| Glycobiology | Pages |
Transcriptional and posttranscriptional regulation of [alpha]1,3-galactosyltransferase in activated endothelial cells results in decreased expression of Gal[alpha]1,3Gal
Introduction
Results
Discussion
Material and methods
Acknowledgments
Abbreviations
Abbreviations
Transcriptional and posttranscriptional regulation of [alpha]1,3-galactosyltransferase in activated endothelial cells results in decreased expression of Gal[alpha]1,3Gal
Gal[alpha]1,3Gal carbohydrate residues are present in the glycoproteins and glycolipids of lower mammals, and appear to be involved in the binding specificity of several membrane receptors. We report here that endothelial cells stimulated with lipopolysaccharide or inflammatory cytokines modulate their expression of UPD-Gal:[beta]-d-Gal [alpha]1,3-galactosyltransferase ([alpha]1,3GT), the Golgi enzyme that attaches a galactose in [alpha]1,3 configuration to an N-acetyllactosamine acceptor. Upon activation, the steady state level of mRNA is transiently increased, the modifications being paralleled by a transcriptional regulation of the gene. Cell-associated enzyme activity, on the other hand, falls rapidly after activation, before being up- and downregulated with kinetics that parallel those of the mRNA, and after 3 days reaches a level representing 40-60% of the activity in cells before activation. Overall Gal[alpha]1,3Gal expression at the cell surface follows enzyme activity, except that it is insensitive to the rapid and transient reduction of activity occurring shortly after activation. This reduced [alpha]1,3GT activity in stimulated EC is correlated with lower stability of the protein, and with a switch in the expression of the isoform pattern, isoform 1 being predominant in resting cells whereas after activation it is isoform 2 that predominates. The two isoforms, however, appear to have similar intrinsic stability, so that the reduced stability of the enzyme in activated EC probably results from an induced proteolytic degradation pathway. Key words: activation/[alpha]1,3-galactosyltransferase/endothelial cells/transcription
Introduction
Inflammatory cytokines, LPS, or anti-EC antibodies activate endothelial cells in vitro and in vivo and lead to the rapid expression of secreted and membrane molecules, among which are those involved in inflammation (IL-1[beta]; Sironi et al., 1989), chemotaxis (IL-8; Sica et al., 1990), thrombosis (t-PA inhibitor; Emeis and Kooistra, 1986, PAI-1; Nachman et al., 1986, tissue factor; Conway et al., 1989), and vasoconstriction (endothelin; Yoshizumi et al., 1990). The expression of most of these inducible genes in ECs is transient, because the transcription of some of them is under the control of NF-[kappa]B-which is itself inhibited by the inducible protein I[kappa]B[alpha]. Among the glycosyltransferases, the [alpha]2-6-Sialyltransferase is known to be inducible by cytokines in human ECs, and can therefore participate in the sialylation of activation-dependent molecules (Hanasaki et al., 1994).
In most mammals ECs contain many surface glycoproteins (Parker et al., 1994; Thibaudeau et al., 1996) and glycolipids (Bouhours et al., 1996) harboring Gal[alpha]1,3Gal terminal carbohydrate moieties. This epitope is absent in humans, apes, and Old World monkeys because in these species the UDP-Gal:[beta]-d-Gal [alpha]1,3-galactosyltransferase ([alpha]1,3GT) gene contains several non-sense and frameshift mutations (Joziasse et al., 1989). On the other hand, humans and Old World monkeys have a considerable number of anti-Gal[alpha]1,3Gal natural antibodies, probably as a result of immunization by intestinal microorganisms (Galili et al., 1988). In experimental xenotransplantation, mainly of pig organs into primates, these xenoreactive natural antibodies bind to and activate donor ECs, initiating an inflammatory reaction that leads to an acute form of vascular rejection. Because recognition the Gal[alpha]1,3Gal epitope by xenoreactive natural antibodies causes rejection of xenografts, and because this carbohydrate has been implicated in the regulation of other carbohydrates that play a role in inflammation (Cho et al., 1996), we studied [alpha]1,3GT expression during activation. In a previous report, we described the four isoforms of pig [alpha]1,3GT mRNA (Vanhove et al., 1997), apparently homologous to those reported in the mouse (Joziasse et al., 1992), which result from the differential splicing of exons 5 and 6 coding for the intra-Golgi stem region of the molecule, separating the trans-membrane stretch from the catalytic domain. In the pig, the expression of these isoforms is tissue-specific. For instance, the ovary contains only the shorter form (isoform 4) whereas the heart contains two long forms (isoforms 1 and 2), and the thymus contains all four isoforms. All four molecules appear to be similarly localized in the Golgi compartment and confer [alpha]1,3GT activity after transfection into [alpha]1,3GT-negative cells. It has not yet been possible, however, to ascribe a specific role to individual isoforms. In the present report we extend our previous analysis (Vanhove et al., 1997) of pig [alpha]1,3GT mRNA isoforms in the endothelium: we investigated the expression of [alpha]1,3GT upon activation, and report that in addition to up- and downregulation of mRNA and enzyme activity, there is a marked shift away from the expression of one dominant isoform (long isoform 1) toward the expression of another isoform (isoform 2). These modifications are paralleled by a reduction in overall [alpha]1,3GT activity, this reduction not being due, however, to a difference in intrinsic stability between the two isoforms.
