Characterization of β-1,2-Mannosyltransferase in Candida guilliermondii and Its Utilization in the Synthesis of Novel Oligosaccharides*

A particulate insoluble enzyme fraction containing mannosyltransferases from Candida guilliermondiiIFO 10279 strain cells was obtained as the residue after extracting a 105,000 × g pellet of cell homogenate with 1% Triton X-100. Incubation of this fraction with a mannopentaose, Manα1→3(Manα1→6)Manα1→2Manα1→2Man, in the presence of GDP-mannose and Mn2+ ion at pH 6.0 gave a third type of β-1,2 linkage-containing mannohexaose, Manβ1→2Manα1→3(Manα1→6)Manα1→2Manα1→2Man, the structure of which was identified by means of a sequential NMR assignment. The results of a substrate specificity study indicated that the β-1,2-mannosyltransferase requires a mannobiosyl unit, Manα1→ 3Manα1→, at the nonreducing terminal site. We synthesized novel oligosaccharides using substrates possessing a nonreducing terminal α-1,3-linked mannose unit prepared from various yeast mannans. Further incubation of the enzymatically synthesized oligosaccharide with the enzyme fraction gave the following structure, Manβ1→2Manβ1→2Manα1→3(Manα1→6)Manα1→ 2Manα1→2Man, which has been found to correspond to antigenic factor 9. Incubation of Candida albicans serotype B mannan with the enzyme fraction gave significantly transformed mannan, which contains the third type of β-1,2-linked mannose units.

A particulate insoluble enzyme fraction containing mannosyltransferases from Candida guilliermondii IFO 10279 strain cells was obtained as the residue after extracting a 105,000 ؋ g pellet of cell homogenate with 1% Triton X-100. Incubation of this fraction with a mannopentaose, Man␣133(Man␣136)Man␣132Man␣132Man, in the presence of GDP-mannose and Mn 2؉ ion at pH 6.0 gave a third type of ␤-1,2 linkage-containing mannohexaose, Man␤132Man␣133(Man␣136)Man␣132Man-␣132Man, the structure of which was identified by means of a sequential NMR assignment. The results of a substrate specificity study indicated that the ␤-1,2-mannosyltransferase requires a mannobiosyl unit, Man␣13 3Man␣13, at the nonreducing terminal site. We synthesized novel oligosaccharides using substrates possessing a nonreducing terminal ␣-1,3-linked mannose unit prepared from various yeast mannans. Further incubation of the enzymatically synthesized oligosaccharide with the enzyme fraction gave the following structure, Man␤132Man␤132Man␣133(Man␣136)Man␣13 2Man␣132Man, which has been found to correspond to antigenic factor 9. Incubation of Candida albicans serotype B mannan with the enzyme fraction gave significantly transformed mannan, which contains the third type of ␤-1,2-linked mannose units.
We have studied the structures of the cell wall mannans of medically important Candida species for several years (1)(2)(3) and demonstrated that there are three types of ␤-1,2 linkagecontaining side chains in its cell wall mannans. One is the ␤-1,2-linked mannooligomer, which is located in a phosphodiesterified oligosaccharide moiety, as the common epitope in the mannans of several Candida species (4 -9). The second type is ␤-1,2-linked mannose units attached to the nonreducing terminal of the ␣-1,2-linked oligomannosyl side chains in the mannans of Candida albicans serotype A (10,11), Candida tropicalis (12), and Candida glabrata (13). These two ␤-1,2 linkagecontaining epitopes have been identified as corresponding to antigenic factors 5 (14) and 6 (15), respectively. The third type of ␤-1,2 linkage-containing side chains can be observed in the mannan of Candida guilliermondii. This type of oligosaccharide contains ␤-1,2-linked mannose units attached to an ␣-1,3linked mannose unit, the presence of which has been demon-strated in the cell wall mannans of Saccharomyces kluyveri (16,17), C. guilliermondii (18), and Candida saitoana (19) and have been identified to correspond to antigenic factor 9 (18).
The ␤-1,2 linkage-containing oligomannosyl side chains are the specific epitopes of the mannans of the genus Candida and have not been found in mammalian cells. Therefore, the detection of such an antigen in the sera from patients using immunological procedures is useful for the diagnosis of invasive candidiasis. Furthermore, several workers reported that the ␤-1,2 linkage-containing side chains participate in the adherence of fungal cells to host cells during the initial step of Candida infection (20,21). Consequently, it is meaningful to characterize the ␤-1,2-mannosyltransferases responsible for the formation of the functional side chain.
