Regulation of the Biosynthesis ofN-Acetylglucosaminylpyrophosphoryldolichol, Feedback and Product Inhibition*

The assembly of the core oligosaccharide region of asparagine-linked glycoproteins proceeds by means of the dolichol pathway. The first step of this pathway, the reaction of dolichol phosphate with UDP-GlcNAc to formN-acetylglucosaminylpyrophosphoryldolichol (GlcNAc-P-P-dolichol), is under investigation as a possible site of metabolic regulation. This report describes feedback inhibition of this reaction by the second intermediate of the pathway,N-acetylglucosaminyl-N-acetylglucosaminylpyrophosphoryldolichol (GlcNAc-GlcNAc-P-P-dolichol), and product inhibition by GlcNAc-P-P-dolichol itself. These influences were revealed when the reactions were carried out in the presence of showdomycin, a nucleoside antibiotic, present at concentrations that block the de novo formation of GlcNAc-GlcNAc-P-P-dolichol but not that of GlcNAc-P-P-dolichol. The apparent K i values for GlcNAc-P-P-dolichol and GlcNAc-GlcNAc-P-P-dolichol under basal conditions were 4.4 and 2.8 μm, respectively. Inhibition was also observed under conditions where mannosyl-P-dolichol (Man-P-dol) stimulated the biosynthesis of GlcNAc-P-P-dolichol; the apparent K i values for GlcNAc-P-P-dolichol and GlcNAc-GlcNAc-P-P-dolichol were 2.2 and 11 μm, respectively. Kinetic analysis of the types of inhibition indicated competitive inhibition by GlcNAc-P-P-dolichol toward the substrate UDP-GlcNAc and non-competitive inhibition toward dolichol phosphate. Inhibition by GlcNAc-GlcNAc-P-P-dolichol was uncompetitive toward UDP-GlcNAc and competitive toward dolichol phosphate. A model is presented for the kinetic mechanism of the synthesis of GlcNAc-P-P-dolichol. GlcNAc-P-P-dolichol also exerts a stimulatory effect on the biosynthesis of Man-P-dol, i.e. a reciprocal relationship to that previously observed between these two intermediates of the dolichol pathway. This network of inhibitory and stimulatory influences may be aspects of metabolic control of the pathway and thus of glycoprotein biosynthesis in general.

the mechanisms that regulate this complex series of reactions, however, is still limited. To this end we have directed our attention to the initial reaction of the pathway, the reaction between dolichol phosphate and UDP-GlcNAc producing Glc-NAc-P-P-dolichol, 1 catalyzed by the enzyme, UDP-GlcNAc:dolichyl-phosphate N-acetylglucosamine 1-phosphate transferase (GPT-1). Factors that modulate the formation of GlcNAc-P-Pdolichol could have an effect on the rate of synthesis of the other intermediates of the dolichol pathway and thus influence nascent glycoprotein biosynthesis in general. Several factors have previously been described that could have a regulatory influence on this reaction as follows: hormonal effects (2), genetic factors (3), and topography of enzymes and substrates (4 -6). Previous studies from this laboratory and others (7)(8)(9)(10)(11)(12)(13)(14)(15) have also revealed that another intermediate of the pathway, mannosyl-P-dolichol (Man-P-dol), acts as an allosteric activator of GPT-1, resulting in the stimulation of GlcNAc-P-P-Dol synthesis. A reciprocal relationship has now been revealed whereby Man-P-Dol formation is stimulated by GlcNAc-P-Pdolichol. The present study has revealed other potential aspects of regulation of the initial reaction of the dolichol pathway. Feedback inhibition of the biosynthesis of GlcNAc-P-Pdolichol was demonstrated by the second intermediate of the pathway, GlcNAc-GlcNAc-P-P-dolichol. The formation of the latter compound is catalyzed by a separate GlcNAc-transferase, UDP-GlcNAc:GlcNAc-P-P-dolichol, N-acetylglucosamine transferase (GT-2), the kinetics of which have recently been described (16). Although the reversibility of GPT-1 has previously been demonstrated (17,18), the present report also described the kinetics of inhibition by GlcNAc-P-P-dolichol of its own synthesis. 2 We have examined the effect of these inhibitory influences on the biosynthesis of GlcNAc-P-P-dolichol at the basal level and under stimulatory conditions in the presence of Man-P-dol. On the basis of these relationships a model is suggested as a mechanism of action of UDP-GlcNAc:dolichol phosphate, GlcNAc-1-phosphate transferase.

Enzyme Preparation, Chemicals
Microsomes were prepared from the retinas of 15-16-day-old embryonic chicks as described previously (10). Purified, recombinant yeast mannosyl-P-dolichol synthase was obtained from Dr. John Schutzbach.

