Glycobiology Advance Access originally published online on April 6, 2005
Glycobiology 2005 15(8):735-746; doi:10.1093/glycob/cwi058
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Sequence-variant repeats of MUC1 show higher conformational flexibility, are less densely O-glycosylated and induce differential B lymphocyte responses
2 Department of Obstetrics and Gynaecology, Vrije Universiteit Medical Center, De Boelelaan 1117, 1081 HV Amsterdam, Netherlands; 3 Eppley Cancer Institute, UNMC, 986805 Nebraska Medical Center, Omaha, NE 68198-6805; 4 Institute of Biochemistry II, Medical Faculty, University of Cologne, Joseph-Stelzmann-Str. 52, 50931 Köln, Germany; 5 Institute of Pathology, University Clinic, University of Cologne, Joseph-Stelzmann-Str. 9, 50931 Köln, Germany; and 6 Central Bioanalytics, Center of Molecular Medicine, University of Cologne, Joseph-Stelzmann-Str. 52, 50931 Köln, Germany
1 To whom correspondence should be addressed; e-mail: franz.hanisch{at}uni-koeln.de
Received on September 24, 2004; revised on March 23, 2005; accepted on March 23, 2005
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
The human epithelial cancer mucin MUC1 is able to break tolerance and to induce humoral immune responses in healthy subjects and in cancer patients. We recently showed that clusters of sequence-variant repeats are interspersed in the repeat domain of MUC1 at high frequency, which should contribute to the structural and immunological features of the mucin. Here we elucidated the potential effects exerted by sequence-variant repeats on their O-glycosylation. Evidence from in vitro glycosylation with polypeptide N-acetylgalactosaminyltransferases GalNAc-T1 and GalNAc-T2 in concert with mass spectrometric analyses of in vivo glycosylated MUC1 probes from transiently transfected HEK293 cells indicated reduced glycosylation densities of repeats with three concerted replacements: AHGVTSAPESRPAPGSTAPA. The Pro to Ala replacement in STAPA exerts not only proximal effects on the ppGalNAc-T2 preferred site at -3 and -4, but also more distant effects on the ppGalNAc-T1 preferred site at -15 (TSAPESRPAPGSTAPA). We also examined the conformational changes of MUC1 glycopeptides induced by the concerted DT to ES replacements and revealed a higher conformational flexibility of ES/P peptides compared to DT/P peptides. Differences in conformational flexibilities and in O-glycosylation densities could underlie the observed differential humoral responses in humans. We were able to show that the natural immunoglobulin G (IgG) responses to the repeat domain of MUC1 in sera from nonmalignant control subjects are preferentially directed to variant repeat clusters. In contrast, the IgG response in patients with adenocarcinoma shifted to higher frequencies of preferential DTR peptide binding.
Key words: B cell response / cancer/MUC1 / O-glycosylation / peptide conformation
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
The epithelial type 1 transmembrane mucin MUC1 is established as a tumor marker for monitoring recurrence of breast cancer and is a promising target for immunotherapeutic strategies to treat cancer by active specific immunization. Immunodominant epitopes of MUC1 are localized within the tandem repeat domain, which is comprised of up to 120 repeated 20 amino acid units and exhibits genetic polymorphism in the number and absolute sequence of the repeats. Recent evidence from protein (Müller et al. 1999
) and DNA sequencing (Engelmann et al. 2001; Fowler et al. 2003) has revealed that the repeat peptide sequence contains a high incidence of sequence variants. In particular, the immunodominant DTR motif is replaced in up to 50% of the repeats by an ESR motif, often in concert with a Pro to Ala replacement in position +10 relative to the Thr/Ser (Engelmann et al. 2001). These variant repeats are generally found in clusters (diads, triads, tetrads) interspersed between clusters of nonvariant repeats (Engelmann et al. 2001; Fowler et al. 2003).
Circulating MUC1 induces a humoral response in healthy subjects and in cancer patients (von Mensdorff-Pouilly et al. 2000a; Croce et al. 2001). Patients with early breast cancer and a natural humoral response to MUC1 have a higher probability of freedom from distant metastases and a better disease-specific survival, suggesting that anti-MUC1 antibodies control haematogenous tumor dissemination and outgrowth (von Mensdorff-Pouilly et al. 2000a). Naturally induced anti-MUC1 immunoglobulin (IgG) antibodies recognize more than one minimal epitope within the same individual, with the PDTRP sequence being one of three preferred epitopes present on the tandem repeat peptide (von Mensdorff-Pouilly et al. 2000a). Vaccination with unglycosylated MUC1 tandem repeat peptides induced IgG responses that were largely restricted to two epitopes: PDTRPAP and STAPPAHGV. Interestingly, natural anti-MUC1 antibodies from breast-cancer patients reacted more strongly with GalNAc-modified peptides than with unglycosylated peptides, whereas the immunization-induced response did not react (von Mensdorff-Pouilly et al. 2000b). The latter can be explained by the assumption that peptide vaccines generate antibodies, which do not or only weakly bind to glycosylated MUC1 repeats. Hence the natural MUC1 glycoforms should be more efficiently targeted by antibodies generated to glycosylated repeats.
Attempts to use different MUC1 glycoforms for the experimental induction of cytotoxic-T cell responses (Hiltbold et al. 1999) or major histocompatibility complex (MHC) class II-mediated helper T cell responses (Hiltbold et al. 2000) have indicated that the immunogenicity of these antigens is inversely correlated with the degree of glycosylation. We have previously shown that MUC1 processing by dendritic cells (DCs) is influenced by O-glycosylation (Vlad et al. 2002; Hanisch et al. 2003). Although complex O-linked chains on MUC1 repeats are not removed by DCs (Vlad et al. 2002), the attachment site of the glycans is critical during processing (Hanisch et al. 2003). This explains why native MUC1 glycoforms with glycosylated Thr at the VTSA motif (preferred substrate site of the ubiquitous polypeptide GalNAc-transferase T1) are poor immunogens (Hiltbold et al. 1999; Hiltbold et al. 2000).
