Molecular Cloning and Characterization of UDP-glucose Dehydrogenase from the Amphibian Xenopus laevis and Its Involvement in Hyaluronan Synthesis*

UDP-glucose dehydrogenase (UGDH) supplies the cell with UDP-glucuronic acid (UDP-GlcUA), a precursor of glycosaminoglycan and proteoglycan synthesis. Here we reported the cloning and the characterization of the UGDH from the amphibian Xenopus laevis that is one of the model organisms for developmental biology. We found that X. laevis UGDH (xUGDH) maintained a very high degree of similarity with other known UGDH sequences both at the genomic and the protein levels. Also its kinetic parameters are similar to those of UGDH from other species. During X. laevis development, UDGH is always expressed but clearly increases its mRNA levels at the tail bud stage (i.e. 30 h post-fertilization). This result fits well with our previous observation that hyaluronan, a glycosaminoglycan that is synthesized using UDP-GlcUA and UDP-N-acetylglucosamine, is abundantly detected at this developmental stage. The expression of UGDH was found to be related to hyaluronan synthesis. In human smooth muscle cells the overexpression of xUGDH or endogenous abrogation of UGDH modulated hyaluronan synthesis specifically. Our findings were confirmed by in vivo experiments where the silencing of xUGDH in X. laevis embryos decreased glycosaminoglycan synthesis causing severe embryonic malformations because of a defective gastrulation process.

the initial sugar in GAG synthesis on proteoglycans. In many strains of pathogenic bacteria, such as group A streptoccoci and Streptococcus pneumoniae type 3, UDP-GlcUA is used in the construction of the antiphagocytic capsular polysaccharide (5,6). UGDH is of biochemical mechanistic interest because it belongs to a family of sugar nucleotidemodifying enzymes that catalyze a net four-electron oxidation and serve as both alcohol dehydrogenases and aldehyde dehydrogenases (7).
The importance of UGDH is remarkable considering that its product, UDP-GlcUA, is critical for GAG synthesis. GAG chains of proteoglycans and HA are ubiquitous components of extracellular matrices and pericellular spaces, and an increasing body of information shows the role of GAGs in cell behavior, including signal transduction, cell proliferation, spreading, migration, cancer growth, and metastasis (8 -10).
Proteoglycans and HA have important roles throughout the development (11). As UGDH synthesizes UDP-GlcUA, one of the main UDPsugar precursors, it is not surprising that alteration in UGDH expression causes evident phenotypes in developing embryos, as found by different authors in different model organisms. In particular, in Drosophila melanogaster UGDH is encoded by the sugarless gene and is required for heparan sulfate modification of proteins that control wing formation (12). In Caenorhabditis elegans, UGDH influences GAG synthesis, which is essential for vulval morphogenesis and embryonic development (13,14). In Zebrafish, the enzyme is critical for normal cardiac development (15). In mouse, UGDH mutants arrest growth during gastrulation with defects in migration of mesoderm and endoderm (16). Moreover, a similar phenotype was found in mutants in the fibroblast growth factor pathway, highlighting that proteoglycans and GAGs facilitate signaling by mammalian growth factors. Another well known model organism extensively used in developmental biology is the amphibian Xenopus laevis. Although several papers report the critical role of proteoglycans in X. laevis development (17)(18)(19)(20)(21)(22), very limited information is available on GAG functions in amphibian embryogenesis (23)(24)(25). Therefore, we have characterized the X. laevis UGDH (xUGDH), the key enzyme in GAG biosynthesis in this model organism, as suitable for developmental studies. Here we report the cDNA cloning of xUGDH, its biochemical kinetic parameters, its expression in developing embryos, and its genome organization. We have also extended our observations to Xenopus tropicalis, an organism of the same genus as X. laevis, but with interesting potential for genetic studies (26). Our results show that the xUGDH expression in human cells increases their HA accumulation (and not that of other GAGs), whereas UGDH abrogation by siRNA reduced HA synthesis and cell-associated GAGs. Moreover, we have also demonstrated that * This work was supported by MIUR COFIN Project "Analysis of the Hyaluronan-Proteoglycan-Hyaluronan Receptor System in Embryonic Development and Disease" (to D. V.) and Consorzio Interuniversitario Biotecnologie (to A. P.). 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. The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AY762616. 1 To whom correspondence should be addressed: Dip UGDH has a crucial role during X. laevis development because xUGDH loss of function, by morpholino injection, causes lethal malformation.

MATERIALS AND METHODS
xUGDH Cloning-An X. laevis EST data bank was searched for homologies with human UGDH cDNA (GenBank TM accession number NM_003359) using the World Wide Web-based BLAST search engine (www.ncbi.nlm.nih.gov/blast/). IMAGE EST number 4202110 was found and was obtained from MRC geneservice (Cambridge, UK). DNA sequencing was done by an external service facility (BMR, Padua, Italy).