Results
Activation of PAECs modulates the steady state level of [alpha]1,3GT mRNA
[alpha]1,3GT mRNA level was assessed by northern blotting of total RNA from resting and stimulated PAECs. TNF[alpha] or LPS treatment resulted in a time-dependent up- and downregulation of the steady state level of [alpha]1,3GT mRNA. Induction was detectable by 4 h and continued to rise, reaching a 2-fold increase over 12 h. After 12 h, the mRNA level decreased at a regular rate, returning to its base level of expression by 16 to 24 h, and falling below the base level after longer periods of time. PAECs incubated in control medium showed no increase in [alpha]1,3GT mRNA during this period. The induction of E-selectin mRNA was analyzed to control that the cells were properly activated in these experiments (Figure 1).
Figure 1. Kinetics of [alpha]1,3GT mRNA induction in PAECs during activation. (A) Total RNA was isolated from PAECs incubated for 72 h in control medium (0 h) and PAECs exposed to LPS for increasing periods of time; 10 µg total RNA from each time point was isolated from cells, electrophoretically separated on a 1.3% agarose gel containing 7% formaldehyde, and transferred to nylon filters. The immobilized RNAs were then hybridized with a 32P-labeled [alpha]1,3GT cDNA probe, before being stripped and reprobed with a 32P-labeled pig E-selectin and rat GAPDH probes as controls for cell activation and equal loading of the RNA on the gel. (B) The expression of [alpha]1,3GT obtained after stimulation with LPS and TNF[alpha] was quantified with a PhosphorImager, and the relative signals were standardized to the level of GAPDH message. The levels of expression of the [alpha]1,3GT transcripts in resting PAECs were arbitrarily taken as 100 units. One of three independent experiments giving similar results is represented here.
Transcriptional modifications of the a1,3GT gene in the course of EC activation
Nuclear run-on assays were conducted to address the question of whether the effect of EC activation on [alpha]1,3GT mRNA levels reflected a change in rates of transcription. Nuclear RNA transcripts were synthesized by incubating the nuclei isolated from unstimulated ECs, from ECs stimulated with LPS for 12 h (i.e., at the peak of mRNA accumulation), and from ECs stimulated for 72 h (i.e., when the steady state level of mRNA had fallen). To verify that ECs were effectively activated in our assays, we used Northern blotting in parallel to analyze total [alpha]1,3GT mRNA as shown in Figure 1. As an internal control for gene transcription, we analyzed [beta]-actin transcripts. The results show that the up- and downregulation of the [alpha]1,3GT mRNA level indeed reflects a parallel up- and downregulation of the transcriptional rate of the gene (Figure 2).
