Monoglucosylated oligomannosides are released during the degradation process of newly synthesized glycoproteins.

The Chinese hamster ovary mutant MI8-5 is known to synthesize Man(9)GlcNAc(2)-P-P-dolichol rather than the fully glucosylated lipid intermediate Glc(3)Man(9)GlcNAc(2)-P-P-dolichol. This nonglucosylated oligosaccharide lipid precursor is used as donor for N-glycosylation. In this paper we demonstrate that a significant part of the glycans bound to the newly synthesized glycoproteins in MI8-5 cells are monoglucosylated. The presence of monoglucosylated glycans on glycoproteins determines their binding to calnexin as part of the quality control machinery. Furthermore, we point out the presence of Glc(1)Man(5)GlcNAc(1) in the cytosol of MI8-5 cells. This indicates that part of the monoglucosylated glycoproteins can be directed toward a deglycosylation process that occurs in the cytosol. Besides studies on glycoprotein degradation based on the disappearance of protein moieties, MI8-5 cells can be used as a tool to elucidate the various step leading to glycoprotein degradation by studying the fate of the glycan moieties.

Chinese hamster ovary cells with reduced hexokinase activity maintain normal GDP-mannose levels.

Parental Chinese hamster ovary (CHO) cells were mutagenized and subjected first to a mannose suicide selection technique and second to a screen of individual colonies grown on polyester discs for reduced mannose incorporation into protein. The incorporation of radioactivity for the selection and the screen was conducted at 41.5 degrees C instead of the normal growth temperature of 34 degrees C in order to allow for the isolation of temperature-sensitive lesions. This selection/screening procedure resulted in the isolation of M15-4 cells, which had three- to five-fold lower incorporation of [2-3H]mannose into mannose 6-phosphate, mannose 1-phosphate, GDP-mannose, oligosaccharide-lipid, and glycoprotein at 41.5 degrees C. We detected no difference in the qualitative pattern of mannose-labeled lipid-linked oligosaccharides compared to parental cells. M15-4 cells synthesized dolichol. The defect of M15-4 cells was determined to be in hexokinase activity; crude cytosolic extracts were eight- to nine-fold lower in hexokinase activity in M15-4 cells compared to parental cells. As a result of this defect, incorporation of labeled mannose from the medium was significantly decreased. However, the level of GDP-mannose in M15-4 cells was 70% of normal. The phenotype of M15-4 was a lower specific activity of labeled GDP-mannose, not a substantial reduction in the level of GDP-mannose. Consistent with these results, no alterations in the glycosylation of a model glycoprotein, G protein of vesicular stomatitis virus, were observed. These cells grew slower than parental cells, especially in low-glucose medium.

Nonglucosylated oligosaccharides are transferred to protein in MI8-5 Chinese hamster ovary cells.

A CHO mutant MI8-5 was found to synthesize Man9-GlcNAc2-P-P-dolichol rather than Glc3Man9GlcNAc2-P-P-dolichol as the oligosaccharide-lipid intermediate in N-glycosylation of proteins. MI8-5 cells were incubated with labeled mevalonate, and the prenol was found to be dolichol. The mannose-labeled oligosaccharide released from oligosaccharide-lipid of MI8-5 cells was analyzed by HPLC and alpha-mannosidase treatment, and the data were consistent with a structure of Man9GlcNAc2. In addition, MI8-5 cells did not incorporate radioactivity into oligosaccharide-lipid during an incubation with tritiated galactose, again consistent with MI8-5 cells synthesizing an unglucosylated oligosaccharide-lipid. MI8-5 cells had parental levels of glucosylphosphoryldolichol synthase activity. However, in two different assays, MI8-5 cells lacked dolichol-P-Glc:Man9GlcNAc2-P-P-dolichol glucosyltransferase activity. MI8-5 cells were found to synthesize glucosylated oligosaccharide after they were transfected with Saccharomyces cerevisiae ALG 6, the gene for dolichol-P-Glc:Man9GlcNAc2-P-P-dolichol glucosyltransferase. MI8-5 cells were found to incorporate mannose into protein 2-fold slower than parental cells and to approximately a 2-fold lesser extent.

Transfer of two oligosaccharides to protein in a Chinese hamster ovary cell B211 which utilizes polyprenol for its N-linked glycosylation intermediates.

