Paramecium bursaria Chlorella Virus 1 Encodes Two Enzymes Involved in the Biosynthesis of GDP- L -fucose and GDP- D -rhamnose*

At least three structural proteins in Paramecium bursaria Chlorella virus (PBCV-1) are glycosylated, includ-ing the major capsid protein Vp54. However, unlike other glycoprotein-containing viruses that use host-en-coded enzymes in the endoplasmic reticulum-Golgi to glycosylate their proteins, PBCV-1 encodes at least many, if not all, of the glycosyltransferases used to glycosylate its structural proteins. As described here, PBCV-1 also encodes two open reading frames that resemble bacterial and mammalian enzymes involved in de novo GDP- L -fucose biosynthesis. This pathway, starting from GDP- D -mannose, consists of two sequential steps catalyzed by GDP- D -mannose 4,6 dehydratase (GMD) and GDP-4-keto-6-deoxy- D -mannose epimerase/ reductase, respectively. The two PBCV-1-encoded protein encoding regions for the putative GMD and GMER enzymes, ORFs A118R and A295L, respectively. The forward primers used to amplify GMD (5 (cid:2) -AATTGGATCCATGTTGTCTAAAGTTGCATTAGT) and GMER (5 (cid:2) -AATTGGATCCATGATAATGGATAAGCACTCTAA) contained a Bam HI restriction site, whereas the reverse primers for GMD (5 (cid:2) -AATTCTCGAGTTATTTATTTCCAAACGTATCTGA) and GMER (5 (cid:2) -AATTCTCGAGTTAGTTCGTGTTAGTATATTTTATA) contained a Xho I restriction site. Polymerase chain reactions (50 (cid:1) l) were performed with 1.5 units of Pfu polymerase (Sigma), in the presence of 2 m M MgSO 4, 0.2 m M deoxyribonucleotides triphosphate, and 0.1 (cid:1) M primers. The PCR products were purified from 1.2% agarose gels using the Nucleobond kit (Macherey-Nagel, Duren, Germany), digested with Bam HI and Xho I, and ligated into the pGEX-6-P1 vector (Amersham Biosciences) using a rapid ligation kit (Roche Applied Science). The resulting plasmids were used to transform E. coli K803. DNA sequenc-ing of positive clones was performed by M-Medical (Firenze, Italy). E. coli K803 cells containing vectors with the correct inserts were grown in 2 (cid:3) YT medium (1.6% tryptone, 1% yeast extract, 0.5% NaCl, pH 7.0), containing 200 (cid:1) g/ml ampicillin. Expression of both glutathione S -transferase fusion proteins was induced with 0.1 m M isopropyl- (cid:2) - D -thiogalactopyranoside. Optimal protein recovery was obtained by induction at 20 °C for 16 h for GMD and at 30 °C for 2 h for

Paramecium bursaria Chlorella virus (PBCV-1) 1 is the prototypical member of a group of large polyhedral, plaque-form-ing viruses in the family Phycodnaviridae that replicate in certain unicellular, eukaryotic Chlorella-like green algae (1). The PBCV-1 genome, which has been sequenced, consists of a large (Ͼ330 kb) linear, non-permuted double-stranded DNA with covalently closed hairpin ends (2). Additional features of the virus have been reviewed recently (3). The PBCV-1 major capsid protein, Vp54, located on the viral surface, is one of three glycosylated viral proteins. The glycan portion of Vp54 contains seven neutral sugars, glucose, fucose, rhamnose, galactose, mannose, xylose, and arabinose (4) that contribute to the protease resistance and antigenicity of the virus. Recently, the Vp54 crystal structure was solved, and four N-linked glycosylation sites and two O-linked oligosaccharides were identified (5). However, the four glycosylated Asn residues are not located in consensus sequences typical of eukaryotic N-linked glycans (5). This finding, together with other results (4,6,7), led to the conclusion that unlike other viruses PBCV-1 encodes at least part, if not all, of the machinery required to glycosylate its structural proteins. Furthermore, glycosylation probably occurs independently of the host endoplasmic reticulum-Golgi system.
