Functional Analysis of Proteoglycan Galactosyltransferase II RNA Interference Mutant Flies*

Heparan sulfate proteoglycan plays an important role in developmental processes by modulating the distribution and stability of the morphogens Wingless, Hedgehog, and Decapentaplegic. Heparan and chondroitin sulfates share a common linkage tetrasaccharide structure, GlcAβ1,3Galβ1,3Galβ1,4Xylβ-O-Ser. In the present study, we identified Drosophila proteoglycan galactosyltransferase II (dβ3GalTII), determined its substrate specificity, and performed its functional analysis by using RNA interference (RNAi) mutant flies. The enzyme transferred a galactose to Galβ1,4Xyl-pMph, confirming that it is the Drosophila ortholog of human proteoglycan galactosyltransferase II. Real-time PCR analyses revealed that dβ3GalTII is expressed in various tissues and throughout development. The dβ3GalTII RNAi mutant flies showed decreased amounts of heparan sulfate proteoglycans. A genetic interaction of dβ3GalTII with Drosophila β1,4-galactoslyltransferase 7 (dβ4GalT7) or with six genes that encode enzymes contributing to the synthesis of glycosaminoglycans indicated that dβ3GalTII is involved in heparan sulfate synthesis for wing and eye development. Moreover, dβ3GalTII knock-down caused a decrease in extracellular Wingless in the wing imaginal disc of the third instar larvae. These results demonstrated that dβ3GalTII contributes to heparan sulfate proteoglycan synthesis in vitro and in vivo and also modulates Wingless distribution.

Previously, we reported on the Drosophila inducible RNA interference (RNAi) knockdown system for analyses of the basic physiological functions of glycans (38) and applied this system to two types of glycosyltransferases, namely, d␤4GalT7 and Drosophila protein O-mannosyltransferases, and two transporters, that is, the Drosophila PAPS-transporters slalom (sll) and dPAPST2 (34, 39 -41). In the present study, we first performed the biochemical characterization of d␤3GalTII. We then generated inducible d␤3GalTII RNAi flies by using the GAL4-UAS system and subsequently performed the functional analysis.
pVL1393-FLAG-d␤3GalTII was co-transfected with Baculo-Gold TM viral DNA (Pharmingen, San Diego, CA) into Sf21 insect cells according to the manufacturer's instructions, and the cells were incubated for 3 days at 27°C to produce recombinant viruses. The Sf21 cells were infected with the recombinant virus at a multiplicity of infection of 5 and incubated for 72 h to yield conditioned media containing recombinant d␤3GalTII proteins fused with the FLAG peptide. An 8-ml aliquot of the culture medium was mixed with 100 l of Anti-FLAG M1 Affinity gel (Sigma). The protein/gel mixture was washed twice with 50 mM TBS (50 mM Tris-HCl, pH 7.4, and 150 mM NaCl) containing 1 mM CaCl 2 and eluted with 100 l of 100 g/ml FLAG Peptide in 10 mM TBS (Sigma).
Western Blot Analysis-To determine the concentration of the purified enzyme, the enzyme purified above was subjected to 12.5% SDS-polyacrylamide gel electrophoresis followed by Western blot analysis. The separated protein was transferred to a Hybond-P membrane (GE Healthcare). The membrane was probed with the anti-FLAG M2-peroxidase conjugate (Sigma) and stained with Konica Immunostaining HRP-1000 (Konica, Tokyo, Japan). The intensity of a positive band obtained using Western blotting was measured by a densitometer to determine the amount of the purified enzyme, using the FLAG-BAP control protein (Sigma).
For the Western blot analysis using the monoclonal antibody 3G10, the protein extract (see the subsection "sample preparation") was subjected to 15% SDS-polyacrylamide gel electrophoresis. The separated protein was transferred to an Immobilon-P membrane (Millipore Corp., Bedford, MA). The membrane was blocked with 5% skim milk in TBST (TBS with 0.1% Tween 20) for 30 min at room temperature and washed with TBST. The blocked membrane was treated with 20 milliunits of heparitinase I (Seikagaku) for 1 h at 37°C and washed with TBST. After heparitinase treatment, the membrane was first probed with the monoclonal antibody 3G10 (1:2000, Seikagaku) and then with horseradish peroxidase-conjugated sheep anti-mouse secondary antibody (1:20000, GE Healthcare). The blots were visualized with the ECL plus Western blotting detection kit (GE Healthcare), according to the manufacturer's instructions.
