Compared to other organs, the mouse thymus exhibits a high level of sialidase activity in both the soluble and crude membrane fractions, as measured at neutral pH using 4MU-Neu5Ac as a substrate. The main purpose of the present study was to identify the sialidase with a high level of the activity at neutral pH in the crude membrane. Several parameters were analyzed using the soluble (S) fraction, N and D fractions that were obtained by NP-40 or DOC/NP-40 solubilization from the thymus crude membrane. The main sialidase activity in the N fraction exhibited almost the same pI as that of soluble Neu2 and 60% of the activity was removed from the membrane by three washes with 10 mM Tris-buffer, at pH 7.0. The N fraction preferentially hydrolyzed the sialic acid bond of glycoprotein and exhibited sialidase activity with fetuin at pH 7.0 but not at pH 4.5. The same activity was observed in a plasma membrane-rich fraction. To date, the removal of sialic acid from fetuin at pH 7.0 was reported only with soluble Neu2 and the membrane fraction from Neu2-transfected COS cells. We analyzed the gene that controls the sialidase activity in the crude membrane fraction at pH 7.0 using SMXA recombinant mice and found that compared with other three genes, Neu2 presented the best correlation with the activity level. We suggest that Neu2 is most likely responsible for the main activity in the N fraction, due to its association with the membrane by an unknown mechanism.
CD4–CD8– double negative
4-methylumbelliferyl 5-acetyl neuraminic acid
Cell-to-cell interactions are important for the exchange of information among cells, especially in the immune system. The removal of sialic acid from the cell surface strongly affects this interaction. We have previously reported that sialidase treatment of Epstein–Barr virus (EBV)-transformed B cells resulted in large cell aggregation [1, 2]. We speculated that a natural sialidase might exist on the surfaces of immune cells and searched for a membrane-associated sialidase that acts at a neutral pH in immune organs. We found a high level of sialidase activity at a neutral pH in the mouse thymus not only in the soluble fraction  but also in the crude membrane fraction . We also determined that a unique cell, a ‘Neu-medullocyte’ (Mac-1 and immunoglobulin-positive cell), exists in the thymus of Neu1b and Neu1c mice [4, 5], whereas this cell and soluble Neu2 sialidase activity are absent in the SM/J (Neu1a) mice . The SM/J mouse is a naturally occurring strain that exhibits a partial deficiency in neuraminidase  due to a point mutation in the Neu1 locus [7, 8]. We sequenced the Neu2 mRNA from the A/J (Neu1b) mouse , discovered a novel sequence at the 5′ terminus (Neu2B) compared to the muscle Neu2 (Neu2C) , and studied several characteristics of these sialidases .
However, the membranous sialidase that acts at neutral pH in the thymus remains unidentified. The sialidases that have been cloned from vertebrates include four kinds (Neu1, 2, 3, and 4), and comprehensive reviews have been published concerning these properties [11, 12]. Neu1, 3, and 4 have optimal activity at acidic pH whereas Neu2 is optimal at neutral pH. The neutral activity in the thymus crude membrane cannot be explained by these acidic sialidase activities.
The main purpose of the present study was to identify the sialidase with a high level of the activity at neutral pH in the crude membrane of the thymus. We attempted to clarify the characteristics of the sialidase and determined that a plasma membrane-rich fraction from the mouse thymus exhibited a sialidase activity that removed sialic acid from fetuin at pH 7.0 but not at pH 4.5. We concluded that the molecule that is primarily responsible for the main activity at neutral pH in the crude membrane fraction is most likely the Neu2 molecule, which is associated with the membranous fraction by an unknown mechanism.