Previous investigations of humoral immune responses to MUC1 were exclusively based on the nonvariant repeat peptide sequence, which had been originally described by four independent groups (Lan et al. 1990; Gendler et al. 1990; Ligtenberg et al. 1990; Wreschner et al. 1990). No information is available on the antigenicity and immunogenicity of variant repeat clusters in MUC1. Although the DT to ES replacement is conservative in nature and a structural similarity of the two motifs can be assumed, many of the DTR-specific hybridoma antibodies generated in mice do not cross-react with peptides containing the variant ESR motif.
In this study the attempt was made to define structural differences of sequence-variant MUC1 repeats and to link information on glycopeptide conformations and peptide O-glycosylation with the relative immunogenicity and antigenicity of variant versus nonvariant repeats in humoral responses. Even subtle differences in the conformational propensities of ES versus DT repeats expectedly can lead to a modulation of anti-MUC1 responses. A higher flexibility of ES repeats—as revealed in this study—could be linked to more easily overcoming tolerance with variant peptides. Moreover, a reduced glycosylation density of ES/A sequence-variant repeats (containing the sequence ESRPAPGSTAPA) compared to the invariant DT/P repeats (containing the corresponding sequence DTRPAPGSTAPP) should further differentiate the immune response to subdomains of the mucin. A structure-based knowledge on the immunogenicity of natural MUC1 glycoforms, of its subdomains, and their immunodominant motifs expectedly will aid the design of efficient tumor vaccines.
|
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Glycosylated variant and nonvariant MUC1 repeat peptides differ in their conformational propensities
Many murine hybridoma antibodies are directed to the immunodominant DTR motif in the MUC1 repeat peptide. The lack of cross-reactivity of several of these antibodies with variant ESR repeat peptides (see below) indicates a structural distinctness of DTR- versus ESR-containing peptides. We defined these structural differences by investigating the detailed conformational aspects of the (glyco) peptides in the vicinity of the sequence variations. Two series of AHG21 peptides were used containing either the sequence DTRPAPGSTAPP or the variant sequence ESRPAPGSTAPP, where only the immunodominant motif varied. Whenever possible, the peptides were designated by using the configuration of the three variable positions, DT/P or ES/P (for structural details on glycopeptides used in this study refer to Table I). Numbering of varied or glycosylated amino acids refers to the AHG starting motif of the AHG21 peptides.
|
Analysis of the nuclear magnetic resonance (NMR) data obtained for the ES/P peptide series has shown that most of the proton resonance assignments for amino acid residues at the same positions are very similar to the chemical shifts for the DT/P peptides presented in Tables I and II in Kinarsky et al. (2003). Distinct differences were observed for DT- and ES-containing regions of the peptide. The proton chemical shifts for the ESR segments within the ES/P peptide series are shown in Table II. The nuclear Overhauser enhancement (NOE) connectivities within glycopeptides, AHG21-ES(T5), AHG21-ES(S10), and AHG21-ES(T17), are shown in Figure 1. Several NOE cross-peaks related to sugar–peptide interactions were observed near glycosylated Thr5*, Thr17*, and Ser10* (Table III). The coupling constants for all amino acid residues ranged from 8.1 to 8.6 Hz, which is consistent with an absence of ordered secondary structure. Many strong consecutive d
N (i, i+1) connectivities suggested mostly extended backbone conformations. Sequential dNN connectivities within the range of 3.8–4.0 Å, which are indicative of the presence of inverse 
-turn-like conformations, were observed near all three sites of O-glycosylation: Thr5, Ser10, and Thr17.
|
|
H. An asterisk indicates ambiguous cross-peaks.
|
To determine the conformational features of the peptide backbone and the effect of O-glycosylation at proximal and distant sites, the NMR-derived structures of peptide fragments, VTSA, ESR, and GSTA, from all three glycopeptides, AHG21-ES (T5*), AHG21-ES (E10*), and AHG21-ES (T17*), were selected for cluster analysis. Briefly, the results obtained for the VTSA and GSTA fragments within the ES/P series were similar to previously described results for the DT/P glycopeptides (Tables IV and VI in Kinarsky et al. [2003]). The most populated structural clusters comprised extended ß-strand-like, polyproline II-like, or inverse -turn-like conformations near threonine residues to be glycosylated. (In this context, terms "ß-strand-like, polyproline II-like, or inverse -turn-like conformations" have been used to emphasize that it is not specific secondary structures, but rather conformations of the distinct residues that fall into the regions of the Ramachandran plot that include the ß-strand, polyproline II, or inverse -turn conformations.) However, conformational ensembles of energy-minimized structures for the VTSA and GSTA fragments from the ES/P series, calculated with the same parameters for the analogous segments, demonstrated larger variations than were observed for the DT/P glycopeptides. For example, the number of representative structural clusters for the VTSA fragment was increased from six to nine root-mean-square-deviation (RMSD=0.7 Å). For the GSTA fragment, the number of clusters was increased from seven to nine (RMSD=0.9 Å).
|
|
For the ESR fragment, nine structural clusters were obtained with the RMSD=0.7 Å in comparison to five major clusters that were defined for the DTR fragment of the DT/P peptides. Nevertheless, a significant population of the inverse -turn-like conformations for Glu9 in AHGVTSAPESRPAPGSTAPPA and more extended polyproline II or ß-strand-like structures for Arg11 resembled the structural ensemble described previously for the backbone conformations of the DTR fragment (Kinarsky et al. 2003). Glycosylation affected conformational propensities and slightly diminished structural deviations within the peptide backbone.
Distinct differences were observed in a conformational variability of side chains of Ser10 in AHGVTSAPESRPAPGSTAPPA and Thr10 in AHGVTSAPDTRPAPG STAPPA. All three conformational rotamers of the side chain (
1
–60°, 180°, and +60°) were observed in structural ensembles generated for both nonglycosylated and glycosylated Ser10. For the nonglycosylated Ser10, 160° and –60° were the most populated rotamers among the side chain conformations of low-energy structures; whereas, for the glycosylated Ser10*, the most populated rotamers of the side chain conformations were 1 –60° and 1 –180°. The side chain flexibility of threonine residues was significantly more restricted than the flexibility of the serine side chain. Side chain conformations of the nonglycosylated Thr5 and Thr17 were populated by only two rotamers of the 1 angle: –60° and +60°. After glycosylation, the side chain conformations of Thr5* and Thr17* demonstrated the only populated rotamer of the 1 angle: +60°.
