α-KDOase Activity in Oyster and Synthesis of α- and β-4-Methylumbelliferyl Ketosides of 3-Deoxy-d-manno-octulosonic Acid (KDO)*

Although α- and β-linked 3-deoxy-d-manno-octulosonic acid (KDO) is found in lipopolysaccharides (LPSs) of Gram-negative bacteria, capsular polysaccharides of microorganisms, and plants, very little is known about its degradation. Using both thin-layer chromatography and the periodate-thiobarbituric acid reaction, we found that the hepatopancreas of oyster (Crassostrea virginica) contained an enzyme (α-KDOase) capable of releasing α-linked KDO from LPSs. To facilitate the studies of α-KDOase, we have carried out the synthesis of 4-methylumbelliferyl-α-KDO (α-KDO-MU) by conjugating the glycosyl chloride of the per-O-acetylated methylester of KDO with methylumbelliferone by the SN2 type reaction and the catalyzed phase-transfer. In both cases, the β-anomer was obtained as the major product with a yield of about 80%, whereas the yield of α-anomer was only about 7%. Attempts to increase the yield of α-anomer were not successful. α-KDO-MU was used as substrate to follow the purification of α-KDOase from oyster hepatopancreas. The pH optimum for oyster α-KDOase was determined to be 4.5 using Re-LPS as substrate and 3.0 using α-KDO-MU as substrate. The enzyme was found to be stable in the pH range of 3–8. This enzyme released KDO from different LPSs, including Re-LPS from Escherichia coliand Salmonella minnesota, Rd-LPS from S. minnesota, and de-O-acyl-Re-LPS (Kiang, J., Szu, S. C., Wang, L.X., Tang, M., and Lee, Y. C. (1997)Anal. Biochem. 245, 97–101).


General Methods
Melting points were determined with a Fisher-Johns apparatus and were not corrected. NMR spectra were recorded at 25°C with a Bruker AMX-300 spectrometer at 300 and 100 MHz for 1 H NMR and 13 C NMR, respectively. Mass spectra were recorded with a VG 70-S mass spectrometer in a chemical ionization mode (reagent gas, NH 3 ) or a fast atom bombardment-positive mode (matrix: 3-nitrobenzyl alcohol). TLC was carried out on precoated silica gel 60 F 254 plates, and the carbohydrate components were detected by charring at 140°C after spraying the plates with 15% H 2 SO 4 in 50% ethanol or by UV absorption. Column chromatography was performed on silica gel. Ratios of solvents for TLC and column chromatography were expressed by volume. All evaporation and concentration was carried out below 40°C under reduced pressure using a water aspirator unless specified otherwise. the eluent, gave compound 1 (18.8 g, 52% from KDO) as a white solid: R F 0.30 (toluene:ether, 1:1); m.p. 154 -156°C; m.p. (literature) 155-158°C (9); MS (m/z, chemical ionization-positive mode) 480 ((M ϩ NH 4 ) ϩ , 100%), 420 ((M Ϫ AcOH ϩ NH 4 ) ϩ , 8%), and 403 ((M Ϫ AcOH ϩ H) ϩ , 26%); 1 H NMR (CDCl 3 ) ␦ 3.818 (s, 3 H, OCH 3 ), 2.151, 2.120, 2.057, 2.010, and 2.006 (each s, each 3 H, 5 CH 3 CO). The data for sugar protons are listed in Tables I and II. To prepare compound 2, dry hydrogen chloride gas was bubbled into a solution of 1 (3.85 g, 7.96 mmol) in acetyl chloride (40 ml) in a 100-ml flask at 0°C for 30 min, then the flask was sealed and kept at 4°C. After 24 h the solution was evaporated and the residual solvent was co-evaporated with toluene to give compound 2 (3.6 g, quantitative) as a colorless oil, R F 0. 55 (toluene: ether, 1:1). This glycosyl chloride of 1 was used as the glycosyl donor (2) without further purification.