Characterization of the xUGDH Gene-Introns of the X. laevis xUGDH gene were amplified with a pair of primers (Table 1) designed on the basis of mouse and human UGDH genomic sequences deposited in public data bases. PCR parameters were as follows: denaturation at 94°C for 30 s, annealing for 30 s at the temperature indicated in Table 1, and elongation at 72°C for 5 min for 35 cycles using 1.5 units of La-TAQ (a proofreading DNA polymerase from Takara) following the manufacturer's conditions. Amplification of intron 8 was done using Phusion TM polymerase (Finnzymes) in 6% Me 2 SO. PCRs were done on genomic DNA. Amplified intron lengths were determined by gel electrophoresis, and exon/intron boundaries were determined by sequencing the extracted bands from the gels.
In silico determinations of the intron/exon structure of the human, mouse, and X. tropicalis UGDH genes were done by BLAST searches on the genomic data bases. Sequence manipulations and contig constructions were done using Vector NTI Suite 6 software.
Recombinant xUGDH Expression and Purification-Cloning and transformation techniques were done essentially as described by Sambrook et al. (27). A histidine-tagged xUGDH construct was generated by PCR. Briefly, the xUGDH ORF was amplified using Phusion DNA polymerase (Finnzymes) with the following primers: CCAAGGCTC-GAGATGTTTCAGATTAAGAAGATTT and CCTTGGCTCGAGT-TAAACTCTTTGTTTCTTATGAGGC (the sequences corresponding to XhoI recognition sites are underlined). The amplified product was digested with XhoI (Takara) and cloned into the XhoI-linearized pET-19b (Novagen) plasmid. This construct (pHisxUGDH) was then sequenced; it encodes the full-length xUGDH protein with an additional His 6 tag sequence at the N terminus.
For protein expression, the plasmid pHisxUGDH was transferred to the host BL21(DE3)pLysS E. coli strain (Promega). E. coli cells carrying the recombinant plasmid were cultivated at 37°C in LB medium containing ampicillin and chloramphenicol (100 and 34 g/ml final concentrations, respectively). After an overnight growth (A 600 nm Ͼ2.5), IPTG was added at a final concentration of 1 mM. The temperature was reduced to 30°C, and the cells were collected after 24 h. Cell extraction and His-tagged xUGDH purification were done using the B-Ter purification kit (Pierce) following the manufacturer's instructions. Some determinations were done on crude extracts that were prepared lysing the bacterial cells in B-Ter solution (Pierce) and recovering the supernatant after centrifugation.
xUGDH Assay-The enzymatic activity of purified His-tagged xUGDH was determined by monitoring the change in absorbance at 340 nm that accompanies reduction of NAD ϩ to NADH. The assay conditions were the same as those described previously by Sommer et al. (28).
Gene Expression Studies-Total RNA from 21 X. laevis embryos (7 from three different females) at different stages of development were extracted using Trizol reagent (Invitrogen) following the manufacturer's protocol. One g of total RNA from each extract was retrotranscribed using 200 units of Moloney murine leukemia virus reverse tran-scriptase (Invitrogen) and a (dT) 16 primer at a concentration of 500 g/ml. The reaction was done in 50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgCl 2 , 10 mM dithiothreitol, and 500 M of a dNTP mixture at 42°C for 50 min. PCR amplifications of the cDNA samples were done using the first strand cDNA synthesis mixture, 25 pmol of primer xUGDH upper primer (CCCTTTGTGAGGCTACAGGA) and xUGDH lower primer (CGGTGCAGATAACCATAGCA), 200 M dNTPs, and 1 unit of RedTaq polymerase (Sigma) in its own buffer. Reaction mixtures were subjected to eight touchdown cycles with annealing temperatures from 58 to 54°C and subsequently with cycles using the following parameters: denaturation at 94°C for 30 s, annealing at 53°C for 30 s, and elongation at 72°C for 40 s. As a control for genomic DNA contamination, all reactions were established with the control sample lacking reverse transcriptase. Normalization was done detecting cytoskeletal actin (upper primer, CTGAGTTCATGAAGGATCAC; lower primer, AAATTTACAGGTGTACCTGC) (29) with the above described conditions.