In summary, these NMR data indicate that DT substitution by ES increased conformational variability of the peptide backbone. The differences were found in conformational distributions for the 1 angle of side chains of Ser10 or Thr10. The side chain of serine was more flexible than the branched side chains of threonines. Glycosylation slightly affected conformational preferences for the side chain of serine and diminished conformational flexibility of the side chains of threonine residues.
The sequence variation of TAPP to TAPA is associated with reduced glycosylation densities of variant repeats
The clusters of variant repeats within the repeat domain of MUC1 can be specifically cleaved with the Glu-C specific V8 protease from Staphylococcus aureus resulting in 20-mers with the sequences SRPAPGSTAPPAHGVTSAPE or SRPAPGSTAPAAHGVTSAPE (Figure 2). We were able to demonstrate that most of proteolytic fragments derived from the MUC1 repeat domains (MUC1 from pooled human milk fat membranes of >10 individual donors) exhibit a concerted replacement of Pro to Ala in the PAH motif (Figure 2, refer also to ref. 1 and 2). Current knowledge of the influences of proximal amino acids on the substrate qualities of putative glycosylation sites suggests that Pro in +3 relative to the Thr/Ser is favorable or even essential, whereas other amino acids in that position can abolish activity of some polypeptide GalNAc-transferases.
|
We tested peptides with the concerted replacements (DT/P and ES/P versus ES/A) as substrates for O-glycosylation by polypeptide GalNAc-transferases rGalNAc-T1 and rGalNAc-T2 (Table IV), which are the ubiquitously expressed isoforms. The variant peptide AHG21-ES/A was not an effective substrate for either enzyme, because no GalNAc incorporation was measured after 24h reaction. In contrast, the corresponding nonvariant repeat peptide AHG21-DT/P, the variant control with two concerted replacements AHG21-ES/P, and the extended TAP25-DT/P yielded diglycosylated (rGalNAc-T1) or triglycosylated products (rGalNAc-T2) during the same time period (Table IV). Accordingly, the negative effect on the substrate qualities of ES/A peptide can be assigned to the Pro to Ala replacement. To exclude that location of putative glycosylation positions with respect to number of flanking residues within the peptide might influence enzyme activities, we tested an alternative variant peptide starting with the GST motif (GST20-A/ES). Also this peptide showed reduced O-glycosylation by yielding only monosubstituted product (rGalNAc-T1) or product with traces of GalNAc incorporation (rGalNAc-T2). Interestingly, the Pro to Ala replacement in STAPA exerts not only the expected negative effect on the proximal sites at -3 and -4 preferred by ppGalNAc-T2, but also more distant effects on the ppGalNAc-T1 preferred site at -15 in peptide AHGVTSAPESRPAPGSTAPAA. This distant effect on ppGalNAc-T2 activity is consistently found for peptide substrates with different configuration of the three sequence variations (A/ES vs. ES/A), whereas ppGalNAc-T1 activity is affected only by the ES/A configuration. Nevertheless, both enzymes showed reduced catalytic activities with the variant MUC1 peptides, supporting the hypothesis that variant repeat clusters with the ES/A configuration could be less glycosylated than those with the DT/P configuration, if only ppGalNAc-T1 and/or ppGalNAc-T2 and isoforms with similar substrate specificities were making up the enzymatic repertoir of a cell.
To prove that also in vivo glycosylated MUC1, which was expressed in epithelial cells, displays reduced glycosylation densities of variant repeats, we performed structural analyses of glycopeptides derived from the repeat domain by proteolysis. MUC1 from natural sources, however, contains undefined numbers of variant repeats. We therefore analysed a recombinant MUC1 probe with six sequence-defined repeats (Müller and Hanisch, 2002), which had been expressed in the virally transformed human epithelial kidney cell line HEK293. Of the six repeats three contain the motif AAH (A repeats), two the motif PAH (P repeats) and one the motif QAH (Q repeat). After limited trifluoromethane sulfonic acid treatment which results in quantitative removal of peripheral and backbone sugars, but retains more than 80% of the core-GalNAc, effective tryptic cleavage of the repeat domain occurred at Arg-C with formation of glycosylated PAP20 peptides PAPGSTAP(A,P,Q)AHGVTSAP (DT,ES)R (Figure 3). Although up to 20% of the core-GalNAc is lost during limited TFMSA treatment, the cleavage occurs randomly at various Ser/Thr positions as demonstrated with synthetic glycopeptides (data not shown). Hence, the remaining core-GalNAc can be regarded to authentically reflect the O-glycosylation densities of sequence-variant repeats. The PAP20 glycopeptides were analysed by MALDI-MS (using the mass increment of -26 for PAP20 with a P to A replacement, and +31 for PAP20 with a P to Q replacement) to determine the molar ratio of A:P:Q repeats in the signal triplets (Figure 3). Although the nominal ratio of A:P:Q PAP20 peptides derived from the repeat domain of the construct is 1:1:0.5 (according to five PAP20 peptides cleavable out of a six repeat domain), the measured ratios of peak intensities were 1:3:0 (for PAP20 with 4 GalNAc residues attached) and 1:1:0.25 (for the unglycosylated PAP20), respectively (Figure 3). This corresponds to reduced glycosylation densities of both A and Q repeats (irrrespective of the DT to ES variation, which can not be discriminated by MS). Similar experiments had previously been performed with the same recombinant MUC1 probe expressed in a panel of breast-cancer cell lines, however, the partial de-O-glycosylation had been achieved by exoglycosidase treatment (Müller and Hanisch, 2002). The ratio of peak intensities corresponding to the three sequence variations agreed largely with the expected ratio of A:P:Q repeats, which implies that the cancer cells did not express A and Q repeats with reduced O-glycosylation densities. In summary, some ppGalNAc-Ts, like the ubiquitously expressed ppGalNAc-T1 and ppGalNAc-T2, apparently are unable to effectively O-glycosylate variant MUC1 peptides with the ES/A configuration, whereas others, like those expressed in some cancer cell lines, seem to be less affected by the sequence variations. Accordingly, both the tissue-specific and differentiation-dependent ppGalNAc-T repertoirs and the enzyme-specific sequence context necessary for efficient GalNAc transfer are likely to account for the differential O-glycosylation of ES/A versus DT/P repeats.
|
Natural IgG responses in healthy subjects show preferential binding to variant repeat clusters
Binding to oligorepeat peptides.