S N 2 coupling in DMF-To a solution of 2 (2.5 g, 5.7 mmol) in DMF (20 ml), 4-methylumbelliferone (sodium salt) (1.70 g, 8.5 mmol) was added. The mixture was stirred at 25°C for 4 h, and TLC (toluene:ethyl acetate, 1:1) indicated the disappearance of 2. DMF was evaporated in vacuo, and the residue was partitioned between water (20 ml) and CHCl 3 (80 ml). The organic layer was separated and washed with aqueous NaHCO 3 and water, dried (Na 2 SO 4 ), and filtered. The filtrate was evaporated, and the residue was subjected to silica gel column chromatography using toluene:ethyl acetate (3:1) as the eluent to give a mixture of two UV-absorbing and carbohydrate-containing compounds that migrated closely on TLC (toluene:ethyl acetate, 2:1 or toluene: ether, 1:1). The mixture was further fractionated by silica gel column chromatography using toluene:ether (1:1) as the eluent to obtain the ␤-anomer 3 (2.5 g, 76%) and the ␣-anomer 4 (224 mg, 6.8%). Alternatively, the two anomers could be simply separated by agitation of the mixture in toluene. The ␤-anomer 3 is quite soluble in toluene but the ␣-anomer 4 is not. Filtration of the suspension gave pure 3 in toluene and pure 4 as a solid.

Preparation of 4-Methylumbelliferyl 3-Deoxy-␤-D-manno-2octulopyranosidonic Acid (5, ␤-KDO-MU)
MeONa/MeOH (0.5 M, 2 ml) was added to a solution of 3 (1.0 g, 1.73 mmol) in MeOH (40 ml), and the mixture was stirred at 25°C. After 1.5 h, the reaction mixture was concentrated to 10 ml and diluted with water (50 ml). The solution was adjusted to pH 11 and maintained at this pH by adding 2 M NaOH. TLC (CHCl 3 :MeOH:H 2 O:AcOH, 65:25:4:1) indicated the complete saponification of the methyl ester after 2 h. The solution was neutralized by Dowex 50W-X8 (H ϩ form) to pH 4 and filtered. The filtrate was then adjusted to pH 8 by adding 0.1 M NH 4 OH and evaporated. The residue was applied onto a Sephadex G-10 column (2.5 ϫ 95 cm), which was equilibrated and eluted with 50 mM NH 4 Tables I and II.

Enzyme Assays
When the KDO-cleaving activity of ␣-KDOase was assayed using Re-LPS or DeOA-LPS as substrate, 40 g of LPS were incubated with an appropriate amount of enzyme in 70 l of 50 mM sodium acetate buffer pH 4.5 at 37°C for a predetermined time. The amount of KDO released was determined by high performance anion exchange chromatography (8) and/or the periodate-thiobarbituric acid method (7). For detection of the free KDO released from LPS or ␣-KDO-MU by TLC, the reaction mixture contained 5 nmol of LPS or 20 nmol of ␣-KDO-MU in 40 l of 50 mM sodium formate buffer pH 4.5 or 3.0 and an appropriate amount of the enzyme. After incubation at 37°C for a preset time, the reaction was stopped by adding 40 l of ethanol followed by brief centrifugation. The supernatant was evaporated to dryness using a SpeedVac, dissolved in 12 l of methanol:water (1:2) and analyzed by TLC using chloroform:methanol: 12 mM MgCl 2 (45/40/12) as the developing solvent. The plate was sprayed with diphenylamine reagent (10) and heated at 120°C for 20 min to reveal glycoconjugates. The fluorometric assay of ␣-KDOase activity using ␣-KDO-MU as substrate was carried out according to the procedure described by Potier et al. (11). The enzyme was incubated with 0.5 mM ␣-KDO-MU in 50 mM sodium formate buffer, pH 3.0, in a total volume of 100 l at 37°C. After a preset time, 1.5 ml of 0.2 M sodium borate buffer, pH 9.8, was added to the reaction mixture to stop the reaction. The released MU was determined using a Sequoia-Turner Model 450 fluorometer. One unit of ␣-KDOase is defined as the amount of the enzyme that liberates 1 nmol of KDO per min at 37°C.