Quantitative real time RT-PCR was used in transfection experiments with aortic smooth muscle cells (AoSMC) (see below). Forty eight hours after transfections, total RNA samples were extracted with Trizol (Invitrogen) and retrotranscribed using the High Capacity cDNA synthesis kit (Applied Biosystems) for 2 h at 37°C. Quantitative RT-PCR was performed on an ABi Prism 7000 instrument (Applied Biosystems) using the Taqman Universal PCR Master Mix (Applied Biosystems) following the manufacturer's instructions. Probe and primers were developed from TaqMan gene expression assay reagents (Applied Biosystems). The following human TaqMan gene expression assays were used: HAS1 (Hs00155410_m1), HAS2 (Hs00193435_m1), HAS3 (Hs00193436_m1), UGDH (Hs00163365_m1), and ␤-actin (Hs99999993_m1). Fluorescent signals generated during PCR amplifications were monitored and analyzed with ABi Prism 7000 SDS software (Applied Biosystems). Comparisons of the amounts of each gene transcript among different samples were made using ␤-actin as a reference. Standard curves were generated by serial dilution of cDNA, and as calculations of PCR efficiency were very similar (about 90%) for each gene assayed, the relative quantitative evaluation of target gene levels was determined by comparing ⌬Ct (Applied Biosystems user bulletin number 2).
xUGDH Expression in Mammalian Cells-The xUGDH ORF was amplified using Phusion TM DNA polymerase (Finnzymes) with the following primers: TACTCGAGACCATGTTTCAGATTAAGAA-GATTT and ATCTCGAGTTAAACTCTTTGTTTCTTATGAGGC. The boldface sequences correspond to XhoI recognition sites, and the underlined adenosine represents the typical Ϫ3 purine of the Kozak consensus sequence. The amplified product was purified, A-tailed using RedTaq (Sigma), and cloned into a pTarget (Promega) vector. Expression plasmids were selected to have the insert in the sense orientation to synthesize the complete xUGDH protein (pTarget-xUGDH sense) or to have the insert in the opposite direction used as the control vector (pTarget-xUGDH antisense). DNA sequencing and in vitro transcription and translation (TNT in vitro transcription/translation system, Promega) were done to check the constructs. AoSMCs (Cambrex) were grown in SmGm2 complete (supplemented with 5% fetal bovine serum) culture medium (Cambrex). The cultures were maintained in an atmosphere of humidified 95% air, 5% CO 2 at 37°C. 1 ϫ 10 6 cells between passages 2 and 5 were transiently transfected by means of a Nucleofector apparatus (Amaxa) and the human aortic smooth muscle cells Nucleofector TM kit using 5 g of either the sense or antisense expression plasmid. After 48 h, transfected cells and conditioned cell culture media were collected. Cells were lysed in 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100 supplemented with a protease inhibition mixture (Roche Applied Science) using a cell scraper. Protein contents in the cell lysates were determined using the Bradford method, and UGDH activities were measured as described above.
Abrogation of Human UGDH-siRNA was used to reduce human UGDH expression in AoSMCs. UGDH siRNA (5Ј-GGACUAAAA-GAAGUGGUAGGtt-3Ј) and negative control siRNA 1 kit (scramble, code 4611) were purchased from Ambion. Transfections were done using a Nucleofector apparatus (Amaxa) and the human AoSMC Nucleofector TM kit using an siRNA concentration of 50 M of either UGDH siRNA or scramble siRNA. After 48 h of incubation, conditioned cell media were assayed for GAG contents by FACE analysis. The transfected cells were used for determination of UDP-sugars contents and UGDH activities. Quantitative real time RT-PCR was used to verify the reduction of UGDH mRNA expression.
GAG Disaccharides and UDP-sugar Precursor Determinations-HA and chondroitin sulfate disaccharides were determined by FACE analysis as described previously (30). UDP-Glc and UDP-GlcUA were quantified by capillary zone electrophoresis using the methodology outlined by Lehmann et al. (31). Briefly, 10 6 AoSMCs were lysed in 0.1 ml of phosphate-buffered saline containing 0.5% Triton X-100. After centrifugation, the supernatants were deproteinized by acetonitrile treatment, lyophilized, and resuspended in 90 mM borate buffer, pH 9. Analyses were done using a 75 cm ϫ 50 m column, 25-kV voltage, and detecting UDP-sugar absorbances at 262 nm. The peak identity was assessed by co-injecting the extracts with commercial UDP-Glc and UDP-GlcUA.