Sample Preparation for Western Blotting by the Monoclonal Antibody 3G10-Twenty to 30 third instar larvae were homogenized in 10 mM Hepes-Tris (pH 7.4) containing 0.25 M sucrose, 1 mM phenylmethylsulfonyl fluoride, and 1 g/mg each of leupeptin, aprotinin, and pepstatin A. The extract was centrifuged at 1000 ϫ g for 10 min to remove the debris. The concentration of protein was assayed by the Bradford method.
Each PCR fragment was inserted as an inverted repeat (IR) sequence into the pSC1 vector. The IR-containing fragments were then subcloned into the transformation vector pUAST, and these vectors were introduced into Drosophila embryos of the w 1118 mutant stock that were used as hosts to construct UAS-IR fly lines according to the procedure reported by Spradling (42). Each line was mated with the appropriate driver fly lines, and the F 1 progeny were raised at 28°C to observe the phenotypes.
Quantitative Analysis of the d␤3GaTII, sotv, dOXT, dHs3st-B, or dHS6ST Transcripts in Knockdown Flies by Realtime PCR-To measure the knockdown efficiency, we cross Act5C-GAL4 flies with UAS-IR flies of each gene. Total RNA was extracted from the third instar larvae of each F 1 progeny by using the TRIzol reagent (Invitrogen). First-strand cDNA was synthesized using a Superscript II first-strand synthesis kit (Invitrogen). Real-time PCR was performed using qPCR Mastermix QuickGoldStar (Eurogentec, Seraing, Belgium) and the ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Foster, CA). The gene-specific primer pairs and TaqMan probes that were used for each gene are as follows. For the quantification of d␤3GalTII, the forward primer 5Ј-GGCACGATCCCAGATTCG-3Ј, the reverse primer 5Ј-GTGCAGCACCATGTGATAGGA-3Ј, and the probe 5Ј-CACCTCCTACGCACCGCGCA-3Ј were used. For the quantification of sotv, the forward primer 5Ј-GCCTGAC-CTATGGGCTACTAGTCA-3Ј, the reverse primer 5Ј-GCT-GCTGGAGAGATCCCAAAGCA-3Ј, and the probe 5Ј-ACC-CGCGCAGAGCGCCA-3Ј were used. For the quantification of dOXT, the forward primer 5Ј-GCGGATTCTATGCCAT-GAAC-3Ј, the reverse primer 5Ј-GGTGGCCGCCAGTTGA-3Ј, and the probe 5Ј-ATACGAGACCGGAATAGCCAAGT-TCACCG-3Ј were used. For the quantification of dHs3st-B, the forward primer 5Ј-GCTTCAATGCCTGCCTCCT-3Ј, the reverse primer 5Ј-GCCAGTAGATTGCGTAAGTTGCT-3Ј, and the probe 5Ј-CCAGCATCAGTCGCAGTCTGCAGC-3Ј were used. For the quantification of d␤3GalTII, the forward primer 5Ј-GGAATTCAACTGGGACAGCAA-3Ј, the reverse primer 5Ј-GGTAATATTTGGTTTCGAGTTCAATG-3Ј, and the probe 5Ј-TGGAGGATGGGCTCAGCACGG-3Ј were used. For the quantification of the ribosomal protein L32 (RpL32), the forward primer 5Ј-GCAAGCCCAAGGG-TATCGA-3Ј, the reverse primer 5Ј-CGATGTTGGGCATCA- GATACTG-3Ј, and the probe 5Ј-AACAGAGTGCGTCGC-CGCTTCA-3Ј were used. The probes were labeled with a reporter dye FAM and quencher dye TAMRA at the 5Ј-and 3Ј-ends, respectively. Relative amounts of the d␤3GalTII, sotv, dOXT, dHs3st-B, or dHS6ST transcripts were normalized to those of the RpL32 transcripts in the same cDNA. Scanning Electron Microscopy-Adult flies were separated in new food vials and allowed to age for several days. The preparation was subjected to triethylamine anesthesia prior to mounting on stubs covered with carbon tapes and viewed under a JEOL 5600LL scanning electron microscope (Japan Electron Optics Laboratory Co., Ltd., Tokyo, Japan) under low vacuum (30 Pa) conditions.