Preparation of GlcNAc-P-P-dolichol
Enzymatic Synthesis-Large scale preparations were performed by incubating microsomes from the retina of the embryonic chick with UDP-[ 14 C]GlcNAc and dolichol phosphate, as described previously (10). The incubations were performed in the presence also of Man-P-Dol and the antibiotic, showdomycin. As described previously, the former stimulates the production of GlcNAc-P-P-dolichol (7)(8)(9)(10)(11)(12), and the latter both inhibits the formation of GlcNAc-GlcNAc-P-P-dolichol and also brings about an increase in the production of the mono-GlcNAc derivative (19). The product was isolated by solvent partitioning according to the procedure of Folch et al. (20) and purified by chromatography on DEAEcellulose, as described previously (10). Its concentration was determined by Dionex chromatography, as described below.
Chemical Synthesis-By using dolichol kindly provided by Dr. Tadeusz Chojnacki of the Institute of Biochemistry and Biophysics, Warsaw, Poland, GlcNAc-P-P-dolichol was initially synthesized by the method of Imperiali and Zimmerman (21), which requires azeotropic drying (toluene or pyridine) at the stage of the oxalyl chloride-mediated coupling of dolichol phosphate to the per-O-acetyl-GlcNAc-P pyridinium salt. Acceptable yields of the latter intermediate could be obtained only by careful azeotropic drying of all commercial ingredients (including tetraethylammonium chloride and dibenzylphosphate) and use of freshly distilled solvents. In the published method (21), oxalyl chloride activation of dolichol phosphate generates the highly activated dolichylphosphoryl dichloride, which actually is an intermediate in the POCl 3 -mediated preparation of dolichol phosphate but which could not successfully be coupled directly to the protected GlcNAc-P, and was instead hydrolyzed to dolichol phosphate, which was then purified and reactivated with oxalyl chloride. During this study, it was found that the requisite dolichylphosphoryl dichloride could in fact be generated from dolichol in a directly usable form using excess POCl 3 in hexane, followed by a brief extraction with water to remove HOPOCl 2 , and then evaporation of hexane and remaining POCl 3 under high vacuum. Following coupling, purification of the protected pyrophosphate prior to NaOMe-mediated deacetylation was achieved by silica gel 60 column chromatography using CHCl 3 /MeOH/H 2 O (65:25:4, by volume) as the eluant. These modifications, which avoid the need to isolate and purify dolichol phosphate, resulted in the best overall yield of GlcNAc-P-Pdolichol from dolichol.
The enzymatically and chemically synthesized GlcNAc-P-P-dolichol functioned in a similar manner in these studies.
Large Scale Preparation of Mannosyl-P-dolichol-Large scale preparations of [ 14 C]Man-P-Dol were made by incubating dolichol phosphate, GDP[ 14 C]mannose (1.7 dpm/pmol), buffer, and metal ions as described previously (12) with extracts from Micrococcus luteus provided by Dr. Charles J. Waechter and Dr. Jeffrey S. Rush of the Department of Biochemistry, University of Kentucky College of Medicine, Lexington, KY, or a purified, recombinant yeast dolichyl-P-synthase from yeast provided by Dr. John S. Schutzbach, Department of Biochemistry, University of Alabama, Birmingham, AL. Purification was carried out by chromatography on DEAE-cellulose acetate as described previously (22,23), followed by preparative thin layer chromatography at 4°C on 0.5-mm thick plates of Silica Gel 60 (Merck) as described previously (10). As indicated by guide strips, the Man-Pdolichol region was scraped from the plates and recovered by leaching the gel in the cold with C/M (chloroform/methanol)/water (10:10:3) followed by C/M (2:1). After solvent partitioning by the procedure of Folch et al. (20), the concentration of [ 14 C]Man-P-Dol was calculated from the specific activity of GDP[ 14 C]mannose used in its preparation.

Incubation Conditions and Assay, Kinetic Studies
GlcNAc Lipid Biosynthesis-Incubations were carried out for 10 -20 min at 37°C in the presence of dolichol phosphate (16 -20 M), Triton X-100 (0.15%), UDP-[ 3 H]GlcNAc (52 M; 169 dpm/pmol), MgCl 2 (27 mM), showdomycin, or N-benzyl-2Ј-deoxyshowdomycin, as indicated, and enzyme (0.2-0.25 mg of protein) in a total volume of 0.15 ml (basal conditions). Reactions were also carried out in the presence of Man-P-Dol (2.1-6.4 M) (stimulatory conditions). The incubations were performed in the absence or presence of exogenously added GlcNAc-P-Pdolichol or GlcNAc-GlcNAc-P-P-dolichol (as indicated in the tables and figures). The dolichol derivatives were evaporated to dryness with nitrogen, vortexed vigorously with 0.015 ml of 1.5% Triton X-100, after which the other components of the reaction mixture were added. The incubations, carried out at 37°C, were started by the addition of the enzyme preparation and stopped by the addition of C/M (2:1). After solvent partitioning by the Folch procedure (20) the radioactivity in the washed lower phase was determined by scintillation spectrometry, as described previously (7,10).
Dionex Chromatography-After incubation followed by solvent partitioning as above, the material in the washed organic phase was evaporated to dryness and subjected to mild acid hydrolysis in 1 ml of 0.1 N HCl in 80% tetrahydrofuran for 100 min at 50°C as described previously (26). After evaporation to dryness and redissolving in water, the material was applied to a column containing 0.5 ml each of AG-2-X8 (200 -400 mesh) acetate and AG-50-X8 Hϩ (200 -400 mesh), the column eluted with 20 ml of water, the eluate evaporated to dryness, and the residual material redissolved in water as described previously (16). To an aliquot was added 5 nmol of fucose to serve as an early eluting reference marker and 10 nmol each of GlcNAc and GlcNAc-GlcNAc to serve as internal standards for analysis by high pH anion exchange chromatography (Dionex Corp., Sunnyvale, CA). The mixture was injected onto a CarboPAc 1 column (4 ϫ 250 mm) with a Carbo PAc guard column (3 ϫ 25 mm) and eluted isocratically with a mixture of 25% of 100 mM NaOH plus 75% 1 mM NaOH at a flow rate of 1 ml/min. The elution of the standards was followed by pulsed amperometric detection (high pH anion exchange chromatography-pulsed amperometric detection), and the products of the reactions by measuring their radioactivity by scintillation spectrometry of 0.5-ml fractions collected from the pulsed amperometric detection cell, as described previously (27).