Figure 2. Rate of [alpha]1,3GT gene transcription in resting and activated PAECs determined by nuclear run-on assay. Nuclei isolated from PAECs or PAECs stimulated with LPS for 12 or 72 h were used as a source of nascent transcripts for the preparation of riboprobe in the nuclear run-on assay, as described in Material and methods. Signals emitted from the radiolabeled riboprobe hybridized with each template DNA for [alpha]1,3GT and [beta]-actin cloned in pSK+ are shown in (A) To test nonspecific binding, the riboprobe was applied to empty pSK+ vector DNA. The signals were quantified with a PhosphorImager, and the intensities of the signal, standardized for [beta]-actin transcription, are presented in (B)
Activation of PAECs modulates [alpha]1,3GT activity
Endothelial [alpha]1,3GT activity was assessed by incorporating the label from UDP-(14C)-galactose into [beta]-lactose. Activation by TNF[alpha] or LPS resulted in a time-dependent variation in cell-associated activity. Enzyme activity at first fell to about half of its base level within 1 h. After 10 h, the level of activity rose again to approach its initial level by 24 h. After 24 h, enzyme activity fell at a regular rate down to half the level found in resting cells after LPS stimulation, and down to 70% the level found in resting cells after TNF[alpha] stimulation (Figure 3). No [alpha]1,3GT activity was found in the supernatant of the PAECs, either resting or stimulated (data not shown).
Figure 3. Time course of changes in [alpha]1,3GT activity during activation of PAECs. [alpha]1,3GT specific activity in crude extracts of PAEC treated with LPS or TNF[alpha] for the time indicated was measured as described in Materials and methods. Typical incorporation of 14C-Gal into lactose using extracts from 106 resting PAECs was of 3000 c.p.m.. Results are representative of three independent experiments.
Activation of PAECs decreases the expression of Gal[alpha]1,3Gal at the cell surface
To address the question of whether the overall reduction of [alpha]1,3GT activity in activated ECs affected the phenotype of cells, expression of the Gal[alpha]1,3Gal terminal carbohydrate on the membrane was assessed by binding of antibodies from rabbit red cell-immunized hens, affinity-purified on an immunoabsorbent containing the Gal[alpha]1,3Gal epitope. Activation results in a continuous decrease in the expression of Gal[alpha]1,3Gal sugar which reaches about 30% at 72 h (Figure 4). The expression of E-selectin, assayed in parallel as a control for EC activation, shows the up and down profile typical of inducible genes in ECs.
Figure 4. Cell surface expression of the Gal[alpha]1,3Gal epitope in the course of PAEC activation. PAECs were grown to subconfluence in microtiter plates and stimulated with LPS and TNF[alpha] for the indicated period of time. Cells were fixed with formaldehyde and the reactions of anti-Gal[alpha]1,3Gal IgY from hens and of an anti-ELAM-1 antibody were assessed by ELISA. Values (mean ± SD of three analyses) are expressed as a percentage of the signal given by resting cells.
Modification of [alpha]1,3GT mRNA isoform pattern during activation of PAECs
We have previously shown that four [alpha]1,3GT mRNA isoforms are expressed in pig tissues, the four isoforms not, however, being simultaneously expressed in all tissues. In PAECs, isoforms 1, 2 and 4 are detectable (Vanhove et al., 1997) but isoforms 1 and 2 represent more than 90% of the mRNA. In order to address the question of whether the relative expression of the different isoforms varies during EC activation, the expression of isoforms 1 and 2 was assessed by semiquantitative RT-PCR analysis, using PCR conditions that we chose as being in the exponential amplification phase. The results show that in resting PAECs, isoform 1 is dominant with an isoform1/isoform2 ratio of 1.5. Upon activation with TNF[alpha] or LPS, this ratio decreases to 1 after 10 h, and to 0.4 after 24 h, after which it is stabilized at that value (Figure 5).
Figure 5. Selective regulation of the two isoforms of [alpha]1,3GT mRNA expressed in PAECs during stimulation with LPS and TNF[alpha]. Semiquantitative RT-PCR was used to amplify the cDNA region encompassing the two exons of the stem domain that are alternatively spliced (corresponding to exons 5 and 6 of the mouse) The 299 bp signal corresponds to isoform 1, containing exons 5 and 6, and the 263 bp signal corresponds to isoform 2, lacking exon 5. At the bottom of the figure are shown the ratios of isoform 1/isoform 2 at each point of the time course.