Analysis of the PBCV-1 genome revealed several ORFs that are related to enzymes involved in glycan formation. For example, PBCV-1 encodes three enzymes, UDP-glucose dehydrogenase, glucosamine synthase, and hyaluronan synthase, involved in the biosynthesis of the extracellular matrix polysaccharide hyaluronan (8 -10). Other ORFs resemble glycosyltransferases and enzymes involved in the metabolism of nucleotide sugars (2,3). In the present study we characterize two PBCV-1 ORFs, A118R and A295L, that encode proteins that resemble GDP-D-mannose 4,6-dehydratase (GMD) and GDP-4-keto-6-deoxy-D-mannose epimerase/reductase (GMER), respectively. GMD and GMER are highly conserved proteins involved in GDP-L-fucose formation in bacteria, plants, and animals (11)(12)(13)(14)(15). Enzyme analyses, performed on both the recombinant proteins and extracts from PBCV-1-infected Chlorella cells, indicate that both proteins are functionally active, leading to the production of two nucleotide sugars, GDP-Lfucose and GDP-D-rhamnose. Therefore, PBCV-1 is the first known virus to encode enzymes involved in nucleotide sugar metabolism.

EXPERIMENTAL PROCEDURES
Cloning and Expression of Recombinant GMD and GMER-Production of virus PBCV-1 and DNA isolation were performed as described (16,17). Gene-specific primers were used for PCR amplification of the * This work was supported by the Italian Consiglio Nazionale delle Ricerche (CNR) Target Project "Biotechnology," CNR Agenzia 2000, CNR/Ministry for Education, University, and Research Project (Legge 95/95) "Biomolecole per la salute umana" and Grant GM-32441 from the National Institutes of General Medical Sciences (to J. V. E.). 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.
This paper is dedicated to the memory of Prof. Eraldo Antonini, outstanding biochemist, prematurely deceased 20 years ago, March 19th, 1983.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) GMD AY225120 and GMER AY225121.
E. coli K803 cells containing vectors with the correct inserts were grown in 2ϫ YT medium (1.6% tryptone, 1% yeast extract, 0.5% NaCl, pH 7.0), containing 200 g/ml ampicillin. Expression of both glutathione S-transferase fusion proteins was induced with 0.1 mM isopropyl-␤-D-thiogalactopyranoside. Optimal protein recovery was obtained by induction at 20°C for 16 h for GMD and at 30°C for 2 h for GMER. Cell lysis, purification of the fusion proteins, and proteolytic cleavage of the glutathione S-transferase tag were performed as described (18). Protein concentrations were determined by the Bradford assay (Bio-Rad), and protein purity was monitored by SDS-PAGE. Both recombinant proteins were stored at 4°C in 50 mM Tris/HCl, pH 7.0, 150 mM NaCl, 1 mM EDTA, and 1 mM dithiothreitol in the presence of 10% glycerol.
In Vivo Expression of PBCV-1 GMD and GMER-Chlorella NC64A was grown in modified Bold's basal medium and infected with PBCV-1 following described procedures (16). Aliquots were withdrawn at various times postinfection (p.i.), and RNA was isolated, electrophoresed under denaturing conditions on 1.5% agarose/formaldehyde gels, stained with ethidium bromide, and transferred to nylon membranes as described (19). The RNA was hybridized with a118r or a295l probes; the probes were labeled with 32 P using the random primers DNA labeling system (Invitrogen). After hybridization, radioactivity bound to the membranes was visualized using Storm 840 phosphorimaging and ImageQuant software (Molecular Dynamics, Inc.) as described (19).
Protein expression and enzyme activities were assayed on cell extracts obtained at various times after virus infection. Cells were disrupted by resuspending algal pellets in Tris/HCl 50 mM, pH 7.0, 0.1% Nonidet P-40, and 1 l/ml protease inhibitor mixture (Sigma), at 3 ϫ 10 7 cells/ml, and vortexing for 10 min at 4°C with 600 mg/ml of 1-mm glass beads. Supernatants obtained after centrifugation at 12,000 ϫ g for 20 min were used for further analyses. Rabbit antisera raised against recombinant GMD and GMER were produced by PRIMM (Milano, Italy). Western analyses employed standard procedures, and detection was performed with the ECLplus system (Amersham Biosciences).
