Immunoscreening reveals cDNAs encoding lepidopteran-specific valine-rich midgut proteins (Vmps)

To identify proteins that are associated with the PM, we performed an immunoscreening using a midgut λ-ZapII cDNA library from M. sexta and polyclonal anti-PM antibodies that we have generated against the components of purified PM from larval midguts (see Material and Methods). From this screening, we obtained ten phage-clones, which were still immune-reactive after two rounds of re-screening. The cDNA inserts from these plaques were sequenced after in vivo excision of the corresponding pBluescript (SK-) plasmids. A tblastn database search revealed positive hits for cDNAs encoding proteins that are typically found in the lepidopteran midgut, such as trypsin, chymotrypsin and β-glycosidase (Table 1). Most strikingly, three of the identified cDNA sequences showed similarities to putative cuticle proteins from Bombyx mori. A tblastn search of the recently sequenced Manduca sexta genome (assembly v1.0) revealed six additional genes with sequence similarities to these three identified cDNAs. These Manduca genes encode members of a well-defined family of nine proteins. We termed these proteins M. sexta valine-rich midgut proteins 1-9 (MsVmp1-9) due to the high percentage of valine residues (19-31%, Table 2). Besides this high valine content, MsVmps are also characterized by a higher than normal content of proline residues, but they are highly deficient in histidine, cysteine, tryptophan and tyrosine residues (Figure S1, supporting information). The theoretical molecular masses of the proteins range between 16.4 and 28.8 kDa (including the signal peptides), and the isoelectric points between 3.3 and 3.8 (Table 2). They all contain several xPA(V/I)xx motifs in the central region and the highly conserved sequence (S/A)PLVQIIV(N/K) (Figure 1). In contrast to PM proteins (PMPs) and invertebrate intestinal mucins, they lack chitin-binding domains belonging to the CBM 14 family, which have six conserved cysteines with a characteristic spacing between these residues. Due to the high content of hydrophobic amino acids (GRAVY indices: 0.64-0.93, Table 2), we analysed these sequences for the presence of putative transmembrane helices. In none of the MsVmps, however, transmembrane helices were computationally predicted by TMHMM 2.0. All MsVmps contain signal peptides with predicted cleavage sites at amino acid positions 15 for Vmp1, position 16 for Vmp2, -3, -5, -6 and -8, position 20 for Vmp4 and -9 and position 24 for Vmp7. Moreover, Vmp2, 3 and 6 have potential single N-glycosylation sites, and all Vmps except Vmp9 have multiple O-glycosylation sites (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. ClustalW alignment of valine-rich midgut proteins from M. sexta (MsVmps).

Highly conserved or identical amino acids are highlighted with light grey, grey or black shadings. The consensus sequence is given below. The accession numbers are as follows: MsVmp1 (Msex012254-PA), MsVmp2 (central part of Msex012257-PA), MsVmp3 (C-terminal part of Msex012257-PA), MsVmp4 (N-terminal part of Msex012261-PA), MsVmp5 (N-terminal part of Msex012257-PA), MsVmp6 (Msex011100-PA), MsVmp7 (C-terminal part of Msex012261-PA), MsVmp8 (Msex012260-PA), and MsVmp (Msex012262-PA).