GAG Determinations-To evaluate the amount of each GAG after induction or abrogation of UGDH, AoSMCs were transfected with pTarget-xUGDH constructs or siRNA, respectively, and labeled with [ 3 H]glucosamine (25 Ci/ml) for 48 h in complete SmGm2 medium. Conditioned cell medium was recovered, and the unincorporated [ 3 H]glucosamine was removed by gel filtration using a PD10 column (Amersham Biosciences). The cell layer was washed with phosphatebuffered saline and scraped into 100 l of phosphate-buffered saline containing 1% Triton X-100, centrifuged at 10,000 ϫ g for 10 min at 4°C. The clear supernatant and the PD10 filtrated medium were digested with proteinase K (Finnzymes) at 50°C. After 8 h of incubation, proteinase K was inactivated at 90°C for 10 min. Each GAG solution was separated into 4 aliquots, precipitated by adding 4 volumes of ethanol at Ϫ20°C for 18 h, and recovered by centrifugation. Two aliquots were resuspended in 0.1 M ammonium acetate, pH 7, and digested with 100 milliunits/ml of hyaluronidase SD from Streptococcus dysgalactiae (Seikagaku) and 100 milliunits/ml of chondroitinase ABC (Seikagaku), respectively; one was resuspended in 10 mM Tris-HCl, pH 7, 4 mM CaCl 2 and digested with 100 milliunits/ml of heparanase I, II, and III (Seikagaku). The last aliquot was used as an undigested control. After 16 h of incubation at 37°C, which allow complete degradation of 20 g/ml of standard GAGs, 50 g of chondroitin sulfate A (Seikagaku) was added to each digested sample as carrier, and undigested GAGs were precipitated with ethanol. Specific GAGs were quantified by counting the radioactivity associated with digestion products (in the supernatant) and the undigested GAGs (pellet) with a liquid scintillation counter (Canberra Packard).
xUGDH Silencing during X. laevis Development-Two antisense morpholino-oligonucleotides (MOs; Gene Tools LLC) were generated on the basis of the xUGDH cDNA sequence, xUGDH MO CATGGT-TTATCTTGCTGAGAACAGA, which is complementary to the xUGDH translation start site and the adjacent coding sequence. A 5-mismatch MO, based on the xUGDH MO sequence, was used as a specificity control (CATcGTTTATgTTGgTGAcAAgAGA, where lowercase letters indicate mismatched nucleotides). The optimal morpholino concentration to inject per embryo was established by independent pilot experiments (not shown) and was determined to be in the range of 10 -20 ng/embryo. The MOs were injected bilaterally at the two-cell stage to down-regulate the xUGDH activity in the whole embryo. All the experiments were performed by co-injecting the morpholino-oligonucleotide and the GFP mRNA (300 pg/embryo) as a tracer in order to select only the properly manipulated embryos for the subsequent analysis. Embryos were injected in 0.1ϫ Modified Marc's Ringer solution (MMR) (100 mM NaCl, 1.8 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 5 mM HEPES, pH 7.8) supplemented with 4% Ficoll and then cultured in 0.1ϫ MMR solution until the stage of analysis.

RESULTS AND DISCUSSION
xUGDH cDNA Cloning-In this study we focused our attention on UGDH, a pivotal enzyme in the production of UDP-GlcUA, an important precursor for GAG synthesis. It has been reported that UGDH is crucial in C. elegans and D. melanogaster morphogenesis (12,13) and Zebrafish cardiac valve formation (15) and is also involved in mouse gastrulation (16). We reported previously the characterization of HA synthases (HAS1, HAS2, and HAS3) in X. laevis and showed their regulation during the early stages of development (34,35). It is also possible that UGDH, which provides one of the substrates for HAS, is regulated in the embryogenesis of this species. Therefore, we decided to clone X. laevis UGDH and characterize it.
To obtain the sequence of the xUGDH mRNA, we initially performed a BLAST search using the human UGDH sequence against an X. laevis EST data base. This revealed several ESTs with a high degree of identity with the human sequences. We then obtained the largest EST (dbEST code 4202110) from MRC geneservice (Cambridge, UK); we completed the sequence and deposited in the public data base with GenBank TM accession number AY762616 (data not shown). This sequence has a 42-bp-long 5Ј-UTR and a 1482-bp-long ORF coding for a protein with 494 amino acids, a calculated molecular mass of 55,209.4 Da, and a calculated isoelectric point of 6.49. A PROSITE analysis (www.expasy.org/ prosite/) and a NetPhos analysis (www.cbs.dtu.dk/services/NetPhos/) of the xUGDH amino acid sequence revealed several putative phosphorylation sites. Although we did not further investigate the possibility of posttranslation modifications in this enzyme, previous studies have reported that UGDH may be phosphorylated in prokaryotic cells (36,37).
We searched for known signal peptides using the PSORT II program (www.psort.org/) without identifying any sorting sequence. xUGDH localization was predicted to be cytoplasmic with a 65.2% probability. Such a localization is consistent with a previous report that placed UGDH in the cytosol (13).
