| Glycobiology | Pages |
[beta]1,6 N-Acetylglucosaminyltransferase (core 2 GlcNAc-T) expression in normal rat tissues and different cell lines: evidence for complex mechanisms of regulation
Introduction
Results
Discussion
Materials and methods
Acknowledgments
Abbreviations
References
[beta]1,6 N-Acetylglucosaminyltransferase (core 2 GlcNAc-T) expression in normal rat tissues and different cell lines: evidence for complex mechanisms of regulation
The distribution of the Golgi enzyme [beta]1,6-N-acetylglucosaminyltransferase (core 2 GlcNAc-T for short) has been investigated in several tissue and cell systems by combining the potentials of a polyclonal antibody and a novel, sensitive fluorescent enzyme assay. In normal rat tissues, levels of the protein were found to vary and as a general trend did not correlate with enzyme activities. Additionally, we observed tissue-specific core 2 GlcNAc-T forms of various size: 75 kDa (liver), 70 kDa (spleen), 60 kDA (heart), and 50 kDa (heart and lung). These forms might arise from differential protein modifications; alternatively, the smaller form may be a product of proteolytic cleavage, given the presence of a catalytically inactive 50 kDa species in rat serum. Chinese hamster ovary (CHO), MDAY-D2, PSA-5E, and PYS-2 cell lines consistently displayed a 70 kDa enzyme. When induced to retrodifferentiate in the presence of butyrate + cholera toxin, CHO cells exhibited a 21-fold increase in enzyme activity, while protein levels remained constant. A similar trend was observed in the embryonal endoderm cell lines PSA-5E and PYS-2, where an approximately 100-fold difference in core 2 GlcNAc-T activity was found notwithstanding unchanged amounts of the protein and identical mRNA levels, as evidenced by RT-PCR. In contrast, levels of core 2 GlcNAc-T activity in MDAY-D2 cells correlated well with protein expression. Taken together, these observations demonstrate that core 2 GlcNAc-T expression may be subjected to multiple mechanisms of regulation and suggest that in at least some instances (i.e., PSA-5E and PYS-2 cells) expression may be regulated exclusively via posttranslational mechanism(s) of control.
Key words: [beta]1,6 N-acetylglucosaminyltransferase/antibody/ expression/fluorometric assay/Western blotting
Introduction
It is now acknowledged that the selective expression of specific glycosyltransferases plays an important role in the appearance of tissue-specific oligosaccharide patterns (Rademacher et al., 1988; van den Eijnden and Joziasse, 1993; Colley, 1997). Many different theories have been advanced regarding the functional roles of carbohydrates during processes of development, differentiation, transformation and cell-cell recognition (Varki, 1993). An understanding of the mechanism(s) by which glycosyltransferases are controlled should yield information regarding oligosaccharide function and permit further studies via targeted modification of specific structures.
Studies at a molecular level have shown that these enzymes may be subjected to tissue- and cell type-specific regulation at the level of gene expression and via posttranslational modifications. Some glycosyltransferases are transcribed from multiple promoter elements; others undergo alternative mRNA splicing or utilize variant translational initiation or termination sites which may lead to the production of different proteins with potentially distinct activities or substrate specificities (Kleene and Berger, 1993; van den Eijnden and Joziasse, 1993). Tissue-specific regulation may also be achieved via modulation of mRNA levels (Svensson et al., 1990). Recent reports that activity levels of glycosyltransferases do not always correlate with levels of message suggest that expression might be modulated translationally or posttranslationally or at points further downstream (Kudo and Narimatsu, 1995; Baum et al., 1996). Of the latter, possible mechanisms include control of subcellular enzyme localization (Colley, 1997; Skrincosky et al., 1997), interenzyme competition for common substrates, modulation of substrate synthesis/transport or control of acceptor accessibility through conformational changes (Schachter, 1986; Watkins et al., 1988; Schachter and Brockhausen, 1992).
The Golgi enzyme UDP-GlcNAc:Gal[beta]1,3GalNAc-R (GlcNAc to GalNAc) [beta]1,6-N-acetylglucosaminyltransferase (i.e., core 2 GlcNAc-T, EC 2.4.1.102) defines a branch point in the O-linked glycan biosynthesis pathway by converting core 1 (i.e. Gal[beta]1,3GalNAc[alpha]-O) to core 2 structures (i.e. Gal[beta]1,3[GlcNAc[beta]1,6]GalNAc[alpha]-O; Williams and Schachter, 1980; Schachter and Brockhausen, 1992). The action of core 2 GlcNAc-T represents an important regulatory step for the extension of O-linked sugars with poly(N-acetyllactosamine) (Yousefi et al., 1991), repeats of Gal[beta]1,4GlcNAc[beta]1,3 which have been associated with malignant transformation (Yousefi et al., 1991) and proliferative activation of lymphocytes (Higgins et al., 1991). Nonetheless, it was recently shown that entry to the synthesis of branched, complex core 2-based O-linked structures is controlled by the relative levels of activity of core 2 GlcNAc-T and [alpha]-2,3 sialyl-T (Whitehouse et al., 1997), both of which compete for the same core 1 acceptor substrate. Polylactosamine structures have been shown to affect cellular adhesion (Zhu and Laine, 1985; Laferté and Dennis, 1988) and may act as ligands for mammalian lectins (Merkle and Cummings, 1988). Thus, core 2 GlcNAc-T may be a key enzyme in the modulation of cell-cell interactions through glycosylation of specific target molecules. In this regard, core 2 GlcNAc-T-mediated O-glycosylation of PSGL-1 was found to be a critical step for binding to P-selectin (Kumar et al., 1996; Li et al., 1996).
Core 2 GlcNAc-T activity appears to be regulated by factors which have an impact on intracellular signaling and developmental status of the cell. For example, marked increases in enzyme activity are associated with T cell activation via the T cell receptor complex in vitro (Piller et al., 1988; Higgins et al., 1991; Datti et al., 1994) and all-trans retinoic acid-induced differentiation of F9 teratocarcinoma cells (Heffernan et al., 1993). Additionally, it has been postulated that specific upregulation of core 2 GlcNAc-T activity in the cardiac tissue of diabetic rats might be induced through the diacylglycerol-protein kinase C pathway (Nishio et al., 1995). Developmental regulation is evidenced by the clear temporal and spatial patterns of expression seen via in situ RNA hybridization studies in postimplantation mouse embryos (Granovsky et al., 1995).