Distribution of Radioactivity in the Glucosaminyl Residues of the Biosynthesized GlcNAc Lipids
The relative contributions of GPT-1 and GT-2 was investigated by analyzing the distribution of the tritium label in the GlcNAc residues of their respective products, GlcNAc-P-P-dolichol and GlcNAc-GlcNAc-P-P-dolichol, as described previously (28,29). In short, this involves the following procedures carried out sequentially. After incubation, the GlcNAc lipids extracted into the chloroform-rich layer after solvent partitioning are subjected to mild acid hydrolysis. GlcNAc and GlcNAc-GlcNAc thus formed are separated by paper chromatography, recovered from the chromatogram, and reduced with NaBH 4 . After mixed bed ion exchange chromatography, strong acid hydrolysis N-deacetylates the products and cleaves chitobiose. The products are then subjected to high voltage paper electrophoresis in 1% sodium borate buffer, the electrophoretogram cut into 1-cm zones, and the radioactivity determined by scintillation spectrometry. The mobilities of standard [ 3 H]GlcNH 2 and [ 3 H]GlcNH 2 OH were determined in the same manner. By these procedures glucosaminitol would have been derived from GlcNAc-P-P-dolichol after hydrolysis and glucosamine from the non-reducing end of GlcNAc-GlcNAc-P-P-dolichol after hydrolysis and glucosaminitol from the reducing end.

Other Analytic Procedures
The concentration of GlcNAc-P-P-dolichol was determined after mild acid hydrolysis, as above, by quantitative Dionex chromatography of the liberated GlcNAc and by the Morgan-Elson reaction as described by Reissig et al. (30). The concentration of GlcNAc-GlcNAc-P-P-dolichol, provided by Dr. B. Imperiali, Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA, was determined by Dionex chromatography of N,NЈ-diacetylchitobiose liberated by mild acid hydrolysis. The concentration of dolichol phosphate and the GlcNAc-P-P-dolichol used in the early phases of this work (provided by Dr. Imperiali) was determined by analysis for total phosphate as described previously (10). Total microsomal phospholipids were determined in a similar manner after solvent partitioning (20) of retina microsomes. Thin layer chromatography was performed using 20 ϫ 20-cm glass plates precoated with a 0.25-or 0.50-mm layer of Silica Gel 60 without fluorescent indicator. The following solvent systems were used: 1) chloroform/methanol/acetic acid/water (25:15:4:2, by volume); 2) chloroform/methanol/ water (65:25:4, by volume). The location of radioactive material was accomplished by measuring the radioactivity by scintillation spectrometry of 1 ϫ 3-cm zones scraped from the chromatogram. The migration of non-radioactive material was detected by the anisaldehyde spray reagent or by exposure to iodine vapor, as described previously (24).

Rationale for Using Showdomycin
Incubations carried out using dolichol phosphate and UDP-[ 3 H]Glc-NAc as substrates would result in the formation of [ 3 H]GlcNAc-P-Pdolichol and [ 3 H]GlcNAc-[ 3 H]GlcNAc-P-P-dolichol. Thus, in assays carried out by solvent partitioning, any effect added compounds might have specifically on the new synthesis of [ 3 H]GlcNAc-P-P-dolichol would be masked by the accompanying formation of the labeled chitobiosyl product. This difficulty was resolved by carrying out the incubations in the presence of the nucleoside antibiotic, showdomycin, which inhibits the formation of the chitobiosyl compound, enhancing the formation of GlcNAc-P-P-dolichol (19). The N-benzyl-2Јdeoxy derivative was shown to have the same effect (19). Thus, using concentrations of the showdomycins that extensively inhibited the formation of GlcNAc-GlcNAc-P-P-dolichol, the effect of exogenously added non-radioactive GlcNAc-P-P-Dol on the de novo synthesis of the labeled compound could be readily determined.

Kinetics of Inhibition
Apparent K i and V max values were calculated from a non-linear least squares analysis of the data using GraFit (31) fit to an expression for general inhibition derived from Equation 1 for mixed inhibition with a constant substrate concentration.
Analyses of the types of inhibition were carried out in the following manner. The steady state kinetic data were fitted using the unweighted non-linear least squares method implemented in the computer program GraFit (31). Inhibition constants were determined by a global fit of the data from an entire experiment. The best fit of the data is one that minimizes the sum of the square of the observed and calculated velocities for an entire data set (32). The standard deviations of the data points were within 3% of the highest velocity. The data were initially fitted to Equation 2 for non-competitive or mixed inhibition. By using the criterion that if the standard error of the slope or intercept inhibition constants were over 50% of the fitted value, that parameter was excluded from the final fit. If so, the data were then fit to Equation 3 or Equation 4 for competitive and uncompetitive inhibition, respectively. The more restrictive mechanism was accepted if the -squared value was not increased by omitting the poorly defined inhibition constant.

Kinetics of Activation by GlcNAc-P-P-dolichol of Man-P-Dol Synthesis
Incubations were carried out under initial rate conditions for the yeast and retina microsomal enzymes. Apparent K a and V max values were calculated from Lineweaver-Burk double-reciprocal plots of the data after analysis by computer using the Kcat program (BioMetalics, Princeton, NJ).

Inhibition by GlcNAc-P-P-dolichol of GlcNAc Lipid Synthesis, Product Identification
Incubations were carried out under optimal conditions of GlcNAc lipid synthesis as described under "Experimental Procedures" in the presence or absence of exogenously added Glc-NAc-P-P-dolichol. The incubation mixtures additionally contained concentrations of showdomycin that inhibited 93-99% of the formation of GlcNAc-GlcNAc-P-P-dolichol. As seen in Fig.  1A, with increasing concentrations of GlcNAc-P-P-dolichol there was increasing inhibition (over 70%) of GlcNAc lipid synthesis. This effect on GlcNAc-P-P-dolichol biosynthesis occurred at the basal level and when the reaction was performed in the presence of Man-P-Dol shown previously to greatly stim- ulate its formation (10). In these studies, either under basal or stimulatory conditions, exogenously added GlcNAc-P-P-dolichol was present over a range from 2-to 86-fold over the [ 3 H]GlcNAc-P-P-dolichol formed in the absence of the inhibitor. (The curves in Fig. 1 were drawn in accord with an analysis of the data by Equation 1 as described under "Experimental Procedures.") The nature of the products formed under these conditions after analysis by Dionex chromatography is seen in Fig. 2. Fig.  2A shows a typical example of the distribution of the mono-GlcNAc and chitobiosyl products formed under basal conditions in the absence of showdomycin. As seen in Fig. 2B, under basal conditions in the presence of showdomycin, there was now extensive inhibition of the formation of the chitobiosyl product (solid line) and an increase in the formation of GlcNAc-P-Pdolichol (solid line) as described previously (19). When the reaction was carried out in the presence also of exogenously added GlcNAc-P-P-dolichol (18 M), there was now an 86% decrease (diamond symbols) in the de novo formation of [ 3 H] GlcNAc-P-P-dolichol. Exogenously added GlcNAc-P-P-dolichol was present in 360-fold molar excess over the [ 3 H]GlcNAc-P-P-dolichol formed in its absence.
Similar results were obtained when the reactions were performed under stimulatory conditions, i.e. in the presence of Man-P-dolichol (Fig. 2C). The formation of GlcNAc-GlcNAc-P-P-dolichol under these conditions was reduced 97% in the presence of showdomycin (filled circles). When the reaction was carried out in the presence also of exogenously added GlcNAc-P-P-dolichol (18 M), the formation of [ 3 H]GlcNAc-P-P-dolichol was decreased 88% (diamonds). The added GlcNAc-P-P-dolichol was present in 56-fold molar excess over that formed in its absence.