In a previous study (22), we detected the ␤-1,2-mannosyltransferase II, which is responsible for the transfer of an additional ␤-1,2-linked mannose unit to the nonreducing terminal ␤-1,2-linked one attached to the ␣-1,2-linked mannotetraose. Although we tried to detect the ␤-1,2-mannosyltransferase I responsible for the transfer of the first ␤-1,2-linked mannose unit to the ␣-1,2-linked mannotetraose to synthesize antigenic factor 6, we could not detect the activity in the homogenate of C. albicans serotype A cells. Therefore, we tried to detect another ␤-1,2-mannosyltransferase responsible for the formation of the antigenic factor 9, which have been found in the mannan of C. guilliermondii (18).
In the present study, we tried to detect and characterize the ␤-1,2-mannosyltransferase IV using several oligosaccharide substrates prepared from various yeast mannans.
Preparation of Pyridylamino-oligosaccharides-The pyridylamination of the oligosaccharide was performed using the method of Yamamoto et al. (29) as follows: to an oligosaccharide (1 mg), 600 l of a 2-aminopyridine solution prepared by dissolving 1 g of 2-aminopyridine in 0.65 ml of concentrated hydrochloric acid was added. After being sealed, the tube was heated at 90°C for 10 min. The tube was then opened, and 60 l of the supernatant of a mixture of 100 mg of sodium cyanoborohydride and 60 l of water, as the reducing reagent, was added. The tube was resealed and heated at 90°C for 1 h. The reaction mixture was diluted with 2 ml of water and applied to a column (1 ϫ 50 cm) of Toyopearl HW-40 and separated from free 2-aminopyridine by elution with 0.01 M ammonium acetate, pH 6.0.
Enzyme Preparation-The preparation of the mannosyltransferase fraction of C. guilliermondii strain cells was carried out by the method described in a previous paper (22) as follows. The cells were grown in the YPD medium (0.5% yeast extract, 1% peptone, and 2% glucose) at 28°C until the mid-logarithmic growth phase (A 600 ϭ ϳ6). The cells were then harvested and washed with 5 mM Tris/HCl, pH 7.5, by centrifugation. The cells (about 40 g of wet cells) were resuspended in 15 ml of 5 mM Tris/HCl, pH 7.5, containing 3 mM MgCl 2 , 0.5% glycerol, 1.0% 2-mercaptoethanol, and 1 mM phenylmethylsulfonyl fluoride and homogenized with a Bead Beater (Biospec Products) with 50 g glass beads. The homogenate was centrifuged for 20 min at 5,000 ϫ g, and the supernatants were centrifuged for 20 min at 15,000 ϫ g. Then the supernatants were recovered and were centrifuged for 1 h at 105,000 ϫ g. The pellet was resuspended in 1 ml of 5 mM Tris/HCl, pH 7.5, containing 1% Triton X-100 and 1 mM phenylmethylsulfonyl fluoride and extracted for 2 h at 4°C. The mixture was centrifuged for 60 min at 105,000 ϫ g. The pellet (fraction P) and supernatant (fraction S) were kept at Ϫ90°C and were both assayed for protein contents and mannosyltransferase activity.
Mannosyltransferase Assay-The assay mixture containing fraction P (300 g of protein), 5 mM pyridylamino-oligosaccharide, 20 mM GDPmannose donor, 50 mM Tris/maleate, pH 6.0, 20 mM MnCl 2 , and 0.3% Triton X-100 in a total volume of 25 l was incubated for 1 h at 30°C (standard assay). The reaction was initiated by the addition of GDPmannose and terminated by heating the mixture for 10 min at 100°C. After removal of the denatured protein by centrifugation, each reaction mixture was analyzed by HPLC as will be described below. The amount of product was estimated by its fluorescence intensity using pyridylaminomannose as a standard.
Analysis of Enzyme Reaction Products by HPLC 2 -An Amide-80 column was used for the normal phase HPLC. The flow solvent was a 35:65 mixture of 3% acetic acid/triethylamine, pH 7.3, and acetonitrile, and the flow rate was 1.0 ml/min at 40°C. Detection of the pyridylaminooligosaccharides was fluorospectrometrically conducted with excitation and emission wavelengths of 320 and 400 nm, respectively.