Although the mechanism(s) controlling core 2 GlcNAc-T expression is currently unclear, several observations have suggested that the enzyme might be regulated at the posttranslational level. For example, the significant elevation of core 2 GlcNAc-T activity exhibited by acute myeloblastic leukemic cells, as compared to their normal counterparts, has been associated with changes in the kinetic properties of the enzyme (Brockhausen et al., 1991). A similar observation was made during upregulation of the enzyme in CHO cells subjected to retrodifferentiation in the presence of sodium butyrate and cholera toxin (Datti and Dennis, 1993). In this system, it was found that large increases in enzyme expression were dependent upon de novo transcription/translation as well as activation of protein kinase(s).
In this study, we have examined the expression of core 2 GlcNAc-T at the protein level and compared enzyme levels with activities in a variety of normal rat tissues and cell lines to determine whether regulation is achieved through posttranslational mechanisms of control. To this end, we produced a specific, affinity-purified polyclonal antibody against recombinant human core 2 GlcNAc-T protein. The results provide several lines of evidence that core 2 GlcNAc-T is subjected to different mechanisms of regulation, acting either separately or in concert. Nonetheless, it is suggested that control of enzyme expression at the posttranslational level represents an important event leading to dramatic changes in activity.
Results
Production and purification of anti-human core 2 GlcNAc-T antibodies
To produce polyclonal antisera against human core 2 GlcNAc-T, a 58 kDa fusion protein was generated, encoding amino acids 84-372 of the human protein fused to bacterial glutathione-S-transferase (GST). Synthesis of the fusion protein was induced by IPTG treatment of host cells, as described in Material and methods. SDS-PAGE analysis of bacterial culture pre- and postinduction and supernatant and pellet fractions indicated that over 95% of the chimeric protein was recovered as insoluble inclusion bodies (not shown). The protein yield was estimated to be 30 µg/ml of bacterial culture. After immunization of rabbits with the purified fusion protein, antibody titer in the whole serum was monitored by immunoblotting analyses using the fusion protein as the target. A second chimeric protein, T7 gene 10-core 2 GlcNAc-T (molecular weight 33 kDa), was included on the blots in order to distinguish anti-core 2 GlcNAc-T antibodies from reactivity against GST epitopes. To eliminate any spurious antibodies present in serum, antisera were subjected to immunoaffinity purification via glycine elution from nitrocellulose strips containing T7-core 2 GlcNAc-T fusion protein, as outlined in Materials and methods. The purity and utility of the putative anti-core 2 GlcNAc-T antisera were assessed via Western blotting using an HL60 lysate, which is known to exhibit relatively high levels of enzyme activity (~1.70 nmol/mg/h; Palmerini et al., 1995). As expected, the immunopurified preparation contained fusion protein-specific antibodies (Figure
Figure 1. Specific detection of core 2 GlcNAc-T protein in a complex lysate of HL60 cells by Western blotting analysis. In panel I, affinity-purified core 2 GlcNAc-T antibody was tested for its ability to detect core 2 GlcNAc-T protein in a complex lysate of HL60 cells on Western blots by employing a sensitive chemiluminescent method (ECL immunodetection kit, Pharmacia). Lane D contained 200 µg of HL60 cell lysate; in lane A, 100 µg of BL21(DE3)plysS bacterial lysate was run as a negative control for immunoreactivity; in lanes B and C, 30 ng of GST-core 2 GlcNAc-T fusion protein and T7-core 2 GlcNAc-T fusion protein, respectively, were run as positive controls for immunoreactivity. In panel PI, the presence of nonspecific reactivity in serum was tested by performing immunodetection using a 1:1000 dilution of preimmune serum subjected to the core 2 GlcNAc-T antibody purification protocol. Lane E contained 200 µg of HL60 cell lysate and lane F was loaded with 30 ng of GST-core 2 GlcNAc-T fusion protein. Positions of protein standard markers are shown.
Distribution and expression of core 2 GlcNAc-T in normal rat tissues
The purified antibody was employed to study the distribution of core 2 GlcNAc-T in various rat tissues via Western blot analysis (Figure
Figure 2. Distribution and activity of core 2 GlcNAc-T protein in normal rat tissues. In (A) 200 µg of protein from testis, spleen, kidney, lung, liver, heart, and brain lysates were subjected to Western blotting analysis using affinity-purified core 2 GlcNAc-T antibody. Tissue lysates were prepared as described under Materials and methods. Positions of protein standard markers are shown. Results are of a representative experiment performed in duplicate. In (B) core 2 GlcNAc-T activity was measured in normal rat tissues via a sensitive, fluorometric assay method (Palmerini et al., 1996), as described under Materials and methods. Activities are expressed as the mean of specificity activity ± the range of triplicate determinations. The activity of the constitutive enzyme [beta]1,3 Gal-T was measured in all samples as a control of tissue lysate quality and observed to be robust in all preparations (i.e., testis, 4.24 nmol/mg/h; spleen, 4.01 nmol/mg/h; kidney, 12.78 nmol/mg/h; lung, 5.13 nmol/mg/h; liver, 10.99 nmol/mg/h; heart, 1.84 nmol/mg/h; brain, 2.74 nmol/mg/h). The intensities of the bands were quantified via densitometric scanning. Using the signal in lung for normalization (=1), relative amounts of core 2 GlcNAc-T in the other organs were 1.06 (spleen), 1.21 (heart), and 1.78 (liver). When compared to assays of core 2 GlcNAc-T activity in the tissue lysates (Figure In a preliminary attempt to discern possible relationships between the various core 2 GlcNAc-T forms, we investigated whether the larger species (70-75 kDa) might represent N-glycosylated proteins; however, the size of the HL60 core 2 GlcNAc-T protein was unchanged on Western blots after treatment of cell lysate with N-glycosidase F to remove N-linked sugar moieties (not shown). Successful N-carbohydrate removal was verified by a lack of coloration in a subsequent probe of the blot with ConA-HRP (Hawkes, 1982). Alternatively, given the existence of proteolytically cleaved, soluble forms of a number of glycosyltransferases (Paulson and Colley, 1989), it was possible that the smaller forms of core 2 GlcNAc-T could be truncated molecules. Accordingly, an aliquot of rat serum was subjected to Western blot analysis in order to assess whether a cleaved, soluble form of the enzyme existed (Figure Figure 3. Presence of core 2 GlcNAc-T protein in rat serum. An aliquot of rat serum containing 80 µg of protein was subjected to Western blotting analysis using affinity-purified core 2 GlcNAc-T antibody (lane A). A lysate of rat heart (200 µg protein) was run as a size reference (lane B). Serum was prepared as described under Materials and methods. Positions of protein standard markers are shown.