https://doi.org/10.1371/journal.pone.0082015.g001

Database searches of various genomes using blastp and a conserved region corresponding to amino acid positions 70-146 of MsVmp1 as query revealed highly significant hits (E values <1e-10) only for lepidopteran species, with the lowest E value of 4e-14 detected for the hypothetical cuticle protein Cph38 from Bombyx mori (Table S2, supporting information). Similar results were obtained when using tblastn. No significant hits were obtained for nucleotide or protein sequences of species belonging to any of the other insect orders, or to crustaceans, fishes, reptilians, birds or mammals. The best hits obtained for non-lepidopteran insects had e-values >0.5e-3. Thus, VMPs constitute a family of lepidopteran-specific proteins. Homologs of these genes were found in the genomes of several lepidopteran species including B. mori (13 genes) Spodoptera frugiperda (11 genes), Helicoverpa armigera (6 genes), Trichoplusia ni (5 genes), and Plodia interpunctella (2 genes). The deduced protein sequences of all of these genes contained xPA(V/I)xx motifs and a region similar to the (S/A)PLVQIIV(N/K) consensus sequence. Phylogenetic analyses of lepidopteran VMPs using the most conserved region (amino acid positions 70 to 146 of MsVmp1) revealed that the VMP genes have greatly diversified throughout lepidopteran evolution presumably starting with the duplication of an ancestral gene (Figure 2). Generally, the B. mori VMP genes were most closely related to the MsVMPs, reflecting the close phylogenetic relatedness between the Bombycidae and Sphingidae families, to which B. mori and M. sexta belong, respectively. Homologs of MsVMPs were not found in clades of the Noctuidae and Pyralidae families. Interestingly, five of the 13 B. mori genes share greatest similarities with VMP genes from S. frugiperda, H. armigera and T. ni (Noctuidae). These B. mori genes may have been derived from gene duplication events, as they form a monophyletic clade within a group of different lepidopteran species. The finding that M. sexta does not have orthologs for any of these five B. mori genes further suggests that their common ancestor has been lost in M. sexta during evolution of the lepidopteran lineage, because more distantly related species do have orthologous relationships.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Phylogenetic tree of lepidopteran VMPs.

The un-rooted maximum likelihood tree was calculated on the basis of a ClustalW alignment of VMPs from different lepidopteran species. Bootstrap values are given in percentages at the internodes; Accession numbers refer to the entries of the dbEST database of Butterflybase. MS, Manduca sexta; BM, Bombyx mori; TN, Trichopulsia ni; HA, Helicoverpa armigera; SF, Spodoptera frugiperda; PI, Plodia interpunctella.

https://doi.org/10.1371/journal.pone.0082015.g002

VMP genes are expressed in the larval midgut during feeding stages

To analyze the expression of MsVMP genes in M. sexta, we performed RT-PCR using total RNA from different developmental stages and various tissues. While significant MsVMP gene expression was observed in larvae of different ages, no transcripts were detectable in eggs, pupae and adults (Figure S2, supporting information). The MsVMPs were expressed specifically in larval midgut tissues, while no expression was detected in hindgut, trachea, epidermis, fat body, salivary glands or Malpighian tubules (Figure 3). While the expression of MsVMP6 was highest in the anterior midgut, expression of MsVMP4, -7, and -8 was highest in the median midgut, expression of MsVMP1, -2, -3, and -5 was highest in the posterior midgut, and that of MsVMP9 was equally high in median and posterior midguts. The low levels of transcripts detected for MsVMP2, -5 and -7 in the hindgut and for MsVMP6 in the Malpighian tubules are likely due to contamination during tissue dissection. To test whether the nutritional state or molting affects MsVMP expression, we compared mRNA levels in feeding, starving (36 h) and molting larvae (metamorphic molt) by qPCR. All MsVMP genes appeared to be expressed in the feeding stage, while molting or starvation leads to a significant drop in the detected mRNA levels (Figure S3, supporting information). Thus, MsVMP gene expression appears to be regulated depending on the nutritional state and/or molting, when the larvae also discontinue feeding.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Tissue specific expression of MsVMP genes in fifth instar larvae of M. sexta.