B211, a glycosylation mutant isolated from Chinese hamster ovary cells, synthesizes 10- to 15-fold less Glc3Man9GlcNAc2-P-P-lipid, the substrate used by the oligosaccharide transferase in the synthesis of asparagine-linked glycoproteins. B211 cells are also 10- to 15-fold deficient in the glucosylation of oligosaccharide-lipid. Despite these properties, protein glycosylation in B211 cells proceeds at a level similar to (50% of) parental cells. We asked whether the near wild-type level of glycosylation was due to the transfer of alternative oligosaccharide structures to protein in B211 cells. The aberrant size of [35S]methionine-labeled VSV G protein and the increased percentage of endoglycosidase H-resistant tryptic peptides as compared to parental cells supported this hypothesis. B211 cells were labeled with [2-3H]mannose either for 1 min or for 1 h in the presence of glycoprotein-processing inhibitors so that the oligosaccharides initially transferred to protein could be analyzed. In addition to Glc3Man9GlcNAc2, a second, endoglycosidase H-resistant oligosaccharide was transferred whose structure was determined by alpha-mannosidase digestion, gel filtration chromatography, and HPLC to be Glc0,1Man5GlcNAc2. Finally, since the synthesis of reduced amounts of Glc3Man9GlcNAc2-P-P-lipid was also a phenotype seen in another glycosylation mutant, Lec9, we analyzed the long-chain prenol in B211 cells. B211 cells synthesized and utilized polyprenol rather than dolichol for all N-linked glycosylation intermediates as determined by HPLC analysis of [3H]mevalonate-labeled lipids. Cell fusions analyzed by similar techniques indicated that B211, originally isolated as a concanavalin A-resistant cell line, is in the Lec9 complementation group.

Cytosolic deglycosylation process of newly synthesized glycoproteins generates oligomannosides possessing one GlcNAc residue at the reducing end.

Recent studies on the mechanism of degradation of newly synthesized glycoproteins suggest the involvement of a retrotranslocation of the glycoprotein from the lumen of the rough endoplasmic reticulum into the cytosol, where a deglycosylation process takes place. In the studies reported here, we used a glycosylation mutant of Chinese hamster ovary cells that does not synthesize mannosylphosphoryldolichol and has an increased level of soluble oligomannosides originating from glycoprotein degradation. In the presence of anisomycin, an inhibitor of protein synthesis, we observed an accumulation of glucosylated oligosaccharide-lipid donors (Glc3Man5GlcNAc2-PP-Dol), which are the precursors of the soluble neutral oligosaccharide material. Inhibition of rough endoplasmic reticulum glucosidase(s) by castanospermine led to the formation of Glc3Man5GlcNAc2(OSGn2) (in which OSGn2 is an oligomannoside possessing two GlcNAc residues at its reducing end), which was then retained in the lumen of intracellular vesicles. Thus they were protected during an 8 h chase period from the action of cytosolic chitobiase, which is responsible for the conversion of OSGn2 to oligomannosides possessing one GlcNAc residue at the reducing end (OSGn1). In contrast, when protein synthesis was maintained in the presence of castanospermine, glucosylated oligomannosides (Glc1-3Man5GlcNAc1) were recovered in cytosol. Except for monoglucosylated Man5 species, which are potential substrates for luminal calnexin and calreticulin, the pattern of oligomannosides was similar to that observed on glycoproteins. The occurrence in the cytosol of glucosylated species with one GlcNAc residue at the reducing end implies that the deglycosylation process that generates glucosylated OSGn1 from glycoproteins occurs in the cytosol.

A functional link between N-linked glycosylation and apoptosis in Chinese hamster ovary cells.

Seven different Chinese hamster ovary (CHO) cell mutants, isolated in different ways and having biochemical defects that were expressed at 34 degrees C, were found to be temperature sensitive for growth at 40.5 degrees C. Six of the mutants had five different lesions in N-linked glycosylation; two mutants were in the same complementation group. The temperature-sensitive phenotype in three mutants appeared by cell fusion studies to be linked to the glycosylation phenotype. In some of the glycosylation mutants [B4-2-1 (Lec15.1), Lec9, Lec1, and Lec24], but not in all of them (MI5-4 and MI8-5), incubation at 40.5 degrees C induced apoptosis, as determined by appearance of DNA fragmentation. Tunicamycin (TM) also induced apoptosis in both parental and Lec9 cells. There was a direct correlation between inhibition of glycosylation by TM treatment and induction of apoptosis. Induction of apoptosis by TM was inhibited by cycloheximide. These studies suggest that specific alterations in N-linked glycosylation in CHO cells are endogenous inducers of apoptosis.