Table I.
| Cells | Specific activity (pmol/mg/h) |
| CHO | 14 |
| CHO + SB | 160 |
| CHO + CT | 16 |
| CHO + SB + CT | 295 |
| PYS-2 | 6250 |
| PSA-5E | 57 |
| MDAY-D2 | 15,400 |
Expression of core 2 GlcNAc-T in different cell systems
Chinese hamster ovary (CHO). In a previous study (Datti and Dennis, 1993) it was shown that core 2 GlcNAc-T activity was dramatically induced in CHO cells after retrodifferentiation mediated by treatment with sodium butyrate alone or in combination with cholera toxin. This system was exploited to determine whether the large elevations in enzyme activity reflected increased levels of core 2 GlcNAc-T protein, by using cell lysates for both measurement of enzyme activity and Western blot analyses. As can be seen in Table I, core 2 GlcNAc-T levels increased ~12-fold upon a 24 h treatment with 2 mM sodium butyrate and 21-fold upon simultaneous exposure to sodium butyrate and 100 ng/ml cholera toxin, while treatment with cholera toxin alone had no effect upon enzyme activity. The Western blot shown in Figure
Figure 4. Expression of core 2 GlcNAc-T protein in CHO cells cultured in the presence of differentiating agents. CHO cell cultures were treated with 2 mM sodium butyrate and 100 ng/ml cholera toxin, alone or in combination, as described under Materials and methods. Lysates containing 200 µgof protein were subjected to Western blotting analysisusing affinity-purified core 2 GlcNAc-T antibody. Lysates were: lane A, untreated CHO cells; lane B, CHO cells treated with sodium butyrate; lane C, CHO cells treated with cholera toxin; lane D, CHO cells treated with sodium butyrate + cholera toxin. Positions of protein standard markers are shown.PYS-2 and PSA-5E cells. Retinoic acid ([beta]-All-trans)-induced differentiation of F9 teratocarcinoma cells into distinct types of endodermal cells has been shown to be associated with changes in specific glycosyltransferase activities and expression of polylactosaminoglycans (Cummings and Mattox, 1988; Heffernan et al., 1993). Similar observations have also been described in the murine embryonal endoderm lines PYS-2 and PSA-5E, which represent parietal (Lehman et al., 1974) and visceral (Adamson et al., 1977) phenotypes, respectively. In particular, both cell lines exhibit elevated levels of [beta]1,6 GlcNAc-T V, [beta]1,3 Gal-T and [beta]1,4 Gal-T (Heffernan et al., 1993). Interestingly, core 2 GlcNAc-T activity was observed to be approximately 100-fold higher in PYS-2 than in PSA-5E cells (Datti et al., 1992), likely reflecting their distinct differentiation states. To determine whether the activities correlated with levels of core 2 GlcNAc-T protein, PYS-2 and PSA-5E cell lysates were analyzed for enzyme activity and subjected to Western blotting using anti-core 2 GlcNAc-T antibodies. Lysates of CHO and MDAY-D2 cells were also run for comparison purposes, since core 2 GlcNAc-T activities in these cells are to some extent comparable to those in PSA-5E and PYS-2 cells, respectively (Datti et al., 1992). Figure Figure 5. Expression of core 2 GlcNAc-T protein in the murine embryonal endoderm cell lines PSA-5E (visceral) and PYS-2 (parietal). Lysates containing 200 µg protein from PSA-5E cells (lane B) and PYS-2 cells (lane C) were subjected to Western blotting analysis using affinity-purified core 2 GlcNAc-T antibody. Lanes A and D contained, respectively, 200 µg of CHO and MDAY-D2 cell lysates, run for comparison. The lack of an immunoreactive band in the CHO cell lysate is due to the low exposure time necessary to obtain an optimal core 2 GlcNAc-T signal in the MDAY-D2 cells. Positions of protein standard markers are shown. As a means of confirming that the differences in enzyme activities manifested by PSA-5E and PYS-2 cells were not due to enhanced transcription of the core 2 GlcNAc-T gene, mRNA levels were assessed in both cell lines via RT-PCR and Southern blotting analysis. This technique has been previously exploited to study the expression of rare GlcNAc-T V transcripts in a variety of rat tissues (Perng et al., 1994). Pilot experiments established that 25 cycles of PCR amplification yielded nonsaturating product. Figure Figure 6. RT-PCR analysis of core 2 GlcNAc-T expression in murine embryonal endoderm cell lines. Total RNA (1 µg) from PSA-5E cells (lane A) and PYS-2 cells (lane B) was subjected to RT-PCR and Southern blotting analysis as outlined in Materials and methods. The position of the primers predicted an amplification product of 560 bp. Lane C represents a positive controlamplified from 1 ng of plasmid containing the 1700 bp murine core 2 GlcNAc-T probe. Positions of the 1 kb DNA ladder standards are shown.
Discussion
The expression of the Golgi glycosyltransferase core 2 GlcNAc-T has been studied, for the first time, at the protein level. To achieve this end, we produced polyclonal, affinity purified antibodies against the human enzyme. The antibodies did not exhibit narrow species-specificity as shown by their cross-reactivity with rat, murine and hamster enzymes. Indeed, such behavior is inconsistent with previous observations for antibodies to other mammalian glycosyltransferases (Ulrich et al., 1986; Burke et al., 1992; Teasdale et al., 1992) and can probably be attributed to the high amino acid sequence conservation across species. In this regard, both rat and mouse core 2 GlcNAc-Ts share ~85% amino acid sequence identity with their human counterpart (Nishio et al., 1995; Sekine et al., 1997).
