties, these derivatives were identified as dolichyl phosphate, glucosylphosphoryldolichol, mannosylphosphoryldolichol, and oligosaccharylpyrophosphoryldolichol. Both short term (4-h) and steady state (4-day) labeling experiments with mevalonate, rather than sugars as previously used, were performed to assess the level of these intermediates. The results of these studies, using a precursor common to all the intermediates, reveal that the early intermediates, N-acetylglucosaminylpyrophosphoryldolichol and N,N'-diacetylchitobiosylpyrophosphoryldolichol, are present at very low levels (less than 5%) relative to the other intermediates on the pathway to oligosaccharylpyrophosphoryldolichol. The total amount of dolichol intermediates remained essentially constant during the chase phase of pulse-chase experiments, indicating the absence of a major catabolic pathway for the polyisoprenoid backbone. As expected, however, the sugar moiety, studied with mannosylphosphoryldolichol, underwent rapid turnover. These results are discussed in the context of our current understanding of the pathway whereby dolichol derivatives participate in glycoprotein synthesis.

The biosynthesis of dolichol and dolichol-linked saccharide intermediates in glycoprotein synthesis was studied in an embryonic Drosophila cell line (K) that lacks the squalene-cholesterol branch of the polyisoprenoid biosynthetic pathway. K, cells were labeled with mevalonic acid and the radioactive lipids formed were analyzed. Although the major labeled product was coenzyme Q, dolichol and a variety of dolichol derivatives could be readily detected. On the basis of their chromatographic and chemical properties, these derivatives were identified as dolichyl phosphate, glucosylphosphoryldolichol, mannosylphosphoryldolichol, andoligosaccharylpyrophosphoryldolichol. Both short term (4-h) and steady state (4-day) labeling experiments with mevalonate, rather than sugars as previously used, were performed to assess the level of these intermediates. The results of these studies, using a precursor common to all the intermediates, reveal that the early intermediates, Nacetylglucosaminylpyrophosphoryldolichol and N,N'diacetylchitobiosylpyrophosphoryldolichol, are present at very low levels (~5 % ) relative to the other intermediates on the pathway to oligosaccharylpyrophosphoryldolichol. The total amount of dolichol intermediates remained essentially constant during the chase phase of pulse-chase experiments, indicating the absence of a major catabolic pathway for the polyisoprenoid backbone. As expected, however, the sugar moiety, studied with mannosylphosphoryldolichol, underwent rapid turnover. These results are discussed in the context of our current understanding of the pathway whereby dolichol derivatives participate in glycoprotein synthesis.
Among higher organisms, insects are unique because they cannot synthesize cholesterol, a result of the absence of the enzyme that converts farnesylpyrophosphate to squalene. Nlinked polymannose glycoproteins have been shown to be synthesized in a mosquito cell line (l), and the detection of sugar-linked dolichol derivatives (2) indicates that the glycoprotein assembly process in insects is analogous to that found in higher organisms. Given these facts and the extensive studies by Watson and co-workers (3)(4)(5)(6)) that showed that Drosophila Kc cells utilize the mevalonate pathway for synthesis of dolichol and coenzyme Q, it appeared that these cells * This work was supported by Grant GM 33184 (to W. J. L.) from the National Institutes of Health. 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.
$ A Robert A. Welch Professor of Chemistry. To whom correspondence should be addressed. would be ideal to study the synthesis and metabolism of dolichol and its glycosylated derivatives that are involved in glycoprotein synthesis.
Numerous studies (reviewed in Ref. 7) have examined the synthesis and metabolism of these derivatives based on sugarlabeling experiments rather than on measurements of the polyisoprenoid chain to which the saccharides are attached. It has been necessary to use sugars rather than mevalonate for labeling these compounds, because in most higher organisms dolichol and its compounds share a common biosynthetic pathway with cholesterol, which often exceeds their amount by 100-1000-fold. Consequently, it was expected that in Drosophila cells the difficulty of isolation and quantification of the various glycosylated derivatives containing label in the dolichol chain would be obviated.