We determined the relative IgG binding of individual serum samples to 61-meric peptides with the DT/P or ES/P configurations and focussed in this way on B cell responses to the immunodominant DTR epitope and its sequence variant. The relative reactivity of individual samples was calculated as the ratio of binding activity toward ESR oligorepeats to that on DTR oligorepeats (100%). Following this definition a relative reactivity <100% indicates a stronger binding to the DTR than to the ESR peptide, and a relative reactivity >100% indicates a stronger binding to the ESR than to the DTR peptide.
The relative reactivity of the control IgG and IgM samples (mean ± SD) was 85 ± 3,2 % and 103 ± 2.4%, respectively. The interassay coefficient of variation of the control IgG and IgM sample was, respectively, 3.8 and 2.3%. IgM responses in the total study population gave relative reactivities (95 ± 6%) that were normally distributed and did not differ significantly among control subjects (healthy subjects, pregnant women, and patients with benign breast tumors) and patients with adenocarcinoma (results not shown). In contrast, IgG responses were not normally distributed. The relative reactivity of the samples for IgG responses is contained in Table V and Figures 4A and B. Relative binding ranked higher in samples from control subjects than in samples from patients with adenocarcinoma, but these differences were not significant (P = .084). As the 99.7% confidence interval (mean ± 3 SE) for the IgG control sample mean ranged from 75.94 to 94.86%, we defined samples with a relative reactivity of 100 ± 10% as having a neutral or indifferent binding, recognizing equally well the DTR and the variant ESR sequences. IgG responses from control subjects were more frequently directed to the ES/P peptide, whereas in the group of patients with adenocarcinoma the frequencies of preferential binding to one of the alternative repeat clusters shifted to the DT/P peptide (P = .003) (Figure 4B).
|
|
Binding to monorepeat glycopeptides.
IgG from patients with adenocarcinoma showed higher relative binding to the AHG21 glycopeptides than to the unglycosylated control peptides (Table V), but these differences were statistically significant only for glycopeptide H14 (P = .013). Relative binding to the glycopeptide H12 was higher (P = .025) for the samples classified as having preferential binding to the DTR motif (N = 14) than for the samples with preferential binding to the ESR motif (N = 20). This significant difference in binding was still present for control samples (P = .022), but not in adenocarcinoma samples. Six samples showed extreme low binding to one or more glycopeptides, five of these were control samples; only one was an adenocarcinoma sample. Whereas individual control samples varied widely in the pattern of reactivities with the different glycopeptides, adenocarcinoma samples tended to recognize all glycopeptides equally well (Figure 4C).
Cross-reactivity of monoclonal DTR-specific mouse antibodies to ESR-peptides
A panel of murine hybridoma antibodies (Price et al. 1998) was tested for their relative binding to DT/P and ES/A repeat peptides (Table VI). Two groups of antibodies were included in this panel: (1) MAbs generated to synthetic DT/P peptides or to recombinantly expressed fusion proteins lacking ESR motifs (VA1, VA2, BCP7, BCP8, BCP9) and (2) MAbs generated to MUC1 from natural sources (DF3, SM3, HMFG-1, BC2, B27.29, E29, b-12, C595). In the first group, only one antibody (VA1) bound exclusively to the DT/P peptide (TAP25) and showed no cross-reactivity to ES/A peptides (AHG21-ES/A, GST20-A/ES), whereas two were weakly cross-reactive (BCP8, VA2). BCP7 and BCP9, two further antibodies with strong binding to ES/A peptides, differ from the other antibodies in this group, in that they define proximal epitopes common to variant and nonvariant peptides (Table VI). In the second group about half of the antibodies showed significant cross-reactivity to ES/A peptides (B27.29, HMFG-1, b-12, C595), while the others were specific for the DT/P peptide (DF3, BC2, E29, SM3).
|
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
According to genetic sequence polymorphisms, the MUC1 gene contains clusters of variant sequences within its repeat domain, the incidence of which can fluctuate individually, but can reach up to 50% of all repeats (Engelmann et al. 2001
; Fowler et al. 2003). On the protein level, these extended stretches of peptides (repeat diads to pentads) were known to exhibit concerted amino acid replacements at the immunodominant motifs, but neither the hereby induced alterations of peptide conformation and O-glycosylation nor the related changes in immunogenicity of the tumor marker were investigated. Structural data presented in this study now revealed that variant repeat clusters in the MUC1 repeat domain are distinct from nonvariant repeat clusters by two major features: (1) The ESR motif within the variant repeat peptide exhibits a higher conformational flexibility compared to the DTR motif. (2) According to in vitro studies with two ppGalNAc-Ts and in vivo O-glycosylation of a recombinant probe in HEK293 cells, most of variant ES/A repeat clusters should be less densely O-glycosylated than the corresponding DT/P peptide clusters (Figure 5).
|
Although only two out of a panel of 15 reported ppGalNAc-Ts were tested in this study, ppGalNAc-T1 and ppGalNAc-T2 are the ubiquitously and constitutively expressed isoforms. Cancer cells may express a different repertoir of ppGalNAc-Ts compared to normal resting epithelium, as shown for ppGalNAc-T3 in well-differentiated adenocarcinoma cells (Sutherlin et al. 1997). In accordance with this, structural analyses of MUC1 probes secreted by four breast-cancer cell lines (T-47D, MCF-7, MDA-MB231, ZR75-1) did not reveal evidence for reduced glycosylation densities of variant ES/A peptides compared to DT/P peptides (Müller and Hanisch, 2002). In this study, when the same probe was expressed in a virally transformed human epithelial kidney cell line, a reduced glycosylation density of repeats with a AAH and QAH motif became apparent. The AAH motif is often found in concert with the ES sequence variation in a representative mixture of endogenous MUC1 from pooled human milk samples, which implies that in natural glycoforms the clusters of ES/A repeats can be less densely O-glycosylated than the invariant DT/P repeats as depicted in Figure 5.