The stability of [alpha]1,3GT enzyme activity decreases in PAECs after activation
The switch from one dominant isoform (number 1) to another with a shorter stem region (number 2) suggests that the intrinsic properties of these two isoforms, and particularly their stability, may be different. To determine whether activation modified the stability of [alpha]1,3GT in ECs, we measured cell-associated activity at various time points after treatment with the protein synthesis inhibitor cycloheximid (Figure 6). In resting PAECs, cycloheximid treatment induced a straight decrease of enzyme activity of about 2% per h (half-life 22 h), whereas in PAEC pretreated with TNF[alpha] for 12 h, enzyme activity fell by 6% per h (half-life 8 h). To address the question of whether the decreased stability of enzyme activity in activated PAECs was a direct effect of the switch from the expression of isoform 1 to isoform 2, we transfected HeLa cells (human cells, therefore [alpha]1,3GT-negative) with expression vectors containing one single isoform, as previously described (Vanhove et al., 1997). Enzyme stability was assessed by measuring cell-associated activity at different time points after treatment with cycloheximid. We observed (data not shown) no difference in stability between isoforms 1 and 2 in that expression system, the half-lives of the two proteins being about 7 h, suggesting that intrinsic characteristics of isoforms are not directly responsible for the reduced enzyme stability in activated ECs.
Figure 6. Stability of [alpha]1,3GT enzyme activity in resting and activated PAECs, determined by cycloheximid-chase experiment. PAECs were left unstimulated, or were stimulated for 12 h with TNF[alpha]. Cycloheximid (10 µg/ml) was added to each culture and the cells were harvested after the indicated period of time. [alpha]1,3GT enzyme activity was measured as described in Material and methods. Similar results were obtained in one other independent experiment.
Discussion
Whereas in some instances glycosyltransferases have been found essential for the correct functioning of adhesion molecules (Kanda et al., 1995; Maly et al., 1996; Salmi and Jalkanen, 1996), a deficiency of [alpha]1,3GT does not appear to affect any physiological system, as [alpha]1,3GT (-/-) mice are essentially normal (Thall et al., 1995). Gal[alpha]1,3Gal epitopes are involved in the recognition of ZP3 glycoproteins on the mouse oocyte by specific receptors expressed on sperm cells, but [alpha]1,3GT(-/-) mice are fertile, indicating that the Gal[alpha]1,3Gal moiety is not essential to sperm-oocyte interaction (Thall et al., 1995). It thus appears clear that functions supported by [alpha]1,3-galactosylation may be taken over by other redundant molecules supporting the same functions. [alpha]1,3GT is a Golgi enzyme using N-acetyllactosamine as acceptor molecule. As such, it competes for its substrate with other glycosyltransferases such as [alpha]1,2-fucosyltransferase, [alpha]1,3-fucosyltransferase, or [alpha]2,3/6-sialyltransferases, which also use the N-acetyllactosamine acceptor molecule. In [alpha]1,3GT(-/-) mice, however, the analysis of surface glycosylation revealed that the complete elimination of the [alpha]Gal epitope has little effect on the expression pattern of other carbohydrates, including fucose and sialic acid. Nevertheless, this analysis of the overall expression of carbohydrates does not eliminate the possibility that individual glycoproteins involved in cell to cell interaction or in cell adhesion may modify their affinity for their counter-receptor when Gal[alpha]1,3Gal structures are absent. Indeed, the very rapid loss of enzyme activity reported here after EC activation has no consequence on overall Gal[alpha]1,3Gal membrane glycosylation, as measured with anti-Gal[alpha]1,3Gal antibodies, possibly because most membrane glycoproteins and glycolipids have a slower turnover than the period of time during which enzyme activity is reduced. It may, however, modify the glycosylation of newly synthesized membrane glycoproteins, especially those that are rapidly upregulated after EC activation, and may possibly modulate their function. In this regard, it has been reported that disturbed [alpha]1,3-GT activity modifies the action of other glycosyltransferases that also use the N-acetyllactosamine as acceptor substrate (Cho et al., 1996; Sandrin et al., 1995; Smith et al., 1990), and modulate adhesion mechanisms in which carbohydrate epitopes other than Gal[alpha]1,3Gal alone are involved. In mouse F9 cells, for instance, retinoic acid-mediated differentiation induces enhanced [alpha]1,3GT activity, the result being that Lex antigens are masked by the upregulated Gal[alpha]1,3Gal residues (Cho et al., 1996). In the present context of EC activation, the decrease in [alpha]1,3GT activity may thus result in an increased expression of Lex antigens that are normally masked by Gal[alpha]1,3Gal residues in resting cells, and may thereby modulate the functioning of adhesion molecules (selectins) in which Lex determinants play a key role. Our recent observation that Gal[alpha]1,3Lex determinants are expressed in pig kidneys (Bouhours et al., 1997) indicates that regulated masking of Lex may also occur in that species.