Total RNA was prepared from various tissues and cDNAs were synthesized. RT-PCR was carried out with primers specific to the indicated genes. PCR products of indicated sizes were separated by agarose gel electrophoresis and stained with ethidium bromide. Products for the ribosomal protein MsRpS3 were used as a loading control. Expression was found exclusively in the midgut.

https://doi.org/10.1371/journal.pone.0082015.g003

As qPCR showed highest transcript levels for MsVMP1 compared to the other MsVMP genes, we analysed this protein in greater detail. To investigate the localization of MsVmp1, we generated polyclonal antibodies to the recombinant protein. For this purpose we expressed Vmp1 in bacteria and purified it by Ni-NTA chromatography. Recombinant MsVmp1 exhibited an anomalous migration behaviour in SDS-PAGE, as it migrates at approximately 50 kDa, more than twice the theoretical molecular mass of 21.7 kDa calculated for MsVmp1 (Figure S4, supporting information). This unusual migration behavior was not influenced by omitting reducing agents during SDS PAGE, denaturing the protein with deionized urea or incubating the sample at room temperature instead of boiling it before loading. Thus, the reason for this unusual migration behavior remains elusive. As we were absolutely sure about the protein identity, we used the recombinant MsVmp1 to immunize guinea pigs. When we tested the obtained anti-Vmp1 antibodies, they reacted strongly with the recombinant protein as well as with two midgut proteins of about 37 and 39 kDa in a crude midgut extract from feeding 5th instar larvae (Figure S4, supporting information). In line with our results from RT-PCR, western blots using protein extracts from the anterior, median and posterior midgut showed that MsVmp1 is highly abundant only in the posterior midgut, and that protein amounts were highest in feeding larvae, and decreased in starving and molting larvae (Figure 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Immunoblotting detects MsVmp1 in the midgut of feeding, starving and molting larvae.

Crude protein extracts from the anterior, median and posterior midgut were separated by SDS-PAGE, blotted onto nitrocellulose and stained with anti-VMP1 antibodies. The given molecular mass was estimated using standard proteins of known molecular masses.

https://doi.org/10.1371/journal.pone.0082015.g004

MsVmp1 is modified by O-glycosylation

The observation that the proteins detected with the anti-Vmp1 antibodies in midgut extracts were significantly larger than the theoretical molecular mass for MsVmp1 suggested that they are modified by glycosylation, significantly increasing their apparent molecular masses in SDS-PAGE. As bacteria lack the eukaryotic glycosylation machinery, we expressed MsVMP1 in insect cells using a baculoviral expression system, and purified the protein by Ni-NTA chromatography. This recombinant version of MsVmp1 exhibited a molecular mass of about 39 kDa in SDS-PAGE, which was in the range of molecular masses of midgut proteins detected with the anti-Vmp1 antibodies. To test for glycosylation, we performed lectin staining using Western blots of MsVmp1 purified from insect cells. Out of several lectins that we have tested, only peanut agglutinin (PNA) reacted with the recombinant protein indicating that MsVmp1 is O-glycosylated, because PNA detects the core disaccharide galactose-β(1–3)-N-acetylgalactosamine found in many O-glycans. The lack of a positive reaction for Galanthus nivalis agglutinin may indicate the absence of high mannose N-glycans. The lack of a reaction with Sambucus nigra, Maackia amurensis and Datura stramonium agglutinins may further indicate the absence of sialic acid linked either (2,3) or (2-6) to galactose, and the absence of galactose-β(1–4)-N-acetylglucosamine disaccharides (Figure 5). To provide final proof for MsVmp1 being a glycoprotein, we subjected the recombinant protein from insect cells to enzymatic deglycosylation using PNGase F and O-glycosidase. While the treatment with PNGase F did not significantly affect the molecular mass of MsVmp1 in SDS PAGE, incubation with O-glycosidase resulted in deglycosylation of MsVmp1 with intermediates appearing at 37 kDa and 26 kDa (Figure 5). The smallest protein band corresponds to the completely deglycosylated protein. Its size was about 22 kDa, which is in good agreement with the theoretical molecular mass of 21.7 kDa of MsVmp1. Thus, we provide evidence that MsVMP1 is O-glycosylated and that the sugar moieties account for about 17 kDa of the mature MsVmp1 glycoprotein with a molecular mass of about 39 kDa.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. MsVmp1 is extensively O-glycosylated.