In vitro negative effect of sequence variation on enzyme activity could be clearly assigned to the Pro to Ala replacement in ES/A repeats and was exerted on the Ser/Thr positions in GSTA (at -3 and -4), but also on the more distant Thr position in VTSA (at -15 or +5, respectively). For ppGalNAc-T2 this effect was independent of the N- or C-terminal location of the Pro to Ala replacement (A/ES vs. ES/A peptides) and in agreement with previous findings (Hanisch et al. 2001). Good substrates of ppGalNAc-T2, like the DT/P peptides tested in this study, are initially glycosylated at Thr in STAPP, which triggers further O-glycosylation at less preferred sites via a postulated lectin domain-mediated mechanism. This would explain why glycosylated ES/A peptide, like AHGVT(GalNAc)SAPESRPAPGSTAPAA, was found to be a good substrate for ppGalNAc-T2, which was glycosylated at Ser and Thr in STAPA, irrespective of the Pro to Ala replacement (Hanisch et al. 2001).
Structural evidence for the relatively weak glycosylation of variant ES/A versus invariant DT/P repeats in vivo is based on a single cellular model, it tells however, that a cell-specific repertoir of ppGalNAc-Ts can account for differential O-glycosylation densities of sequence-variant repeats. Further studies will have to elucidate, which enzyme repertoirs characterize the normal, resting breast epithelium and breast carcinoma cells and which of these enzymes are able to add GalNAc to variant ES/A repeats.
The structural distinctness of variant and nonvariant peptides may also explain their antigenicity as is reflected in the selective binding of several murine DTR-specific monoclonal antibodies to MUC1. Besides obvious side chain differences, the increased conformational variability of ESR-containing peptides may enable a population of specific backbone conformations that are characteristic for recognition and binding of the DTR-containing peptides by DTR-specific antibodies. This may partially explain results summarized in Table VI. For antibody B27.29, which recognizes the RPAP segment adjacent to DT(ES) residues (Price et al. 1998), TAP25 was bound best among the peptides tested here, presumably because of a less flexible DTR sequence and a longer N-terminal segment. As suggested by Kirnarsky et al. (2000), the lengths of N-terminal segments of peptide "ligands" might significantly affect their structural flexibility and conformational propensities.
The structural distinctness of variant repeat clusters and commensurate reduced glycosylation densities could explain the predominance of responses to the ESR motif, and the variable binding to the glycosylated peptides observed in the natural responses to circulating MUC1 glycoforms in control subjects. According to studies on the processing of O-glycosylated MUC1 in the MHC class II pathway (Vlad et al. 2002; Hanisch et al. 2003), the activation of helper T-hybridoma cells is dependent on site-specific effects of glycans attached to the major cathepsin L processing site (Hanisch et al. 2003). No difference was found in cathepsin L-catalyzed cellular processing of ES/A and DT/P repeat peptides. Hence, the peptide-directed humoral response to the variant and invariant repeat clusters should be largely dependent on the sites and average density of O-glycosylation. The shift to increased frequencies of preferential binding to DTR-peptide clusters, as well as an equal recognition of all glycoforms by IgG from patients with adenocarcinoma can be explained by the tumor-associated changes in O-glycosylation of MUC1: a general reduction of the glycan chain lengths. This could lead to facilitated accessibility of the nonvariant peptide backbone and may enable immune responses to it. Results of previous studies (von Mensdorff-Pouilly et al. 2000b) have demonstrated that humoral responses to MUC1 are polyclonal. Results from this study suggest that this polyclonality is more extensive in carcinoma patients than in controls.
Interestingly, the natural responses to the variant repeats found in normal control subjects may show that overcoming tolerance to ES repeats occurs more easily than to nonvariant DT repeats. Increased conformational variability of ES/P peptides versus DT/P peptides along with the IgG binding data may link the tolerance and conformational flexibility of B-cell epitopes: more flexible peptide epitopes might more easily overcome tolerance to such epitopes at the level of B cells. A possible explanation of this phenomenon may involve decreased binding affinity of ES-containing epitopes because of peptide flexibility. This may result in more diverse antibodies with ES than DT reactivity. Easy brake of tolerance to variant repeat peptides does not mean that ESR motifs—as possibly low-affinity epitope—could be better antigens for immune interventions than the nonvariant DT/P repeats. Whether DT-specific responses are more efficient because of increased affinity or lower-clonal heterogeneity remains to be determined in mouse models by using transgenic MUC1 mice as well as in clinical settings.
|
Materials and methods |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Controls, patients, and serum samples
Fifty-two serum samples with high levels of IgG and/or immunoglobulin M (IgM) antibodies to MUC1 were obtained from 21 control subjects (one healthy man, 10 healthy women, four pregnant women, and six benign breast-tumor patients) and from 31 patients with adenocarcinoma before primary treatment (28 breast, two ovarian, and one gastric carcinoma), selected from a screened population of more than 2000 subjects. Screening for humoral responses to a 61-meric MUC1 tandem repeat peptide had been performed as previously described (von Mensdorff-Pouilly et al. 1996
) at initial serum dilutions of 1:100 for IgG and 1:500 for IgM determinations. The selected samples had a median optical density units (OD) of 1.093 (range 0.447–2.047) and 0.971 (range 0.224–1.655) for IgG and IgM determinations, respectively, equally distributed among patients and controls. Serum samples were collected, aliquoted, and stored at –80°C until analyzed.
Peptides and glycopeptides
The MUC1 oligorepeat peptides were synthesized in local facilities (HGV61-DT/P) or kindly provided by Dr Olivera Finn, University of Pittsburgh Medical School, Pittsburgh, USA (AHG61-ES/P). The monorepeat peptide TAP25-DT/P was obtained from Dr Joyce Taylor-Papadimitriou, Imperial Cancer Research Fund, London, UK. Variant monorepeat peptides, AHG21-ES/A and GST20-A/ES, were synthesized by a local facility at the Center of Biochemistry in Cologne, Germany. Pepceuticals Ltd, Leicester, UK, synthesized the monorepeat peptides AHG21-DT/P and the variant monorepeat AHG21-ES/P. All glycopeptides based on the AHG21 sequence with two concerted replacements (H11–H16: AHGVTSAPESRPAPGSTAPPA) and with GalNAc monosubstitution at Thr5, Ser10, or Thr17 were synthesized by Prof Dr Hans Paulsen, Institute of Organic Chemistry, University of Hamburg, Germany, according to previously published protocols (Karsten et al. 1998). For structural details refer to Table I.