Purification of ␣-KDOase
All operations were carried out at a temperature between 0 and 5°C. Centrifugations were routinely carried out at 20,000 ϫ g for 30 -40 min using a Sorvall RC5C refrigerated centrifuge. Ultrafiltration was carried out with an Amicon stirred cell using a PM10 membrane.
The hepatopancreas (450 g) dissected from 1 gallon of fresh oysters was homogenized with 3.2 liters of cold acetone using a Polytron homogenizer, quickly filtered through a Buchner funnel, and immediately dried under vacuum to obtain 90 g of the acetone powder of oyster hepatopancreas. The acetone powder was extracted with 2.25 liters of water containing 1 mM EDTA and 1 mM phenylmethylsulfonyl fluoride as protease inhibitors using a Polytron homogenizer, followed by centrifugation. The pH of the water extract was adjusted to 4.3 with a saturated solution of citric acid, and the precipitate was removed by centrifugation. The supernatant was brought to 45% saturation with solid ammonium sulfate (277 g/liter). After standing for 2 h, the pre-cipitate was removed by centrifugation, and the supernatant was brought to 85% saturation with ammonium sulfate (295 g/liter). After standing overnight, the precipitate was collected by centrifugation, dissolved in 40 ml of 50 mM sodium acetate buffer, pH 4.2, and applied onto a Sephacryl S-200 column (5 ϫ 100 cm) equilibrated with 50 mM sodium acetate buffer, pH 4.2, containing 0.15 M NaCl. The column was eluted with the same buffer at 1 ml/min, and 20-ml fractions were collected. The fractions containing ␣-KDO-MU cleaving activity, as shown in Fig. 2A, were pooled, concentrated to 5.0 ml, and applied onto a SP-Fractogel column (1.5 ϫ 10 cm) equilibrated with 25 mM sodium phosphate buffer, pH 6.2. After washing with the same buffer, the column was eluted with a linear NaCl gradient from 0 to 0.5 M (total volume, 170 ml). The fractions containing ␣-KDO-MU-cleaving activity as shown in Fig. 2B were pooled, concentrated, and dialyzed against 25 mM sodium phosphate buffer, pH 7.0. This preparation (1.4 ml) was applied to a Con A-Sepharose column (1.0 ϫ 5 cm) that had been equilibrated with 25 mM sodium phosphate buffer, pH 7.0. The column was washed with the same buffer; the ␣-KDOase activity was then eluted with 0.5 M methyl-␣-mannoside. Fractions with ␣-KDO-MUcleaving activity were concentrated, dialyzed against 1.5 M ammonium sulfate, and applied to an octyl-Sepharose column (1.5 ϫ 9 cm) that had been equilibrated with 1.5 M ammonium sulfate. After washing with the same buffer, the column was successively eluted with 1.0 M ammonium sulfate, 0.5 M ammonium sulfate, and water. The majority of ␣-KDO-MU-cleaving activity was eluted with 1.0 M ammonium sulfate. This ␣-KDOase preparation was concentrated, dialyzed against 25 mM sodium acetate buffer, pH 5.0, and used for subsequent studies. Table III summarizes the purification of ␣-KDOase.  a Not analyzed because of the overlapping of signals.

RESULTS AND DISCUSSION
Preparation of ␣and ␤-KDO-MU-The glycosyl chloride of the per-O-acetylated methylester of KDO (2) was chosen as the glycosyl donor for the synthesis of ␣and ␤-KDO-MU. Two procedures, which have been used for the preparation of 4methylumbelliferyl ketoside of NeuAc, were applied for the coupling of 2 with the sodium salt of 4-methylumbelliferone, namely, the direct coupling in a dipolar aprotic solvent such as DMF (12) and the catalyzed phase transfer reaction (13). In both cases, the methyl (4-methylumbelliferyl 4,5,7,8-tetra-Oacetyl-3-deoxy-␤-D-manno-2-octulopyranoside)onate (3) was obtained as the major product (ϳ80%), together with the ␣anomer (4) as the minor product (ϳ7%). Compounds 3 and 4 could be separated either by repeated column chromatography or by taking advantage of their differential solubility in toluene. The ␤-anomer 3 was more soluble in toluene than the ␣-anomer 4. The structures of 3 and 4 were determined by mass spectrometry and by 1 H and 13 C NMR spectroscopy. Although it is difficult to deduce the anomeric configuration by 1 H NMR analysis, we assigned the major product as the ␤anomer 3 based on the following considerations. First, it is reasonable to assume that due to the anomeric effect, compound 2 should exist predominantly in the ␣-anomeric configuration. Second, the coupling reaction under a given condition should occur by the usual S N 2 mechanism to give the product an anomeric inversion. This assignment was confirmed by the proton-coupled 13 C NMR signals for C-1 in the final products.