Statistical Analysis-t test were used for analysis of the data of Western blotting. Tukey tests were used for analysis of the eye data. All statistical analysis was performed using the public domain R program.

Characterization of the Galactosyltransferase Activity of d␤3GalTII-
We prepared recombinant candidate d␤3GalTII (CG8734), which has been reported to be the Drosophila homolog of SQV-2 (24,27), to identify proteoglycan galactosyltransferase II activity. The soluble form was prepared by replacing the N-terminal region, including the cytoplasmic and transmembrane domains (amino acids 1 to 38) with an Ig signal sequence and FLAG peptide sequence. The secreted enzyme was purified using the anti-FLAG M1 gel and quantitated by Western blotting analysis using the anti-FLAG antibody.
The purified enzyme was used for galactosyltransferase assay with various acceptor substrates (Table 1). d␤3GalTII showed stronger activity toward Gal␤1,4Xyl␤1-pMph than Gal␤-pNph, only slight activity toward -␣-Gal, and no activity toward -␤-Xyl, -␤-GlcNAc, -␤-Glc, and -␤-GalNAc. These results demonstrated that d␤3GalTII was the Drosophila ortholog of h␤3GalTII in view of their activities. Spatiotemporal Expression Pattern of d␤3GalTII-The developmental expression profiles and tissue distribution of d␤3GalTII mRNA were investigated using a quantitative analysis by real-time PCR (Fig. 1, A and B). As shown in Fig. 1A, d␤3GalTII showed higher expression levels in the early embryonic stage (stages 0 -2) than in the other stages. The tissue distribution of d␤3GalTII mRNA in the third instar larvae and adult flies is shown in Fig. 1B. The d␤3GalTII gene showed ubiquitous distribution.
Subsequently, we analyzed the amount of HSPG in these RNAi mutant flies by Western blotting. No signal was detected by 3G10 in the heparitinase-untreated membrane (data not shown). The signal was detected in the heparitinase-treated membrane (Fig. 2C). The ratios of total signal intensity in the Act5C-GAL4/UAS-d␤3GalTII-IR(1), Act5C-GAL4/UAS-d␤3GalTII-IR (4), and Act5C-GAL4/UAS-d␤3GalTII-IR (5) lines were 84, 41, and 49%, respectively, of that in the wild-type fly. Moreover, the signal intensity of the double RNAi mutant flies of d␤3GalTII and d␤4GalT7 was less than that in d␤4GalT7 single RNAi mutant flies (Fig. 2D). These results demonstrated that the knockdown of d␤3GalTII influenced HSPG production.
d␤3GalTII Genetically Interacts with the Drosophila d␤4GalT7 Gene in the Wing and Eye-We investigated the biological functions of d␤3GalTII in the fly. First, we crossed the UAS-d␤3GalTII-IR fly lines with seven GAL4-driver lines: A9-GAL4, MS096-GAL4, hh-GAL4, en-GAL4, sd-GAL4, ptc-GAL4, and ap-GAL4. Each of the driver lines had a different promoter/enhancer that was expressed in a specific region of the wing. When ap-GAL4 was crossed with UAS-d␤3GalTII-IR, curly wings with concavity along the anteroposterior axis were observed (hereafter referred to as "A-P curly wing phenotype") (Fig. 3).
Human and Drosophila ␤4GalT7 proteins are involved in the synthesis of GAGs by transferring a Gal to the Xyl-O-Ser in the tetrasaccharide structure of the protein-linkage region of GAGs. To investigate whether d␤3GalTII is involved in the synthesis of proteoglycans in vivo, we tested for a genetic interaction between d␤3GalTII and d␤4GalT7.
The double RNAi mutant flies have two UAS-IR transgenes. If the control flies have one UAS-IR transgene, theoretically, GAL4 binds the UAS sequence twice in these flies and the double strand RNA of the same gene is produced twice that in the double RNAi mutant flies. Therefore, the decrease in the target mRNA is expected to be half of that in the control flies. To avoid this difference, we used the UAS-EGFP transgene. In our experiment, the control flies possessed one UAS-IR and one UAS-EGFP gene, and we obtained the same number of UAS sequences in both the control and double RNAi mutant flies.