Analysis of Enzyme Activities-GMD and GMER activities were assayed on both recombinant proteins and algal extracts prepared from virus-infected cells. Recombinant GMD dehydratase activity was assayed using either unlabeled GDP-D-mannose (Sigma) or a mixture of unlabeled substrate and GDP-D-[U-14 C]mannose (New England Nuclear, final specific activity 231 MBq/mmol). Unless otherwise specified, the standard assay contained 100 M substrate in 50 mM Tris/HCl, pH 7.0, and 1 mM NADP ϩ ; products were monitored by reverse phase and anion exchange HPLC (see below).
To prepare the GMER substrate GDP-4-keto-6-deoxy-D-mannose, 5 mM GDP-D-mannose was incubated with 200 g/ml of recombinant viral GMD for 30 min. Conversion of GDP-D-mannose to GDP-4-keto-6-deoxy-D-mannose was determined by HPLC (see below). At the end of the reaction GMD was removed from the incubation mixture by ultrafiltration, using Microcon YM-10 microconcentrators (Millipore). Because of its chemical instability, GDP-4-keto-6-deoxy-D-mannose was used immediately after production. Recombinant GMER activity was assayed by monitoring GDP-L-fucose production by reverse phase HPLC. The standard reaction contained 100 M substrate in 50 mM Tris/HCl, pH 7.0, and 1 mM NADPH. The products were identified as nucleotide FIG. 2. Reverse phase HPLC analysis of the products from PBCV-1 GMD and GMER enzyme activities. The eluate was monitored by continuous flow scintillation counting. A, GDP-D-mannose (peak 1) was incubated with GMD and NADP ϩ , leading to the formation of the intermediate compound GDP-4-keto-6-deoxy-D-mannose (peak 2). B, GDP-4-keto-6-deoxy-D-mannose (peak 2) incubated with NADPH in the absence of enzyme. Essentially identical results were obtained when GDP-4-keto-6-deoxy-D-mannose was incubated with either GMD or GMER in the absence of NADPH. C, GDP-4-keto-6-deoxy-D-mannose (peak 2), incubated with GMER and NADPH, was converted to GDP-Lfucose (peak 3). D, incubation of GDP-D-mannose with GMD and both NADP ϩ and NADPH resulted in the formation of a new compound (peak 4), whose retention time differed from the expected nucleotide sugars. E, incubation of GDP-D-mannose (peak 1) with both GMD and GMER in the presence of NADP ϩ and NADPH led to the formation of both GDP-L-fucose (peak 3) and a new unidentified compound (peak 4).
sugars by comparing HPLC retention times with known standards and by electrospray mass spectrometry analyses. Monosaccharides obtained after acid hydrolysis were analyzed by TLC and GC-MS (see below).
Cell extracts from uninfected and virus-infected algae were assayed for enzyme activity in Tris/HCl, pH 7.0, 100 M 14 C-labeled substrates, and either 1 mM NADP ϩ or NADPH at a protein concentration of 5 mg/ml. Products were monitored by both reverse phase and anion exchange HPLC.
Chromatographic Analyses of GMD and GMER Products-Products formed from both recombinant proteins and cell extracts were compared To improve separation of the products resulting from GMD activity, a new anion exchange procedure was developed (see "Experimental Procedures"). The eluate was monitored at 260 nm. The peak numbers are the same as in Fig. 2: peak 1 is GDP-D-mannose, peak 2 is GDP-4-keto-6-deoxy-D-mannose, peak 4 is the new GDP sugar. with known standards by TLC and HPLC. TLC was performed on the 14 C-radiolabeled monosaccharides after acid hydrolysis, as described in Ref. 11. To analyze the intermediate compound GDP-4-keto-6-deoxy-Dmannose, samples were reduced with NaBH 4 prior to hydrolysis (11). The radiolabel was detected with the Cyclone system (Canberra Packard).