The results of these studies revealed that the major mevalonate-derived lipid product formed in Drosophila Kc cells and extractable by CHCb:CH30H (2:l) was coenzyme Q. The dolichol-related compounds in this fraction were found to be free dolichol, dolichyl phosphate, mannosylphosphoryldolichol, and glucosylphosphoryldolichol. Little or no N,N'diacetylchitobiosylpyrophosphoryldolichol could be detected. Oligosaccharylpyrophosphoryldolichol was identified in the CHC13:CH30H:H20 (10:103) extract from mevalonate-labeled cells. Pulse-chase experiments suggested that dolichyl phosphate is the initial product and that dolichol is slowly formed by its hydrolysis. The relative level of dolichol-linked saccharide derivatives was determined in short and long term labeling experiments. In addition, the effect of several agents on the synthesis of dolichol and its derivatives were studied. The results are discussed in the context of the relative levels and the dynamics of dolichol-linked intermediates in glycoprotein synthesis in these cells. Thin-layer Chromatography-Thin-layer chromatography (TLC) was carried out on Silica Gel 60 precoated glass plates or plastic plates (Merck) or on reversed-phase RP-18 precoated glass plates (Merck). The solvent systems used for development were as follows: solvent system A, petroleum ether:benzene:methanol (50:48:2); solvent system B, ch1oroform:methanokconcentrated ammonia:water (65:35:4:4); solvent system C, propanokconcentrated ammonia:water (63:l); solvent system D, diisobuty1ketone:glacial acetic acidwater (60:45:6); solvent system E, benzene:ethyl acetate (4:l); solvent system F, acetone; solvent system G , ch1oroform:methanokwater (10103). The position of migration of authentic standards was visualized with iodine vapor. Radioactivity was determined on air-dried plates by scraping (glass plates) or cutting (plastic plates) 0.5 X 2-cm sections, which were then placed in scintillation vials containing 5 ml of Liquiscint (National Diagnostics, Somerville, NJ). For autoradiography, the plate was sprayed three times with EN3HANCE and exposed on Kodak X-Omat x-ray film at -80 "C for a week. The relative intensities of the bands of 3H-or 14C-labeled compounds were measured on film with a scanning densitometer.
Incorporation of [3H]Mevalonic Acid into Lipids-Specific experimental conditions are presented in the appropriate figure or tables. The general method for labeling was as follows: Drosophila K, cells ( 106-107 cells/ml) in suspension were harvested by centrifugation and the cell pellets were washed once with Echalier's D22 medium. The washed cells were suspended in fresh Echalier's medium containing 1% fetal calf serum. The cells were labeled with [3H]mevalonic acid at room temperature with stirring at 100 rpm and bubbled for 1 min every 20 min with air through a syringe fitted with a sterile filter. After being labeled for 4 h, the suspension was chilled on ice, centrifuged, and washed with ice-cold Robb's physiological saline (12). CHC13:CH30H (21) was added to the cell pellets, and the suspension was incubated at 37 "C for 10 min. The extraction with CHC&:CH30H was repeated three times. The combined CHCI3:CH3OH extracts were back-washed once with 0.9% NaCl and then twice with 0.9% NaC1:CH30H (1:l). The chloroform extracts were dried with a stream of N1 and redissolved in a small volume of CHC13:CH30H. After CHC13:CH30H extraction, the cell pellets were washed with water at least five times and then further extracted with CHC13:CH30H:H20 (10103) three times.
For analysis, an aliquot of the CHC13:CH30H extract was subjected to ion-exchange chromatography on a DEAE-cellulose column equilibrated with CHC13:CH30H. After the column was washed with the same solvent, radioactive metabolites were eluted stepwise by 50,100, and 200 mM CH3COONH4 in CHC13:CH30H. The uncharged fraction that did not bind to the DEAE-cellulose column and the charged fraction that was eluted by 50 mM CH3COONH4 in CHC13:CH30H were further analyzed by silica gel TLC.