The deduced amino acid sequences of xUGDH and other known UGDH from several species are aligned (data not shown). The N-terminal region has the consensus sequence for NAD binding (i.e. GXGXXG) that is maintained with a high degree of identity among all the UGDH analyzed. The central region of the protein (i.e. GFGGSCFQKDVLN) has been proposed to be the catalytic domain (38), and all the aligned sequences show a high degree of identity with only a minor amino acid substitution in the UGDH belonging to C. elegans, D. melanogaster, and soybean. Moreover, the critical cysteine residue, number 276 of xUGDH, is conserved in all sequences, and it has been shown to be involved during the second half-reaction of UGDH (i.e. conversion of UDP-aldehydroglucose to UDP-GlcUA) (28). With respect to function/structure considerations, proline residues at positions 92 and 160 are believed to represent main bends in the protein structure (39 -41), and they are conserved for UGDH from all species. Two lysine residues at positions 220 and 339 are also conserved, which correspond to lysine 219 and 338 of bovine UGDH. One of these lysine residues is probably catalytically involved in the first half-reaction of the enzyme (i.e. conversion of UDP-Glc to UDP-aldehydoglucose) (39,42). Moreover, using Conseq analysis (freely available at the Internet site conseq.bioinfo.tau.ac.il/) (43), regions spanning from amino acids 41-54, 128 -135, 340 -350, and 461-464 have been identified as important for the structure or function of the protein (results not shown). Although this analysis is only a virtual prediction, it was done on all the known UGDH sequences deposited in the data bases. Such regions could be useful to understand not only the catalytic mechanism of the reaction but also to identify critical residues involved in a hypothetical regulatory mechanism that a pivotal enzyme such as UGDH could possess. Recently, for example, Conseq analysis has been successfully used to isolate important regions for protein-protein interactions (44).
Genomic Organization of xUGDH-By comparing the xUGDH transcript to the human and mouse genomic sequences available in public data bases, we have inferred the positions of the introns of the xUGDH gene. For each hypothetical intron, we have designed, on the two flanking exons, a pair of specific primers (Table 1) to amplify by long amplification PCR a region expected to contain the intron. PCR products were separated by gel electrophoresis, and the exon-intron junctions were sequenced. As reported in Table 2, top and bottom portions, the structure of the xUGDH gene is composed of 11 exons and 10 introns, and it matches that of mammalian UGDH genes (45). Notably, in the 5Ј-untranslated region of the human and mouse UGDH mRNA, we and others (45) found an additional intron of about 5700 bp. As this intron is located far upstream from the ATG start codon, we did not further investigate its position in the xUGDH gene. The major variability was detected in intron lengths, although all of them conserve their intron phase, indicating that the intron insertion points in the cDNA sequence have remained constant during xUGDH gene evolution. Most interestingly, the consensus gt-ag (5Ј-3Ј) splice site sequences are not always conserved in the xUGDH gene as shown in the 5Ј splice acceptor sequence of intron 5. However, this noncanonical splice site has been reported to be quite frequent and does not cause any deleterious mutations (46,47). Moreover, this noncanonical splice site may be the main 5Ј intron acceptor site used during embryogenesis. In fact, no alternative splice isoforms of xUGDH transcripts have been found during X. laevis development by RT-PCR using two primers (i.e. xUGDH upper primer and xUGDH lower primer) flanking the intron 5 insertion point in the cDNA sequence.
X. laevis is certainly a well recognized "model organism" for cell and developmental biology, but it has the disadvantage of being tetraploid, which greatly complicates the creation of mutants and the analysis of gene regulation. In contrast, X. tropicalis, an amphibian of the same genus, is diploid and possesses a relatively small genome, whose sequencing is currently in progress. Therefore, X. tropicalis is destined to complement X. laevis as a model organism (26). Thanks to available preliminary data on the X. tropicalis genome (www.genome.jgipsf.org), we have determined the X. tropicalis UGDH genomic organization ( Table 2, bottom portion), and as expected, the two genes are indeed very similar.
Recombinant xUGDH Expression and Characterization-To study the biochemical properties of xUGDH, we have generated the pHisx-UGDH plasmid that codes for a His 6 -tagged xUGDH. This plasmid was used to transform Escherichia coli BL21(DE3)pLysS cells, and the recombinant protein expression was induced by IPTG treatment. Experiments were done to optimize the expression conditions, analyzing the effect of temperature, time of induction and collection, and IPTG concentration on xUGDH expression. The best conditions were obtained using E. coli cells grown overnight at 37°C, induced with 1 mM IPTG, cultured at 30°C, and harvested 24 h after induction. The recombinant xUGDH overexpressed under these conditions was completely soluble and thus fully recovered in the crude extract. The specific UGDH activity in the crude extract of induced E. coli BL21(DE3)pLysS was about 50-fold higher than that of crude extracts of uninduced cells (data not shown).
The crude extract from IPTG-induced E. coli cells was affinity-purified using a commercial spin column kit (Pierce). After the three elution steps suggested by the manufacturer, recombinant His-tagged xUGDH was eluted as a nearly single resolved band, and the final preparation was at least 90% homogeneous as judged by SDS-PAGE analysis (result not shown). The 54-kDa band corresponding to recombinant His-tagged xUGDH is in agreement with the molecular weight theoretically calculated from the translated xUGDH cDNA sequence. After the affinity chromatography step, the specific activity increased about 12-fold (data not shown).