After stimulation of PAECs with LPS or TNF[alpha], there is an initial fall in enzyme activity observed as early as 1 h (Figure 2), which is not the reflection of a downregulation in the level of mRNA. This suggests that [alpha]1,3GT activity is under the control of preexisting factors capable of responding very rapidly to stimulation signals. This reduced activity lasts for 8 h, after which activity rises again, reaching at 24 h a level still below that of resting cells. This increase may result from the upregulation of mRNA observed at 12 h, which is itself the result of an increase in transcription of the gene. After 24 h, levels of [alpha]1,3GT mRNA and enzyme activity fall off at a regular rate. Membrane expression of Gal[alpha]1,3Gal residues, on the other hand, decreases without showing the 'down-up-down" profile of the activity level. This dissociation between enzyme activity and sugar expression adds to the argument in favor of a regulation system involving other factors than the amount of [alpha]1,3GT protein produced.
That a single gene may generate a family of differing transcripts has been reported for the expression of many genes, including glycosyltransferases. For instance, alternative transcriptional initiation in [beta]1,4-galactosyltransferase results in mRNAs that putatively use one of two in-frame AUGs for translational initiation (Shaper et al., 1988). The two forms of this enzyme, differing in the 13 amino acids at the N-terminus, are either located in the Golgi apparatus or are targeted at the plasma membrane (Russo et al., 1990; Lopez et al., 1991). In [alpha]2,6-sialyltransferase, an alternative expression of the 5[prime]-untranslated region has been implicated in the regulation of translation (Aasheim et al., 1993), and a shorter, soluble form of that enzyme secreted in the liver results from proteolytic cleavage of the catalytic domain from its membrane anchor (Weinstein et al., 1987; Paulson and Colley, 1989). For [alpha]1,3GT in the mouse, Joziasse et al. (Joziasse et al., 1992) have demonstrated how alternative splicing of exons 5 and 6 results in the production of four transcripts. In the pig, we have found a similar configuration (Vanhove et al., 1997), whereas cattle express only one [alpha]1,3GT transcript (Joziasse et al., 1992). In this paper, we describe one example of an [alpha]1,3GT splicing isoform being preferentially expressed upon activation of PAECs. It was tempting to speculate that this switch by itself might induce-or at least participate-in the reduction of overall enzyme stability observed in activated PAECs (Figure 6). Analysis of [alpha]1,3GT isoforms 1 and 2 expressed individually in recombinant HeLa cells, however, revealed no significant difference in intrinsic stability, suggesting that the reduction of enzyme stability after PAEC activation is due instead to an activation-dependent proteolytic mechanism. Whether such a proteolytic mechanism affects one isoform more than another is unknown. Several questions remain to be answered before we have a comprehensive overview of the regulation of [alpha]1,3GT: How is the splicing variant isoform 2 preferentially expressed upon activation; Does the reduced [alpha]1,3GT activity in activated ECs modify the glycosylation phenotype of particular molecules involved in mechanisms related to inflammation; and Do these potential modifications directly modulate the function of these molecules?
Material and methods
Endothelial cell culture and stimulation
Enzymatic isolation and culture of porcine aortic endothelial cells (PAEC) was performed as described previously (Warren, 1990). PAECs were maintained in DMEM medium containing 10% FCS and used between passages 3 and 14. PAEC were stimulated with 100 U/ml human TNF[alpha] (Genzyme, Cambridge, MA), or 50 ng/ml lipopolysaccharide (Sigma, St. Quentin Fallavier, France), added to the medium. Cycloheximid was used at 10 µg/ml.
