Báo cáo khoa học: The N-glycans of yellow jacket venom hyaluronidases and the protein sequence of its major isoform in Vespula vulgaris

The N-glycans of yellow jacket venom hyaluronidases and the protein sequence of its major isoform in Vespula vulgaris

Tác giả: Daniel Kolarich, Renaud Léonard, Wolfgang Hemmer, Friedrich Altmann

Lĩnh vực: Hóa học, Miễn dịch học, Sinh học phân tử

Nội dung tài liệu: Nghiên cứu này tập trung vào việc phân tích cấu trúc N-glycans của enzyme hyaluronidase trong nọc ong bắp cày Vespula vulgaris, một trong những dị nguyên chính gây dị ứng. Các nhà khoa học đã sử dụng phương pháp MALDI-TOF MS và HPLC để xác định các cấu trúc glycan phức tạp, chủ yếu là các glycan paucimannosidic được difucosyl hóa. Nghiên cứu cũng xác định được một isoform mới của hyaluronidase, được gọi là Ves v 2b, là dạng chiếm ưu thế trong nọc ong bắp cày Vespula vulgaris. Phân tích chuỗi protein của isoform này cho thấy sự tương đồng về trình tự với các hyaluronidase đã biết khác, đồng thời xác định được các vị trí glycosyl hóa. Kết quả nghiên cứu cung cấp hiểu biết sâu sắc hơn về vai trò của cấu trúc carbohydrate trong phản ứng dị ứng chéo giữa nọc ong bắp cày và ong mật, cũng như sự khác biệt giữa các isoform của hyaluronidase.

Mục lục chi tiết:

• Keywords
• Correspondence
• Note
• Hyaluronidase (E.C. 3.2.1.35), one of the three major allergens of yellow jacket venom, is a glycoprotein of 45 kDa that is largely responsible for the cross-reactivity of wasp and bee venoms with sera of allergic patients.
• The asparagine-linked carbohydrate often appears to constitute the common IgE-binding determinant.
• Using a combination of MALDI MS and HPLC of 2-aminopyridine-labelled glycans, we found core-difucosylated paucimannosidic glycans to be the major species in the 43-45 kDa band of Vespula vulgaris and also in the corresponding bands of venoms from five other wasp species (V. germanica, V. maculifrons, V. pensylvanica, V. flavopilosa and V. squamosa).
• Concomitant peptide mapping of the V. vulgaris 43 kDa band identified the known hyaluronidase, Ves v 2 (SwissProt P49370), but only as a minor component.
• De novo sequencing by tandem MS revealed the predominating peptides to resemble a different, yet homologous, sequence.
• cDNA cloning retrieved a sequence with 58 and 59% homology to the previously known isoform and to the Dolichovespula maculata and Polistes annularis hyaluronidases.
• Close homologues of this new, putative hyaluronidase b (Ves v 2b) were also the major isoform in the other wasp venoms.
• Almost 50% of patients with suspected hymenoptera allergy turn out to be sensitized to both honeybee and yellow jacket venom [1].
• Contrasting with these in vitro results, only very few patients show adverse reactions to both venoms.
• In two recent studies, protein-bound carbohydrate was shown to be a major, but not the only, cause for the simultaneous reactivity of patients’ IgE with honeybee and yellow jacket venoms [2,3].
• Apart from cases where true double sensitization to both venoms or antibodies against cross-reactive protein appears to be involved, a frequently found situation is that patients sensitized to one of the venoms cross-react with the other venom on the sole basis of protein-linked carbohydrate, which forms the so-called cross-reactive carbohydrate determinants (CCDs) [2].
• Probably, CCDs also cause some of the cross-reactivity observed for hyaluronidases from hornet (Dolichovespula sp.), paper wasp (Polistes sp.) [4] and even bumble bee and fire ant venom [5,6].
• The protein responsible for most of the carbohydrate based reactivity on immunoblots of yellow jacket venom migrates at 43 kDa under denaturing conditions [2].
• A rabbit serum raised against plant glycoproteins (oilseed rape extract) likewise primarily bound to the 43 kDa protein [2].
• This band is believed to represent hyaluronidase for which, in the case of Vespula vulgaris, the primary structure is known from cDNA sequencing (Swissprot P49370) [4] and which is known to be glycosylated [7].
• However, in a recent investigation, by MALDI-TOF MS, of the 43 kDa band, this hyaluronidase P49370 constituted only a minor component and the strong signals could not be assigned to a known protein sequence [2].
• In this study, a mixture of venoms from V. vulgaris and V. germanica was employed and thus the unknown peptide peaks were interpreted as possibly arising from V. germanica hyaluronidase.
• This explanation was wrong, as shown below.
• Irrespective of the nature of the protein, protein-linked glycans can bind IgE, which turns many proteins, especially those of higher molecular mass, into apparent allergens.
• Even though some studies have shown in vitro histamine release by plant glycoprotein glycans [8-10], the contribution of such carbohydrate determinants to clinical symptoms is unclear and is believed, by many researchers, to be negligible [2,3,5,11,12].
• However, patients’ sera containing anti-glycan immunoglobulin can bind to a variety of plant and insect glycoproteins and even to human proteins unrelated to any allergen in the peptide part when the glycan has been modified with core α1,3-fucose, as shown for patients 9-14 in a previous study [11].
• The glycan-based cross-reaction of honeybee-allergic patients’ sera with wasp venom, and vice versa for that of wasp venom-allergic patients’ sera with bee venom, could be inhibited by oilseed rape pollen extract [2].
• Furthermore, the inhibition could be performed with BSA, to which small glycopeptides containing core α1,3-fucose (and xylose) had been attached [2].
• Thus, it can be speculated that yellow jacket hyaluronidase (or, more precisely, the 43 kDa protein) carries core α1,3-fucosylated N-glycans, such as the honeybee venom hyaluronidase [13].
• However, Vespula venom and the proteins therein have never been subjected to any structural analysis of protein-linked carbohydrate.
• Moreover, not even the polypeptide of hyaluronidase has been analyzed, to date, and the data on its sequence have been obtained exclusively from cDNA cloning [4].
• Here we report on the identification of the major polypeptide in the 43 kDa band of V. vulgaris as a new isoform of hyaluronidase.
• Furthermore, we analyzed the hyaluronidase band in five other common yellow jacket species and we investigated the N-glycan structures of hyaluronidase from all six species.
• Results
• Immunological detection of venom glycoproteins
• Electrophoretic separation of V. vulgaris venom under reducing and denaturing conditions yielded three major bands (Fig. 1).
• Antigen 5 migrated at 28 kDa, phospholipase at 34 kDa and the putative hyaluronidase formed a double band at 43-45 kDa.
• On immunoblots with a rabbit serum that binds complex plant N-glycans (CCDs), one major band (at 43 kDa), believed to be hyaluronidase, became visible.
• Faint bands of lower molecular mass (34-40 kDa) were believed to represent degradation products of hyaluronidase.
• A sharp band at c. 105 kDa was shown by MS to be a glycoprotein (data not shown) but could not be identified.
• Only the 43 kDa band was subjected to N-glycan analysis.
• Analysis of N-glycans of the 43 kDa band and of whole venom
• The major glycan species found by MALDI-TOF MS had masses indicating them to consist of two GlcNAc, either two or three mannose residues and two fucose residues (Fig. 2).
• Hence, they resembled the difucosylated paucimannosidic N-glycans (Fig. 3), known from honeybee venom phospholipase A2 and hyaluronidase [13,14].
• Apart from quantitative differences, essentially the same results were obtained when this analysis was conducted with the hyaluronidase bands from V. maculifrons (Table 1).
• In this case, the structure of the presumed N-glycan structure MMF3F6 (Fig. 3) was verified by 2D HPLC and fucosidase digestion.
• However, for this purpose, whole venom from V. maculifrons was employed in order to obtain the necessary amount of glycans, as this venom is the least expensive.
• As hyaluronidase is the dominating glycoprotein in the venom, we believe that the glycans from whole venom essentially stem from, and therefore credibly represent, those from hyaluronidase.
• First, pyridylaminated glycans were fractionated according to size (Fig. 4).
• A large peak eluting at 21.7 min was collected and further analyzed.
• Its elution time on reverse-phase HPLC matched that of MMF3F6 from honeybee phospholipase.