Identification of Schizosaccharomyces pombe prenol as dolichol-16,17.

The identity of the prenol involved in N-linked glycosylation in the fission yeast Schizosaccharomyces pombe was unknown. In order to determine the identity of the prenol, S. pombe cells were incubated with a metabolic precursor of prenol, tritiated mevalonolactone. The cells incorporated only a modest amount of label, about 1000 dpm per million cells, into base-stable lipid and only 1% of that radioactivity was incorporated into prenol. We found by normal phase silica HPLC and more directly by the lack of reactivity with MnO2 that the labeled lipid was predominantly dolichol, not polyprenol. Reverse phase HPLC demonstrated that in S. pombe dolichol ranged between 14 and 18 isoprene units with dolichol-16,17 being the most abundant prenol. This dolichol is of an intermediate length, between the dolichol of S. cerevisiae and that of mammalian cells.

Reduced utilization of Man5GlcNAc2-P-P-lipid in a Lec9 mutant of Chinese hamster ovary cells: analysis of the steps in oligosaccharide-lipid assembly.

Recently we reported that CHB11-1-3, a Chinese hamster ovary cell mutant defective in glycosylation of asparagine-linked proteins, is defective in the synthesis of dolichol [Quellhorst et al., 343:19-26, 1997: Arch Biochem Biophys]. CHB11-1-3 was found to be in the Lec9 complementation group, which synthesizes polyprenol rather than dolichol. In this paper, levels of various polyprenyl derivatives in CHB11-1-3 are compared to levels of the corresponding dolichyl derivatives in parental cells. CHB11-1-3 was found to maintain near normal levels of Man5GlcNAc2-P-P-polyprenol and mannosylphosphorylpolyprenol, despite reduced rates of synthesis, by utilizing those intermediates at a reduced rate. The Man5GlcNAc2 oligosaccharide attached to prenol in CHB11-1-3 cells and to dolichol in parental cells is the same structure, as determined by acetolysis. Man5GlcNAc2-P-P-polyprenol and Man5GlcNAc5-P-P-dolichol both appeared to be translocated efficiently in an in vitro reaction. Glycosylation of G protein was compared in vesicular stomatitus virus (VSV)-infected parent and mutant; although a portion of G protein was compared in vesicular stomatitus virus (VSV)-infected parent and mutant; although a portion of G protein was normally glycosylated in CHB11-1-3 cells, a large portion of G was underglycosylated, resulting in the addition of either one or no oligosaccharide to G. Addition of a single oligosaccharide occurred randomly rather than preferentially at one of the two sites.

Synthesis of dolichol in a polyprenol reductase mutant is restored by elevation of cis-prenyl transferase activity.

CHB11-1-3 is a glycosylation mutant of Chinese hamster ovary (CHO) cells, isolated by screening mutagenized cells for those with decreased intracellular lysosomal enzyme activity [C. W. Hall et al. (1986) Mol. Cell. Biochem. 72, 35-45]. CHB11-1-3 synthesizes the lipid polyprenol, the metabolic precursor of dolichol, rather than dolichol, indicating a defect in polyprenol reductase. This defect was demonstrated previously in Lec9 CHO mutants, and cell fusion experiments confirmed that CHB11-1-3 is a member of this complementation group. A revertant of CHB11-1-3, CHBREV, isolated for its ability to grow at 39 degrees C, synthesizes dolichol at near-normal levels. CHBREV is probably a second-site revertant, because it synthesizes three to four times as much polyprenol as CHB11-1-3 and exhibits a similar elevation in the specific activity of cis-prenyl transferase. This higher activity appears to reflect an increase in enzyme molecules rather than the presence of an activator or absence of an inhibitor. These results suggest that CHB11-1-3 is a "K(m)" mutant, because synthesis of higher amounts of the substrate of polyprenol reductase obviates the defect.

Aspartic acid 252 and asparagine 185 are essential for activity of lipid N-acetylglucosaminylphosphate transferase.