Sample extractions and reverse phase HPLC, using a C18 column (Waters), followed published procedures (11). Better separation of GDPmannose, GDP-4-keto-6-deoxy-D-mannose, and GDP-D-rhamnose was achieved by anion exchange chromatography. For this analysis, 100-l aliquots of the extracted samples were injected onto a Wesca Anion/R column (Alltech) and eluted with a 100 -500 mM NH 4 HCO 3 gradient at a flow rate of 2 ml/min. An additional step, using 1 M NaCl, was required to elute NADPH, which otherwise remained tightly bound to the column. For both reverse phase and anion exchange HPLC, eluates were monitored by UV absorbance at 260 nm and by continuous flow scintillation counting (Packard Radiomatic 500TR).
Electrospray MS Analysis of GDP Sugars-Electrospray analysis was performed on an Agilent1100 MSD ion trap instrument (Agilent Technologies, Palo Alto, CA) by flow injection of the samples with an infusion pump (KD Scientific, New Hope, PA). The reaction mixtures were diluted 1:19 with water to reduce salt concentration; spectra were acquired in the negative ion mode in the mass range of the expected m/z ratios. The ion source parameters were set to obtain optimal signal to noise ratio for molecules of interest (20).
Monosaccharide Analysis by GC-MS-GDP sugars obtained with the recombinant enzymes were purified by anion exchange chromatography as described above. The GDP-D-rhamnose standard was kindly provided by Dr. Paul Messner, Universitat fü r Bodenkultur, Wien, Austria (21). Samples were derivatized according to Henry et al. (22) and Merkle and Poppe (23) with slight modifications. Briefly, the nucleotide sugars were hydrolyzed with 2 N HCl for 90 min at 100°C, dried in a Speed-Vac, and the resulting monosaccharides were reduced to their corresponding alditols with NaBH 4 (10 mg/ml in 1 M ammonia solution). The products were acetylated in 1:1 pyridine/acetic anhydride, resuspended in 100 l of CH 2 Cl 2 , and 5-l aliquots of this solution were injected into an HP 5890 series II gas chromatograph coupled with an HP 5889A engine mass spectrometer equipped with an electron impact ion source (Hewlett Packard). Separation was performed on a S.E.-54 (Alltech) capillary column; the helium gas flow was 45 ml/min. The oven temperature gradient was as follows: initial temperature 80°C, isothermal for 10 min, 80 to 190°C (rate 15°C/min), 190 to 250°C (rate 4°C/min), isothermal to 54 min. The MS analysis was performed in single ion monitoring detection mode (ion 115), monitoring characteristics and specific ions derived from the scan analysis of preliminary MS spectra.
To determine the D or L nature of the rhamnose monosaccharide, (S)-octyl glycoside derivatives were obtained from the dried hydrolyzed GDP-rhamnose according to Leontein et al. (24). Samples were dissolved in (S)-2-octanol and heated at 120°C overnight with a catalytic amount of trifluoroacetic acid. The solution was dried, acetylated, and analyzed as described above. The single ion monitoring chromatograms of the 273 ion corresponding to the D-and L-rhamnose standards were compared with the unknown epimeric form.

Expression of PBCV-1 Recombinant GMD and GMER Enzymes-PBCV-1 ORFs A118R (345 amino acids) and A295L
(317 amino acids) (accession number U42580 for PBCV-1 genome) encode proteins that resemble GMD and GMER, respectively, from prokaryotic and eukaryotic organisms. Resequencing both genes indicated that the putative GMD sequence was identical to that initially reported, whereas GMER differed at amino acid positions 142 and 143, with AK instead of ER. The nucleotide and the deduced amino acid sequences have been deposited in GenBank TM , accession codes AY225120 for GMD and AY225121 for GMER. No other sequences bearing homology to enzymes involved in the metabolism of nucleotide sugars were detected by either BLAST or TFASTA analyses of the PBCV-1 genome.
Recombinant PBCV-1 GMD and GMER proteins with the expected molecular weights were obtained in high yield and good purity (ϳ2 mg/liter of initial bacterial culture for GMD and 0.8 mg/liter for GMER) (Fig. 1). The lower yield for GMER resulted from precipitation of the protein during purification.