Incorporation To measure protein-linked oligosaccharides, 10% (v/v) of trichloroacetic acid was added to the pellets. The mixture was incubated at 100 "C for 10 min, chilled, and stored overnight at 4 "C. The precipitates were collected by centrifugation, washed two times with cold 10% trichloroacetic acid, and then washed four times with diethyl ether. The precipitates were dried with a stream of NB and then digested with Pronase for 48 h as previously described (13). The insoluble material was removed by centrifugation and the supernatant was analyzed by gel filtration on a Sephadex G-50 column equilibrated with 50 mM HCOONH,. The radioactive fraction with VJV, of 1.8 was collected and further analyzed by affinity chromatography on a concanavalin A-agarose column.
For determination of dolichol-linked saccharides, CHC13:CH30H was added to the cell pellet. The extraction was repeated three times and the combined extracts were back-washed with 0.9% NaCl and with 0.9% NaCl:CH30H (1:l) twice and analyzed by ion-exchange chromatography on a DEAE-cellulose column equilibrated with CHCL:CH3OH. The fractions eluted directly by 0.1 M CH3COONH, in CHC13:CH30H were pooled and analyzed by silica gel TLC in solvent system B. The pellets insoluble in CHC13:CH30H were dried with a stream of NP and washed with water at least five times until no radioactivity in the supernatant was detected. Then the wet pellet was dried with NZ and extracted with CHC13:CH30H.H20 three times in order to recover the oligosaccharide-lipid. lA). The fraction containing uncharged components not bound to DEAE-cellulose was analyzed by TLC. The major radioactive component (I) was found to correspond to coenzyme Q (Fig. lB), as previously reported (3, 4). A second uncharged radioactive component (ZI), migrating slower than coenzyme Q, had the mobility of dolichol, whereas a third one coincided with authentic farnesol. Two other minor components, one with a mobility between that of components I and I1 and another remaining at the origin, were not identified.

Characterization of Metabolites
The charged fraction, containing components bound to DEAE-cellulose, was analyzed by TLC. Two radioactivity peaks were observed; the slower moving one (V) migrated with the mobility of dolichyl phosphate ( Further analysis was performed as follows on each of the following compounds after their isolation by preparative TLC as described under "Experimental Procedures." Dolichol-The putative labeled dolichol fraction (11) cochromatographed with authentic dolichol by TLC in solvent system E. Furthermore, it could be resolved into a family of dolichol homologues by reversed-phase TLC (Fig. 2). In contrast to the profile of pig liver dolichol, the insect dolichol consists of three homologues with the major carbon chain length ranging from C75 to C, . To test the possibility that the insect dolichol was actually unsaturated, i.e. didehydrodolichol, it and authentic 14C-labeled dolichol were separately treated with mild acid (14,15) and analyzed by TLC in solvent system E. Neither authentic dolichol nor the insect dolichol were altered in chromatographic mobility after acid treatment, indicating that the cy-isoprene unit was not unsaturated (14, 15).
Dolichyl Phosphate-The major product formed after wheat germ acid phosphatase treatment of putative labeled dolichyl phosphate (V) had the mobility of dolichol. To test for the possible presence of unsaturated didehydrodolichyl phosphate, both the insect dolichyl phosphate fraction (V) and authentic 14C-labeled dolichyl phosphate were treated with mild acid and then analyzed by TLC in solvent system B. presence of a phosphodiester bond in compound IV.
To identify the sugar moiety in the putative glycosylphosphoryldolichol, the metabolite was compared with authentic [3H]mannosylph~sphoryyldolichol by TLC in solvent system B. Surprisingly, the unknown metabolite was found to consist of two radioactive bands (Fig. 3A, lane 2, and Fig. 3B, lane l ) , with the lower, faint band corresponding in mobility to 13H] mannosylphosphoryldolichol (Fig. 3A, lane 3). As shown in Fig. 3B (lane 2), the upper, major band of compound IV was found to coincide with authentic [3H]glucosylphosphoryldolichol. As shown in lane 4, [3H]mannosylphosphoryldolichol and [3H]glucosylphosphoryldolich~l are separated from each other in this system using solvent system B. The ability to separate mannosylphosphoryldolichol from glucosylphosphoryldolichol by TLC, with the latter migrating faster than the former, has been reported (16).