We must consider the possibility, although remote, that the anti-core 2 GlcNAc-T antibodies recognize epitopes present on enzymes that are functionally and/or structurally related to core 2 GlcNAc-T. In fact, core 2 GlcNAc-T and the I antigen branching enzyme (another [beta]1,6 GlcNAc-T), both of the O-linked oligosaccharide synthetic pathway, have three regions of extensive homology within their catalytic domains (Bierhuizen et al., 1993). There are also indicia which suggest the existence of other glycosyltransferase genes belonging to the same family. For example, low-stringency Southern blotting analyses of human genomic DNA using core 2 and I antigen GlcNAc-T probes revealed at least one distinct, related sequence (Bierhuizen et al., 1993). In this regard, a possible candidate might be the gene encoding the enzyme which catalyses the addition of the [beta]1,6 GlcNAc branch to form the mucin-type core 4 structure GlcNAc[beta]1,3(GlcNAc[beta]1,6)GalNAc- (i.e., core 4 GlcNAc-T). Notwithstanding the existence of related enzymes, the antibody's specificity for core 2 GlcNAc-T is demonstrated by several observations. The antibody was able to mediate isolation of a single core 2 GlcNAc-T recombinant phage in a large-scale screening of a human cDNA expression library. Additionally, there was a general lack of background in all Western blotting analyses.
Our investigation of core 2 GlcNAc-T expression at the protein level suggests that the enzyme is subject to complex mechanism(s) of control, a conclusion which emerged upon comparison of amounts of protein and catalytic activities of the enzyme in different biological systems. The reliability of this approach was based on pilot experiments which established that the Western blotting analyses were quantitative under the experimental conditions used.
In normal rat organs, the presence of several proteins having molecular weights ranging from 50-80 kDa is indicative of tissue-specific expression. The different forms might reflect distinct preferences towards endogenous substrates. In fact, at least two core 2 GlcNAc-T isoforms exist, each exhibiting particular substrate specificities: one present in mucin-producing cells (mucin-type), catalyzing the addition of the [beta]1,6 GlcNAc branch to produce core 2 and core 4 structures and the blood group I antigenic determinant (i.e., GlcNAc[beta]1,3[GlcNAc[beta]1,6]Gal[beta]-) (Brockhausen et al., 1986; Ropp et al., 1991), the other present in leukocytes (leukocyte-type) expressing only core 2 activity (Bierhuizen and Fukuda, 1992).
The sequence of the leukocyte-type, cloned human cDNA predicts a protein of 428 amino acids in length, with a molecular mass of 49,790 Da (Bierhuizen and Fukuda, 1992), and in fact, a core 2 GlcNAc-T enzyme purified from mouse kidney, thought to be encoded by the homolog of the cloned human gene, was found to have a molecular mass of 50 kDa (Sekine et al., 1994). By contrast, a mucin-type enzyme isolated from bovine tracheal epithelium was 69 kDa (Ropp et al., 1991). Taken together, these observations are in agreement with our findings of differently-sized tissue-specific forms.
How the various proteins arise is still unclear. It is doubtful that the different forms represent multiple proteins translated from alternatively spliced mRNAs, since the coding region of the human and mouse core 2 GlcNAc-T genes is restricted to a single exon (Bierhuizen et al., 1995; Sekine et al., 1997). Indeed, putative alternatively spliced core 2 GlcNAc-T mRNAs found in mouse kidney and submandibular gland differed only in their 5[prime]-untranslated sequence (Sekine et al., 1997).
It may be that the various enzyme forms reflect different glycosylation states of the protein. A preliminary investigation of core 2 GlcNAc-T in HL60 cells indicated that neither of the two consensus N-glycosylation sites on the protein are utilized. Although N-oligosaccharide removal was overall quite successful, it is possible that the core 2 GlcNAc-T enzyme is especially resistant to digestion. Interestingly, a recent study of the expression of a truncated form of the human enzyme in the baculovirus/Sf9 insect cell system verified the presence of N-glycans at both predicted sites (Toki et al., 1997). As an alternative to the presence of N-oligosaccharides, it is possible that the larger core 2 GlcNAc-T species may reflect other forms of posttranslational change, such as O-glycosylation.
The smaller core 2 GlcNAc-T form seen in heart and lung tissues could also represent a proteolytically cleaved, soluble enzyme, given the presence of a 50 kDa immunoreactive protein in rat serum. Analogous forms of several glycosyltransferases have been purified from milk, serum and other body fluids and are thought to result from release of membrane-bound enzymes by proteolytic cleavage between transmembrane and catalytic domains (Paulson and Colley, 1989; Schachter, 1991; Joziasse, 1992).Whereas these soluble enzymes generally retain catalytic activity, the putative soluble core 2 GlcNAc-T was found to be inactive, a fact which might be due to the presence of inhibitory factors in serum. Alternatively, it is possible that the protein plays some nonenzymatic role, for example by acting as a lectin. Such a potential function was previously suggested for a catalytically inactive, truncated form of [alpha]2,6 sialyltransferase found in rat kidney (O'Hanlon and Lau, 1992).
In normal rat tissues, the general lack of correlation between specific activity and protein levels indicates that regulation of core 2 GlcNAc-T expression is generally more complex than simple transcriptional/translational control. In this regard, another group reported that rat brain and kidney contained 3-fold higher levels of core 2 GlcNAc-T transcripts than lung (Nishio et al., 1995); however, our study revealed that the protein was undetectable in brain and kidney, but well-expressed in the lung. The same authors also described that the hearts of diabetic rats exhibited an 8-fold elevation of core 2 GlcNAc-T mRNA levels, as compared to normal rats, but only a 2-fold increase in enzyme activity. In combination with the data of the present study, it is therefore clear that the expression of core 2 GlcNAc-T in rat tissues is controlled posttranscriptionally in some tissues.