(A) MsVMP1 was expressed in insect Sf21 cells and purified by Ni-NTA chromatography. After SDS-PAGE, the proteins were transferred to a nitrocellulose membrane and stained with Ponceau Red (left) and subsequently with different lectins (right lanes). GNA, Galanthus nivalis agglutinin; PNA, peanut agglutinin; SNA, Sambucus nigra agglutinin; DSA, Datura stramonium agglutinin; MAA, Maackia amurensis agglutinin. (B) For deglycosylation, MsVmp1 (expressed in insect cells) was treated with PNGase F or O-glycosidase in the presence of protease inhibitors. The reaction products were separated by SDS PAGE and stained with Coomassie Blue. Left lane, MsVmp1 input without addition of glycosidase; Middle lane, PNGase F treatment; Right lane, O-glycosidase treatment. Std, standard proteins with indicated molecular masses in kDa. Arrows point to reaction intermediates and the terminal deglycosylation product with indicated molecular masses.

https://doi.org/10.1371/journal.pone.0082015.g005

MsVmp1 is loosely associated with the PM and does not bind chitin

MsVmp1 was originally identified in an immunoscreening using polyclonal anti-PM antibodies indicating that the protein is associated with the PM. Indeed, MsVmp1 was detectable with the anti-Vmp1 antibodies in western blots of proteins that were extracted from PM by SDS-treatment (Figure 6A). Also native PM dissected from the posterior midgut and washed only with PBS buffer could be visualized by treatment with anti-Vmp1 antibodies followed by fluorescent secondary antibodies (Figure 6B). In agreement with the expression pattern of MsVMP1 along the midgut, PM prepared from the anterior midgut did not react with these antibodies. These results suggest that in the posterior midgut, MsVmp1 is associated at least weakly with the PM even though the PM has been washed extensively. The amino acid sequence of MsVmp1, however, does not contain regions similar to known chitin-binding domains, such as CBM14 domains or R&R domains, which have been shown (and/or predicted) to mediate tight binding to chitin fibrils. To test whether MsVmp1 protein is able to bind chitin, we carried out chitin-binding assays using colloidal chitin beads and recombinant MsVmp1 expressed in insect cells. The flow-through after washing the beads contained about as much MsVmp1 as the input material indicating that MsVmp1 did not bind to chitin under these conditions. Consistent with this interpretation, after washing or elution with CFW or SDS, no MsVmp1 was detectable in the collected wash and eluate fractions, even in western blots stained with anti-Vmp1 antibodies (Figure S5, supporting information).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 6. Immunodetection of MsVmp1 in PM preparations from anterior and posterior midguts of feeding larvae.

(A) PM proteins were extracted by SDS treatment and separated by SDS-PAGE. Then the proteins were transferred to nitrocellulose and reacted with polyclonal anti-Vmp1 antibodies. Std, standard proteins with molecular masses indicated in kDa. (B) Immunodetection of Vmps using anti-Vmp1 antibodies. The PM preparations from the anterior and posterior parts of the midgut were washed several times with PBS buffer, blocked with bovine serum albumin and stained with the anti-Vmp1 antibodies. Cy3-conjugated anti-guinea pig IgGs were used as secondary antibodies. The PM was transferred to a microscope slide and mounted with Vectashield under a cover slip. The specimens were viewed under a fluorescence microscope using appropriate excitation an emission filters.

https://doi.org/10.1371/journal.pone.0082015.g006

MsVmp1 is produced by columnar cells of the posterior midgut of feeding larvae and found in the gut lumen during starvation and molting