To confirm the identity of the insect cell mannosylphosphoryldolichol, cells were labeled with [3H]mannose in glucose-free Echalier's medium. The TLC profile of the fraction that bound to DEAE-cellulose is shown in Fig. 4. The mobility of the fastest moving mannose-labeled compound (lane 3) corresponded to that of the minor peak obtained with mevalonate labeling (lane 1); this conclusion was confirmed when a mixture of the two were analyzed ( l a n e 2). It is noteworthy that there was no significant difference in the relative amounts of mevalonate-labeled glycosylphosphoryldolichol and dolichyl phosphate synthesized in glucose-free and in regular Echalier's medium (Fig. 4, lane 1, and Fig. IC). In a separate experiment, cells were labeled with [3H]galactose, which can serve as a precursor of UDP-[3H]Glc in other cells. Analysis of the charged lipids by TLC revealed the presence of a product with the mobility of authentic glucosylphosphoryldolichol (data not shown). Taken together these results indicated that peak IV consisted of a mixture of glucosylphosphoryldolichol and mannosylphosphoryldolichol, with the former being the major component.
N-Acetylglucosamine-containing Dolichl Intermediutes-Neither N-acetylglucosaminylpyrophosphoryldolichol nor N,N'-diacetylchitobiosylpyrophosphoryldolichol could be detected in the fraction that bound to DEAE-cellulose, as shown in Fig. 5. To exclude the possibility that these two lipids were formed but degraded during extraction, a mixture of authentic 3H-labeled glucosaminylpyrophosphoryldolichol and N,N'diacetylchitobiosylpyrophosphoryldolichol was added to a pellet of K, cells. The cell pellets were then immediately extracted with CHC13:CH30H and the extract was chromatographed on a DEAE-cellulose column as described under "Experimental Procedures." Over 95% of the added radioactivity was recovered in the CHC13:CH30H extract. Moreover, greater than 90% of the radioactivity in the CHC13:CH30H extract bound to the DEAE-cellulose column and was subsequently eluted with ammonium acetate.

Characterization of Metabolites Extractable with
CHCl3:CH30H:H20 The putative labeled oligosaccharide-lipid derived from [3H]mevalonic acid and recovered in the CHCI3:CH30HH20 fraction was subjected to silica gel TLC in solvent system C. It exhibited a single radioactivity peak with an R F value of 0.16. The compound did not migrate from the origin in solvent systems B or G. The putative oligosaccharide-lipid had the 15614 Dolichol Metabolism in Drosophila --

+-Chito-PP-Dol
-- same chromatographic mobility in solvent system C as ["C] isopentenyl pyrophosphate, however, unlike this compound, it was resistant to alkaline phosphatase treatment. When it was treated with mild acid (0.1 M HCl, 50 "C, 30 min), two radioactivity peaks that migrated faster than the starting compound were observed upon TLC in solvent C. The faster moving one coincided in mobility to dolichyl phosphate; the slower moving one was tentatively identified as dolichyl pyrophosphate because 1) this chromatographic system separated polyisoprenyl monophosphate from polyisoprenyl pyrophosphate (15) and 2) stronger acid treatment (0.1 M HC1, 95 "C, 30 min) increased the amount of dolichyl phosphate and decreased the amount of putative dolichyl pyrophosphate. The insect oligosaccharylpyrophosphoryldolichol bound to DEAE-cellulose equilibrated with CHCI3:CH30HH20 and eluted more slowly than authentic [3H]mannosylphosphoryldolichol with a linear gradient of 0-0.1 M ammonium acetate in CHCl3:CH30H:HZO.