In all of the cell lines analyzed, core 2 GlcNAc-T exhibited a molecular mass of 70 kDa. An earlier report (Datti and Dennis, 1993) had shown that a dramatic increase in core 2 GlcNAc-T activity during butyrate-induced retrodifferentiation of CHO cells was paralleled by de novo gene transcription/translation, activation of protein kinase(s) and changes in core 2 GlcNAc-T kinetic parameters. These observations provided strong evidence that the enzyme was regulated, at least in part, at the posttranslational level, likely through mechanism(s) of phosphorylation/ dephosphorylation mediated by cAMP-dependent kinase(s). The current Western blotting analyses support such a theory, given the constant protein expression during CHO cell retrodifferentiation and induction of core 2 GlcNAc-T activity. The same trend was observed in the embryonal cell lines PYS-2 and PSA-5E, which display similar patterns of glycosyltransferase activities with the exception of core 2 GlcNAc-T. As in CHO cells, a >100-fold difference in enzyme activity did not reflect changes in protein levels. Furthermore, both cell lines showed comparable amounts of core 2 GlcNAc-T transcripts, thereby affirming that in this cell system, the catalytic efficiency of the enzyme is controlled only posttranslationally. Accordingly, populations of translated core 2 GlcNAc-T molecules may be turned on completely and simultaneously, giving rise to large changes in specific activity and, in turn, abnormal expression of core 2-related carbohydrates. Nevertheless, comparison of core 2 GlcNAc-T protein in CHO, PSA-5E and MDAY-D2 cells did reveal a correlation with enzyme activities: MDAY-D2>>PSA-5E>CHO. This result suggests that, at least in some cells, the expression of core 2 GlcNAc-T may also be controlled by transcriptional/translational means.
To date, posttranslational mechanisms of regulation have been seldom reported for glycosyltransferases. It is known that an allosteric cofactor, [alpha]-lactalbumin, can interact with [beta]1,4 galactosyltransferase and alter both its acceptor (Brew et al., 1968; Yadav and Brew, 1991) and donor (Do et al., 1995) substrate specificities. Additionally, [beta]1,4 galactosyltransferase activity was found to increase 2.5-fold in COS cells following phosphorylation mediated by a p58 protein kinase (Bunnell et al., 1990). Evidence to support cAMP-mediated phosphorylation as a likely event leading to enhanced mannosylphosphodolichol synthase activity has also been reported (Banerjee et al., 1987).
In conclusion, the results of this study show that core 2 GlcNAc-T is likely subjected to multilevel mechanisms of regulation. It will now be of considerable interest to characterize the different tissue-specific enzyme forms present in rat organs and to define the precise type(s) of posttranslational expression controls. Special attention should also be directed toward the identification of possible enzyme inhibitors, whose action could cause dramatic effects on core 2 GlcNAc-T activity despite normal transcription/translation of the gene.
Materials and methods
Chemicals
Gal[beta]1,3GalNAc[alpha]-pNp and GalNAc[alpha]-pNp were purchased from Toronto Research Chemicals. Dansyl chloride, UDP-GlcNAc, GlcNAc, buffer salts, Triton X-100, UDP-Gal, BSA, Brij-35, sodium butyrate, cholera toxin, SDS, Nonidet P-40, aprotinin, PMSF, IPTG, Tween 20, and 4-chloro-1-naphtol were from Sigma. UDP-6-[3H]galactose (18.9 Ci/mmol) was provided by NEN. Methanol and HPLC-grade acetonitrile were from BDH. All other reagents were of the highest purity commercially available.
Cell culture
The human promyelocytic cell line HL60 (Collins et al., 1977) was cultured in RPMI 1640 medium supplemented with 10% fetal calf serum. Chinese hamster ovary (CHO) cells (Stanley, 1984) were grown in [alpha]-modified Eagle's minimum essential medium ([alpha]-MEM) supplemented with 10% fetal calf serum. For treatment of CHO cells with differentiating agents (Datti and Dennis, 1993), sodium butyrate was prepared as a 1 M stock solution in PBS and added to cultures at a final concentration of 2 mM. Cholera toxin was dissolved at 0.5 mg/ml in sterile H2O and used at 100 ng/ml. Passaged cells were cultured for 24 h prior to drug addition and treatment was of 24 h duration. CHO cell retrodifferentiation was verified by monitoring morphology and core 2 GlcNAc-T activity. PYS-2 (Lehman et al., 1974) and PSA-5E (Adamson et al., 1977), murine endodermal cell lines displaying parietal and visceral differentiation, respectively, were cultured in [alpha]-MEM supplemented with 10% fetal calf serum. MDAY-D2, a highly metastatic, murine DBA/2 strain lymphoid tumor cell line (Kerbel et al., 1980), was grown in [alpha]-MEM containing 5% fetal calf serum. All cultures were maintained at 37°C in a 95% O2,5% CO2 humidified atmosphere.
Recombinant human core 2 GlcNAc-T fusion proteins
A 864 bp sequence of the human core 2 GlcNAc-T cDNA (Bierhuizen and Fukuda, 1992), encoding amino acids 84-372, was obtained via PCR using a human T cell cDNA library and ligated into pBluescript KS+, previously digested with SmaI and subjected to T-tailing (Marchuk et al., 1991). The fragment having the correct 5[prime]->3[prime] orientation was excised with BamHI and PstI and inserted into the bacterial expression vector pGEX-3X (Pharmacia). The resulting plasmid encoded an in-frame fusion protein consisting of a glutathione-S-transferase (GST) carboxyl terminus and a core 2 GlcNAc-T amino terminus, with a predicted molecular mass of 58 kDa.
The fusion vector pet3a-core 2 GlcNAc-T was prepared by employing the same 864 bp PCR product. After excision from the correct pBluescript KS+ vector as a BamHI/EcoRI fragment, the insert was partially ligated into BamHI-digested pet3a (Studier et al., 1990). The unligated sticky ends were end-filled with Klenow enzyme and then blunt-ligated, resulting in an out-of-frame fusion vector construct. This vector was recovered after transformation of TG2 cells and subjected to digestion with BamHI, followed by end-filling with Klenow fragment and religation. The product was an in-frame fusion construct encoding a chimeric protein with a T7 gene 10 carboxyl terminus and core 2 GlcNAc-T amino terminus, with a predicted molecular mass of 33 kDa.