MATERIALS AND METHODS
Chemicals and sialidase substrates
The following materials were obtained commercially: 4MU-Neu5Ac was purchased from Nacalai Tesque (Kyoto, Japan); fetuin, sialyllactose, mucin, and a ganglioside mixture from the bovine brain (type II) were obtained from Sigma (St Louis, MO, USA). PNA was obtained from Seikagaku Kogyo, Co., Ltd (Tokyo, Japan), biotin-labeled lectins were obtained from J-OIL MILLS, Inc. (Tokyo, Japan), and Ampholine was obtained from LKB (Stockholm, Sweden). Total gangliosides for the natural substrate of sialidases were prepared from 18 g pooled murine spleens. Briefly, total lipids were extracted with 200 mL each of chloroform (C)/methanol (M) (2:1) and C/M/water (W) (1:2:0.8). The total ganglioside fraction was obtained using DEAE-Sephadex and eluted with C/M/0.8 M sodium acetate and condensed after dialysis. We detected two purple spots by resorcinol reagent in the middle and lower parts of the thin layer chromatography (TLC) developed with C/M/0.2% CaCl2 aq. (55:45:10). The sialic acid content was quantified by the thiobarbituric acid method. The crude membrane fraction was prepared from the mouse thymuses for use as an endogenous substrate as described below and was divided into small tubes and frozen until use.
Mice, antibodies, and transfection
A/J and C57BL/6 mice were purchased from Shizuoka Laboratory Animal Center Hamamatsu, Shizuoka, Japan. MRL/lpr, MRL/n, B6/lpr, and B6/n mice were purchased from Jackson Laboratory (Bar Harbor, ME, USA). SMXA recombinant mice were kindly donated through Hamamatsu Medical School and Nagoya University (Japan) by Dr M. Nishimura, who established these mice when he was at Hamamatsu Medical School . After the mice were killed, the thymus, spleen, and peripheral lymph nodes were carefully dissected and were either used immediately or frozen at –70°C until use.
PE-labeled anti-CD8a (BD Biosciences, San Jose, CA, USA) and fluorescein isothiocyanate (FITC)-conjugated anti-mouse CD4 (CALTAG Laboratories, Bangkok, Thailand) were obtained commercially. Anti-Neu2 rabbit polyclonal antibody was produced by immunizing a synthetic peptide corresponding to the Glu39-Ser58 segment of the rat cytosolic Neu2 , and the antisera preparation was purified using an affinity column coupled with the synthetic peptide. Commercial anti-Neu2 (M-13) sc-168736 goat polyclonal antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Rabbit non-immunized serum was prepared in our laboratory. COS-1 cells were transfected with Neu1, Neu2C, Neu3, Neu4S, or Neu4L cDNA, and the homogenates were prepared and used as their respective enzyme sources. COS-7 cells transfected with the Neu2B plasmid or the Neu2C plasmid were prepared as described previously .
Measurement of sialidase activity
Sialidase activity using 4MU-Neu5Ac as a substrate was assayed by the spectrofluorometric measurement of released 4MU in a large-scale method with a final volume of 1 mL (Hitachi spectrofluorometer, Tokyo, Japan)  or in a small-scale method with a final volume of 0.2 mL using a SpectraMax M5 multimode micro-plate reader (Molecular Devices Japan, Tokyo, Japan) . With substrates other than 4MU-Neu5Ac, the amount of released sialic acid was determined by the thiobarbituric acid method  or by fluorometric high-performance liquid chromatography with malononitrile .
SDS-PAGE and Western blot analysis
NP40-solubilized protein from the crude membrane fraction from normal or lpr thymus was applied for SDS-PAGE under reducing conditions with 10% acryl amide gel with the size marker (APRO Science, Naruto, Tokushima, Japan) and transferred to a polyvinylidene fluoride (PVDF) membrane (Hybond-P; Amersham Pharmacia Biotech, Little Chalfont, UK) . A sheet was treated with 0.1 N trifluoroacetic acid at 65°C for 2 hr  to remove sialic acid and all membranes were subsequently blocked with 1% BSA in Tris-Tween-saline (TTS, 10 mM Tris-HCl, pH 7.5/0.05% Tween/0.15 M NaCl). After washing with TTS, the blot was stained with biotin-labeled lectin and StAv-horseradish peroxidase. Enhanced chemiluminescence (ECL) generated with ECL plus reagent was detected using X-ray film.
Extraction and analysis of ganglioside from thymus
The frozen thymuses (500 mg wet weight) were homogenized and lipids were extracted with C/M by an orthodox method and applied to DEAE column chromatography. After the alkaline methanolysis of the acid glycolipid fraction, gangliosides were separated by TLC with C/M/0.5% CaCl2 aq. (55:45:10) solution and observed with resorcinol reagent. The residue after C/M extraction was solubilized by alkaline treatment, and the protein amounts were determined.