To support the conclusion that the product isolated in the CHC13:CH30H:H20 fraction was oligosaccharylpyrophosphoryldolichol, authentic [14C]glucose-labeled oligosaccharylpyrophosphoryldolichol was synthesized in vitro in hen oviduct microsomes in the presence of UDP-N-acetylglucosamine, GDP-mannose, UDP-[14C]glucose, and dolichyl phosphate (10). In Fig. 6 are shown the TLC profiles of a mixture of insect cell 3H-labeled oligosaccharylpyrophosphoryldolichol and authentic [14C]glucose-labeled oligosaccharylpyrophosphoryldolichol before and after treatment with 0.1 M HCl at 95 "C for 30 min. Before acid hydrolysis (Fig. 6A), both the 'H-labeled compound and ['4C]oligosaccharylpyrophosphoryldolichol migrated with the same mobility, although the insect preparation contained an additional minor component of high mobility that probably is a breakdown product. TLC analysis (Fig. 6B) after acid treatment revealed that the majority of the 3H-labeled compound then migrated with a mobility identical to dolichyl phosphate, whereas the I4C-labeled product, as expected for a free oligosaccharide, remained at the origin.

Dynamics of [3HlMevalonic Acid-and [3HlManmse-labeled Dolichol Intermediates
Having characterized the major dolichol-linked intermediates synthesized by K, cells, we turned to measurements of

FIG. 6. TLC of the [SH]mevalonic acid-labeled compound in
the CHCls:CHsOH:H20 fraction. This fraction (-) was mixed with authentic [14C]glucose-labeled oligosaccharylpyrophosphoryldolichol (---) and then analyzed by TLC in solvent system C before (Panel A ) and after (Panel B ) mild acid treatment. In the case of acid treatment, samples were treated with 0.1 M HCI at 95 "C for 30 min, neutralized with 2 M Na2C03, and directly subjected to TLC. The recovery of radioactivity for 3Hand for "C-labeled compounds was 74 and 94% before and 76 and 82% after acid treatment, respectively. Dol, dolichol. their metabolism. The results of a pulse-chase experiment are summarized in Fig. 7. As shown in panel A, following short term labeling (4 h) when it is unlikely that the steady state has been achieved, one of the major products, dolichyl phosphate, apparently is converted to dolichol during the chase. Overall, the total amount of labeled dolichol derivatives remained constant. After long term labeling (panel B), the relative levels of each of the intermediates is quite different from that shown in panel A. Moreover, relatively little change in their levels occurred during the chase. The relative levels of the intermediates during short term (4-h) and steady state (4-day) labeling are summarized in Table I. It is evident from the chase experiments that the cellular level of the glycosyl dolichol derivatives remained relatively constant. If these compounds are involved as saccharide donors, one would expect that, even if the dolichol backbone is stable, the glycose moiety should exhibit rapid turnover. To test this idea, K, cells were pulse-labeled with ['Hlmannose and the mannosylated lipids were analyzed by TLC over the course of the chase. The results of this experiment revealed a dramatic decrease in the amount of [:'H]mannosylphosphoryldolichol observed during the chase (Fig. 8A). An unknown, charged metabolite derived from [3H]mannose, which was not labeled with mevalonic acid and therefore does not contain dolichol, decreased more slowly over the course of the chase. Quantitative analysis of the turnover of the mannosyl moiety of mannosylphosphoryldolichol indicated that its half-life (tu) was 3 h (Fig. 8B).