Verification of both fusion vectors was carried out by restriction mapping and double-stranded DNA sequencing (Sanger method).
Induction and purification of GST-core 2 GlcNAc-T and T 7-core 2 GlcNAc-T fusion protein
pGEX-3X-core 2 GlcNAc-T and pet3a-core 2 GlcNAc-T fusion vectors were transformed into E.coli BL21(DE3)plysS cells, a protease-negative strain harboring a chloramphenicol-selectable plasmid which encodes lysozyme, thereby permitting bacterial lysis via a single freeze/thaw cycle (AMS Biotech. Ltd., Witney, Oxon, UK). Cells were cultured in 50 ml of Luria broth to an O.D. of 0.4, followed by induction of fusion protein with 0.5 mM IPTG for 3 h at 37°C. Cells were recovered by centrifugation at 3000 r.p.m. for 10 min at 4°C, washed once in 10 ml of cold STE buffer ( 0.1 M NaCl, 10 mM Tris-HCl pH 8.0, and 1 mM EDTA, pH 8.0) and repelleted. Cells were then resuspended in 10 ml cold TE buffer (50 mM Tris-HCl, pH 8.0 and 2.5 mM EDTA, pH 8.0) containing 0.1 mM PMSF and frozen at -20°C. Frozen cell suspensions were thawed at 37°C and subjected to sonication for two brief bursts of 45-60 s. Both fusion proteins were recovered as inclusion bodies, which were pelleted by centrifugation at 6000 r.p.m. for 15 min at 4°C and subjected to washing; once in 1 ml 1.0 M NaCl, three times in TxNE (5% Triton X-100, 1.0 M NaCl, and 50 mM EDTA pH 8.0) and twice in H2O. All washes contained 1 mM PMSF. Yields of the fusion proteins were approximated by comparing Coomassie blue staining intensities after SDS-PAGE with those of known quantities of BSA standard.
Production of anti-core 2 GlcNAc-T antibodies
Approximately 100 µg of GST-core 2 GlcNAc-T fusion protein was loaded onto a large preparative well of a 10% SDS-polyacrylamide gel. Following electrophoretic separation, the fusion protein was localized by staining in 0.05% Coomassie blue in H2O and excised in a minimum gel slice. Gel slices were crushed and emulsified in Freund's complete adjuvant. Two rabbits were immunized by subcutaneous injection and given monthly boosts of 200 µg of purified fusion protein emulsified in Freund's incomplete adjuvant, injected intramuscularly over a 6 month period. Rabbits were bled 14 days after each boost. Blood samples were clotted at 37°C for 1 h and then stored at 4°C overnight. Serum was aspirated from clotted blood and spun at 10,000 r.p.m. for 10 min at 4°C to remove insoluble matter. Aliquots were stored at -20°C. Presence of anti-core 2 antibodies was determined by testing serum reactivity against GST-core 2 GlcNAc-T and T 7-core 2 GlcNAc-T fusion proteins on Western blots. Blood samples were taken from both rabbits 'preimmunization" for use as controls for nonspecific serum reactivity on Western blots.
Purification of anti-core 2 GlcNAc-T antibodies from whole serum
The T 7-core 2 GlcNAc-T fusion protein was purified as inclusion bodies as described above. Approximately 100 µg of fusion protein was loaded onto a single preparative well of a 12.5% polyacrylamide gel. A gel slice containing the fusion protein was excised and blotted to nitrocellulose. The blot was stained with 0.5% Ponceau S in 1% acetic acid and the region containing the fusion protein excised in the minimum possible strip. The strip was blocked for 1 h in 5% low fat milk in TBS (50 mM Tris-HCl pH 7.6, 150 mM NaCl) containing 0.2% Tween 20 (Blotto-Tween) and then incubated overnight with 100 µl rabbit serum diluted in 5 ml of Blotto-Tween containing 0.02% sodium azide. Strips were washed extensively in Tween-TBS and then rinsed twice in 10 mM sodium phosphate buffer, pH 6.8, for 5 min. Antibodies were eluted in 300 µl 0.1 M glycine, pH 2.3, neutralized with 30 µl 1 M Na2HPO4 and stored at 4°C in 1% BSA and 0.02% sodium azide. Since the purified antibody preparation is diluted 3.7-fold with respect to initial serum concentration, for use as primary antibody, the preparation was further diluted 27-fold, thereby resulting in a final dilution of 1:1000.
Lysate preparations and Western blotting
Cells were pelleted, washed three times in PBS and resuspended in 10 volumes of Laemmli sample buffer (Laemmli, 1970). After boiling for 5 min, lysates were sonicated briefly and clarified by centrifugation at 10,000 r.p.m. for 10 min at 4°C. Supernatants were recovered and stored at -20°C.
Tissues were excised from male Sprague-Dawley rats, frozen in liquid nitrogen and stored at -80°C until processed. Tissues were weighed and dispersed in 5 volumes of ice-cold suspension buffer (0.1 M NaCl, 10 mM Tris/HCl pH 7.6, 1 mM EDTA pH 8.0, 1 µg/ml aprotinin, and 100 µg/ml PMSF) using an Ultra-Turrax homogenizer. After adding an equal volume of 2× Laemmli sample buffer, homogenates were boiled for 10 min, sonicated briefly and centrifuged at 10000 r.p.m. for 10 min at 4°C. Supernatants were recovered and stored at -20°C.
Rat serum was isolated from whole blood which was clotted at 37°C for 1 h and then stored at 4°C overnight. After removal of the serum, insoluble material was eliminated by centrifugation at 10,000 r.p.m. for 10 min at 4°C and aliquots stored at -20°C.
Protein concentrations were determined with Bio-Rad reagent using BSA as the standard.