Large Scale Enzyme Reaction for NMR Analysis-For the NMR analysis, the enzyme reaction was carried out in a total volume of 500 l containing 5-10 mg of oligosaccharide or mannan, fraction P (about 12 mg of protein), 20 mM GDP-mannose, 50 mM Tris/maleate, pH 6.0, 20 mM MnCl 2 , and 0.3% Triton X-100. After incubation for 24 -48 h at 30°C, the reaction was stopped by boiling, and the denatured protein was removed by centrifugation. The enzyme-modified oligosaccharide was fractionated by HPLC and lyophilized. The enzyme-modified mannan was dialyzed against water and lyophilized.
NMR Spectroscopy-All NMR experiments were performed using a JEOL JNM-GSX 400 spectrometer at 400 MHz for 1 H in D 2 O at a probe temperature of 45°C. Acetone (2.217 ppm) was used as the internal standard for the 1

H NMR.
Protein Determination-The protein was determined using the bicinchoninic acid protein assay kit (Pierce) (30) with bovine serum albumin as the standard.

RESULTS
Enzyme Preparation-The total amounts of protein in the membrane preparations, fractions P and S, prepared from the C. guilliermondii IFO 10279 strain cells were 48 and 38.5 mg, respectively. Transferase activities of the two fractions were assayed by incubation with GDP-mannose and pyridylamino-SaMan 5 . Although the two reaction systems gave the same single product corresponding to pyridylaminomannohexaose by HPLC, the total activity observed in fraction P (2597 nmol⅐h Ϫ1 ) was about three times higher than that in fraction S (923 nmol⅐h Ϫ1 ). Therefore, we used fraction P as the enzyme preparation for further studies. To determine the structure of the enzyme reaction product by NMR, a large scale reaction mixture with free SaMan 5 was incubated for 36 h. The HPLC profile of the reaction products indicated that approximately 50% of the SaMan 5 was transformed into a hexaose.
Sequential NMR Assignment of the Enzyme Reaction Product-The linkage sequence of the mannohexaose obtained by the enzyme reaction, abbreviated as SagMan 6 , was analyzed by a sequential NMR assignment method using rotating frame nuclear Overhauser effect spectroscopy (ROESY). This nonempirical assignment method has been demonstrated to give satisfactory results (8,18,24,25,31). The H-1 signal of the reducing terminal ␣-mannose unit (␦ ϭ ϳ5.35 ppm) was able to be empirically assigned. Therefore, we started the sequential assignment of SagMan 6 from the reducing terminal mannose unit. The boxed regions in Fig. 2 indicate intraresidue H-1-H-2 or H-1-H-3 connectivities, which were confirmed by relayed coherence transfer spectroscopy (relayed COSY) and two-dimensional homonuclear Hartmann-Hahn spectroscopy (HO-HAHA). On the other hand, cross-peaks labeled with primed letters indicate interresidue H-1-H-2Ј or H-1-H-3Ј connectivities between two adjacent mannose units. The numbers on the labels indicate the corresponding ring protons. Fig. 2, B and C, shows partial two-dimensional HOHAHA and ROESY spectra, respectively, of SagMan 6 . Since the H-1-H-2-correlated crosspeak A2 indicates the H-2 chemical shift of Man-A, the NOE cross-peak A2Ј between the H-2 of Man-A and the H-1 of Man-B was easily assigned. Similarly, the NOE cross-peak B2Ј between the H-2 of Man-B, which was assigned from cross-peak B2, and the H-1 of Man-C was assigned. Since Man-C is substituted by an ␣-1,3 linkage, the NOE cross-peak C3Ј was found through the H-1-H-3-correlated cross-peak C3. Usually, an ␣-1,3 linkage gives weak H-1-H-2Ј NOE cross-peak in addition to strong H-1-H-3Ј NOE cross-peak. Therefore, we can also assign through the H-1-H-2-correlated cross-peak C2 and the NOE cross-peak C2Ј between the H-2 of Man-D and the H-1 of Man-C. Additionally, we could find the NOE cross-peak D2Ј between the H-2 of Man-D and the H-1 of Man-E, of which the signal, 4.764 ppm, corresponds to the ␤-1,2-linked mannose unit (10,11,17). Using this procedure, we could sequentially assign the H-1 signal from Man-A to Man-E as A2-A2Ј-B2-B2Ј-C3-C3Ј-(or -C2-C2Ј-)-D2-D2Ј-E2. Therefore, we assumed that Man-D of SagMan 6 is substituted by a ␤-1,2-linked mannose unit, Man-E. It has been shown that the H-1 signal of an ␣-linked mannose unit substituted by a single ␤-1,2-linked unit causes a downfield shift (⌬␦ ϭ 0.09 -0.12 ppm) (10,11,17) and that the ␤-1,2-linked mannose units gave H-1-H-5-correlated cross-peaks in a characteristic region (14,17,18). In the spectrum of SagMan 6 , the H-1 proton of Man-D at 5.239 ppm appeared at 0.1 ppm more downfield than that of SaMan 5 at 5.129 ppm, and there is a ␤-mannose specific H-1-H-5-correlated cross-peak. Moreover, SagMan 6 resisted digestion with a jack bean ␣-mannosidase. These data also supported the fact that Man-D of SagMan 6 is substituted by the ␤-1,2-linked mannose unit. Therefore, we determined the structure of Sag-Man 6 to be the following. The same hexaose has been found in O-linked oligosaccharides (16) and acetolysate (17) of the mannan of S. kluyveri.