Effect of Ecdysone and Tunicamycin on Dolichol-related
Intermediates Although the K, cells responded to ecdysone (the molting hormone) by undergoing morphological changes identical to

TABLE I Composition of dolichol and its derivatives from Drosophila Kc cells labeled with [3HJmeva~nic acid
The K, cells were labeled with [3H]mevalonic acid as described in the legend to Fig. 7. The radioactivity found in coenzyme Q after labeling for 4 h or 4
days was 648,000 and 3,219,000 cpm, respectively. those observed in other reports (17-20), in the presence of this hormone no change in the synthesis of dolichol derivatives or in coenzyme Q (Table 11) was observed. Tunicamycin had little effect on coenzyme Q synthesis but, as expected, oligosaccharylpyrophosphoryldolichol synthesis was inhibited. It was surprising, however, that this drug caused an actual increase in total incorporation into the total dolicholrelated intermediates (Table 11). A more detailed study (data not shown) of the dose response of K, cells to tunicamycin established that the maximum effect was observed at the tunicamycin concentration (1 pg/ml) used in the experiment shown in Table 11. As the concentration of tunicamycin was increased from 0 to 1 pg/ml, incorporation into dolichyl phosphate remained nearly constant, whereas incorporation into both mannosylphosphoryldolichol and glucosylphosphoryldolichol progressively increased (3-and 2.1-fold, respectively). As expected, the amount of oligosaccharylpyrophosphoryldolichol decreased. These results suggest that the labeled dolichyl phosphate that accumulates as a result of inhibition of its conversion to N-acetylglucosaminylpyro-

15615
-El  phosphoryldolichol by tunicamycin is utilized in the formation of glycosylphosphoryldolichol, especially mannosylphosphoryldolichol. However, the mechanism whereby the total incorporation into these dolichol derivatives is increased in the presence of tunicamycin is not clear.

Uptake of ['T]Dolichol into Cultured Cells
Cells (2.9 X lo6 cells/ml) were incubated with 30&i of [l-Yldolichol during a 4-day period. Cells were then harvested, washed with Eschalier's medium four times, and then extracted with CHCl,:CHsOH and with CHC13:CH30H:Hz0 as described under "Experimental Procedures." As a control, an equivalent amount of ['Vldolichol was added to cells that were immediately harvested and extracted with CHC&:CH,OH and with CHC13:CH30H:Hz0. Analysis by DEAE-cellulose chromatography, followed by silica gel TLC in solvent system B and radioautography, revealed that exogenous [l-'4C]dolichol had been converted to dolichyl phosphate, g1ucosy1pyrophosphoryldolichol, and mannosylphosphoryldolichol. Although only 0.1% of the radioactivity recovered in the CHC13:CH30H fraction was present in the form of phosphorylated derivatives, this finding indicated that K, cells contain a kinase capable of phosphorylating free dolichol.
ized by a variety of chromatographic and chemical methods described under "Results." The CHC13:CH30H-extractable products were separated into charged and uncharged lipids, The uncharged lipids were shown to be coenzyme Q, farnesol, and dolichol; in agreement with an earlier study (3), no fatty acid esters of dolichol were found. The major charged isoprenoid compounds were shown to be dolichyl phosphate, glucosylphosphoryldolichol, and mannosylphosphoryldolichol. No labeled polyisoprenoid derivatives corresponding in mobility to N-acetylglucosaminylpyrophosphoryldolichol or N,N'diacetylchitobiosylpyrophosphoryldolichol were detected. In the CHC13:CH30H:H20 fraction, the major labeled product was found to be oligosaccharylpyrophosphoryldolichol.

Glycoproteins
To confirm that under the conditions of these in viuo experiments the K, cells were involved in assembly of Nlinked glycoproteins, the proteins from cells labeled with either [3H]glucosamine or [3H]mannose were precipitated with trichloroacetic acid and digested with Pronase. When the resulting glycopeptides were analyzed by Sephadex G-50 gel filtration three classes of labeled materials were separated; V,, primarily consisting of glycosaminoglycans excluded from the column (21,22), glycopeptides with a V,/V, of 1.8, and an included fraction (Vi). In the case of [3H]glucosamine labeling, 37% of radioactivity in the Pronase digest was associated with the glycopeptide fraction, whereas in the case of [3H]mannose, 76% of the radioactivity was found in the glycopeptide fraction. The two labeled glycopeptide fractions recovered from the gel filtration column were analyzed further on a concanavalin A-agarose column. About 80% of each of the glycopeptide fractions was bound to this lectin, indicating that they both contained a-mannosyl units in their oligosaccharide chains.