To localize MsVmp1 protein, we generated cryosections of the posterior midgut from second instar larvae that were feeding, starving or molting, and stained them with the anti-Vmp1 antibodies. In well fed larvae, MsVmp1 was detected predominantly in the cytosol of columnar cells of the posterior midgut. The signals inside the cells appeared to be in punctate structures (presumably transport vesicles) that are somewhat enriched at the apical site (Figures 7 and S6). The second major cell type in this single-layered gut epithelium, the goblet cells, was essentially free of fluorescence signal (Figure S6, supporting information). In addition, signals were observed in the ectoperitrophic space, indicating secretion of some of the MsVmp1 (Figure 7). In some regions, MsVmp1 was found in the outer region of the PM, as indicated by partial overlap of signals from anti-Vmp1 antibodies and CFW staining in feeding larvae. The endoperitrophic space of the gut lumen was essentially free of MsVmp1 under these conditions. In contrast, in starving and molting larvae, MsVmp1 was not detected in columnar cells, but in the endoperitrophic gut lumen. Apparently, MsVmp1 is produced in columnar cells during the feeding periods, and secreted into the ectoperitrophic space but initially restrained by the PM (Figure 7). To reach the endoperitrophic lumenal space in starving or molting larvae, MsVmp1 has to penetrate the PM, whose permeability appears to be compromised under these conditions compared to the well-fed state. In contrast to feeding larvae, we usually fail to detect the PM in starving or molting larvae in cryosections using CFW, indicating major structural changes of the PM and/or turnover. During periods of starvation or molting, we typically observe macroscopic swelling of the PM, which additionally turns into a highly fragile structure.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 7. Immunodetection of MsVmp1 in the posterior midgut of M. sexta larvae at different physiological conditions.

Cryosections of posterior midguts were stained with CFW and immune-labeled with anti-VMP1 antibodies. Primary antibodies were detected with ALEXA 488-conjugated anti-guinea pig IgGs. Brightfield and fluorescence images (for anti-Vmp1 and CFW), and overlays of fluorescence images are shown. Cryosections were obtained from feeding 2^nd instar larvae (top), starving 2^nd instar larvae (middle) and larvae molting from the 2^nd to the 3^rd instar. C, cuticle; EC, ectoperitrophic space; EN, endoperitrophic space; PM, peritrophic matrix. Size bar, 100 µm.

https://doi.org/10.1371/journal.pone.0082015.g007

Chemicals

Routine chemicals were purchased in P. A. quality from Sigma-Aldrich (Hamburg, Germany). Primers (Tab. S1, supporting information) were synthesized by Eurofins MWG Operon (Ebersberg, Germany).

Experimental animals

Manduca sexta larvae were reared under long-day conditions (16 h of light at 27 °C) using a synthetic diet modified according to [31].

Immunoscreening

Immunoscreening of a λ-Zap-cDNA library from the posterior midgut of M. sexta was performed as described previously [32]. About 1 x 10⁵ plaque-forming units were screened using a 1:150 dilution of the polyclonal anti-PM antiserum (see below), and a 1:10,000 dilution of the secondary anti-guinea-pig antibody (Sigma-Aldrich). Following three re-screening steps with stepwise reduced phage titers, several antibody-positive clones were isolated. The corresponding cDNAs were sub-cloned by in vivo excision into pBlueskript (SK-) according to [33].

Gene expression studies

Total RNA was isolated by using the TRIZOL reagent according to the manufacture’s protocol (Invitrogen, Carlsbad, CA, USA). cDNA synthesis was performed with Superscript III first strand synthesis system (Invitrogen). The expression profile for each gene was determined with pairs of gene-specific primers (Tab. S1, supporting information). The constitutively expressed house-keeping gene encoding the M. sexta ribosomal protein S3 (MsRpS3) was used as a loading control. The PCR reactions were carried out at the following conditions: denaturation at 92 °C for 30 s, annealing at 57 °C for 45 s and elongation at 72 °C for 1 min for 30 cycles. PCR products were separated by agarose gel electrophoresis and stained with ethidium bromide. qPCR was carried out as described previously [34].