Preparation of the thymus homogenate
The thymus homogenates were prepared according to the method described previously . Briefly, frozen or fresh thymuses (6- to 9-week-old mice) were homogenized with a Dounce homogenizer using 100 volumes of 1 mM NaHCO3/0.1 mM CaCl2. After 5 min of slow mixing, the filtrate that had been passed through three layers of nylon mesh was used as the total homogenate and was centrifuged at 450g for 10 min; next the supernatant was used as the thymus homogenate.
Preparation of the S, N, and D fractions from the thymus homogenate
The thymus homogenate described above was centrifuged at 105,000g for 1 hr, and the soluble (S) fraction and the precipitate were obtained. The precipitate was designated the crude membrane fraction. The crude membrane fraction was solubilized using sonication buffer, DNase, and NP-40-solubilizing buffer (10 mM Tris-HCl, pH 7.4, 0.15 M NaCl containing 0.5% NP-40, 200 µg phenylmethylsulfonyl fluoride/mL)  on ice with stirring for 30 min. After centrifugation at 90,000g for 30 min, the supernatant was designated the N fraction. The insoluble precipitate was solubilized again using NP40/DOC solubilizing buffer (2.5 mM DOC in the NP-40-solubilizing buffer), and the solubilized fraction after the centrifugation described above was designated the D fraction.
Subcellular fractionation of the thymus crude membrane
The crude membrane fraction described above was suspended in PBS by sonication, overlaid on 35% sucrose in PBS and centrifuged at 24,000g for 1 hr using a Hitachi swing rotor SW28 (Hitachi, Tokyo, Japan). Four fractions, the upper, middle, lower, and precipitate fractions, were separated and mixed with PBS and centrifuged at 105,000g for 1 hr. Each precipitate was then suspended in PBS and sialidase activity was assayed with 4MU-Neu5Ac at pH 7.0 and pH 4.5 in citrate-phosphate buffer. The sialic acid released from the other substrate was assayed by the thiobarbituric acid method .
Washing of the crude membrane
The crude membrane fraction was washed according to the method described previously . The sialidase activity in the precipitate and the supernatant fractions was determined using 4MU-Neu5Ac as a substrate.
Preparative isoelectric focusing
The S, N, and D fractions prepared from pooled frozen thymuses described above were individually added to a solution containing 0.8 mL Ampholine pH 3 − 10 and 0.4 mL each of Ampholine pH 4 − 6, pH 6 − 8, and pH 8 − 9.5. The volume of each sample was adjusted to 40 mL with water and subjected to IEF using a Rotofor cell system (BioRad, Hercules, CA, USA) at 4°C for 4 − 4.5 hr at 12 W with 1 M H3PO4 as the anode solution and 1 M NaOH as the cathode solution. The solution in the Rotofor cell was separated into 20 tubes, the pH was measured and the protein concentration was determined using the Bio-Rad Protein Assay Kit. The first five and last four fractions were dialyzed against 10 mM Tris buffer (pH 7.4) for 4 hr to remove H3PO4 or NaOH, concentrated using dialysis tubing covered with dried Sephadex beads (Pharmacia Fine Chemicals, Uppsala, Sweden), and dialyzed against PBS for 30 min before assaying for sialidase activity.
Preparation of T cells, non-T cells, PNA-aggregated and PNA-unaggregated T cells
The freshly removed thymus was gently crushed by hand using a loose frosted glass homogenizer in RPMI medium and the cell suspension was passed through three layers of nylon mesh. The passed cells were denominated T cells. The cells remaining on the nylon mesh were designated non-T cells. The T cells were divided into PNA-aggregated and PNA-unaggregated T cells by the PNA method as described previously [5, 19]. PNA-aggregated cells consisted of immature T (CD4+CD8+ DP) cells and PNA-unaggregated T cells consisted of mature T (CD4+ SP and CD8+ SP) cells and immature CD4–CD8– DN cells .