Having identified the isoprenoid products of metabolic labeling, we undertook to estimate the relative level of each during short and long term labeling. It is important to note that these labeling experiments provide a true estimate of the relative amounts of the labeled dolichol derivatives because all arise from a common precursor. This is in marked contrast to the many studies carried out over the past decade using sugar labeling of glycosyl moieties attached to dolichol. Quantification of these sugar-labeled intermediates has been difficult because of possible differences in (a) the rate of sugar uptake and (b) the pool sizes of the different sugars and sugar nucleotides. Under these circumstances the relative levels of the different radioactive glycosylated derivatives of dolichol may have little resemblance to the actual amount synthesized. DISCUSSION In a series of studies Watson and his co-workers (3-6) have demonstrated that a variety of insect cell lines, including Drosophila K cells, do not synthesize squalene or sterols. The major nonsaponifiable products synthesized from mevalonate in this cell line were shown to be coenzyme Q and dolichol; analysis for polar derivatives of dolichol was not carried out. Studies in a mosquito cell line showed that polymannose Nlinked glycoproteins are present (1) and that their assembly appeared to be mediated via dolichol intermediates (2). Based on these observations and the likelihood that studies on synthesis of dolichol and its derivatives would be facilitated by the absence of squalene and cholesterol biosynthesis, we undertook to determine if cultured K, cells would provide new insights into dolichol metabolism and the dolichol-linked pathway for glycoprotein synthesis. These cells are an established line of unknown tissue origin that was isolated from Drosophila embryos and shown to be responsive to the steroid hormone ecdysone (20). Cells were labeled with [3H]mevalonic acid and the labeled lipids, isolated by sequential extraction with CHCl,:CH,OH and CHC13:CH30H:H20, were character-Using mevalonic labeling, dramatic differences in the relative amounts of the labeled intermediates was observed between short and long term labeling. After 4 h of labeling, the major products were glucosylphosphoryldolichol and dolichyl phosphate. However, after long term labeling (4 days), glycosylphosphoryldolichol, oligosaccharylpyrophosphoryldolichol, and dolichol were all present at approximately equivalent levels and dolichyl phosphate was very low. Under neither labeling condition could significant amounts of N-acetylglucosaminylpyrophosphoryldolichol or N,N'-diacetylchitobiosylpyrophosphoryldolichol be detected. The rate and extent of loss of dolichyl phosphate corresponded to the rate and the extent of increase in dolichol, suggesting that dolichyl phosphate is the precursor of dolichol. After either short or long term labeling, the total level of labeled polyprenoid chains did not decrease during the chase, indicating the absence of a major catabolic pathway. Similar stability was observed in the level of glycosylphosphoryldolichol. However, this finding does not mean that this fraction is static. In fact, when the sugar moiety of the mannosylphosphoryldolichol component was labeled and then subjected to a chase, a halflife of approximately 3 h was observed. It is clear, then, that this pool is in a metabolically active state with respect to the glycosyl unit, although the total amount of glycosylphosphoryldolichol remains constant. This, coupled with the fact that the Ievel of free dolichyl phosphate only slowly declined over a 30-h period, suggests that the dolichyl phosphate moiety in mannosylphosphoryldolichol is rapidly reutilized for mannosylation.
These results, which for the first time in any biological system provide an estimate of the relative levels of a variety of newly synthesized dolichol derivatives, can be considered in the context of the biosynthetic pathway for glycoprotein synthesis shown in Scheme 1. The results suggest that the end product of de nouo synthesis from mevalonate is dolichyl phosphate. Some of this dolichyl phosphate is converted to dolichol while the rest is used for oligosaccharide-lipid assembly. Much earlier work has shown that assembly of oligosaccharylpyrophosphoryldolichol is initiated by conver-