First, we used ap-GAL4 for the genetic interaction analysis. However, we could not observe the genetic interaction because of the strong phenotype in the control flies (Table 2).   We then screened the GAL4 drivers to induce the weaker phenotype. The F 1 fly line A9-GAL4/ϩ;UAS-d␤4GalT7-IR[N2]/ϩ;UAS-EGFP/ϩ showed a weak A-P curly wing phenotype (22% penetrance in this genotype, Table 2) and we used this driver as the test for the genetic interaction between d␤3GalTII and d␤4GalT7. Neither of the three, that is, A9-GAL4/ϩ; UAS-d␤3GalTII-IR [ [5];ϩ/ϩ, respectively, exhibited the A-P curly wing phenotype ( Table 2). These results demonstrated that d␤3GalTII genetically interacted with d␤4GalT7 in wing development.
Decrease in Extracellular Wingless Expression by d␤3GalTII Knockdown in the Wing-Previous studies have demonstrated that HSPG is involved in the development of the wing via the signaling events mediated by the Wg, Hh, and Dpp morphogens (6 -8). We examined Wg, Hh, and pMad expression in the wing of the d␤3GalTII RNAi fly line. The expression level of Hh and pMad, which is the activated form of Mad and activated by Dpp receptor Thickveins (Tkv) in response to Dpp signaling, did not change in the d␤3GalTII RNAi region in the wing disc of the third instar larvae (data not shown).
On the other hand, when we induced RNAi by en-GAL4, the amount of Wg in the d␤3GalTII RNAi region decreased. Wg is normally expressed at the wing margin, that is, the boundary between the dorsal and ventral sides. We used en-GAL4, which expresses the GAL4 protein, in the posterior region of the wing in the third instar larva (Fig. 6A). The d␤3GalTII RNAi fly line, namely, en-GAL4/ UAS-d␤3GalTII-IR (5), showed decreased amounts of extracellular Wg in the posterior region of the wing (Fig. 6, B-D). This result indicated that the decreased HSPG expression with a decrease in the linkage structure formation affected extracellular Wg distribution.

DISCUSSION
Human ␤3GalTII (h␤3GalTII) has been cloned as ␤3GalT6, the sixth member of the ␤1,3-galactosyltransferase family, which transfers Gal to the GlcNAc or GalNAc residue (22). This d␤3GalTII shares a homology with Brainiac, that is, ␤1,3-N-acetylglucosaminyltransferase, which transfers GlcNAc to ␤-linked mannose (Man), with a preference for the disaccharide Man␤1,4Glc, the core of arthro-series glycolipids (44,45). In human, the ␤1,3-galactosyltransferase family shares a high homology with the ␤1,3-N-acetylglucosaminyltransferase family (46) but not with the ␤1,4-galactosyltransferase family, although both of the galactosyltransferases transfer the same Gal residue to ␤-linkage. This phenomenon is conserved between human and Drosophila.
d␤3GalTII transfers Gal to ␤-linked Gal, with a four times higher preference for the disaccharide Gal␤1,4Xyl␤-than for the monosaccharide Gal␤- (Table 1). Its activity is completely consistent with h␤3GalTII (22), showing that d␤3GalTII is an ortholog of h␤3GalTII in terms of activity. C. elegans ␤3GalTII showed no activity for the monosaccharide Gal␤-, whereas it showed considerably high activity for the disaccharide Gal␤1,4Xyl␤- (27). In a previous study, we cloned and characterized d␤4GalT7, which synthesizes the substrate of d␤3GalTII, the Gal␤1,4Xyl␤-structure (34). We assayed the d␤4GalT7 activity using a protocol that was almost the same as that adopted for d␤3GalTII. The purified recombinant d␤4GalTI transferred Gal to Xyl␤-, at a rate of 6.3 nmol/nmol of protein/min at 25°C. This transfer rate is of the same order as that of d␤3GalTII, which is 5.5 nmol/nmol of protein/min at 25°C. Thus, in the Golgi apparatus, both enzymes transfer Gal at almost equal rates for smooth synthesis of the linkage tetrasaccharide structure GlcA␤1,3Gal␤1,3Gal␤1,4Xyl␤1-O-Ser. HSPG are cell-surface, extracellular matrix molecules composed of HS and core proteins (47)(48)(49). In mammals, several core proteins in HSPG have been demonstrated. On the other hand, only four core proteins in HSPG are known in Drosophila; these include Syndecan, Dally, Dally-like, and Trol (50 -53). In this study, we showed the possibility that there are several other core proteins. More than 10 signals were observed approximately between 30 and 80 kDa (Fig. 2C). Because the putative molecular masses of Syndecan, Dally, Dally-like, and Trol are 23 and 40, 63, 77, and 94 kDa, and, 320 -460 kDa, respectively, novel core proteins in HSPG possibly exist in the larval Drosophila proteins. The RNAi knockdown flies of d␤3GalTII will be useful in analyzing these novel core proteins.