Inhibition by GlcNAc-GlcNAc-P-P-dolichol, Product Identification
The reaction between GlcNAc-P-P-dolichol and UDP-GlcNAc catalyzed by GT-2 results in the formation of the chitobiosyl derivative. As seen in Fig. 1B, when GlcNAc lipid synthesis was examined in the presence of exogenously added GlcNAc-Glc-NAc-P-P-dolichol, inhibition of about 90% of that formed in the absence of the inhibitor was attained. As with the mono-Glc-NAc derivative (Fig. 1A), inhibition was also induced and to a similar extent when the reaction was stimulated by the addition of Man-P-Dol as seen also in Fig. 1B. Shown in Fig. 2D is an identification of the products from experiments of this type using Dionex chromatography. The presence of showdomycin resulted in the essentially complete (96%) inhibition of the formation of the chitobiosyl derivative. When the incubation was performed in the presence also of exogenously added Glc-NAc-GlcNAc-P-P-dolichol (diamond symbols), the formation of GlcNAc-P-P-dolichol was inhibited 94% of that formed in its absence. In this experiment, a 230-fold molar excess of GlcNAc-GlcNAc-P-P-Dol was added over the GlcNAc-P-P-dolichol formed in its absence.
The identification by Dionex chromatography of the products of the reactions under stimulatory conditions (ϩMan-P-dol) is seen in Fig. 2E. As under basal conditions the presence of showdomycin inhibited the formation of the chitobiosyl derivative 99% of that formed in its absence. In the presence also of exogenously added GlcNAc-GlcNAc-P-P-dolichol, the synthesis of GlcNAc-P-P-dolichol was inhibited 67% (diamonds). In this experiment a 42-fold molar excess of GlcNAc-GlcNAc-P-P-dolichol was added over the GlcNAc-P-P-dolichol formed in its absence.
A summary of the apparent K i and V max values 4 generated by these inhibition studies is shown in Table I.

Other Characteristics of the System, Inhibition Studies in the Absence of Showdomycin
A series of experiments was performed to see if inhibition would also occur in the absence of showdomycin. As described previously, the simultaneous production of radioactive GlcNAc-P-P-dolichol and GlcNAc-GlcNAc-P-P-dolichol would prevent merely looking for a loss of dpm in the organic phase after solvent partitioning as an assay to reveal the influence of added GlcNAc lipids on the formation of the mono-GlcNAc derivative. In order to investigate this without blocking the formation GlcNAc-GlcNAc-P-P-dolichol (as affected by showdomycin in these studies), the isolation and analysis of the individual components of the reactions was performed as described under "Experimental Procedures." The net accumulation of [ 3 H]Glc-NAc-P-P-dolichol reflects the difference between its rate of formation and the rate of conversion to the chitobiosyl derivative. As seen in Table II, part A, at the basal level, in the absence of showdomycin, the addition of GlcNAc-GlcNAc-P-Pdolichol (19 M) resulted in extensive net inhibition (74%) of GlcNAc-P-P-dolichol formation. Inhibition was also brought about by the chitobiosyl derivative under stimulatory conditions (ϩMan-P-dol) in which similar and extensive inhibition occurred in the presence or absence of the antibiotic as seen in Table II ( (Table II, part D, lines a and b), exogenously added GlcNAc-P-P-dolichol inhibited the stimulatory response 62% (Table II, part D, lines b and c). The latter response, however, may not indicate a direct effect on GPT-1 at the basal level. Indeed, at the basal level using exogenously added GlcNAc-P-P-dolichol (18 M) as the test substance, apparent stimulation rather than inhibition was detected in the absence of showdomycin, as seen in Table II (part C, line 1, lines a and b). Similar results were encountered using 2.9 and 9 M GlcNAc-P-P-dolichol (data not shown). The reason for this apparent stimulation is, at least in part, accounted for by the isotope dilution effect of exogenously added unlabeled GlcNAc-P-P-dolichol (300 -700-fold) preventing the conversion of [ 3 H]GlcNAc-P-P-dolichol to [ 3 H]GlcNAc-[ 3 H]Glc-NAc-P-P-dolichol. The chitobiosyl derivative formed under these conditions is predicted to be largely [ 3 H]GlcNAc-GlcNAc-P-P-dolichol. Because only a small fraction of the total GlcNAc-P-P-dolichol is converted to the chitobiosyl derivative, the isotope dilution results in almost all of the enzymatically synthesized [ 3 H]GlcNAc-P-P-dolichol being trapped. These effects can be confirmed by kinetic modeling of the reactions and numerical integration (33). We have modeled the GPT-1 and GT-2 system using the kinetic constants determined in this study to show that the increase in radioactivity observed in Table II, part C, line 1, lines a and b, is predicted by isotope dilution and fully consistent with inhibition of GPT-1 by Glc-NAc-P-P-dolichol (data not shown). 5 In addition to this theo- 4 Among the problems associated with studying enzyme kinetics in mixed micelle systems is whether the kinetic constants should be reported by concentration or as mole fractions. In these experiments, the large excess of Triton X-100 (about 2.5 mM) relative to the dolichol substrates/inhibitors or phospholipid added with the microsomes (0.24 mol of total lipid P i /mg protein) ensures that their mole fractions are proportional to their concentration, with a 1 M concentration corresponding to a mole fraction of 4 ϫ 10 Ϫ4 . retical examination, this situation was examined further experimentally, as described below.