Samples were subjected to SDS-PAGE under reducing conditions (10% gels; Laemmli, 1970). Separated proteins were transferred electrophoretically onto nitrocellulose membranes (Hybond ECL, Amersham) using 25 mM Tris, 190 mM glycine, 20% methanol as the transfer buffer. Blots were blocked in 5% low fat milk in TBS (50 mM Tris/HCl, pH 7.6, 150 mM NaCl) containing 0.2% Tween 20. Human core 2 GlcNAc-T specific antibody was applied overnight at room temperature at 1:1000 dilution. After washings, nitrocellulose sheets were incubated 1 h with HRP-goat anti-rabbit secondary antibody (Bio-Rad; 1:3000 dilution). Immunodetection was carried out by employing an enhanced chemiluminescence method (ECL kit, Amersham) according to the manufacturer's instructions. For each blot, a number of time exposures were performed in order to ensure that the results were obtained in the linear response range of the film. Additionally, pilot experiments showed that results of Western blotting analyses were proportional in the range of 0-500 ng of recombinant human core 2 GlcNAc-T protein. Densitometric analysis was performed using a VideoImage Analyzer, model MCID-M4 (Imaging Research Inc., St. Catharine's, Ontario).
N-Glycosidase treatment of cell lysates
An extract of HL60 cells was prepared in a single detergent buffer (50 mM Tris/HCl pH 8.0, 150 mM NaCl, 0.02% sodium azide, 100 µg/ml PMSF, 1 µg/ml aprotinin, and 1% Nonidet P-40). Aliquots containing 100 µg of protein were denatured by boiling for 3 min in the presence of 1% SDS and then diluted 10-fold with the above buffer prior to incubation in the presence or absence of 4 units of N-glycosidase F (Boehringer Mannheim) overnight at 37°C. Proteins were precipitated with 4 volumes acetone for 1 h at -20°C, recovered by centrifugation, and boiled in Laemmli sample buffer.
Reverse transcription-PCR (RT-PCR) and Southern blotting
One microgram of total cellular RNA from both PSA-5E and PYS-2 cells was subjected to reverse transcription in a reaction containing 1× buffer (Promega), 10 mM dNTPs, 50 pmol reverse primer, 10 mM dithiothreitol, 1 mg/ml BSA, and 10 units avian myeloblastosis virus reverse transcriptase (Promega) in a total volume of 20 µl. After purification by phenol/chloroform extraction and ethanol precipitation, cDNA was resuspended in a volume of 20 µl H2O, of which 2 µl was used as the template for a single round of amplification. PCR was performed in a 100 µl volume using 200 µM dNTPs, 5 units of Taq polymerase (Promega), and 50 pmol of both forward and reverse oligonucleotide primers. The reaction mixtures were denatured at 94°C for 5 min, followed by 25 cycles of 94°C, 30 s; 60°C, 2 min; 72°C, 2.5 min; and a final extension at 72°C for 10 min. Primers (21-mer each) were designed based on the murine cDNA sequence for core 2 GlcNAc-T (C. E. Warren and J. W. Dennis, personal communication). The amplified products were resolved on a 1% agarose gel and transferred to a BA85 nitrocellulose membrane (Schleicher and Schuell). The hybridization protocol is as outlined in Sambrook et al. (1989). After incubation with a 32P-labeled 1700 bp murine core 2 GlcNAc-T probe (kindly provided by Drs. Charles E. Warren and James W. Dennis, Mount Sinai Hospital, Toronto), the blot was washed 2 × 10 min in 2× SSC, 0.5% SDS at room temperature, 1 ×20 min in 1× SSC, 0.5% SDS at 60°C and finally for 1 × 15 min in 0.2× SSC, 0.5% SDS at 60°C prior to autoradiography. Several exposures were generated in order to ensure that the result obtained was in the linear response range of the film.
Immunological screening of a human cDNA library
An HL60 (undifferentiated human peripheral blood promyelocytic leukemia subclone) cDNA library constructed in the [lambda]ZAPII expression vector was purchased from Stratagene. Approximately 4.5 × 105 pfu were plated on E.coli XL-1 bacteria. Infected plates were kept at 42°C for 3.5 h prior to overlaying with IPTG-impregnated nitrocellulose filters overnight at 37°C. After incubation, filters were removed, washed several times at room temperature in a large volume of TNT buffer (10 mM Tris/HCl pH 8.0, 150 mM NaCl, 0.05% Tween 20) and then blocked for 1 h in the same solution containing 5% low fat milk. The putative purified anti-human core 2 GlcNAc-T antibody was applied to the filters as the primary antibody at a 1:1000 dilution and incubated overnight at room temperature, after which filters were extensively washed in TNT buffer, incubated with HRP-coupled goat anti-rabbit secondary antibody (1:3000 in blocking solution) and rinsed. Immunodetection was carried out colorimetrically using 4-chloro-1-naphtol reagent according to the manufacturer's instructions (Bio-Rad). After three rounds of immunoscreening, one positive plaque was purified to homogeneity. The cDNA insert of the phage was recovered as an EcoRI fragment, subcloned into pBluescript, and sequenced by the dideoxy chain termination method using the Sequenase Version 2.0 kit (United States Biochemical).
Glycosyltransferase assays
As the source of glycosyltransferase activities, rat tissues were homogenized in 5 volumes of 10 mM Tris/HCl, pH 7.5, 250 mM sucrose, using an Ultra-Turrax device. Specimens were passed several times through 22 gauge needles and centrifuged at 3000 × g for 30 min at 4°C. Supernatants were collected and employed as crude enzyme preparations. To obtain cell lysates, pellets were washed in 0.9% saline solution and lysed in 0.315% Triton X-100, 0.15 M NaCl for 10 min at 4°C.
The fluorometric assay for core 2 GlcNAc-T was essentially performed as described previously (Palmerini et al., 1996). The core 2 GlcNAc-T acceptor substrate Gal[beta]1,3GalNAc[alpha]-pNHdansylphenyl was obtained from commercially available Gal[beta]1,3GalNAc[alpha]-pNO2phenyl through catalytic reduction of the p-nitrophenyl group to p-aminophenyl (Palmerini et al., 1995) and derivatization with dansyl chloride. The labeling reaction was carried out (1 h at 65°C) in the presence of 20 µl 0.25 M TES buffer, pH 7.0, 60 µl water, and 100 µl dansyl chloride in acetone (5 mg/ml) (Palmerini et al., 1996).