A key step in the assembly of oligosaccharide-lipid intermediates in N-linked glycosylation is the transfer of N-acetylglucosamine 1-phosphate to dolichyl phosphate, catalyzed by the enzyme UDP-N-acetylglucosaminyl:dolichyl phosphate N-acetylglucosaminyl phosphoryl transferase (L-G1PT). Comparison of the amino acid sequences of L-G1PT from five diverse species showed 75 amino acids identical in all five proteins. Using site-directed mutagenesis, we analyzed the importance of a number of these conserved residues to the enzymatic activity of L-G1PT using a plasmid shuffling procedure in Schizosaccharomyces pombe. S. pombe cells containing a chromosomal deletion of the essential gpt+ gene are rescued by a plasmid containing the S. pombe gpt open reading frame. Replacement of that plasmid by a plasmid encoding a mutated hamster L-G1PT cDNA sequence indicated that the mutated protein provided sufficient enzyme activity to permit cell growth. Mutations of aspartic acid 252 and asparagine 185 did not allow plasmid shuffling, indicating these residues were essential for activity. A combination of mutations at asparagine 182 and tryptophan 122 did not allow plasmid shuffling, although the single mutations did. Overexpression of the mutant proteins in S. pombe conferred tunicamycin (TM) resistance, indicating that the mutant proteins had a conformation necessary for binding TM, a substrate analog. The mutant proteins were also detected in Western blots and were correctly localized to the membrane fractions. However, the overexpressed proteins did not increase the endogenous level of enzymatic activity in these cells, indicating they were enzymatically inactive.

Asparagine-linked glycosylation in Schizosaccharomyces pombe: functional conservation of the first step in oligosaccharide-lipid assembly.

The gene gpt encoding uridine diphosphate N-acetyl-D-glucosamine:dolichol phosphate N-acetylglucosaminylphosphoryltransferase (L-G1PT) was isolated by screening a Schizosaccharomyces pombe genomic DNA library in lambda phage under low-stringency hybridization using the Saccharomyces cerevisiae gene ALG7 as probe. Sequencing 2.4 kb of S. pombe DNA revealed a 1338-bp open reading frame (ORF) encoding a hydrophobic protein of 446 amino acids with a predicted molecular weight of 49,852. The S. pombe protein was 50% identical to the S. cerevisiae protein and 43% identical to the protein from Chinese hamster ovary (CHO) cells. Overexpression of the gpt gene in S. pombe cells increased resistance to tunicamycin 25-fold and increased the specific activity of the enzyme in isolated cell membranes 13-fold. This was accompanied by a 50-fold increase in poly(A)+ RNA hybridizing to the gpt probe. Northern analysis indicated a single 1.8-kb message is transcribed from the gpt gene. The gpt gene is essential for viability of S. pombe. Cells containing a disrupted ORF could be rescued by an expression plasmid containing either the intact S. pombe gpt ORF or the CHO L-G1PT cDNA. The S. pombe gpt gene was mapped to chromosome 2 near top1 and ade1.

Genomic organization and expression of hamster UDP-N-acetylglucosamine:dolichyl phosphate N-acetylglucosaminyl phosphoryl transferase.

The hamster gene for uridine diphosphate N-acetyl-D-glucosamine:dolichyl phosphate N-acetylglucosaminyl phosphoryl transferase (L-G1PT) was found to extend over 6.5 kb and to contain nine exons. The exons ranged in size from 63 to 214 bp, encoding the 408 amino acid protein. The introns ranged from 85 bp to 1.4 kb. Upstream 5' sequences included two possible TATA boxes, one possible CCAAT box and at least two potential GC boxes. Heterologous expression was successful in Schizosaccharomyces pombe, and resulted in cells that were tunicamycin resistant and had 12-fold more L-G1PT activity than wild-type cells. Antiserum prepared to a hydrophilic peptide (residues 300-320) of the L-G1PT protein reacted with a 35-36 kDa protein in membrane samples from Chinese hamster ovary (CHO) cells and S. pombe cells that had increased levels of L-G1PT activity. In both cases, antigenic peptide competed with the 35-36 kDa protein detected by the antiserum.

Three enzymes involved in oligosaccharide-lipid assembly in Chinese hamster ovary cells differ in lipid substrate preference.