Distribution of Label
The trapping of [ 3 H]GlcNAc-P-P-dolichol by isotope dilution was examined by determining the distribution of radioactivity in the glucosaminyl residues of the GlcNAc lipids formed in the absence of showdomycin. The results from these experiments are presented in Fig. 3 showing the patterns obtained by high voltage paper electrophoresis after processing the products as described under "Experimental Procedures." Fig. 3, A and B, refers to the labeling that occurred from incubations that contained exogenously added GlcNAc-P-P-dolichol. The material in Fig. 3A is derived from the GlcNAc region of the paper chromatogram and shows glucosaminitol as the major labeled component. This would be the derivative formed from GlcNAc-P-P-dolichol synthesized de novo during the incubation. The relatively small amounts of labeled GlcNH 2 in Fig. 3A most likely are due to the presence of contaminating GlcNAc-GlcNAc that accompanied the GlcNAc region when cut and eluted from the paper chromatogram. The source of the labeled peak material in Fig. 3B is chitobiose recovered from the paper chromatogram showing the greatly enhanced labeling in GlcNH 2 derived from the non-reducing end of [ 3 H]GlcNAc-GlcNAc-P-Pdolichol formed by the addition of [ 3 H]GlcNAc to the exogenously added GlcNAc-P-P-dolichol. Fig. 3, C and D, indicates the labeling that occurred from incubations carried out under basal conditions, i.e. in the absence of exogenously added Glc-NAc-P-P-dolichol. The material in Fig. 3C comes from processing the GlcNAc region of the paper chromatogram. As with Fig.  3A, labeled glucosaminitol in Fig. 3C arises from de novo synthesized GlcNAc-P-P-dolichol and the small amount of GlcNH 2 from incompletely resolved chitobiose after paper chromatography. The material in Fig. 3D is derived from processing the material in the GlcNAc-GlcNAc region of the chromatogram. As seen in Fig. 3D, under basal conditions glucosamine derived from the non-reducing end of the chitobiosyl derivative and glucosaminitol from the reducing terminus were labeled to a similar extent, consistent with both sugar residues originating directly from the same donor source, i.e. UDP-[ 3 H]GlcNAc, and the reaction approaching isotopic equilibrium under the conditions of this experiment. In contrast, in the presence of exogenously added GlcNAc-P-P-dolichol (Fig. 3B), no discernible peak was detected in the GlcNH 2 OH region as would be predicted if the exogenously added unlabeled GlcNAc-P-P-dolichol was the major, if not only, species converted to the chitobiosyl

FIG. 2. Analysis by Dionex chromatography of the products of the reactions after inhibition by GlcNAc-P-P-dolichol and Glc-NAc-GlcNAc-P-P-dolichol.
Incubations (2-4-fold over that described under "Experimental Procedures") were carried out for 20 min in the presence of dolichol phosphate, UDP-[ 3 H]GlcNAc (52 M; 185 dpm/ pmol), Mg 2ϩ , buffer, Triton X-100, and microsomes. A is a typical pattern after incubation under these conditions. In the other panels are the patterns obtained after incubation in the presence of showdomycin (0.17 mg/ml) and exogenously added GlcNAc-P-P-dolichol or GlcNAc-GlcNAc-P-P-dolichol and also, where indicated, under stimulatory conditions in the presence of Man-P-dol. After solvent partitioning, mild acid hydrolysis, and mixed bed ion exchange chromatography, the products were analyzed by Dionex chromatography as described under "Experimental Procedures." The arrows indicate the elution times of standard GlcNAc and GlcNAc-GlcNAc added as internal standards. The incorporation of tritium into the products was determined by scintillation spectrometry using double labeling techniques where appropriate. B, absence (q) or presence (छ) of exogenously added GlcNAc-P-P-

TABLE I Inhibition by GlcNAc-P-P-dolichol and GlcNAc-GlcNAc-P-P-dolichol of the initial reaction of the dolichol pathway
The kinetic values were calculated by Eq. 1 from a non-linear least squares analysis of the data plotted in Fig. 1 derivative. Thus, the experimental evidence supports the previous suggestion that the apparent lack of inhibition of formation of [ 3 H]GlcNAc-P-P-dolichol that occurred in the presence of exogenously added GlcNAc-P-P-dolichol under basal conditions in the absence of showdomycin reflects the competition by the large excess of unlabeled exogenously added GlcNAc-P-Pdolichol to serve as a substrate for the synthesis of the chitobiosyl derivative. When the synthesis of the chitobiosyl derivative is inhibited as with the use of showdomycin, these additional influences are eliminated, and the inhibitory effect is directly seen. The presence of detergent in the incubation mixtures argues against the possibility that the difference in response to showdomycin at the basal level was due to the presence of different pools of added versus newly synthesized GlcNAc-P-P-dolichol.