Binding assays of human natural IgG/IgM to MUC1 peptides and glycopeptides
Binding to oligorepeat peptides.
Serum samples were tested for binding to a 61-meric MUC1 tandem repeat peptide (HGV61-DT/P), and the oligorepeat peptide AHG61-ES/P, with an enzyme-linked immunosorbent assay (ELISA) as previously described (von Mensdorff-Pouilly et al. 1996; von Mensdorff-Pouilly et al. 2000a). The assay was performed for each serum sample/peptide in duplicate, and results were calculated as the mean difference between the readings in OD in peptide-coated and BSA-coated wells. The relative reactivity of the samples with the AHG61-ES/P peptide was calculated as the ratio of the ES versus DT peptide binding within the same plate. A MUC1 IgG-positive serum sample from a healthy control and a MUC1 IgM-positive serum sample from a breast-cancer patient were tested correspondingly in every plate to define interassay variability.
Binding to glycopeptides.
Serum samples were tested for IgG binding to the glycopeptides H11 to H16 and to the corresponding unglycosylated monorepeat peptides, AHG21-DT/P and AHG21-ES/P. The following modifications were introduced in the assay: The two peptides and one set of glycopeptides (H11 to H13, or H14 to H16), each conjugated to BSA, and BSA were adsorbed in duplicate columns in 96-well ELISA plates. The samples were incubated at a dilution of 1:100 one sample per row in sets of two plates. Results were calculated as the mean difference between the OD measured in experimental wells (BSA-conjugated peptide/glycopeptide) and BSA-coated control wells. The relative reactivity of the samples with the glycopeptides was calculated as the ratio of glycopeptide versus peptide binding within the same plate.
Statistical methods
Statistical analysis was performed using SPSS software (Version 9.0, SPSS Inc, Chicago, IL, USA). Assay results were tested for normality of distribution. The correlation among assay results obtained with the peptides and glycopeptides was evaluated by linear regression analysis. Distribution of variables in contingency tables was analyzed with a 2 test or, when the sample size was small, a Fisher’s exact test. The percentage of reactivity with the peptides and glycopeptides in the different groups was analyzed using the Mann–Whitney U/Wilcoxon rank sum W Test. In all cases, a two-tailed P = .05 was considered significant.
Binding assays of murine hybridoma antibodies to MUC1 peptides
The assays were performed in triplicate on 96-well microtitration plates by coating with 10 µg/ml of the respective peptides as described previously (Karsten et al. 1998). Out of a panel of murine antibodies to the MUC1 peptide core a selection with reactivity to the peptide TAP25-DT/P (Price et al. 1998) was tested for cross-reactivity to variant peptides (AHG21-ES/A, GST20-A/ES). Cross-reactivity of antibodies relative to the reference peptide TAP25-DT/P was calculated on the basis of averaged optical densities and expressed as +++ (>70%), ++ (30–70%), + (<30%) or – (no binding to variant repeat peptides).
600 MHz nuclear magnetic resonance spectroscopy of glycosylated MUC1 repeat peptides
The NMR experiments used to acquire structural data included total correlation spectroscopy (TOCSY), nuclear Overhauser enhancement spectroscopy (NOESY), rotating frame NOE spectroscopy (ROESY), double quantum filtered correlation spectroscopy, and HSQC (heteronuclear single quantum coherence). All data were acquired on a 600-MHz Varian INOVA spectrometer operating at 14.1 Tesla. The samples were dissolved in 90:10 H2O:D2O (600 µL) to a concentration of 3 mM, and adjusted to pH 4.5; all data were acquired at a sample temperature of 5°C. Water suppression was accomplished by using presaturation (using a 50-Hz field), except for the HSQC experiment, which used pulsed field-gradient solvent suppression. The mixing times used were 75 msec for TOCSY, 150 msec for ROESY, and 100, 200, and 400 msec for the NOESY data sets. Resonance assignments of protons in the glycopeptides were obtained from the TOCSY and NOESY spectra (Wüthrich, 1986). In addition, the HSQC spectrum was used to assist resonance assignments of carbohydrate protons. 3JN
vicinal coupling constants were obtained using the DQF-COSY spectra. The peak intensities in the NOESY spectrum (200 ms) were converted into the upper distance constraints and classified as strong, medium, weak, and very weak corresponding to distance ranges of 1.8–2.5 Å, 1.8–3.0 Å, 1.8–4.0 Å, and 1.8–5.0 Å, respectively. Structure calculation and refinement procedures were described previously (Kinarsky et al. 2003; Kirnarsky et al. 2000). From 200 structures of each glycopeptide, generated by the Dynamics Algorithm for Nmr Applications program (Güntert et al. 1997), 50 conformers with lowest constraint violations were taken for the structural analysis. These structures were energy minimized and 25 low-energy structures of each peptide were pooled together and clustered using a RMSD criterion for the backbone atoms of the VTSA, ESR, and GSTA segments.
Mass spectrometric analysis of proteolytic fragments from MUC1 repeat domains
The peptide and glycopeptide samples (20 µl) contained in 0.1% aqueous trifluoroacetic acid (TFA) or in mixtures with acetonitrile (ACN) were applied to the stainless steel target by mixing one volume of sample with matrix (one volume of saturated solution of -cyano-4-hydroxycinnamic acid in ACN/0.1% aq. TFA, 2:1, or two volumes of 20 mg dihydroxy benzoic acid in ACN/0.1% aq. TFA, 1:2). Mass spectrometric analysis was performed on a Bruker-Reflex IV instrument (Bruker-Daltonic, Bremen, Germany) by positive ion detection in the reflectron mode. Ionization of cocrystallized analytes was induced with a pulsed nitrogen laser beam (337 nm), and the ions were accelerated in a field of 20 kV and reflected at 23 kV ( Müller et al. 1997).