T-cell sorting into DN, DP, CD4+ SP, and CD8+ SP cells
T cells prepared from the four thymuses of C57BL/6 mice (6 weeks old) were labeled with FITC-conjugated anti-mouse CD4 and phycoerythrin (PE)-labeled anti-CD8a in 1.5 mL PBS/0.1% BSA/0.1% NaN3 for 30 min on ice and washed two times in PBS, suspended in 2 mL PBS/0.1% BSA/0.1% NaN3, and then sorted into DN, DP, CD4+ SP, and CD8+ SP cells using a BD FACSAria™ II flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA).
Sialidase activity level in various organs with age
Figure 1 shows the level of sialidase activity with 4MU-Neu5Ac in the various organs and age. The sialidase activity level in the thymus at pH 7.0 was the highest among the organs tested for activity in both of the soluble and crude membrane fractions (Fig. 1a). Although the weight of the thymus did not decrease at 17 weeks (Fig. 1b), the activity level in the soluble fraction at pH 7.0 decreased to less than half of the maximum at 17 weeks (Fig. 1c): levels in the crude membrane fraction at pH 7.0 and 4.5 did not decrease. The spleen and lymph node exhibited little activity (Fig. 1c). Thus, the high levels of sialidase activity at pH 7.0 observed in the soluble and crude membrane fractions are unique in the thymus, and the decrease in the activity level at approximately 17 weeks is unique in the soluble fraction of the thymus.
Natural substrates for the sialidases in the S, N, and D fractions obtained from the thymus homogenate
We obtained the S, N, and D fractions from the thymus homogenate and investigated the activity level of sialidase with five substrates (Table 1). The N fraction exhibited activity with fetuin only at pH 7.0 but not at 4.5, and it was higher than the activity level with sialyllactose, indicating that the N fraction was not a mixture of the S and D fractions. In Figure 2, the crude membrane fraction from the thymus (Fig. 2a) or gangliosides (Fig. 2b) was used as a natural substrate. The N fraction exhibited its highest activity at pH 5.7 (Fig. 2a). The crude membrane fraction was found to contain mainly sialoglycoproteins, as well as sialoglycolipids (gangliosides) to a lesser extent, as available substrates for sialidases. Therefore, we next examined sialidase activity with mixed gangliosides as the substrate. Because mouse gangliosides primarily contain N-glycolyl-sialic acid, and bovine gangliosides mainly contain N-acetyl-sialic acid, we prepared the total mixed gangliosides from the mouse spleen for use as a substrate (Fig. 2b). The D fraction exhibited high activity with double peaks at pH 4.1 and 5.1. In summary, the N and D fractions preferentially hydrolyzed the sialic acid bond of glycoproteins and gangliosides on the thymus-derived crude membrane, respectively, and the S fraction hydrolyzed both at a low level at neutral pH.
|Fraction||Sialidase activity (nmolSA released per mg protein per 2 hr)|
|pH 7.0||24.0 ± 1.6||0.85 ± 0.12||1.74 ± 0.48||0||0|
|pH 4.5||1.6 ± 0.2||0||0.50 ± 0.44||0||0|
|pH 7.0||20.4 ± 0.1||4.4||2.25||5.48||0.90|
|pH 4.5||10.5 ± 2.0||0||10.56||2.16||4.41|
|pH 7.0||17.6 ± 2.0||0||0.92||4.76||0.11|
|pH 4.5||26.5 ± 4.7||0||7.05||1.72||6.84|
Candidates of the natural substrates studied with lymphoproliferation thymus
To investigate the substrates in the crude membrane, we used normal and lpr mice whose FAS gene was destroyed with a transposon [21, 22]. The thymus and lymph nodes of the lpr mice contained abnormal T cells that did not die as a result of the lack of apoptosis. Therefore, the lpr thymus was expected to have more enriched endogenous substrates. We analyzed glycoproteins on Western blots by lectin staining and gangliosides by TLC (Fig. 3). Several bands were clearly stained by SSA (Sambucus sieboldiana agglutinin, recognizes Siaα2-6 structure) in the lpr lane (indicated by arrowheads; Fig. 3a1). However, these bands were also detected by PNA (Arachis hypogaea (peanut) agglutinin, recognizes Galβ1-3GalNAc-Thr/Ser, O-glycan type; Fig. 3a2), most likely as a result of the presence of certain glycoproteins that contain multi-antennae glycans or several glycan chains on a molecule, and one has sialic acid in the terminal of the glycan, whereas the other does not. Therefore, the same molecule is detected with PNA and also with SSA lectins. More bands were