Characterization of PBCV-1 Recombinant GMD and GMER
Enzymes-Enzyme assays established that both proteins have the expected catalytic activities. Dehydration of GDP-D-mannose by viral recombinant GMD produced a compound with an identical retention time on reverse phase HPLC ( Fig. 2A, peak  2) as the GDP-4-keto-6-deoxy-D-mannose, an intermediate in the GDP-L-fucose biosynthetic pathway, produced by human recombinant GMD (18). The identity of this compound was confirmed by TLC after NaBH 4 reduction, which led to the production of both rhamnose and 6-deoxy-talose (11) (not shown), and by electrospray MS, which indicated a m/z of 586.3, consistent with GDP-D-mannose (m/z 604.3) lacking a water molecule (Fig. 3, A and B). GDP-4-keto-6-deoxy-D-mannose (peak 2) incubated with NADPH in the absence of enzyme was stable, thus ruling out any spontaneous chemical reduc-

FIG. 5. GC-MS analysis of the monosaccharides in the GDP-6deoxy-hexoses produced by GMD and GMER.
A and B, the nucleotide sugars were purified by anion exchange chromatography, and the monosaccharides were released by acid hydrolysis, followed by conversion to alditol acetates and GC-MS analysis. Retention times were compared with those of sugar standards: fucose, produced by GMER activity (A), and rhamnose, produced by GMD reductase activity (B). C, rhamnose obtained by GMD reductase activity was derivatized with a chiral compound, (S)-2-octanol. The corresponding (S)-octyl glycoside derivative, analyzed by GC-MS, had the same retention time as the GDP-D-rhamnose standard, indicating that rhamnose synthesized by PBCV-1 GMD has a D configuration. For further details, see "Experimental Procedures." tion of the intermediate (Fig. 2B).
Incubation of recombinant GMER with NADPH and GDP-4keto-6-deoxy-D-mannose produced a compound with a retention time identical to GDP-L-fucose (Fig. 2C, peak 3). Electrospray ionization-MS spectra indicated a m/z of 588.4, which is compatible with GDP-L-fucose (not shown). After hydrolysis, this monosaccharide migrated identically to fucose on TLC. Thus, recombinant viral GMER is a catalytically active, bifunctional enzyme with both epimerase and reductase activities (11).
Unexpectedly, when GMD was incubated with its substrate GDP-D-mannose and both NADP ϩ and NADPH, instead of NADP ϩ alone, an unidentified compound was detected by reverse phase HPLC (Fig. 2D, peak 4). Electrospray MS analysis of this unidentified GMD product indicated a m/z of 588.3, which is consistent with a GDP-6-deoxy-hexose (Fig. 3C). TLC analysis performed after acid hydrolysis (without NaBH 4 reduction) of this C 14 -labeled GMD product indicated a single compound that migrated like the 6-deoxyhexose sugar rhamnose (not shown). This same unidentified compound appeared when purified intermediate GDP-4-keto-6-deoxy-D-mannose was reincubated with GMD and NADPH. Parallel experiments established that formation of the new compound was accompanied by a stoichiometric decrease in NADPH absorbance at 340 nm (not shown). GDP-4-keto-6-deoxy-D-mannose was not converted to any other compounds when GMD or NADPH was omitted from the reaction. These findings indicate that the virus GMD also has a stereospecific NADPH-dependent reductase activity. This activity was confirmed by reconstructing the complete pathway, i.e. by adding GMD and GMER to an incubation mixture containing GDP-D-mannose, NADP ϩ , and NADPH. These conditions produced both GDP-L-fucose (peak 2) and the unknown compound (Fig. 2D, peak 4). The amounts of GDP-L-fucose and the new compound depended on the relative amounts of GMD and GMER (i.e. a high GMER/GMD ratio resulted in more GDP-L-fucose formation and vice versa).
To confirm the identity of all the nucleotide sugars formed by GMD and GMER, we developed an anion exchange HPLC procedure that improved separation of the nucleotide sugars (Fig. 4). Spectra analyses indicated that each peak contained a guanine moiety. GC-MS analysis on the monosaccharides, performed after acid hydrolysis of the nucleotide sugars and conversion to the corresponding alditol acetates, confirmed that GMER activity produced GDP-L-fucose (Fig. 5A). GMD reductase activity produced only one compound with the same retention time as rhamnose (Fig. 5B). This finding establishes that GMD also acts stereospecifically on the 4-keto group of the intermediate nucleotide sugar that is generated from GDP-Dmannose. The absolute configuration of rhamnose was determined by a GC-MS procedure that differentiates chiral compounds (23). The rhamnose produced by the GMD reductase activity had a retention time identical to that of the GDP-Drhamnose standard (21) (Fig. 5C).