Initial steps in N-linked glycosylation involve formation of a large oligosaccharide structure on a lipid carrier, dolichyl phosphate. We have previously characterized Chinese hamster ovary (CHO) glycosylation mutants (Lec9 cells) that utilize the polyisoprenoid lipid polyprenyl phosphate rather than dolichyl phosphate in these glycosylation reactions. Polyprenyl phosphate differs from dolichyl phosphate only in the degree of saturation of its terminal isoprenyl unit. Our goal was to determine whether the glycosylation defect of Lec9 cells could be explained simply by knowing lipid substrate preferences of the enzymes involved in the assembly of oligosaccharide-lipid (OSL) intermediates. In this study, we have used in vitro assay systems to compare the ability of dolichyl phosphate and polyprenyl phosphate to act as substrates for three glycosyl transferase enzymes involved in OSL assembly. In order to insure that we were only examining lipid substrate preferences of the enzymes and not other potential defects present in Lec9 cells, we used membranes prepared from wild-type cells in these in vitro reactions. Our results indicate that one of the enzymes, mannosylphosphoryldolichol (MPD) synthase, exhibited a significant preference for the dolichol substrate. Glucosylphosphoryldolichol (GPD) synthase, on the other hand, showed no binding specificity for the dolichol substrate, although the enzyme used the dolichol substrate at a twofold higher rate. N,N'-diacetyl-chitobiosylpyrophosphoryldolichol (CPD) synthase was able to use either lipid substrate with equal efficiency. These results suggest that not all glycosyl transferases in this pathway show a preference for dolichol derivatives.(ABSTRACT TRUNCATED AT 250 WORDS)

Induction of dolichyl-saccharide intermediate biosynthesis corresponds to increased long chain cis-isoprenyltransferase activity during the mitogenic response in mouse B cells.

There are large increases in the rates of Glc3-Man9GlcNAc2-P-P-Dol (Oligo-P-P-Dol) biosynthesis and protein N-glycosylation during the proliferative response of murine B lymphocytes (B cells) to bacterial lipopolysaccharide (LPS). To learn more about the regulation of dolichyl-saccharide biosynthesis, the possible relationships between developmental changes in specific steps in dolichyl phosphate (Dol-P) and N-acetyl-glucosaminylpyrophosphoryldolichol (GlcNAc-P-P-Dol) biosynthesis and the induction of Oligo-P-P-Dol biosynthesis were investigated. These studies describe an impressive induction of long chain cis-isoprenyltransferase (cis-IPTase) activity, an enzyme system required for the chain elongation stage in de novo Dol-P synthesis, which corresponded to the striking increase in the rate of Oligo-P-P-Dol biosynthesis in LPS-activated B cells. The cellular level and specific activity of cis-IPTase increase 15-fold in LPS-treated cells with relatively unaltered affinity for isopentenyl pyrophosphate. The rates of Dol-P and Oligo-P-P-Dol synthesis increased substantially when cis-IPTase activity was induced, suggesting a regulatory relationship between the level of cis-IPTase activity and lipid intermediate synthesis. Distinctly different developmental patterns were observed for cis-IPTase and HMG-CoA reductase activity, and when sterol biosynthesis was drastically inhibited by lovastatin, the rate of synthesis of Dol-P was slightly higher in the presence of the drug. Modest elevations in the cellular levels of dolichol kinase, Dol-P phosphatase, and UDP-GlcNAc:Dol-P N-acetylglucosaminylphosphoryltransferase (L-G1PT) activities were also observed, but these changes were relatively small compared with the increases in cis-IPTase activity and the rates of Dol-P, Gl-cNAc-P-P-Dol, and Oligo-P-P-Dol synthesis. The expression of the L-G1PT gene is an early event in the developmental program for Oligo-P-P-Dol synthesis, but GlcNAc-P-P-Dol formation is apparently not rate-limiting. In summary, large increases in cis-IPTase activity and the rate of Dol-P biosynthesis appear to play a key regulatory role in the induction of Oligo-P-P-Dol biosynthesis during the proliferative response of B cells to LPS, and the biosynthetic pathways for Dol-P and cholesterol are regulated independently in dividing B cells.

Mammalian glycosyltransferases prefer glycosyl phosphoryl dolichols rather than glycosyl phosphoryl polyprenols as substrates for oligosaccharyl synthesis.

We have studied the effectiveness of polyprenyl-P-mannose and polyprenol-P-glucose as donor substrates for the dolichyl-P-mannose:Man5(GlcNAc)2-PP-dolichol mannosyltransferase and the dolichyl-P-glucose:Man9(GlcNAc)2-PP-dolichol glucosyltransferase, respectively. The polyprenol moiety differs from dolichol only in the unsaturation of the terminal isoprene unit of the molecule. Based on the kinetics of the reactions, we have found that both glycosyltransferases have higher apparent Kms and lower apparent Vmaxs using polyprenyl-P-monosaccharides as substrates rather than the dolichyl-P-monosaccharides. The products formed with the polyprenyl-P-sugars were the same as those formed by the dolichol-linked sugars, indicating that the polyprenol substrates could be utilized by the glycosyltransferases in vitro. The results also indicate that the dolichyl-P-sugars and the polyprenyl-P-sugars compete for the same binding site on the enzyme. These findings are significant in terms of understanding the glycosylation phenotypes of Chinese hamster ovary cell mutants of the Lec9 complementation group, which lack the ability to convert polyprenol into dolichol.