also detected in lpr than in normal mice with RCA (Ricinus communis agglutinin recognizes terminal Galβ1-4 bond; Fig. 3a3), and the band marked by an arrowhead in Figure 3a4 was strongly stained after the removal of sialic acid by an acid treatment of the sheet. Thus, several candidates were demonstrated to be natural glycoprotein substrates for the sialidase. Next, we analyzed the ganglioside components of the thymus from normal and lpr mice (Fig. 3b). Bands 1 − 4 were detected more strongly in lpr than in normal mice. Next, using the crude membrane fraction from normal and lpr thymus as a substrate, we evaluated the activity level of the five types of sialidase using the plasmid-transfected COS-1 cells (Fig. 3c) to determine whether or not the lpr sample is indeed a better substrate than the normal one. All sialidases exhibited a higher level of activity against the crude membrane fraction from lpr than that from normal mice except Neu4a (L) (Fig. 3c1). The results with alternative transcripts Neu2B and Neu2C are shown in Figure 3c2. Thus, the candidate glycoproteins and gangliosides in the crude membrane described above were used as the substrates. Figure 3c3 shows the activity levels of five types of plasmid-transfected COS-1 cells with 4MU-Neu5Ac or with GM3 ganglioside. Neu2 preferred 4MU-Neu5Ac over other sialidases. When we compared the results in Figure 3c1 and c3, it is not necessarily true that the activity level with 4MU-Neu5Ac reflects the level that would be observed with a natural substrate.
Characterization of the S, N, and D fractions by preparative IEF
The S, N, and D fractions were analyzed by IEF to characterize the sialidase in the N fraction. A representative example is shown in Figure 4. All of the sialidase-positive fractions in the S fraction exhibited activity at pH 7.0 (Fig. 4S). Thus, the S fraction contains only soluble cytosolic Neu2 sialidase that exhibits optimal activity at neutral pH. The N fraction (Fig. 4N) exhibited two components. The major fraction exhibited higher activity at pH 7.0 than at pH 4.5 and exhibited a similar or slightly higher pI than Neu2 (tubes #12 − 16), and the minor fraction exhibited higher activity at pH 4.5 than at pH 7.0 and exhibited a more acidic pI than Neu2 (tubes #6 − 7 and #10). Although the D fraction also contained two different groups, the major activity in this case was in the acidic fraction (tubes #6 − 11). After IEF fractionation, we investigated the sialidase activity using fetuin, sialyllactose, mucin, and mixed gangliosides from bovine as substrates. Fractions #9 and #10 of the S fraction, fractions #12 − 14 of the N fraction and fractions #10 and #11 of the N fraction, and #7 − 10 of the D fraction were pooled and used as S, N basic, N acidic, and D enzymes, respectively. No activity (after the removal of substrate and enzyme blanks) was detected with these substrates, although we were able to detect the activity with 4MU-Neu5As using S, N basic, N acidic, and D enzyme at 37.6, 19, 1.7, and 1.9 nmoles/mg per 2 hr at pH 7.0 and at 8.5, 3.6, 2.4, and 1.5 nmoles/mg per 2 hr at pH 4.5, respectively.
The pH of the same tube number of each fraction was not the same. Tube #9 of the S fraction and tube #12 of the N fraction had almost the same pH at 6.6 and 6.8, respectively. Based on the amino acid sequence of the four cloned murine sialidases [3, 7, 23, 24], the theoretical pI values were calculated as 5.67, 7.65, 6.28, and 6.06 for Neu1, Neu2, Neu3, and Neu4, respectively, by the ‘Swiss plot’ computer method based on the pKa values of amino acids . If we assume that the pH values at the maximum sialidase activities separated by IEF (e.g. tube #9 in S and tubes #10 and #7 in D) correspond to the theoretical pI values, the enzyme of tube #9 in S is Neu2 (the most basic protein), and the enzymes in tubes #10 and #7 in D are Neu3/Neu4 and Neu1 (most acidic), respectively. It seems to be reasonable that tube #12 from the N fraction is Neu2. However, it is important to note that the theoretical pI may not be the actual pI of the enzymes. We conclude, however, that the basic sialidase in the N fraction that differed from Neu1, 3, and 4 is the main enzyme responsible for the membrane-associated activity toward 4MU-Neu5Ac observed in the thymus at a neutral pH.