In vitro glycosylation of variant and nonvariant MUC1 repeats
The enzymes rGalNAc-T1 and rGalNAc–T2 were kindly provided by Dr Henrik Clausen, School of Dentistry, Faculty of Health Sciences, University of Copenhagen, Denmark. Peptide substrates (100–500 µM) were dissolved in 25 mM cacodylate buffer, pH 7.4 containing 10 mM MnCl2 and mixed with UDP-GalNAc (200 µM) and rGalNAc-Ts to a total volume of 20 µl as previously described (Hanisch et al. 2001). After 24 h incubation at 37°C aliquots of the reaction mixtures were diluted 1:20 in 0.1% aqueous TFA and mixed with the double volume of dihydroxy benzoic acid (20 mg/ml 0.1% aqueous TFA/acetonitrile, 2:1) on the target. The dried samples were analyzed by MALDI mass spectrometry on a Bruker Reflex IV as described above.
Structural analysis of proteolytic fragments from endogenous and recombinant MUC1
The recombinant MUC1 construct MFP6 was transiently expressed in the human epithelial kidney cell line HEK293 (American type culture collection) and isolated from the culture supernatant as described previously (Müller and Hanisch, 2002). The mucin probe (100 µg) was partially de-O-glycosylated by limited treatment with trifluoromethane sulfonic acid (Sigma, Unterhaching, Germany) in the presence of anisol (10% vol.) at 0°C for 60 min ( Müller et al. 1997) and desalted after neutralization with pyridine/water (60% vol.) on a NAP-5 column (Amersham Biosciences Europe, Freiburg, Germany). The dried sample was solubilized in ammonium hydrogencarbonate (0.1 M) containing 1 mM CaCl2 and digested with trypsin (Promega, Mannheim, Germany) at an enzyme/substrate ratio of 1% for 18h at 37°C. After vacuum evaporation the sample was taken up in 0.1% aq. TFA and chromatographed on a narrow-bore C18 HPLC column (Ultrasphere C18, 2 x 150 mm, Beckman Coulter, Unterschleißheim, Germany) using a gradient of the solvent mixtures A (2% acetonitrile/0.1% TFA) and B (80% acetonitrile/0.1% TFA; 6 to 26% solvent B during 30 min). The vacuum dried fractions were analysed by MALDI-MS.
Endogenous MUC1 from milk fat membranes (50 µg) was partially de-O-glycosylated ( Müller et al. 1997) and further treated in 0.1 M citrate–phosphate buffer, pH 3.8 with 3.3 µg/ml -N-acetylgalactosaminidase (Sigma) for 16h at 37°C. The deglycosylated mucin was solubilized in 0.1 M ammonium hydrogencarbonate, pH 7.8 and incubated with 1.25 µg of endoproteinase Glu-C from Staph. aureus (Roche) for 16h at 37°C.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
This investigation was supported by NIH grant R01 CA84106 to M.A.H., S.S., and F.G.H. and by a grant from the Deutsche Krebshilfe (70–2975-Ba 2) to S.E.B. and F.G.H. The authors acknowledge the skilful technical assistance of K. van Uffelen and K. Ottenberg. The authors acknowledge Dr. G. Wang, Dr. P. Keifer, and G. Suryanarayanan for the NMR data collection and processing.
|
Abbreviations |
|---|
ACN, acetonitrile; BSA, bovine serum albumin; HPLC, high-performance liquid chromatography; HSQC, heteronuclear single quantum coherence; Ig, immunoglobulin; MALDI, matrix-assisted laser desorption ionization; NMR, nuclear magnetic resonance; NOE, nuclear Overhauser enhancement; NOESY, nuclear Overhauser enhancement spectroscopy; OD, optical density units; ppGalNAc-T, UDP-GalNAc, polypeptide N-acetylgalactosaminyl-transferase; RMSD, root mean square deviation; TOCSY, total correlation spectroscopy; TFA, trifluoroacetic acid
|
References |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Croce, M.V., Isla-Larrain, M.T., Capafons, A., Price, M.R., and Segal-Eiras, A. (2001) Humoral immune response induced by the protein core of MUC1 mucin in pregnant and healthy women. Breast Cancer Res. Treat., 69, 1–11.[Medline]
Engelmann, K., Baldus, S.E., and Hanisch, F.-G. (2001) Identification and topology of variant sequences within individual repeat domains of the human epithelial tumor mucin MUC1. J. Biol. Chem., 276, 27764-27769.
Fowler, J.C., Teixeira, A.S., Vinall, L.E., and Swallow, D.M. (2003) Hypervariability of the membrane-associated mucin and cancer marker MUC1. Hum. Genet., 113, 473–479.[Medline]
Gendler, S., Lancaster, C.A., Taylor-Papadimitriou, J., Duhig, T., Peat, N., Burchell, J., Pemberton, L., Lalani, E.N., and Wilson, D. (1990) Molecular cloning and expression of the human tumor-associated polymorphic epithelial mucin PEM. J. Biol. Chem., 265, 15286–15293.
Güntert, P., Mumenthaler, C., and Wüthrich, K. (1997) Torsion angle dynamics for NMR structure calculation with the new program DYANA. J. Mol. Biol., 273, 283–298.[CrossRef][ISI][Medline]
Hanisch, F.-G., Reis, C.A., Clausen, H., and Paulsen, H. (2001) Evidence for glycosylation-dependent activity of polypeptide N-acetylgalactosaminyl-transferases rGalNAc-T2 and-T4 on mucin glycopeptides. Glycobiology, 11, 731–740.
Hanisch, F.-G., Schwientek, T., von Bergwelt-Baildon, M.S., Schultze, J.L., and Finn, O.J. (2003) O-Linked glycans control glycoprotein processing by antigen-presenting cells: a biochemical approach to the molecular aspects of MUC1 processing by dendritic cells. Eur. J. Immunol., 33(3242–32), 54.