Mutants in dolichol synthesis: conversion of polyprenol to dolichol appears to be a rate-limiting step in dolichol synthesis.

Chinese hamster ovary (CHO) cells of the Lec9 recessive complementation group display a distinctive profile of resistance to a variety of toxic lectins. In addition, they accumulate cis-alpha-unsaturated polyprenol and use mainly polyprenol rather than dolichol to synthesize the glycosylated lipids used in asparagine-linked glycosylation of proteins. The primary defect in these cells is thought to result from a deficiency in polyprenol reductase activity. Three new mutants were isolated and determined to have qualitatively, although not quantitatively, similar lectin resistance profiles to Lec9 cells. Two of these mutants (AbrR and RicR) also contained polyprenol rather than dolichol. The lectin resistance profile of an independent mutant which accumulates polyprenol, F2A8, was also found to be qualitatively similar to the Lec9 pattern. The relationship among these mutants was analysed in more detail by construction of cell-cell hybrids. Lectin resistance profiles of the hybrids demonstrated that AbrR, RicR and F2A8 fell into the Lec9 complementation group. Analysis of prenols in the hybrids also showed that F2A8 was a member of the Lec9 group. Surprisingly, a significant fraction of the prenols found in Lec9 x Parent hybrids was polyprenol (up to 30% of the neutral fraction), whereas the prenols found in Parent x Parent hybrids were nearly exclusively dolichol (97% of the neutral lipid fraction). Therefore, reduction of polyprenol to dolichol appears to be a rate-limiting step in the synthesis of dolichol since hybrids with differing numbers of wild-type alleles can be biochemically distinguished.

Lec15 cells transfer glucosylated oligosaccharides to protein.

B4-2-1 cells (Lec15 cells) are Chinese hamster ovary cells deficient in mannosylphosphoryldolichol synthase activity. They synthesize the truncated lipid intermediate Man5GlcNAc2-P-P-dolichol rather than the Glc3Man9GlcNAc2-P-P-dolichol synthesized by wild-type cells. In this report we present evidence that these cells did synthesize glucosylated Man5GlcNAc2-P-P-dolichol, but this species represented only a minor fraction of the labeled oligosaccharide-lipid. On the other hand, glucosylated oligosaccharides were a major species transferred to protein in these cells, showing that in vivo, glucosylated oligosaccharides are preferentially transferred to protein. The truncated oligosaccharides found in B4-2-1 cells were removed from the protein by N-glycanase treatment, since they were resistant to both endo-beta-N-acetylglucosaminidase H and F activity. B4-2-1 cells processed the glucosylated, truncated oligosaccharides transferred to G protein of vesicular stomatitis virus, leading to infectious virus.

Substrate specificity of N-acetylglucosamine 1-phosphate transferase activity in Chinese hamster ovary cells.

The assembly pathway of the oligosaccharide chains of asparagine-linked glycoproteins in mammalian cells begins with the formation of GlcNAc-PP-dolichol in a reaction catalysed by the enzyme N-acetylglucosamine 1-phosphate transferase. We have investigated the efficiency of two lipid substrates for the transferase activity in an in vitro assay using Chinese hamster ovary (CHO) cell membranes as an enzyme source. Experiments were carried out with varying concentrations of dolichyl phosphate or its precursor, polyprenyl phosphate. We determined that enzyme activity was optimal at pH 9, where the enzyme exhibited a 3-fold higher Vmax and a 2-fold lower Km for the dolichol substrate. At pH 7.4, the Km and Vmax differences between the two lipids were 10-fold. Under all assay conditions tested, we found that GlcNAc-PP-lipid was the only product formed. We conclude from these results that dolichyl phosphate rather than polyprenyl phosphate is the preferred substrate for the transferase enzyme in CHO cells. This observation is significant in light of the fact that we have previously isolated CHO glycosylation mutants which fail to convert polyprenol into dolichol, and hence utilize polyprenyl derivatives for glycosylation reactions. Thus, these results contribute to our understanding of the glycosylation defects in the mutant cell lines.