As UGDH reduces two molecules of NAD to NADH during the conversion of a UDP-Glc substrate to a UDP-GlcUA, we followed the reaction by measuring NADH absorbance at 340 nm. To obtain kinetic constants (K m and V max ) for substrate and cofactor, we used the same experimental buffers and conditions described previously (28). Moreover, we did not remove the His 6 -tagged sequence as it has been reported that it does not alter the property of the enzyme (28). The purified enzyme showed no reduction of NAD if incubated with either UDP-galactose, UDP-GlcUA, or UDP-N-acetylglucosamine, indicating the purity and specificity of the enzyme (results not shown). The steady-state kinetic parameters of His-tagged xUGDH have been calculated by incubating the enzyme with increasing concentrations of NAD in the presence of saturating UDP-Glc substrate (Fig. 1A). Similarly, the dependence of reaction kinetics on the substrate was measured by increasing UDP-Glc concentration in the presence of saturating NAD (Fig. 1C). For both the substrates, saturation kinetics were observed. A double-reciprocal plot of the initial velocities revealed a K m of 0.9 mM for UDP-Glc (Fig. 1B) and a K m of 0.3 mM for NAD (Fig. 1D). Both sets of conditions yielded a similar V max of about 10 nmol of NADH/min/mg of enzyme. These data are comparable with those of UGDH from other species (Table 3). xUGDH Expression during Development-The amphibian X. laevis has become one of the most studied organisms in developmental biology together with C. elegans and D. melanogaster. As no information was available, we investigated the regulation of xUGDH during X. laevis development. RT-PCR was done on RNA extracted from the same number of embryos at different developmental stages. To visualize differences in the expression, different PCR cycles were done as outlined in Fig. 2. Although the 531-bp-long band corresponding to the amplicon of xUGDH was detectable in all the tested developmental stages, a more pronounced signal was visualized in the sample from 30 h post-fertilization (hpf) that corresponds to an embryo at the tail bud stage. Such an expression pattern is related to the crucial function of UDP-GlcUA. In fact, this molecule is a precursor for all GAGs (with the exception of keratan sulfate), including GAGs of proteoglycans and HA that are known to be essential for a proper development (11). Most interestingly, the increase in xUGDH in the 30-hpf embryos corresponded with our previous finding that HA is dramatically elevated in embryos at this developmental stage (34). Moreover, the enzymes involved in HA synthesis (i.e. HAS2 and HAS3) are also up-regulated in this embryonic stage (34). Although the precise role of HA in X. laevis development is still to be defined, HA is critical during embryonic development of mammals. For example, mice unable to produce HA die in the uterus from severe cardiac malformation (49,50).
xUGDH Expression and HA Synthesis-The data reported in this study strongly support the hypothesis that coordinated expression of UGDH and HAS may have a pivotal role in regulating HA levels. To elucidate the relationship between UGDH and HA synthesis and accumulation, we transfected human primary aortic smooth muscle cells (AoSMCs) with an xUGDH expression plasmid pTarget-xUGDHsense. We also transfected the plasmid pTarget-xUGDH-antisense that did not code for any protein as a control. We hypothesized that an increase in UGDH activity that augments UDP-GlcUA synthesis may increase the synthesis of HA. As shown in Fig. 3A, UGDH activity increased in AoSMCs transfected with pTarget-xUGDH-sense,   whereas cells transfected with pTarget-xUGDH-antisense maintained a lower UGDH activity comparable with that of untransfected cells. The increment of UGDH activity of pTarget-xUGDH-sense transfected cells was in agreement with that reported for other mammalian cell lines transfected with mouse or bovine UGDH (42,51). To verify that a UGDH activity increase could actually increase UDP-GlcUA, we quantified the UDP-GlcUA:UDP-Glc ratio by capillary zone electrophoresis (Fig. 4). AoSMCs transfected with pTarget-xUGDH-sense, pTarget-xUGDH-antisense vectors, or not transfected showed UDP-GlcUA: UDP-glucose ratios of 4, 0.8, and 0.7, respectively. This clearly indicates that the increase of UGDH activity augmented the cellular UDP-GlcUA levels by about 4-fold with respect to control levels.
The conditioned culture media of the AoSMC cultures were collected after 48 h of transfection, and their HA concentrations were measured by FACE analysis (Fig. 3B). AoSMCs produce high levels of HA (52,53). Most interestingly, the HA disaccharide band derived from 10 6 pTarget-xUGDH-sense transfected AoSMCs was significantly stronger than that derived from the same number of control cells (i.e. transfected with pTarget-xUGDH-antisense or not transfected). The UGDH activity detected in control cells (Fig. 3A) could explain the basal HA signals in control cell culture medium (Fig. 3B). On the other hand, the increase of HA production in pTarget-xUGDH-sense transfected cells could be ascribed to the augmented UGDH activity in those cells. A similar result was obtained by Magee et al. (54) who discovered that the up-regulation of UDP-glucose pyrophosphorylase stimulates HA synthesis in hypertrophic chondrocyte cultures. As UDP-glucose pyrophosphorylase and UGDH belong to the same pathway that forms UDP-GlcUA, it would be reasonable that the up-regulation of both of these enzymes can increase the concentration of HA synthesis precursors. Moreover, the K m values of the three HA synthases are in the range of the cytoplasmic concentration of UDP-sugars (55), making the activity of HA synthetic enzymes responsive to the concentrations and the pool sizes of the cellular sugar nucleotides.