Hiltbold, E.M., Alter, M.D., Ciborowski, P., and Finn, O.J. (1999) Presentation of MUC1 tumor antigen by class I MHC and CTL function correlate with the glycosylation state of the protein taken up by dendritic cells. Cell Immunol., 194, 143–149.[CrossRef][ISI][Medline]
Hiltbold, E.M., Vlad, A.M., Ciborowski, P., Watkins, S.C., and Finn, O.J. (2000) The mechanism of unresponsiveness to circulating tumor antigen MUC1 is a block in intracellular sorting and processing by dendritic cells. J. Immunol., 165, 3730–3741.
Karsten, U., Diotel, C., Klich, G., Paulsen, H., Goletz, S., Müller, S., and Hanisch, F.-G. (1998) Enhanced binding of antibodies to the DTR motif of MUC1 tandem repeat peptide is mediated by site-specific glycosylation. Cancer Res., 58, 2541–25490.
Kinarsky, L., Suryanarayanan, G., Prakash, O., Paulsen, H., Clausen, H., Hanisch, F.-G., Hollingsworth, M.A., and Sherman, S. (2003) Conformational studies on the MUC1 tandem repeat glycopeptides: implication for the enzy-matic O-glycosylation of the mucin protein core. Glycobiology, 13, 929–939.
Kirnarsky, L., Prakash, O., Vogen, S., Nomot, O.M., Hollingsworth, M.A., and Sherman, S. (2000) Structural effects of O-glycosylation on a, 15-residue peptide from the mucin (MUC1) core protein. Biochemistry, 39, 12076–12082.[CrossRef][Medline]
Lan, M.S., Batra, S.K., Qui, W.-N., Metzgar, R.S., and Hollingsworth, M.A. (1990) Cloning and sequencing of a human pancreatic tumor mucin cDNA. J. Biol. Chem., 265, 15294–15299.
Ligtenberg, M.J., Vos, H.L., Gennissen, A.M.C., and Hilkens, J. (1990) Episialin, a carcinoma associated mucin, is generated by a polymorphic gene encoding splice variants with alternative amino termini. J. Biol. Chem., 265, 5573–5578.
von Mensdorff-Pouilly, S., Gourevitch, M.M., Kenemans, P., Verstraeten, A.A., Van Kamp, G.J., Kok, A., van Uffelen, K., Snijdewint, F.G., Paul, M.A., Meijer, S., and Hilgers, J. (1996) An enzyme-linked immunosorbent assay for the measurement of circulating antibodies to polymorphic epithelial mucin (MUC1). Tumor Biol., 19, 186–195.
von Mensdorff-Pouilly, S., Verstraeten, A.A., Kenemans, P., Snijdewint, F.G.M., Kok, A., Van Kamp, G.J., Paul, M.A., Van Diest, P.J., Meijer, S., and Hilgers, J. (2000a) Survival in early breast cancer patients is influenced by a natural humoral immune response to polymorphic epithelial mucin (MUC1). J. Clin. Oncol., 18, 574–583.
von Mensdorff-Pouilly, S., Petrakou, E., Kenemans, P., van Uffelen, K., Verstraeten, A.A., Snijdewint, F.G.M., van Kamp, G.J., Schol, D.J., Reis, C.A., Price, M.R., and others (2000b) Reactivity of natural and induced human antibodies to MUC1 mucin with MUC1 peptides and GalNAc peptides. Int. J. Cancer, 86, 702–712.[CrossRef][ISI][Medline]
Müller, S., Alving, K., Peter-Katalinic, J., Zachara, N., Gooley, A.A., and Hanisch, F.-G. (1999) High density O-glycosylation on tandem repeat peptide from secretory MUC1 of T47D breast cancer cells. J. Biol. Chem., 274, 18165–18172.
Müller, S., Goletz, S., Packer, N., Gooley, A.A., Lawson, A.M., and Hanisch, F.-G. (1997) Localization of O-glycosylation sites on glycopeptide fragments from lactation-associated MUC1. J. Biol. Chem., 272, 24780–24793.
Müller, S., and Hanisch, F.-G. (2002) Recombinant MUC1 probe authentically reflects cell-specific O-glycosylation profiles of endogenous breast cancer mucin. J. Biol. Chem., 277, 26103–26112.
Price, M.R., Rye, P.D., and, 59 coauthors. (1998) Summary report on the ISOBM TD-4 Workshop: analysis of 56 monoclonal antibodies against the MUC1 mucin. Tumor Biol. 19 (Suppl. 1), 1–20.
Sutherlin, M.E., Nishimori, I., Caffrey, T., Bennett, E.P., Hassan, H., Mandel, U., Mack, D., Iwamura, T., Clausen, H., and Hollingsworth, M.A. (1997) Expression of three UDP-GalNAc: polypeptide GalNAc N-acetylgalactosaminyltransferases in adenocarcinoma cell lines. Cancer Res., 57, 4744–4748.
Vlad, A.M., Müller, S., Cudic, M., Paulsen, H., Otvos, L., Hanisch, F.-G., and Finn, O.J. (2002) Complex carbohydrates are not removed during processing of glycoproteins by dendritic cells: processing of tumor antigen MUC1 glycopeptides for presentation to major histocompatibility complex class II-restricted T cells. J. Exp. Med., 196, 1435–1446.
Wreschner, D.H., Hareuveni, M., Tsarfaty, I., Smorodinky, N., Horev, J., Zaretzky, J., Kotkes, P., Weiss, M., Lathe, R., and Dion, A. (1990) Human epithelial tumor antigen cDNA sequences. Differential splicing may generate multiple protein forms. Eur. J. Biochem., 189, 463–473.[ISI][Medline]
Wüthrich, K. NMR of Proteins and Nucleic Acids. John Wiley and Sons, New York (1986).
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Ninkovic and F.-G. Hanisch O-Glycosylated Human MUC1 Repeats Are Processed In Vitro by Immunoproteasomes J. Immunol., August 15, 2007; 179(4): 2380 - 2388. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Freire, R. Lo-Man, F. Piller, V. Piller, C. Leclerc, and S. Bay Enzymatic large-scale synthesis of MUC6-Tn glycoconjugates for antitumor vaccination Glycobiology, May 1, 2006; 16(5): 390 - 401. [Abstract] [Full Text] [PDF]
|
|
| ||||||||