The core 2 GlcNAc-T reaction mixture contained, in a total volume of 50 µl, 0.1 M TES buffer pH 7.0, 0.1 M GlcNAc, 2 mM UDP-GlcNAc, 1 mM Gal[beta]1,3GalNAc[alpha]-pNHdansylphenyl, and 20 µl tissue or cell lysate (5-500 µg of protein). After a 1-2 h incubation at 37°C, the reaction was stopped by boiling for 3 min. The volume of each tube was brought to 100 µl with water, after which samples were microfuged. Supernatants were collected and immediately analyzed (20 µl) by HPLC. Separation of the core 2 GlcNAc-T reaction product (i.e., Gal[beta]1,3[GlcNAc[beta]1,6]GalNAc[alpha]-pNHdansylphenyl) was carried out by using a 4.6 × 250 mm Spherisorb ODS2 column, which was developed isocratically at 1.2 ml/min in 25% acetonitrile containing 0.1% Brij-35 and whose effluent was continuously monitored at 340 nM excitation/450 nM emission. Quantitation was performed by employing a standard curve generated with known amounts of the pure tetrasaccharide.
The [beta]1,3 Gal-T assay was performed as described elsewhere (Datti and Dennis, 1993), employing a reaction mixture of 50 µl consisting of 20 mM MnCl2, 0.1 M MES buffer, pH 6.7, 0.5% Triton X-100, 1.6 mM UDP-Gal, 0.5 µCi UDP-6-[3H]galactose, 2 mM GalNAc[alpha]-pNp, and 5 µl cell lysate containing 50-100 µg of protein.
Acknowledgments
We are particularly grateful to Drs. Charles E. Warren and James W. Dennis (Mount Sinai Hospital, Toronto) for supplying a murine core 2 GlcNAc-T cDNA probe and its sequence. We would also like to thank Mr. Marcello Coli for his excellent technical assistance in the animal colony. The work was supported in part by the Italian CNR and the University of Perugia Research Funding Programme. A.D. thanks the Canadian MRC and the Italian CNR for providing a Visiting Scientist award under an exchange Programme for Senior Researchers. This article represents part of a doctoral thesis submitted to the University of Perugia, Italy, by I.E.V., who was supported by a fellowship granted by the Italian Ministry of Education and Scientific and Technological Research (MURST).
Abbreviations
UDP, uridine 5[prime]-diphosphate; T, transferase; Gal, d-galactose; GalNAc, d-N-acetylgalactosamine; GlcNAc, D-N-acetylglucosamine; PSGL-1, P-selectin glycoprotein ligand-1; Con A-HRP, horseradish peroxidase-conjugated Concanavalin A;pNp, para-nitrophenyl; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; PMSF, phenylmethylsulfonylfluoride; IPTG, isopropyl-[beta]-d-thiogalactopyranoside; PBS, phosphate-buffered saline; TES, 2-([2-hydroxy-1,1-bis(hydroxymethyl)ethyl]amino)ethanesulfonic acid; MES, 2-[N-morpholino]ethanesulfonic acid; TBS, Tris-buffered saline; pfu, plaque forming unit.
References
1To whom correspondence should be addressed at: GlycoDesign Inc., 480 University Avenue, Suite 900, Toronto, Ontario, Canada M5G 1V2
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 10 Jun 1998
Copyright©Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Tan and P.-W. Cheng Mucin Biosynthesis: Identification of the cis-Regulatory Elements of Human C2GnT-M Gene Am. J. Respir. Cell Mol. Biol., June 1, 2007; 36(6): 737 - 745. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. M. Ben-Mahmud, G. E. Mann, A. Datti, A. Orlacchio, E. M. Kohner, and R. Chibber Tumor Necrosis Factor-{alpha} in Diabetic Plasma Increases the Activity of Core 2 GlcNAc-T and Adherence of Human Leukocytes to Retinal Endothelial Cells: Significance of Core 2 GlcNAc-T in Diabetic Retinopathy Diabetes, November 1, 2004; 53(11): 2968 - 2976. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T.-Y. Yen, B. A. Macher, S. Bryson, X. Chang, I. Tvaroska, R. Tse, S. Takeshita, A. M. Lew, and A. Datti Highly Conserved Cysteines of Mouse Core 2 {beta}1,6-N-Acetylglucosaminyltransferase I Form a Network of Disulfide Bonds and Include a Thiol That Affects Enzyme Activity J. Biol. Chem., November 14, 2003; 278(46): 45864 - 45881. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Chibber, B. M. Ben-Mahmud, G. E. Mann, J. J. Zhang, and E. M. Kohner Protein Kinase C {beta}2-Dependent Phosphorylation of Core 2 GlcNAc-T Promotes Leukocyte-Endothelial Cell Adhesion: A Mechanism Underlying Capillary Occlusion in Diabetic Retinopathy Diabetes, June 1, 2003; 52(6): 1519 - 1527. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. R. Falkenberg, K. Alvarez, C. Roman, and N. Fregien Multiple transcription initiation and alternative splicing in the 5' untranslated region of the core 2 {beta}1-6 N-acetylglucosaminyltransferase I gene Glycobiology, June 1, 2003; 13(6): 411 - 418. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Schwientek, J.-C. Yeh, S. B. Levery, B. Keck, G. Merkx, A. G. van Kessel, M. Fukuda, and H. Clausen Control of O-Glycan Branch Formation. MOLECULAR CLONING AND CHARACTERIZATION OF A NOVEL THYMUS-ASSOCIATED CORE 2 beta 1,6-N-ACETYLGLUCOSAMINYLTRANSFERASE J. Biol. Chem., April 6, 2000; 275(15): 11106 - 11113. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Schwientek, M. Nomoto, S. B. Levery, G. Merkx, A. G. van Kessel, E. P. Bennett, M. A. Hollingsworth, and H. Clausen Control of O-Glycan Branch Formation. MOLECULAR CLONING OF HUMAN cDNA ENCODING A NOVEL beta 1,6-N-ACETYLGLUCOSAMINYLTRANSFERASE FORMING CORE 2 AND CORE 4 J. Biol. Chem., February 19, 1999; 274(8): 4504 - 4512. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||