Several attempts to improve the yield of the ␣-anomer 4 were unsuccessful. For example, silver triflate-catalyzed reaction of 2 with anhydrous 4-methylumbelliferone in CH 2 Cl 2 in the presence of molecular sieves (MS4A) led to the formation of methyl 4,5,7,8-tetra-O-acetyl-2,6-anhydro-3-deoxy-D-mannooct-2-enonate (14) (the glycal derivative of KDO) in 82% yield instead of the coupling product. It is noteworthy that the reaction of 2 with the sodium salt of 4-methylumbelliferone in DMF in the presence of tetrabutylammonium chloride (2 molecular equivalents) did not improve the yield of 4. Therefore, the formation of the ␣-anomer 4 did not result from in situ anomerization of the glycosyl chloride by chloride ion that was generated during the coupling reaction, but instead, the ␣anomer must have come from the contaminated ␤-glycosyl chloride. It should be pointed out that the formation of isomeric products was not reported in the similar preparation of MUketoside of sialic acid (12,13).
De-O-acetylation of 3 and 4, followed by alkaline hydrolysis of the methyl ester and purification by Sephadex G-10 chromatography, gave the ␤-anomer 5 (84%) and the ␣-anomer 6 (91%), respectively, as their ammonium salts. The 1 H NMR spectra (Fig. 3) of 5 and 6 revealed that the sugar protons of the two anomers had distinctly different chemical surroundings. In 5, the signals for H-4, 5, 6, 7, and 8 were overlapping in a narrow area ranging from ␦ 4.04 to 3.82; whereas in 6, the signals for each proton of 4, 5, 6, 7, 8 were completely separated in a wide range (␦ 4.26 -3.12).  Determination of the Anomeric Configuration by Proton-coupled 13 C NMR-Some empirical 1 H NMR rules have been used for deducing the anomeric configurations of KDO derivatives. One such rule is that the difference in chemical shift between H-3a and H-3e in ␤-anomer is usually bigger than that in ␣-anomer (9). However, these empirical rules often lead to ambiguous assignments. For example, in the present case, the difference between ␦ (H-3e) and ␦ (H-3a) in ␤-anomer 5 is bigger than that in ␣-anomer 6, whereas the difference between ␦ (H-3e) and ␦ (H-3a) in ␤-anomer 3 is actually smaller than that in ␣-anomer 4. This is due to the fact that the substituents could greatly influence the chemical shifts of neighboring protons. A definitive determination of the anomeric configurations of KDO derivatives could be achieved by comparison of the proton-coupled 13 C NMR signals of the C-l in ␣and ␤-anomers (9). In a typical 5 C 2 chair conformation of KDO derivatives, the dihedral angles of (C-1)-(C-2)-(C-3)-(H-3a) in ␣and ␤-anomers are nearly 60°and 180°, respectively. Therefore, the ␣-anomer would give a small value for the coupling constant between C-l and H-3a (J C-1, H-3a Ͻ 1 Hz), and the ␤-anomer would give a relatively large coupling constant (J C-1, H-3a ϭ 5-6 Hz), according to the Karplus relationship (15). In the protoncoupled 13 C NMR spectra (Fig. 4), the C-1 signal of 5 appeared at ␦ 173.73 as a doublet (J C-1, H-3a ϭ 5.5 Hz), whereas the C-1 signal of 6 appeared at ␦ 174.18 as a broad singlet (J C-1, H-3a Ͻ 1 Hz). Accordingly, 5 should be in the ␤-D-configuration and 6 in the ␣-D-configuration. This method was previously used for the determination of the anomeric configurations of sialic acid derivatives (16). The present study shows that this assignment is equally applicable for KDO derivatives.