In addition to the biochemical function of d␤3GalTII in vitro, the present study showed its developmental function in vivo. The RNAi in the wing and eye caused the morphological changes (Figs. 3 and 4). Mutations in the genes involved in the synthesis of HSPG, including those encoding the core proteins (51,54,55), glycosyltransferases (6 -8), sulfotransferases (56,57), and nucleotide sugar transporter (58,59), show several phenotypes in the wing and eye. For example, dHs3st-B RNAi in the eye resulted in the rough-eye phenotype (57). These similar phenotypes in comparison to the d␤3GalTII RNAi fly strongly indicate that d␤3GalTII is the gene involved in the synthesis of HSPG in vivo and, moreover, suggest that d␤3GalTII interacts with these genes. In fact, d␤3GalTII interacted with d␤4GalT7 in the wing (Table 2) and with d␤4GalT7, dOXT, ttv, sotv, dHs3st-B, dHS6ST, and papss in the eye (Figs. 4 and 5).
A number of reports have mentioned that HSPG is required in the signaling of growth factors such as Wg, Hh, and Dpp during development (6, 60 -62). During photoreceptor differentiation in the eye of Drosophila, the Hh protein controls the progression of the morphogenetic furrow, and Dpp signaling follows the Hh signaling within and in front of the morphogenetic furrow (63,64). Because dHs3st-B, dHS6ST, and dlp (dally-like) are expressed in the morphogenetic furrow (55,57,65), HSPG is also expected to be involved in the growth factor signaling in the developing eye. The genetic interactions between d␤3GalTII and d␤4GalT7, dOXT, ttv, sotv, dHs3st-B, dHS6ST, and papss in the eye (Figs. 4 and 5) suggest that HSPG also contributes to morphogenesis in the eye via Hh and/or Dpp signaling.
Baeg et al. (56) reported that sfl (sulfateless), one of the genes involved in HSPG synthesis, is involved in Wg signaling during development. The accumulation of extracellular Wg decreases in the sfl mutant clone in the wing discs. The same result was obtained from the mutant of the sll (slalom) gene, which was involved in HSPG synthesis (66). In this study, we also observed a decrease in extracellular Wg expression in the d␤3GalTII RNAi flies (Fig. 6). The wing phenotypes of mutants in the genes involved in the synthesis of HSPG were also attributable to defective Hh and/or Dpp signaling (6 -8, 61). However, the expression of Hh or pMad, a downstream molecule of Dpp, was not altered in d␤3GalTII RNAi flies (data not shown). d␤3GalTII was expressed in the wing and eye discs in the third instar larvae of Drosophila (Fig. 1A) and may be involved in Hh and Dpp signaling. Because the expression level of d␤3GalTII in d␤3GalTII RNAi flies was ϳ30 -50% and that of the total HSPG was ϳ40 -80% of that in wild type (Fig. 2, A and C), the Hh or Dpp signaling pathway might have been slightly altered; therefore, we could not detect changes in these pathways.
The linkage tetrasaccharide structure of GAGs, GlcA␤1, 3Gal␤1,3Gal␤1,4Xyl␤1-O-Ser, is common to HS and CS. d␤3GalTII is also possibly involved in CS synthesis because of the enzyme, which transfers a Gal to a -␤-Gal residue in the linkage structure. Several studies have mentioned that CS is required in the regeneration of the central nervous system in mammal (15, 68 -70) and binds to signaling molecules, which are involved in axon guidance in vitro (67). On the other hand, in Drosophila, although a few studies indicated the existence of CSPG (28,31), the in vitro and in vivo functions of CSPG have yet been clarified. Further investigations will be necessary to clarify the functions of CSPG in Drosophila.