Influence of Mild Acid Treatment of GlcNAc Lipids; Influence of Potential Hydrolytic Products, Other GlcNAc Lipids
These studies have shown that exogenously added GlcNAc-P-P-dolichol and GlcNAc-GlcNAc-P-P-dolichol inhibit the de novo synthesis of GlcNAc-P-P-dolichol. The effect of these compounds was examined after they were subjected to mild acid hydrolysis. This was performed in 0.1 N HCl in the presence of 0.045% Triton X-100. After evaporation to dryness, incubations were performed in the presence of showdomycin as described under "Experimental Procedures." No inhibition of GlcNAc lipid synthesis by hydrolyzed GlcNAc-P-P-dolichol or hydrolyzed GlcNAc-GlcNAc-P-P-dolichol was observed as compared with controls carried out in the same manner.
The influence of potential products of the hydrolytic procedures was also examined. Under conditions where GlcNAc-GlcNAc-P-P-dolichol (0.015 mM) in the presence of showdomycin resulted in 86% inhibition of GlcNAc lipid synthesis, no inhibition occurred in the presence of 3-5-fold higher concen-trations of GlcNAc-GlcNAc or GlcNAc, nor by a mixture of these compounds plus dolichol phosphate (1.4-fold excess) that had been exposed to the conditions of mild acid hydrolysis. Likewise, under conditions where GlcNAc-P-P-dolichol (0.0144 mM) in the presence of showdomycin brought about 75% inhibition of GlcNAc lipid synthesis, 4.6-fold higher concentrations of GlcNAc or GlcNAc-GlcNAc had no effect. The effect of very high concentrations of these compounds was also examined. Thus, in the presence of GlcNAc at a 1500-fold higher concentration over the apparent K i value for GlcNAc-P-P-dolichol (see Table I) and GlcNAc-GlcNAc at 2400-fold higher concentration than the apparent K i value for GlcNAc-GlcNAc-P-P-dolichol, 93-98% of control GlcNAc lipid synthesis was seen (four experiments). Likewise in the presence of dolichol phosphate at 55-86-fold excess over these apparent K i values, 88% of control activity was observed. Thus, little or no inhibition was observed in the presence of these potential hydrolytic products.
Furthermore, the biosynthesis of GlcNAc lipids as described in these studies when carried out in the presence of tunicamycin (0.4 g/ml) was blocked virtually 100% (data not shown). This observation argues against contributions to the effects described by compounds such as GlcNAc-containing glycosphingolipids or GlcNAc/glucosaminylphosphatidylinositides involved in glycophospholipid anchor biosynthesis (34).
These control studies support the proposal that the inhibitions observed were due specifically to the presence of the mono-GlcNAc and chitobiosyl-P-P-dolichol derivatives.

Stability of GlcNAc-GlcNAc-P-P-dolichol during Incubation
The possibility was examined that the inhibition of the formation of GlcNAc-P-P-dolichol by GlcNAc-GlcNAc-P-P-dolichol under these experimental conditions was due to the instability of the chitobiosyl derivative and its formation in situ of Glc-NAc-P-P-dolichol which actually brought about the effect. Glc-  NAc-GlcNAc-P-P-dolichol (11.4 M) was incubated for 20 min in the presence of UDP-[ 3 H]GlcNAc, Tes buffer, Mg 2ϩ , microsomes, and tunicamycin (2 g/ml) but in the absence of dolichol phosphate and showdomycin. The formation of [ 3 H]GlcNAc-P-P-dolichol from the breakdown of the chitobiosyl derivative to dolichol phosphate would be prevented by tunicamycin. The amount of [ 3 H]GlcNAc-GlcNAc-P-P-dolichol newly formed under these conditions would be equivalent to the amount of GlcNAc-P-P-dolichol derived from the degradation of the chitobiosyl substrate. By using this assay, it was demonstrated that 0.0262 M GlcNAc-P-P-dolichol was formed from the initial 11.4 M GlcNAc-GlcNAc-P-P-dolichol started with or only about 0.2% breakdown. In control incubations, exogenously added GlcNAc-P-P-dolichol (2.9 M) incubated in the presence of showdomycin brought about 42% inhibition of de novo Glc-NAc lipid synthesis compared with that formed in the absence of added GlcNAc-P-P-dolichol (data not shown). Thus, the material formed by the breakdown of GlcNAc-GlcNAc-P-P-dolichol (described above) would account for only about 0.4% inhibition.