Distribution of sialidase activity in the subcellular fractions
We next determined the distribution of sialidase activity by centrifugation under conditions used to obtain the plasma membrane. The results obtained with 4MU-Neu5Ac as a substrate at pH 7.0 and pH 4.5 are summarized in Table 2. Before sucrose fractionation, the activity level at pH 7.0 was higher than that of pH 4.5. After sucrose fractionation, the activity at pH 7.0 was lower than that at pH 4.5 in all of the fractions, and the recovery of the activity at pH 7.0 and that of the proteins was low. This result may be explained as follows: a portion of the activity at pH 7.0 with 4MU-Neu5Ac in the crude membrane fraction might result from the interaction between Neu2 and the membrane structure, and Neu2 was dissociated from the membrane in 35% sucrose or the washing process. Conversely, no loss in activity was observed at pH 4.5 because the membrane-bound enzymes exhibited the activity.
|Sialidase activity†||Recovery of activity (%)||Recovery of protein (%; average of two cases)|
|pH 7.0 / pH 4.5||pH 7.0 / pH 4.5|
|Total homog.||14.9 / 10.4||100 / 100||100|
|450P||2.3 / 3.0||3.9 / 5.9||15.7 ± 2.8|
|450S||14.3 / 14.5||88.5 / 105||78.2 ± 10.1|
|S (105k sup)||10.8 / 3.8||51.4 / 21.2||47.3 ± 5.1|
|CM (105k ppt)||20.6 / 18.7||46.5 / 60.4||33.7|
|Upper||4.8 / 11.2||1.3 / 3.7||2.5 ± 0.6|
|Middle||21.5 / 63.4||5.6 / 19.5||2.7 ± 0.1|
|Lower||21.5 / 73.8||1.9 / 7.5||1.2 ± 0.2|
|Ppt||10.8 / 27.8||9.8 / 29.6||12.7 ± 2.7|
Sialidase activity was also analyzed in the middle layer fraction that was expected to enrich with plasma membrane (Table 3). Among the substrates tested, only fetuin exhibited a higher value at pH 7.0 than that at pH 4.5. Exp. #1 in Table 3 has a higher ratio of alkaline/acid phosphatase activity and a higher level of sialidase activity with fetuin at pH 7.0 than Exp. #2. This result might suggest that the enzyme that released sialic acid from fetuin at pH 7.0 existed in the plasma membrane.
|Sialidase activity with ‘middle’ fraction (nmol/mg per 2 hr)|
|Exp. #1||Exp. #2|
|pH 7.0 / pH 4.5||pH 7.0 / pH 4.5|
|4MU-Neu5Ac||11.8 / 38.8||8.6 / 55.1|
|Fetuin||6.6 / 0||3 / 1.2|
|Sialyllactose||3.6/22.6||1 / 54|
|GM3||ND / ND||20.5 / 126.7|
|Oroso mucoid||ND / ND||0 / 0|
|Alk. / Acid P||88.5||18.8|
Main membranous sialidase at neutral pH containing Neu2 is associated with the membrane structure but is not anchored
As mentioned above, Neu2 appeared to interact with the membrane in the crude membrane fraction. We have already reported that the crude membrane fraction from COS-7 cells that were transfected with the Neu2-plasmid released its sialidase activity following three washes with 10 mM Tris-HCl, pH 7.0 . Thus, we carried out the same experiment with the crude membrane fraction from the mouse thymus. The results were similar to those observed with the COS cells; the first wash had little or no effect on the activity, whereas three washes reduced the activity to the residual level (approximately 40%) of the crude membrane fraction after solubilization with NP-40 (Fig. 5a). We also detected sialidase activity in the washed solution (data not shown). These observations suggest that sialidase solubilized with NP40 is associated with the membrane structure of the thymus but not anchored in the membrane by a signal peptide or GPI-anchor, for example, trans-sialidase .