Since these findings indicate that the assay system is able to detect the ␤-1,2-mannosyltransferase IV, we further examined some properties of this enzyme.
Optimum pH-The effect of buffer pH on the transferase activity was studied over the pH range of 3.5-8.5. The ␤-1,2mannosyltransferase IV exhibits maximum activity at about pH 6.0 in 50 mM Tris/maleate (Fig. 3).
Metal Ion Requirement-The effect of several divalent cations, Mn 2ϩ , Mg 2ϩ , Ca 2ϩ , Ni 2ϩ , and Zn 2ϩ , on the enzyme activity was studied using their respective chlorides. As shown in Table I, the enzyme activity was enhanced by the addition of Mn 2ϩ , Mg 2ϩ , or Ca 2ϩ . The enzyme activity was not affected by the addition of EDTA. Furthermore, the enzyme activity was completely inhibited by the addition of 20 mM ZnCl 2 , and the lost activity could not be recovered by the addition of EDTA.
Enzyme Kinetics-The enzyme activity is linear for at least 1.5 h at 30°C under the standard conditions. The Lineweaver-Burk plot of the ␤-1,2-mannosyltransferase IV for pyridylamino-SaMan 5 is shown in Fig. 4. The K m and the V max values for pyridylamino-SaMan 5 calculated from this figure were about 18 mM and 200 nmol⅐mg protein Ϫ1 ⅐h Ϫ1 , respectively.
The substrate specificity study indicates that the ␤-1,2-mannosyltransferase IV requires the nonreducing terminal ␣-1,3linked mannose unit. However, the linkage of the penultimate mannose unit does not affect the substrate activity of the oligosaccharides. A similar substrate specificity has been found on the ␣-1,6-mannosyltransferase responsible for the synthesis of ␣-1,6-branched side chains (23). It is of interest to compare  Man␣133Man␣133Man␣132Man␣132Man-pyridylamine 48 1 6 Man␣1 a The ␤-1,2-mannosyltransferase and ␣-1,6-mannosyltransferase activities were calculated from the amount of the jack bean ␣-mannosidaseresistant and susceptible enzyme reaction products, respectively. the substrate recognition mechanisms of the two transferases or with ␣-1,3-mannosidase (48) or with the ␣-1,3-linked mannose-specific lectin (28). Since C. guilliermondii cells contain the ␤-1,2and ␣-1,6-mannosyltransferases judging from the structure of its mannan (18), it is predictable that if we use an ␣-1,3 linkage-containing linear oligosaccharide as the substrate of the enzyme reaction, we obtain at least two products that contain the ␤-1,2or ␣-1,6-linked mannose unit. Therefore, we compared the substrate specificity using pairs of oligosaccharides with or without the ␣-1,6-linked branching mannose unit.
In this study, we could synthesize a pure ␤-1,2 linkagecontaining oligosaccharide, Man␤132Man␣133(Man␣136)-Man␣132Man␣132Man, using fraction P prepared from C. guilliermondii cells as the enzyme and Man␣133(Man␣136)-Man␣132Man␣132Man as the substrate in the presence of MnCl 2 and GDP-mannose. Since the synthesis of the oligosaccharides using specific glycosyltransferases does not require a complicated protecting procedure and does not produce by-products, it is an effective tool for the synthesis of oligosaccharides.
We will be able to synthesize many novel oligosaccharides by taking advantage of the substrate specificity of each enzyme obtained from several Candida species. The novel oligosaccharides seem to be useful for studying not only the substrate specificity of mannosyltransferases but also the specificity of lectins and mannosidases. Furthermore, this approach will become important for the synthesis of the sugar chains of medically important glycoproteins or glycolipids.