KDO-containing glycoconjugates have been found to contain both ␣and ␤-linked KDO (7). Therefore, ␣and ␤-KDO-MU should be useful for studying ␣and ␤-KDOases, and the availability of these two enzymes will facilitate the studies of the structure and function of KDO-containing glycoconjugates.
␣-KDOase from Oyster Hepatopancreas-Despite the medical importance of LPSs, virtually nothing is known about their degradation. The enzyme (␣-KDOase) that cleaves ␣-linked KDO from LPSs has never been reported. The hepatopancreas of the oyster, C. virginica, was found to be rich in various glycoconjugate-cleaving enzymes. Using the periodate-thiobarbituric acid reaction (7) and TLC, the crude extract of oyster hepatopancreas was found to liberate KDO from Re-LPS prepared from E. coli and S. minnesota. Since KDO residues in LPSs have been shown to be ␣-ketosidically linked (19,20), we used the synthesized ␣-KDO-MU as substrate to follow the enzyme activity to purify ␣-KDOase from the crude extract of oyster hepatopancreas. Before ␣-KDO-MU was synthesized, we used Re-LPS as substrate, and the liberated KDO was determined by the periodate-thiobarbituric acid reaction. Initially, this method was used to monitor the purification of ␣-KDOase during Sephacryl-S-200 gel filtration. We subsequently found that in the column fractions, the enzyme activity detected by the Re-LPS/periodate-thiobarbituric acid reaction coincided well with that detected using ␣-KDO-MU as substrate. The availability of ␣-KDO-MU greatly facilitated the purification of ␣-KDOase.
The oyster ␣-KDOase was found to be stable in the pH range In all cases, 1 unit of the enzyme was used. For the hydrolysis of DeOA-LPS, the incubation was carried out at pH 4.5 for 1 h, whereas for cleaving ␣-KDO-MU, the incubation was carried out at pH 3.0 for 15 min under the assay conditions described under "Experimental Procedures." of 3-8. Interestingly, the pH optima for this enzyme using the synthetic and the natural substrates were found to be significantly different. The pH optimum of this enzyme was determined to be 4.5 for releasing KDO from Re-LPS, whereas that for the hydrolysis of ␣-KDO-MU was 3.0 (Fig. 6). Thus, the optimal pH for the hydrolysis of KDO from KDO-containing glycoconjugates cannot be inferred from the hydrolysis of ␣-KDO-MU.
The time courses of the liberation of KDO from DeOA-LPS by ␣-KDOase and by acid hydrolysis are compared in Fig. 7. The oyster ␣-KDOase was able to completely detach the KDO from DeOA-LPS. The amount of KDO released by the enzyme was found to be comparable to that liberated by acid hydrolysis. Fig.  8 shows the TLC analysis of the release of KDO from DeOA-LPS and ␣-KDO-MU. ␣-KDOase isolated from oyster hepatopancreas was able to cleave KDO efficiently from all LPS substrates tested, including Re-LPS from E. coli, Re-LPS from S. minnesota R595, Rd-LPS from S. minnesota R7, and DeOA-LPS. However, ␤-KDO-MU was found to be refractory to this enzyme. Thus, the specificity of ␣-KDOase also supports the assignment of the anomeric configurations of the two KDO-MU anomers. The oyster ␣-KDOase represents the first ␣-KDOcleaving enzyme, and the presence of such an enzyme in the hepatopancreas of oyster may suggest the wide occurrence of KDO in marine organisms. ␣-KDOase capable of liberating KDO from LPSs should be important and useful for studying the structure and function of LPSs and other ␣-KDO-containing glycoconjugates.