To determine whether Neu2 is responsible for the sialidase activity in the crude membrane fraction at neutral pH with 4MU-Neu5Ac, we carried out an immunoprecipitation experiment. Incubation of anti-Neu2 rabbit polyclonal antibody with the S or N fraction reduced a portion of the activity (12% or 7% reduction, respectively; Fig. 5b), although the extent of the reduction with the S fraction (positive control) was not sufficient. We examined the same experiment with a commercially available anti-Neu2 goat polyclonal antibody purchased from Santa Cruz Biotechnology, but the reduction of the activity was less than that in the above experiment. At this time, we cannot say definitively that Neu2 is responsible for the sialidase activity in the crude membrane fraction at neutral pH with 4MU-Neu5Ac, although we can state that Neu2 is partially responsible for this activity.
SMXA recombinant mouse studies suggest that the Neu2 gene controls sialidase activity in the crude membrane fraction at pH 7.0
Next, we tried to access how the sialidase gene controls the activity in the crude membrane fraction at pH 7.0 using a recombinant inbred strain of SM and A mice (SMXA). We reported that the level of sialidase activity in the soluble fractions from 22 substrains of SMXA mice exhibited good correlation with the D1Mit8/9 marker, which is located close to the Neu2 gene on mouse chromosome 1 . Therefore, in the present study, we determined the activity level of the crude membrane fraction from the 22 substrains at pH 7.0 and observed which gene best correlated with the activity. Although the activity levels in the crude membrane fractions from the parental SM and A/J mice did not exhibit any differences at neutral pH, the membrane-bound activities from 22 substrains of the SMXA were well correlated with the genotype of the D1Mit8/9 locus, except for one substrain (SMXA #19; Fig. 6). In contrast, the activities did not correlate with the genotype of Neu1, Neu3, or Neu4. This finding shows that the sialidase activity of the crude membrane fraction at pH 7.0 is most likely regulated by the Neu2 gene by an unknown mechanism.
Sialidase activity in the sorted cells
Finally, we attempted to identify the sialidase-positive cells in the thymus. First, we separated the cells from the thymus into T cells and non-T cells. Next, T cells were separated into PNA-aggregated and PNA-unaggregated T cells. The sialidase activity at pH 7.0 and pH 4.5 in each of the cell types was investigated using 4M-Neu5Ac as a substrate (Table 4). Non-T cells (‘neu-medullocytes’ (5) are most likely contained in this fraction) exhibited clearly higher activity than T cells at pH 7.0 and 4.5 (Table 4). Next, T cells were sorted into DP cells (corresponding to the PNA-aggregated cells in Table 4), CD4+ SP, CD8+ SP, and DN cells (these three cell types correspond to the PNA-unaggregated cells in Table 4) using a cell sorter, and the activity level of sialidase was determined (Table 5). Few differences in activity were observed in the soluble fractions among the four types of T cells, whereas the crude membrane fractions from CD8+ SP cells exhibited slightly higher activity than the other T cell types.
|Sialidase activity (4MU released nmol/mg protein per 2 hr)|
|Activity at pH 7.0||Activity at pH 4.5|
|Soluble fr.||Membrane fr.||Soluble fr.||Membrane fr.|
|PNA-agg. T||1.78 ± 0.00||1.64 ± 0.30||0.82 ± 0.41||1.28 ± 0.18|
|PNA-unagg. T||0.78 ± 0.40||1.86 ± 0.21||0.78 ± 0.26||1.04 ± 0.76|
|Non-T||9.16 ± 2.10||5.23 ± 0.52||0.60 ± 0.47||8.79 ± 0.34|
|Relative activity (%) to the activity with CD4|
|Exp. #1||Exp. #2|
|Soluble fr.||Membrane fr.||Soluble fr.||Membrane fr.|
The present study aimed to determine whether soluble sialidase Neu2 is responsible for the high level of sialidase activity at neutral pH with 4MU-Neu5Ac in the crude membrane of the thymus. We examined this question from several perspectives and determined that it is likely that the high level of membranous sialidase activity at neutral pH is due to soluble sialidase Neu2 associated with the cellular membranous fraction by some unknown mechanism. This conclusion is supported by the results obtained by IEF analysis (Fig. 4), the membrane fraction washing experiment (Fig. 5a), and the activity-regulating gene results suggested by our experiments with SMXA recombinant mice (Fig. 6). Unfortunately, we could not show a clear-cut result with the anti-Neu2 antibody used in our immunoprecipitation experiments or by Western blot analyses. One point to be stressed is that when using 4MU-Neu5Ac as a substrate for the assay of sialidase activity, the results will primarily reflect Neu2 activity as shown in Figure 3c. Another important point is that we found that the plasma membrane-rich fraction from the mouse thymus contained a sialidase that removed sialic acid from fetuin at pH 7.0 but not at pH 4.5. This activity is unique among the sialidases to date, indicating that this enzyme perhaps removes sialic acid from a glycoprotein on another T cell or the glycoprotein existing near the sialidase on the plasma membrane. This sialidase must have an important physiological function.