The V max of recombinant GMD in the presence of 1 mM NADP ϩ was 34.5 Ϯ 1.2 mol/h/mg protein for the dehydratase activity, which is similar to recombinant human GMD (18). However, unlike GMDs from other organisms, which are reported to be very unstable, PBCV-1 GMD retains complete activity when stored at either 4°C or Ϫ20°C for several weeks. Interestingly, GMD activity decreased about 3-fold if NADP ϩ was omitted from the reaction; NAD ϩ was less effective in stimulating the enzyme activity. This finding suggests that, unlike

FIG. 7. Reverse phase HPLC of GMD and GMER products in extracts prepared from virus-infected cells.
Peak identity is the same as reported in Fig. 2. A, 14 C-labeled GDP-D-mannose (100 M) was incubated with 1 mM NADP ϩ and an algal extract (5 mg/ml of protein) obtained at 180 min p.i. After incubating for 30 min, the substrate was quantitatively converted to GDP-4-keto-6-deoxy-D-mannose (peak 2). B, human GMD (18), NADP ϩ is loosely bound to the virus enzyme. GMD reductase activity (510.3 Ϯ 10.9 mol/h/mg protein) determined on GDP-4-keto-6-deoxy-D-mannose in the presence of NADPH, was about 10-fold higher than the dehydratase activity. The specific activity of GMER epimerase/reductase activity was 17.9 Ϯ 0.7 mol/h/mg protein. Unlike GMD, viral GMER was unstable, exhibiting a significant loss of activity due to precipitation from solution when maintained at both 4°C and Ϫ20°C.
GMD and GMER Expression in Vivo-Expression of PBCV-1 GMD and GMER in virus-infected Chlorella cells was investigated by Northern and Western analyses. Both GMD (A118R) and GMER (A295L) mRNAs were expressed starting at 45 min p.i. and reached a maximum at 60 min p.i (Fig. 6A). The size of the GMD mRNA (ϳ1.4 kb) is about the expected size for the 1038-nucleotide-long A118R ORF. In contrast, the size of GMER mRNA (ϳ2.2 kb) is about twice the size expected for the 954-nucleotide-long ORF A295L. However, complex mRNA expression patterns are often observed for other PBCV-1 genes (25). Because PBCV-1 DNA synthesis begins at 60 -90 min p.i (26), both genes are expressed as early genes. GMD and GMER proteins of the appropriate size first appeared 90 min p.i. and peaked at 180 -210 min p.i. (Fig. 6B).
Cell extracts were assayed for GMD and GMER activities at various times after PBCV-1 infection. GDP-D-mannose was completely converted to the intermediate compound GDP-4keto-6-deoxy-D-mannose when the reaction contained NADP ϩ and no NADPH (Fig. 7A). However, both GDP-D-rhamnose and GDP-L-fucose appeared when NADPH was added to the reaction (Fig. 7B); these sugars result from GMD reductase activity and GMER activity, respectively. The identity of all 14 C-labeled compounds was verified by TLC analysis. The relatively higher amount of GDP-D-rhamnose compared with GDP-L-fucose (Fig.  7B) probably results from the intrinsic instability of GMER as noted above for the recombinant GMER. In fact, GMER activity could only be detected immediately after disrupting the cells, whereas its activity disappeared after incubating the cell extracts at 4°C for a few hours or after freezing at Ϫ20°C. In contrast, GMD was stable under both conditions. Neither GDP-L-fucose nor GDP-D-rhamnose appeared when GDP-mannose or GDP-4-keto-6-deoxy-D-mannose was incubated with extracts from uninfected cells, either in the presence or absence of NADPH. Thus, the enzymatic activities in extracts from the infected cells proved to result from the virus-encoded enzymes and not from host-encoded enzymes. Degradation of GDP-D-mannose to the monosaccharide form began only after 3 h of incubation (ϳ10%) in extracts from healthy cells. The formation of GDP-D-rhamnose and of GDP-L-fucose using cell extracts at various times after virus infection (Table I) correlates with the appearance of GMD and GMER (Fig. 6B). DISCUSSION The results presented here demonstrate that Chlorella virus PBCV-1 encodes the genes for two functionally active enzymes involved in GDP-L-fucose biosynthesis. Enzyme activities were detected with recombinant proteins as well as in extracts from virus-infected cells, i.e. both proteins are expressed and function during viral replication in the host. PBCV-1 GMD catalyzes the dehydration of GDP-D-mannose at the C-4 and -6 positions to form the intermediate metabolite GDP-4-keto-6deoxy-D-mannose (Fig. 8). The virus GMER behaves as a bifunctional enzyme, with an epimerization activity that leads to a change in the sugar configuration, followed by the NADPHdependent reduction of the 4-keto group of the intermediate to form GDP-L-fucose. These GMD and GMER activities are identical to those of GMD and GMER enzymes from prokaryotes, plants, and animals (11)(12)(13)(14)(15).