Most interestingly, densitometric quantification of the HA and the chondroitin 4-sulfate disaccharide bands obtained from FACE analyses revealed that chondroitin 4-sulfate synthesis did not change significantly after the xUGDH transfections (results not shown), indicating  that the UGDH effect was selective for HA. Although UGDH is a central enzyme in the metabolism of GAGs, it may specifically regulate HA accumulation without altering other GAGs. In light of this specific HA increase, we performed quantitative RT-PCR analyses on transfected AoSMC cDNA and found that HAS2 and HAS3 mRNA levels were up-regulated about 2-fold only in pTarget-xUGDH-sense transfected cells, whereas HAS1 mRNA levels were unchanged (Fig. 3C). This suggests that transfection of UGDH leads to an induction of HAS2 and HAS3 gene transcription or to a stabilization of HAS2 and HAS3 mRNAs. The specific increase of HAS2 and HAS3 fits well with the specific increase of HA in transfected AoSMCs (Fig. 3B).
The lack of increase in chondroitin sulfate may result from the fact that precursors for this GAG must enter the Golgi apparatus by using several UDP-sugar transporters that have been shown to possess low K m values for their substrates (i.e. 1-10 M) (56,57). Although cytoplasmic UDP-sugar precursor concentrations could directly affect the activity of HA synthases located in the cell membrane, the UDP-sugar transporters would be saturated at the cellular concentration of the UDP-sugar precursors thereby maintaining constant concentrations inside the Golgi apparatus and thus the constant activity of the Golgi GAG synthetic enzymes. The tight relationship between UGDH activity and GAG synthesis was also recently outlined in in vivo experiments with fibroblast-like synovial lining cells (58) and in vitro on immature and mature human articular cartilage explants (59), although the key importance of UGDH activity in GAG production was already suggested years ago (60,61).
Abrogation of Human UGDH and HA Synthesis-To demonstrate the putative link between UGDH activity and HA synthesis, we did siRNA experiments to reduce the endogenous UGDH in human AoSMCs. The greatest reduction of human UGDH mRNA was achieved by transfecting 50 M of UGDH siRNA and incubating for 48 h. In these conditions quantitative RT-PCR showed a reduction of the human UGDH transcript of about 90% with respect to the controls (i.e. scramble siRNA and no DNA transfected cells) (Fig. 5A). Such a reduction of UGDH mRNA corresponded to a strong inhibition of UGDH activity. In fact, we were not able to detect any signals in the UGDH spectrophotometric assay used in this work indicating that the residual UGDH activity was below the detection limit of the assay (result not shown). To verify that the UGDH activity decrease could actually reduce UDP-GlcUA content, we quantified the UDP-GlcUA:UDP-Glc ratios (data not shown). AoSMCs transfected with UGDH siRNA, scramble siRNA, or not transfected showed UDP-GlcUA:UDP-glucose ratios of 0.25, 1.0, and 0.8, respectively, clearly indicating that the inhibition of UGDH activity reduced the cellular UDP-GlcUA levels by 4-fold with respect to control levels. Furthermore, the concentrations of HA in the conditioned culture media clearly showed a corresponding decrease (Fig. 5B) indicating that HA synthesis could be strictly regulated by the availability of the UDP-sugar precursors at least in AoSMCs. This finding supports the proposed mechanism of the 4-methylumbelliferone inhibition of HA synthesis; in fact, the glucuronidation of 4-methylumbelliferone by endogenous UDP-glucuronyltransferase induces a depletion of UDP-GlcUA (62). Moreover, chondroitin 4-sulfate was not influenced by changing availability of the precursors (see above).
Most interestingly, the inhibition of UGDH by siRNA did not change the levels of transcripts coding for HAS1, HAS2, and HAS3 quantified by quantitative RT-PCR (Fig. 5C). These data indicated that the expression of these enzymes cannot be reduced below a basal level even when one of their substrates is strongly decreased.