Next, we discuss the physiological function of the cytosolic sialidase in the thymus. In terms of catabolism, a cytoplasmic peptide: N-glycanase (PNGase) is reported to release glycans from newly synthesized misfolded glycoproteins . The released glycans are hydrolyzed by cytosolic glycosidases and, very recently, the dual functions of cytosolic glycosidase during glycan catabolism and apoptotic signaling were reported . Although the function of glycosidase as an apoptotic signal was reported [29, 30], this function is unclear in the thymus. In the thymus, apoptosis occurs frequently and the degradation of glycoproteins by cytosolic glycosidases must play an important role. Thus, it might be important that Neu2B sialidase in the thymus has a higher activity with sialyllactose compared to Neu2C sialidase in the muscle (5.8-fold for α2-6sialyllactose and 8.6-fold for α2-3sialyllactose ). Another point that we wish to stress is that a transit state of some cellular proteins may carry out some important physiological function. For example, it may be possible that the association of Neu2 with the membrane structure or Neu2's stability is affected by some conditions; that is, it is metabolically stable under some conditions but becomes unstable upon a change in the cell's state, for example, apoptosis. Biochemically, the labile character of the murine sialidase (e.g. 1.1 − 1.3 hr of half-life) was reported . The excess six amino acids (if they are translated in the enzyme) might affect cytosolic Neu2 during this type of physiological change. For example, it associates with the membrane structure and then dissociates after carrying out a specific function, as we discussed in the case of tetraspanin . Indeed, we demonstrated the expression of active Neu2 on the plasma membrane by X-NANA active staining of Neu2B transfected COS cells . The sialidase activity with fetuin in the plasma membrane may be one such example. The majority of lysosomal sialidase Neu1 relocalizes from the lysosomes to the cell surface during the differentiation of monocytes . Studies of type B Neu2 in the thymus may provide additional information.
More than half of this study was carried out at the Institute of Immunology/Genetic Medicine, Hokkaido University, with the support of a Grant-in-Aid for Scientific Research (C)  (07808069) from the Ministry of Education of Japan (1995 − 1996) until 2001, when S.K.-O, N.D., and M.F. were present. The rest of the study was conducted at the laboratory of Prof. T. Koda (Hokkaido University). Some of the study was conducted at Tohoku Pharmaceutical University (Sendai). We are grateful to Prof. J. Inokuchi (Tohoku Pharmaceutical University, Sendai) for introducing Dr K. Kabayama and Dr S. Go to conduct TLC and Western blot analyses, respectively. We are grateful to Ms H. Matsukawa (deceased) for her work on Figure 6 when she was at the Institute of Immunology, Hokkaido University. We are grateful to Dr K Azumi (homogenizer), Prof. Y. Igarashi, and Dr S. Mitsutake (ultracentrifuge, sonicator), and Prof. S. Nishimura (spectrofluorometer) at Hokkaido University for the use of their laboratory equipment, as well as the Central Institute of Isotope Science for storing materials at –70°C after S.K.-O.'s retirement. We are grateful to Prof. M. Nishimura (retired from Nagoya University) for his kind donation of SMXA recombinant mice and Dr K. Ogasahara (Institute for Protein Research, Osaka University) for suggesting the use of the Swiss Plot method to calculate the pI values.
No authors have any conflict of interest to disclose.