Types of Inhibition
Two substrates, dolichol phosphate and UDP-GlcNAc, participate in the formation of GlcNAc-P-P-dolichol. The steady state kinetics of this reaction in the presence of the two intermediates of the dolichol pathway which these studies have shown to inhibit its activity were examined, and the data are summarized in Fig. 4. As seen in Fig. 4A, in the presence of exogenous GlcNAc-P-P-dolichol added at 7.5-and 15-fold higher concentrations than that produced in the absence of the inhibitor, the data indicate that this inhibitor is competitive versus UDP-GlcNAc, i.e. UDP-GlcNAc and GlcNAc-P-P-dolichol compete for the same site or same form of the enzyme. In contrast, (Fig. 4B) non-competitive or mixed type inhibition was demonstrated versus dolichol phosphate. The kinetics of inhibition by GlcNAc-GlcNAc-P-P-dolichol are seen in Fig. 4, C and D. When examined as a function of variation in the concentration of dolichol phosphate, competitive inhibition is seen (Fig. 4D). Variation in the concentration of UDP-GlcNAc, however, revealed that GlcNAc-GlcNAc-P-P-dolichol is an uncompetitive inhibitor versus UDP-GlcNAc (Fig. 4C) where the apparent V max and apparent K m values are each changed to a similar extent, as indicated by the parallel lines. The inhibition constants obtained from the non-linear least squares analysis are tabulated in Table III.
Retina Microsomes-Incubation of retina microsomes under optimal conditions for Man-P-Dol formation, as described previously (24) in the presence of exogenously added GlcNAc-P-Pdolichol, resulted in an enhanced formation of Man-P-dol, as seen in Fig. 5A. From Lineweaver-Burk analyses of the data, apparent K a and V max values were calculated, as summarized in Table IV. Thin layer chromatography in solvent systems 1 gram formed in the presence of GlcNAc-P-P-dolichol. C, the same as A, i.e. the GlcNAc region, except that the incubation had been performed under basal conditions, i.e. in the absence of GlcNAc-P-P-dolichol. D, the same as B, i.e. the GlcNAc-GlcNAc region, except that the incubation was performed under basal conditions. The arrows indicate the mobilities of authentic [ 3 H]GlcNH 2 and [ 3 H]GlcNH 2 OH.
FIG. 3. High voltage paper electrophoresis of glucosaminyl residues after incubation in the absence of showdomycin. Incubation mixtures, scaled up 4-fold over that described under "Experimental Procedures," contained dolichol phosphate, Tes buffer, Mg 2ϩ , Triton X-100, UDP-[ 3 H]GlcNAc, microsomes, and where indicated, GlcNAc-P-P-dolichol (18 M). After solvent partitioning of the products, mild acid hydrolysis, separation of GlcNAc and GlcNAc-GlcNAc by paper chromatography, and recovery from the chromatogram, followed by reduction with NaBH 4 and strong acid hydrolysis (4 N HCl, 6 h), the products were examined by high voltage paper electrophoresis on Whatman 3MM paper in 1% sodium tetraborate for 60 min at 53 V/cm. The electrophoretogram was cut into 1 ϫ 4-cm zones, and the radioactivity was measured by scintillation spectrometry. A, the pattern obtained after processing the GlcNAc region after paper chromatography of the products formed in the presence of GlcNAc-P-P-dolichol. B, the pattern obtained after processing the GlcNAc-GlcNAc region of the chromato-and 2 of the radioactive product formed under these conditions showed the presence of a single radioactive area that migrated with purified, standard Man-P-dolichol (data not shown).
Yeast Enzyme-The stimulation by GlcNAc-P-P-dolichol of Man-P-Dol synthesis was examined further using a purified, recombinant yeast Man-P-Dol synthase (25). After incubation under initial rate conditions, the same response was obtained as with the retina microsomes as shown in Fig. 5B. From Lineweaver-Burk analysis of the data, apparent K a and V max values were calculated, as summarized in Table IV. Thin layer chromatography in solvent systems 1 and 2 showed a single radioactive component migrating with standard Man-P-Dol (data not shown).

DISCUSSION
In all eukaryotic organisms thus far investigated, the biosynthesis of the core region oligosaccharide of asparaginelinked glycoproteins proceeds via the dolichol pathway. The following Reaction 1 Dolichol phosphate ϩ UDPGlcNAc 7 GlcNAc-P-P-dolichol ϩ UMP REACTION 1 has been described as the first committed step of this complex sequence of reactions (35). Regulation of the formation of Glc-NAc-P-P-dolichol could thus clearly influence the steady state concentration of the final product of the pathway, Glc 3 Man 9 -