Run-on assays
Unstimulated nuclei and nuclei stimulated for 12 h and 72 h were isolated as described previously (Kronke et al., 1984). Briefly, 5 × 107 EC were lysed by two 5 min incubations in 4 ml 10 mM Tris·Cl, pH 7.4, 10 mM NaCl, 3 mM MgCl2, and 0.5% NP-40, on ice. In vitro transcription was carried out according to Groudine et al. (1981) with slight modifications. Nuclei were resuspended in 400 µl transcription buffer containing: 16% glycerol, 20 mM Tris.Cl pH 8.0, 5 mM MgCl2, 150 mM KCl, 0.4 mM of each ATP, GTP, and CTP (Sigma), and 170 µCi 32P-UTP (800 Ci/ mmol, Amersham, Little Chalfont, UK). After 30 min. incubation at 26°C, DNase I (20 µg/ml; Boehringer Mannheim, Meylan, France) was added for 10 min. Digestion (30 min., 42°C) of the reaction mixture with 100 µg/ml of proteinase K (Boehringer Mannheim) in the presence of 25 µg of carrier E.coli tRNA was followed by phenol chloroform extraction. The aqueous phase was ethanol precipitated and the RNA treated once more with DNase I, proteinase K, and phenol/chloroform plus ethanol precipitation. The RNA pellet was resuspended in 10 mM Tris-HCl pH 8.0, 1 mM EDTA and free 32P was eliminated on a Sephadex G-50 spin column. After reprecipitation, the samples were resuspended in 100 µl 10 mM Tris-HCl pH 8.0, 1 mM EDTA, and the incorporated radioactivity was determined by liquid scintillation counting; 6 × 106 c.p.m. (or 3 × 106 c.p.m. for one time point) was added to 2 ml hybridization solution (50% formamide, 5× SSC, 20 mM Na-phosphate buffer, 1× Denhardt's solution, 100 µg/ml denatured salmon sperm DNA, 0.1% SDS). 32P-Labeled nRNA was hybridized for 3 days at 42°C with prehybridized nitrocellulose filters on which 10 µg of denatured plasmid cDNA had been immobilized using a slot-blot apparatus. Filters were washed twice for 10 min with 2× SSC, 0.1% SDS at room temperature and twice for 20 min with 0.2% SSC, 0.1% SDS, at 65°C. Filters were then dried and exposed to a PhosphorImager (Molecular Dynamics, Sunnyvale) for 10 h for quantification, and on Kodak XAR films with intensifying screens for 5 days, at -80°C. Signals yielded by the [alpha]1,3GT transcription were standardized according to the signals given by [beta]-actin transcription in the same nucleus preparation.
Northern blot analysis
Total cellular RNA was isolated from cultured cells essentially as described previously (Zipfel et al., 1989); 10 µg of total RNA was fractionated on agarose/formaldehyde gel, transferred to Hybond N membranes (Amersham) by capillary action in 20× SSC for 16 h, and immobilized by UV cross-linking (Autocrosslink on Appligene cross linker, Pleasanton, CA). The filters were prehybridized for 6 h in a solution containing 50% formamide, 5× SSPE (0.18 M NaCl, 10 mM sodium phosphate, pH 7.7, 1 mM EDTA), 5× Denhardt's, 0.5% SDS, and 100 µg/ml denatured salmon sperm DNA. RNA bound to the membranes was hybridized with 32P-labeled cDNA probes corresponding to the coding region of pig [alpha]1,3GT isoform 1 (Vanhove et al., 1997), and to the housekeeping gene GAPDH. Hybridization was carried out overnight at 42°C in prehybridization solution containing 106 c.p.m./ml of probe. The filters were washed twice in 2× SSPE, 0.5% SDS at room temperature for 15 min, and three times in 0.5× SSPE, 0.5% SDS at 65°C for 15 min. Hybridization was visualized and signals quantified with a PhosphorImager. Signals were standardized for variations in the amount and quality of RNA using GAPDH as internal controls, detected after dehybridization (15 min at 90°C in 10 mM Tris, 1 mM EDTA) and rehybridization of the same filters.