GAG Determinations-To better elucidate the UGDH role in the control of GAG synthesis, we incubated AoSMCs with [ 3 H]glucosamine to determine the GAG amount. We found that the total (i.e. medium and cell layer) incorporated radioactivity associated with GAGs was much higher in pTarget-xUGDH-sense transfected cells than in the controls (i.e. 0.16 versus 0.12 cpm/cell, respectively). On the other hand, the incorporated radioactivity associated with GAGs was lower in siRNA against UGDH-treated cells than in the controls (0.06 versus 0.12 cpm/cell, respectively). The silencing of UGDH affected dramatically the GAG associated to the cell layer fractions in which the radioactivity incorporated in the GAG was undetectable. These data support the critical role of UGDH in promoting GAG synthesis.
Moreover, we quantified the percentage of the specific GAG family in conditioned medium (Table 4) as well as the cell layer (Table 5) samples. The determinations were performed counting the radioactivity in GAG resistant to hyaluronidase SD, chondroitinase ABC, and heparinase I, II, and III digestions as described under "Materials and Methods." For the percentage calculations, the total radioactivity that was digested was considered 100%. The quantifications were done on untransfected AoSMCs (Tables 4 and 5, control row) or transfected with pTarget-xUGDH-sense (Tables 4 and 5, sense row) or with siRNA against human UGDH (Tables 4 and 5, RNA UGDH row). We found that HA is the most produced GAG in the medium as well as in the cell layer. Most interestingly, when we overexpressed the xUGDH, we found a significant increase in both medium and cell-associated HA content, whereas the synthesis of other GAGs was only slightly modified (Tables 4 and 5). On the other hand, when we abolished the expression of UGDH in AoSMC by siRNA, we found a decrease of medium HA without altering the synthesis of other released GAGs. However, in the cell layer-associated GAGs, the incorporated radioactivity was under the detection limit (Tables 4 and 5), indicating that cell surface-associated GAGs could have a more rapid turnover in comparison to GAGs secreted into the medium. Most interestingly, as the sum of percentages indicated in the table rows represents the GAG-associated radioactivity, it is noteworthy that in the medium fractions almost 100% of the radioactivity was GAG-associated, and in the cell fractions the radioactivity accounted for about 80%. This value reached about 100% when UGDH was induced, and this increase (ϳ20%) was specifically caused by HA synthesis.
As observed from our previous results (Figs. 3B and 5B), we showed that HA is the GAG that is more sensitive to UDP-GlcUA concentration inside the cytoplasm, and therefore, the cell could regulate specifically the HA synthesis controlling the UDP-GlcUA availability.
In Vivo xUGDH Silencing-X. laevis embryos were injected bilaterally with xUGDH MO complementary to the 5Ј-UTR of xUGDH mRNA to prevent its translation. Control and xUGDH MO-treated embryos were co-injected with mRNA coding for the green fluorescent protein (GFP), allowing the selection of only those embryos injected in both sides, before the subsequent analysis. Embryos treated with xUGDH MO had severe malformations at the end of gastrulation that led to a failure in the blastopore closure. In the subsequent developmental phases, xUGDH MO-injected embryos are not able to complete the neurulation step correctly as the neural tube remained posteriorly opened (90% of injected embryos n ϭ 120) (Fig. 6A). In contrast control embryos (injected with the same dose of the 5-mismatch MO) showed a normal development (100% of injected embryos n ϭ 130) (Fig. 6B). The embryos progressively died and did not reach the tail bud stage. It has been shown that during gastrulation, both in X. laevis and mouse embryos, mesodermal cells enter and move within an HA-rich environ-ment (23,63). Moreover, the HAS2 is expressed in the involuting mesoderm in X. laevis embryos (35) and appears to be critical in the Zebrafish gastrulation process (32). Therefore, our xUGDH knock down functional data are consistent with a role of HA in the gastrulation movements and suggest an important role of the xUGDH activity during early X. laevis development.
In conclusion, we have reported the cloning of the UGDH cDNA sequence from the amphibian X. laevis, and we found that the gene structure and biochemical properties confirmed the data obtained in other species. We described an up-regulation of xUGDH mRNA during the tail bud stage of X. laevis development that was correlated with the elevated HA production typical of the developing embryo at this stage. We also demonstrated a critical role of xUGDH activity in Xenopus early development. Functional down-regulation of xUGDH, in fact, results in early embryonic lethality. Moreover, we reported that an increase in UGDH may be responsible for HA accumulation at least in an in vitro cellular model. These results support the hypothesis that the synthesis of HA requires the activation of a complex mixture of enzymes, not only the synthetic proteins but also the enzymes involved in metabolism of the UDP-sugar precursors.

TABLE 4 Conditioned medium GAG content in AoSMCs
AoSMCs (control) or AoSMCs transfected with either pTarget-xUGDH-sense (sense) or pTarget-xUGDH-antisense (antisense), scramble siRNA (scramble) or siRNA against human UGDH (siRNA UGDH) were treated as outlined under "Materials and Methods." GAG contents are reported in percent of the total radioactivity digested in the sample. The total GAG column represents the percentage of the total radioactivity sensitive to enzymatic digestions.