However, the virus GMD has an additional unexpected NADPH-dependent, stereo-specific reductase activity on the 4-keto group of the intermediate. This activity creates another nucleotide sugar, GDP-D-rhamnose (Fig. 8). This novel activity indicates that PBCV-1 GMD is a bifunctional enzyme. GMDs from other organisms lack this reductase activity, with the exception of a GMD from Aneurinibacillus thermoaerophilus (21). However, in this latter case the reductase activity only occurred with recombinant His-or glutathione S-transferasetagged GMD and only small amounts of GDP-D-rhamnose appeared. Accordingly, the reductase activity observed with the A. thermoaerophilus enzyme may not occur in vivo. In fact, the bacterium has another enzyme, GDP-deoxy-D-lyxo-4-hexulose reductase (RMD), for GDP-D-rhamnose production (21). In contrast, GDP-D-rhamnose is formed in vivo after PBCV-1 infection, indicating that GMD reductase activity occurs with wild type enzyme and the activity is not an artifact of the recombinant enzyme (e.g. because of incorrect protein folding). The possibility that GDP-D-rhamnose arose in vivo by another enzyme activity encoded by the virus is excluded because no recognizable RMD homologs exist in the PBCV-1 genome. Moreover, the high reductase activity observed with both wild type and recombinant GMD, as well as its kinetic properties 2 suggest that GDP-D-rhamnose production by GMD is physiologically relevant.
Because both fucose and rhamnose are components of the glycan portion of PBCV-1 major capsid protein Vp54 (4), the viral GMD and GMER enzymes may be necessary to provide the virus with the appropriate nucleotide sugars. Indeed GDP-D-mannose metabolizing activities leading to GDP-Drhamnose and GDP-L-fucose were not detected in extracts from uninfected cells. However, rhamnose and trace amounts of fucose are among the seven neutral sugars in uninfected 2 M. Tonetti, L. Sturla, J. L. Van Etten, manuscript in preparation.
FIG. 8. Scheme of the biosynthesis of GDP-L-fucose and GDP-D-rhamnose. PBCV-1 GMD catalyzes both the dehydration of GDP-D-mannose to the intermediate GDP-4-keto-6-deoxy-D-mannose and the NADPH-dependent reduction of this latter compound to GDP-Drhamnose. NADP ϩ serves as cofactor for GMD during the internal oxidoreduction reaction involved in the dehydration process. The epimerization and the NADPHdependent reduction of the 4-keto group leading to GDP-L-fucose are carried out by PBCV-1 GMER.
Chlorella NC64A cell walls (27). Presumably the host is unable to supply enough of these two monosaccharides for virus production.
To summarize, the results in this report have three unique aspects. (i) This is the first report of a virus encoding enzymes involved in the synthesis of nucleotide sugars. (ii) Unlike GMD enzymes from other organisms, the virus-encoded GMD is a bifunctional enzyme. (iii) Unlike GMD enzymes from other organisms, the virus recombinant GMD is very stable.