Heterologous expression of recombinant proteins

For antigen preparation, the cDNA encoding MsVmp1 was expressed in E. coli BL21 Codon plus (DE3)-RIL. The MsVMP1-cDNA was amplified using the primers pet29b-VMP1-F and pET29b-VMP1-R (Tab. S1, supporting information). The resulting PCR product was restricted with BamHI and EcoRI, and ligated into the linearized pET29b vector (Novagen), which encodes a carboxy-terminal hexa-His-Tag for protein purification. Expression in E. coli BL21 cells was induced by adding 1 mM IPTG and continued for 6 h at RT. For baculoviral expression, the MsVMP1-cDNA was amplified using the primers pFBD-VMP1-F and pFBD-VMP1-R (Tab. S1, supporting information). The resulting PCR product included a sequence encoding a carboxyterminal octa-His-Tag. The MsVMP1-cDNA was ligated into the pFast-Bac-Dual vector (pFBD, Invitrogen), which additionally contained a cDNA sequence encoding the enhanced green fluorescence protein (EGFP) under the control of the p10 promoter [35]. The resulting pFBD-MsVMP1 vector was used to transform DH10Bac cells, to allow transposition into the baculovirus bacmid. The bacmid DNA was isolated according to the Invitrogen protocol, and used for the transfection of Spodoptera frugiperda (Sf21) cells to generate a P1 viral stock. After 72 h, budded viruses were harvested from the cell culture medium. A baculoviral stock was used to generate a high-titer P2 stock. Subsequently, the P2 stock was used to infect Sf21 cells, which were assayed for expression of the recombinant protein starting after 3 days. The cells were harvested 5 days after infection, when significant EGFP expression could be observed with a fluorescence microscope (Zeiss Discovery V8.0). Affinity purification of proteins by Ni-NTA agarose column was performed according to the manufacturer’s protocol (His·Bind^® Kit, Novagen/Merck, Darmstadt, Germany).

Generation of antibodies

To obtain anti-PM antibodies, the PM was isolated from ten 5^th instar larvae of M. sexta, and washed several times with PBS (20 mM KH₂PO4, 20 mM NaH₂PO₄, 0.15 M NaCl, pH 7.4) until they appeared completely transparent and devoid of any food material. The washed PM was frozen in liquid nitrogen, thoroughly ground and resuspended in 1 ml deionized H₂O, and finally used to immunize two guinea pigs (Charles River Inc., Sulzfeld, Germany). To generate anti-Vmp1 antibodies, recombinant MsVmp1 purified from E. coli BL21 cell lysates was used for the immunization of two guinea pigs (Charles River, Sulzfeld, Germany), whose pre-immune sera did not cross-react with midgut proteins from Manduca 5^th instar larvae. Four booster injections each involving 150 µg protein were administered. After the fourth booster injection, the antisera were collected and purified by protein-A affinity chromatography.

Cryosectioning and immunohistochemistry

The tissue samples were fixed overnight at 4 °C in a solution of 4 % (w/v) formaldehyde in PBS. After removing the fixation solution, the tissue was washed for 3 x 10 min in 20 mM PBS. Cryosectioning and immunohistochemistry were performed as described previously [36] [9]. For localization of MsVmp1, the cryosections were treated with a 1:100 dilution of the anti-Vmp1 antibodies. Visualization of the primary antibodies was performed with ALEXA Fluor 488 goat anti-guinea pig antibodies at a 1:200 dilution (Invitrogen). To test for unspecific binding of the secondary antibodies, control reactions were carried out without primary antibodies. Next, the sections were rinsed 3 times for 5 min with PBS. For Calcofluor white (CFW, Sigma) staining, the sections were incubated in a 0.01% (w/v) solution for 60 min at room temperature and washed three times with PBS. Finally, the specimens were covered with Vectashield Mounting Medium (Vector Laboratories Inc, Burlingame), the coverslips were sealed with Fixo gum and the sections were viewed under an Olympus IX70 fluorescence microscope using appropriate filter sets.