Semiquantitative RT-PCR analysis
Two micrograms total RNA from PAECs was retrotranscribed with oligo-dT primers, using the M-MLV RT (Gibco, Cergy Pontoise, France) according to the manufacturer's instructions. Five percent of the reaction was used in a PCR reaction mix combining 100 pmol primers 1 and 2 (oligonucleotide 1, 5[prime]-AGGAAGAGTGGTTCTGTC-3[prime] corresponding to nucleotides 12-30 of pig [alpha]1,3GT isoform 1; oligonucleotide 2, 5[prime]-GTTATGGTCACGACCTCT-3[prime] corresponding to nucleotides 324-306 of pig [alpha]1,3GT, isoform 1, as described in Vanhove et al. (1997), 200 µM dNTPs, 1.5 mM MgCl2, 10 mM Tris-HCl buffer, 50 mM KCl, and 2 U of Taq polymerase (Gibco). PCR products were separated on agarose gel and cloned into pSK+ vectors using the SureClone ligation kit (Pharmacia). Both strands were sequenced, confirming that the 299 and 263 bp products correspond to isoforms 1 and 2, as previously described (Vanhove et al., 1997). Conditions for PCR amplification were: 35 cycles consisting in 94°C, 30 s; 54°C, 30 s; 72°C, 30 s. To determine the conditions for semiquantitative amplification, the reaction mixture was submitted to increasing numbers of amplification cycles ranging from 15 to 35. For detection of amplified products, primer 1 was labeled with 33P in a kinasing reaction from [gamma]33P-ATP, and 1.25 pmol labeled primers were added to the PCR mix. PCR reactions were resolved on 6% polyacrylamide gels containing 8M urea, and exposed for 12 h on a PhosphorImager for quantification. An amplification process of 18 cycles was selected as it corresponds to the middle of the linear amplification phase.
[alpha]1,3Galactosyltransferase assay
The assay for [alpha]1,3GT activity was adapted from the method described by Joziasse et al. (1990). In brief, lactose (50 mM) was used as acceptor substrate in a cacodylate buffer (50 mM) containing 40 mM MnCl2, 0.8% Triton X-100, 4 mM ATP, 20 mM [gamma]-galactonolactone, 0.5 mg/ml BSA, and 37.5 nCi UDP-(14C)galactose at 305 mCi/mmol (Amersham), diluted in unlabeled 5 mM UDP-Gal (Sigma). After 1 h incubation with 106 cells, the lysate was chromatographed on a 1 ml column of Dowex 1-X8 (Cl- form), equilibrated in H2O. The column was washed with 1 ml water and the radioactivity in the eluate counted. Values were corrected for the reactivity obtained in a control reaction in which the acceptor substrate was absent.
Cell ELISA
Postconfluent PAECs in microtiter plates were stimulated by LPS and fixed with 0.4% glutaraldehyde for 15 min. at 4°C. Cells were stained for 1 h, 37°C, using IgY from hens immunized with rabbit RBC membranes, and immunopurified on Gal[alpha]1,3Gal immunoabsorbent (Synsorb 14, Chembiomed, Canada; generous gift from Dr. J. F. Bouhours), and with rabbit anti-hen Ig labeled with peroxidase (Sigma). Pig E-selectin expression was detected with the BBA-1 mouse monoclonal antibody (Abingdon, UK) and revealed with anti-mouse peroxidase (Sigma).
Acknowledgments
We are grateful to D. Joziasse for expert assistance in enzyme activity assays and to J. F. Bouhours for helpful advice in the writing of the manuscript. B.V. has an E.U. fellowship.
Abbreviations
EC, endothelial cell; PAEC, porcine aortic endothelial cell; [alpha]1,3GT, UDP-Gal, Gal[beta]1-4GlcNAc [alpha]1-3-galactosyltransferase; Gal[alpha]1,3Gal, galactose[alpha]1,3galactose; Lex, Lewis x.
References
1To whom correspondence should be addressed at: INSERM U437, 30 Bd. J. Monnet, F-44035 Nantes Cedex 01, France
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: 2 Apr 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. H. Chen, A. Kadner, R. N. Mitchell, and D. H. Adams FRESH PORCINE CARDIAC VALVES ARE NOT REJECTED IN PRIMATES J. Thorac. Cardiovasc. Surg., June 1, 2000; 119(6): 1216 - 1220. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||