TABLE III Observed inhibition by GlcNAc-P-P-dolichol and GlcNAc-GlcNAc-P-P-dolichol
The inhibition constants were obtained from the fits of the data plotted in Fig. 4 to Equations 2-4 (see "Experimental Procedures"). K is is the inhibition constant determined from the variation of the slopes, i.e. when the varied substrate is extrapolated to 0, whereas K ii is the inhibition constant determined from the y intercept, i.e. at a saturating concentration of the varied substrate. NA (not applicable) indicates that the best fit of the data was obtained with the single inhibition constant indicated. In accord with the sensitive site in metabolism that this reaction may occupy, down-regulation of GPT-1 has been shown to inhibit the glycosylation and secretion of proteins by Xenopus oocytes (36) and cause defects in the life cycle of yeast (37). It has also been suggested that the accumulation of the large lipid-linked oligosaccharide might also influence the formation of other earlier intermediates of the pathway, such as GlcNAc-P-P-dolichol, by feedback inhibition (38). Several influences of potential regulatory nature have been described concerning the activity of the GlcNAc-transferase that catalyzes the formation of GlcNAc-P-P-dolichol. Treatment of chick oviduct membranes with diethylstilbesterol was observed to increase its production (2). Allosteric stimulation of the synthesis of this compound by Man-P-Dol has been described (7)(8)(9)(10)(11)(12)(13)(14)(15). In this regard, it has been demonstrated that the stimulation by Man-P-Dol was directed specifically to the biosynthesis of GlcNAc-P-P-dolichol, whereas the addition of the second GlcNAc residue resulting in the formation of GlcNAc-GlcNAc-P-P-dolichol was not directly affected, and any increase in its concentration was a secondary consequence of the enhanced formation of its precursor (29). A high degree of specificity was detected concerning the polyprenolog component of the activating compound in terms of chain length, stereospecificity, and saturation of the terminal isoprene unit (12). It has been observed (3) that the ALG7 gene functioning in the biosynthesis of GlcNAc-P-P-dolichol in yeast is selectively regulated and may influence other genes of the dolichol pathway. The structural organization of GTP-1, such as its oligomerization, has been shown to influence its activity (39). Additional factors regulating the assembly of the intermediates of the dolichol pathway are topographical considerations within the cell that could influence the accessibility of substrates and enzymes for reaction. Although major questions remain unresolved in this regard, evidence has been presented supporting a cytoplasmic orientation on the surface of the endoplasmic reticulum of the N-acetylglucosaminyltransferases concerned with the biosynthesis of GlcNAc-P-P-dolichol and GlcNAc-GlcNAc-P-P-dolichol (4 -6). Mechanisms influencing the supply and availability of dolichol phosphate and other transcriptional and post-translational aspects of the formation of GlcNAc-P-P-dolichol have been reviewed (35).
We have now observed other modulating influences on the initiating reaction of the dolichol pathway with the finding that GlcNAc-GlcNAc-P-P-dolichol has a feedback-inhibiting influence on the synthesis of its precursor, GlcNAc-P-P-dolichol, and in addition, GlcNAc-P-P-dolichol itself can inhibit its own formation. These studies were performed in the presence of the antibiotic showdomycin that preferentially blocks the de novo synthesis of GlcNAc-GlcNAc-P-P-dolichol thus allowing ready analysis by solvent partitioning of the influence of the inhibitors on the synthesis of labeled GlcNAc-P-P-dolichol without being masked by the production of the labeled chitobiosyl derivative. Inhibition by these compounds was detected at the basal level of synthesis of GlcNAc-P-P-dolichol and also when the reaction is stimulated by Man-P-dol. In the absence of showdomycin, inhibition was also demonstrated by the chitobiosyl derivative under basal and stimulatory conditions and by GlcNAc-P-P-dolichol under stimulatory conditions. The present studies have also revealed that the formation of Man-P-Dol can be stimulated by GlcNAc-P-P-dolichol. These influences on the initiating reaction of the dolichol pathway are summarized in Scheme 1. The mutual stimulatory relationships between Man-P-Dol and GlcNAc-P-P-dolichol can be modulated by the inhibitory influences of GlcNAc-P-P-dolichol and GlcNAc-GlcNAc-P-P-dolichol. These relationships, in addition to more global influences such as the availability of substrates and cofactors, may play roles in maintaining the steady state concentration of the first intermediate of the dolichol pathway and thus of the assembly of the oligosaccharide lipids required for the biosynthesis of asparagine-linked glycoproteins.
The biosynthesis of GlcNAc-P-P-dolichol involving an exchange of pyrophosphate bonds on either side of the equation should be readily reversible, as has been demonstrated (17,18). Although the equilibrium constant for this reaction has not  Purified yeast enzyme 11.0 Ϯ 3.6 (2) 8.89 ϫ 10 6 Ϯ 1.6 (2) been reported, the results of the present studies have demonstrated that the inhibition of the formation of GlcNAc-P-Pdolichol, however, cannot be a function only of the concentration of the product driving the reaction backward by mass action. Rather, in addition to the latter effect (the quantitative contribution of which cannot be determined as yet), more subtle influences on the kinetics of this reaction are involved, as demonstrated in the present work. A Kinetic Model for the Mechanism of UDP-GlcNAc:Dolichyl Phosphate N-Acetylglucosamine 1-Phosphate Transferase (GPT-1)-Since patterns of product and substrate inhibition can provide a basis for analyzing the order of additions of substrate and products (40), the following model is suggested from the results of the kinetic studies for the mechanism of action of GPT-1. The steady state kinetic studies are all consistent with a sequential bi-bi mechanism with the feedback inhibition by GlcNAc-GlcNAc-P-P-dolichol being caused by the formation of a ternary dead-end complex with UDP-GlcNAc as shown in Scheme 2.
The data that strongly corroborate the ordered addition and release of substrates are the competitive inhibition of GlcNAc-P-P-dolichol with UDP-GlcNAc as the varied substrate. This indicates that the binding of the sugar nucleotide and GlcNAc-P-P-dolichol is a mutually exclusive event, i.e. they compete for the same form of the enzyme. The initial addition of UDP-GlcNAc is further confirmed by the uncompetitive inhibition pattern with GlcNAc-GlcNAc-P-P-dolichol as the inhibitor. Uncompetitive inhibition arises when the inhibitor binding occurs subsequent to the binding of the varied substrate (40,41). The competitive pattern of GlcNAc-GlcNAc-P-P-dolichol versus dolichol phosphate indicates that the inhibitor and the substrate compete to bind to the E⅐UDP-GlcNAc binary complex. The non-competitive inhibition by GlcNAc-P-P-dolichol versus dolichol phosphate is consistent with the mechanism in Scheme 2. This inhibition arises because GlcNAc-P-P-dolichol can bind to form an inhibitory complex whether dolichol phosphate is present at low or high concentrations. The magnitude of the K ii and K is values (Table III) do not necessarily reflect only the affinity of GlcNAc-P-P-dolichol for a given enzyme form but are dependent on the fraction of the enzyme present in that form (40). It is anticipated that the specific K i values determined in these studies will reflect a greater affinity, i.e. lower K i than that determined in the activity assay (Table I) because the K is and K ii terms reflect an extrapolation to substrate concentrations where a single enzyme form predominates, whereas with the activity assay employing substrates near their K m values several different enzyme forms will be present at steady state.
It might be argued that the data from 2.88 M GlcNAc-P-Pdolichol in Fig. 4A would suggest non-competitive inhibition because drawing a line derived from an analysis of each individual data point rather than as an entire data set will not intersect at a common point on the y axis. If we were to have concluded that this inhibition was non-competitive, a non-statistically significant K ii of 5.8 Ϯ 4.3 mM would have been obtained. The two characteristics of being greater than the largest inhibitor concentration employed coupled with the large standard error are the hallmarks of an unnecessary kinetic parameter, i.e. one that is not required to accommodate the data (32). Concerning the magnitude of the deviations of the data points from the computed values, these systematic errors are less than 1, i.e. the systematic error is within the standard error or random variability of the assay. This variability is reflected in the standard errors of the reported inhibition constants of 10 -25%.
The comparative results with GlcNAc-P-P-dolichol and Glc-NAc-GlcNAc-P-P-dolichol could not be predicted a priori based on homology to the two substrates. GlcNAc-P-P-dolichol acts as an analog of UDP-GlcNAc as is apparent by the observed competitive inhibition, whereas GlcNAc-GlcNAc-P-P-dolichol acts as an analog of dolichol phosphate. The K is of 0.16 M for GlcNAc-GlcNAc-P-P-dolichol versus dolichol phosphate (Table  III) suggests a specific binding. As can be seen in Table III comparing the K is values, the additional GlcNAc residue of the chitobiosyl derivative must introduce specific additional interactions that permit it to bind to the dolichol phosphate site over 20-fold more tightly than GlcNAc-P-P-dolichol.
These studies may contribute to our understanding of mechanisms that regulate the early events of glycoprotein biosynthesis, about which much remains to be known.