SDS-Page, immuno- and lectin-labeling

Crude protein extracts and different larval tissues were prepared as described previously [37]. SDS-PAGE was carried out essentially as described by 38, and semidry electroblotting (Biorad) onto nitrocellulose membranes (Millipore, Schwalbach, Germany) was performed according to [39], modified by the addition of 20% (v/v) methanol. Membranes were stained with 0.02% (v/v) Ponceau S (Sigma-Aldrich). Immunostaining was performed as described previously [40]. Polyclonal anti-VMP1 antibodies (1:100) were used as primary antibodies, and anti-guinea pig antibodies conjugated to alkaline phosphatase were used as secondary antibodies (1:10,000; Sigma-Aldrich). Lectin binding was examined with the DIG Glycan Differentiation Kit (Roche Diagnostics, Grenzach-Wyhlen, Germany) following the manufacturer’s recommendations.

Glycosidase treatment

Peptide N-Glycosidase F (PNGase F, Roche) cleaves asparagine linked oligosaccharide from glycoproteins. For deglycosylation, 10 µg recombinant MsVmp1 was treated with 4 U PNGase F in a total volume of 100 µl using a buffer consisting of 20 mM sodium phosphate, pH 7.2. MsVmp1 was denatured by boiling it for 2 min in the presence of 0.1% (w/v) SDS. Next, a 10-fold molar excess of Triton-X-100 (compared to SDS) was added to the sample to avoid inactivation of PNGase F by SDS. Subsequently, PNGase F was added and the reaction mixture was incubated for 6 h at 37 °C. O-glycosidase cleaves unsubstituted gal-β(1-3)galNAc-alpha disaccharides attached to serine or threonine residues of glycoproteins or glycopeptides. Treatment of Vmp1 with O-glycosidase was conducted by incubating 2 µg of MsVmp1 in 50 µL of a 0.02 M sodium phosphate buffer (pH 6.0) containing 0,1% (w/v) SDS, 1% (v/v) Non-ident P-40, 1 mM phenylmethylsulfonylfluoride and 2 mU O-glycosidase (Roche Diagnostics, Rotkreuz, Swiss) at 37 °C for 24 h. Following deglycosylation, the samples were analyzed by SDS-PAGE.

Other methods

Protein concentrations were determined by the Amido Black method [41]. The chitin binding assay was performed using a modified protocol reported by [42]. In this assay, 50 µg purified recombinant MsVmp1 protein purified from insect cell lysates and diluted in 1 ml of a binding buffer containing 0.5 M NaCl, 0.05% (w/v) Triton-X100, 10 mM Tris-HCl (pH 7) and a protease inhibitor cocktail (Roche) was bound overnight at 4°C to 1 ml of washed colloidal chitin beads (New England Biolabs, Ipswich, MA, USA). The suspension was transferred to a 2 ml plastic column, and the beads were washed with 3 column volumes binding buffer and 1 column volvolume 5 mM NaCl, 5 mM Tris-HCl (pH 7). Elution was carried out either with 1% CFW or 2% SDS (each in PBS buffer). Flow-through, wash and elution fractions were collected, and the beads were boiled with 1% SDS. Each fraction was subjected to SDS PAGE. The proteins were either stained with Coomassie blue or transferred to a nitrocellulose membrane and immune-labeled with anti-VMP1 antibodies. Nucleotide sequencing was done by Seqlab GmbH, Göttingen, Germany. Prediction of signal peptide cleavage sites was performed with SignalP 4.0 [43]. GRAVY indices were determined according to [44]. Scanning for potential N- and O-glycosylation sites was done with the NetNGlyc 1.0 and DictyOGlyc 1.1 Servers (http://www.cbs.dtu.dk/services/NetNGlyc and ~DictyOGlyc). Transmembrane helices were predicted with the TMHMM 2.0 server [45]. For phylogenetic analyses, the highly conserved region of MsVmp1 (amino acid positions 70-146) was used for “blastn”, “blastp” and “tblastn” searches of different protein and nucleotide databases (www.ncbi.nlm.nih.gov/blast, www.butterflybase.org, http://butterflybase.ice. mpg.de). Alignment of amino acids sequences was performed with ClustalW (EMBL-EBI), and the maximum-likelihood tree was constructed with the Mega 5.0 software package.