• A moderate dose of bovine kidney α-fucosidase caused a shift to a lower retention time, which is consistent with the removal of an α1,6-linked fucose residue from the reducing end GlcNAc [14,15].
• The glycan mass and the removal of one fucose residue were confirmed by MS (Fig. 4).
• In summary, we conclude that wasp venom hyaluronidases contain the difucosylated paucimannosidic N-glycans MUF3F6 and MMF3F6 (Fig. 3) as the major species.
• In addition, glycans containing one or two GlcNAc residues at the nonreducing end and trace amounts of other, probably hybrid-type, N-glycans were present, but have not been further analyzed.
• Proof that MMF3F6 is, in fact, bound to hyaluronidase came from tandem MS of a glycopeptide with a mass of 2037.0 Da, which at first could not be assigned to any known sequence.
• Only in retrospect was it recognized as having the peptide sequence NGTYEIR (which does not occur in P49370, see below), with a glycan of the composition Man3GlcNAc2Fuc2.
• Among the various glycopeptide and oligosaccharide fragments, the fragment at m/z = 1347.6, which still contained two fucoses but only one GlcNAc residue, exemplified best the difucosylation of the asparagine-linked GlcNAc residue (Fig. 5).
• Proteomic characterization of the 43 kDa band
• As mentioned in a previous publication [2], peptide mapping of the 43 kDa band revealed considerable inconsistency between the data bank entry for hyaluronidase and the tryptic peptides obtained from the 43 kDa band of V. vulgaris venom (Fig. 6).
• Only a few, smaller, signals could be assigned to SwissProt entry P49370 (Fig. 6).
• Hence, the tryptic peptides were subjected to nano-HPLC separation followed by tandem mass spectrometric peptide sequencing.
• BLAST analysis of the sequences deduced from the fragment spectra soon revealed moderate, but obvious, sequence homologies with hyaluronidase P49370 (Fig. 7).
• We concluded that the 43 kDa band contained a second isoform of hyaluronidase – which from the intensity of MALDI and ESI peaks – actually constituted the major form.
• To assess the true relative amount of the isoforms, the band was subjected to Edman sequencing, which revealed one sequence that deviated from the sequence of P49370 (Fig. 7), but no clear indication for the presence of the known hyaluronidase (data not shown).
• It may be added here that both potential glycosylation sites were found as glycopeptides by either MALDI-TOF MS (site 81, data not shown) or ESI MS (site 66, see also previous chapter).
• Furthermore, the two glycopeptides were detected in both the upper and the lower part of the double band, at 43-45 kDa, thus ruling out underglycosylation as a reason for the different migration behaviour.
• Molecular cloning of the new hyaluronidase isoform, Ves v 2b
• Aiming at the molecular cloning of the cDNA of the newly discovered hyaluronidase isoform (referred to as Ves v 2b in the present article, in contrast to the known form, Ves v 2), the N-terminal sequence and internal fragments near the C terminus were translated into degenerate primers for PCR.
• Extension by 3′-RACE led to identification of the entire sequence.
• This new sequence was found to contain two glycosylation sites.
• Site 66 yields a tryptic peptide of mass 851.44 Da, which is identical to the peptide mass observed for the (glyco-)peptide dealt with in Fig. 5.
• It is notable that from the same cDNA we could also isolate a clone of the ‘old’ isoform, Ves v 2 (which probably should be referred to as Ves v 2a from now on).
• Simple BLAST search with the new hyaluronidase protein yielded 58% homology with P49370 and 59% similarity with the hyaluronidases from P. annularis and D. maculata.
• Sequence alignment with the currently known insect venom hyaluronidases is shown in Fig. 7.
• Like its homologues, Ves v 2b contained several hyaluronan-binding motifs (i.e. two basic amino acids spaced by seven other residues) [16].
• Ves v 2b-homologues in other Vespula species
• The MALDI spectra of the hyaluronidase bands from other wasp venoms (V. germanica, V. flavopilosa, V. maculifrons and V. pensylvanica) were almost identical to that from V. vulgaris.
• This corroborates the close relatedness of these wasp species and, at the same time, identifies hyaluronidase b (possibly with a few species-specific differences) as the major isoform in these species.
• In contrast, the hyaluronidase band from V. squamosa yielded a totally different MALDI peptide map, which suggests a low sequence homology with the hyaluronidases from the other species.
• Discussion
• The similarity of the glycan structures on wasp and bee venom hyaluronidases explains at least to a large extent their allergological cross-reactivity.
• The role of venom hyaluronidases as major cross-reactive allergens in honeybee and Vespula venom has been recognized previously [17] and was thought to be a result of the significant sequence identity (≈ 50%) between these allergens [4].
• Although there is some evidence for cross-reactivity between honeybee and Vespula hyaluronidases at the protein level [2,7], its significance in relation to the cross-reactivity mediated by glycans should be redefined in future studies.
• The cross-species survey performed in this study showed that the venoms from all six of the species usually considered (V. vulgaris, V. germanica, V. flavopilosa, V. maculifrons, V. pensylvanica and V. squamosa) contained MUF3F6 and MMFF6 as the major glycan structures (Fig. 3).
• Sera from patients with antibodies directed to the core α1,3-fucose determinant will therefore inevitably react with the hyaluronidase in venoms of all types of hymenoptera, irrespective of the underlying protein.
• These sera may be expected to bind also with other venom components bearing similar glycans, such as phosphatases [18], serine proteases [19], and honeybee and fire ant phospholipases [6,14,20].
• Moreover, in such cases it is not even clear whether the original sensitizing agent was an insect venom or a plant allergen, which also contain this carbohydrate determinant [21].
• Especially in the case of such ‘plant-allergic’ patients, a positive in vitro test against insect venoms can be expected to be unassociated with clinical symptoms from insect stings.
• Likewise, IgE against carbohydrates induced by Hymenoptera stings leads to a positive in vitro test with pollen allergens but is not associated with symptoms of pollinosis [5].
• This shows, once again, the importance of a discrimination between carbohydrate- and protein-based IgE binding.
• Incidentally, analysis of glycans from the wasp venom hyaluronidase led to the identification and molecular cloning of Ves v 2b, the true major protein in the 43 kDa band of V. vulgaris venom.
• It is currently unknown whether Ves v 2b represents a ‘true’ allergen, or binds with IgE only through its carbohydrate determinants.
• The other isoform, Ves v 2 (or Ves v 2a, as we would like to call it), was originally isolated by using primers that had been designed for the hyaluronidase of D. maculata [4].
• In fact, the higher