key: cord-271470-j58mr9xk
authors: Zhu, Feifei; Li, Dong; Song, Dandan; Huo, Shuhao; Ma, Shangshang; Lü, Peng; Liu, Xiaoyong; Yao, Qin; Chen, Keping
title: Glycoproteome in silkworm Bombyx mori and alteration by BmCPV infection
date: 2020-04-29
journal: J Proteomics
DOI: 10.1016/j.jprot.2020.103802
sha: 
doc_id: 271470
cord_uid: j58mr9xk

Abstract The biological functions of protein glycosylation have been increasingly recognized but not yet been very well understood, especially in lower organisms. Silkworm as a model lepidopteran insect and important economic insect, has been widely studied in life science, however, the current knowledge on the glycosylation status of its proteome is not satisfactory, and little is known about how pathogenic infections could affect the glycosylation status. This study performed large scale glycosite mapping for the silkworm Bombyx mori P50 strain, and quantitatively compared with that infected with the Bombyx mori cytoplasmic polyhedrosis virus (BmCPV). Some 400 glycoproteins were mapped in the silkworm, including N- and O-glycoproteins. Upon virus infection, the glycosylation levels of 41 N-glycopeptides were significantly changed, some of them belonging to transmembrane glycoproteins. The O-glycosylation profiles were also affected. In addition, 4 BmCPV-encoded viral proteins were found to be glycosylated for the first time, including polyhedrin, P101, VP3, and the NS protein. This study drafted a silkworm protein glycosylation map and underlined the potential impact of virus infection on glycosylation. Significance This study reveals the characteristics of the glycoproteome in the silkworm strain P50, and quantitatively compared to that infected by the virus BmCPV, which underlines the impact of virus infection on the alteration of protein glycosylation in invertebrate species. Our findings add to the knowledge of the post translational modifications of this model organism, and also uncovered for the first time the glycosylation status of the viral proteins expressed by BmCPV.

Silkworm Bombyx mori is an economically important Lepidopteran insect. Since its genome revealed in 2004 [1] , it has become an increasingly popular model organism for molecular biology [2] . However, compared to the current knowledge on the silkworm's genome and transcriptome, we do not know much at the proteomic level, especially the status of post translational modifications (PTMs) of the proteome. Glycosylation, being one of the most common and diverse PTMs, has been widely acknowledged in vertebrates as an important indicator of their physiological or pathological state [3] , however, very little is known about how glycosylation are altered in lower organisms, like during a pathogen invasion. Such investigations can provide clues to the interactions between glycosylation pathways, inflammatory responses and viral immune escape [4] .

The initial steps of viral infections always involve the attachment of viral particles to the receptors on the host cell surface. An important class of the receptors is transmembrane glycoproteins, which play vital roles cell adhesion and receptor bindings [5] . Therefore, glycosylation patterns of the host cell, especially on the cell surface, can influence the interactions between an infecting virus and the host, and identifying such molecules are critical to the understanding of viral infection mechanisms and host immune responses.

Once a virus enters the host cell, it hijacks the host's cellular machinery for viral replication, altering numerous biological pathways of the cell, which is evidenced by changes in the expression levels of many genes, transcripts, and proteins. Although several studies have performed transcriptomic and proteomic comparisons on the silkworm upon virus infection, such as by the Bombyx mori cytoplasmic polyhedrosis virus (BmCPV) [6, 7] , whether the glycosylation profiles are affected have not been reported.

BmCPV is one of the earliest identified viruses infecting silkworms, and often results in significant economic loss to the sericulture industry. BmCPV is a double stranded RNA virus belonging to the Cypovirus genus, Reoviridae family. Its genome contains 10 discrete copacked RNA segments [8] , but only a few of them has been cloned and characterized so far [9, 10] . The RNA genome is wrapped by 5 structural proteins, namely VP1-5 [11] . The virus forms occlusion bodies in infected silkworm midgut, which is composed of the matrix protein polyhedrin that encloses many visions within the matrix [12] .

In this study, we performed large scale glycosylation site mapping on the silkworm strain P50, and quantitatively compared to that infected by BmCPV. The mapping strategy, which was first introduced by Zielinska et al. [13] , uses high precision mass spectrometer for detection of a +2.988 Da mass shift after enzymatic deamidation of the N-glycosylated asparagine residue in O 18 water. Taking advantage of the greatly reduced glycosylation complexity due to removal of the N-glycans, we were also able to simultaneously map the O-glycosites and their associated glycans by searching against an insect O-glycan panel. Additionally, the glycosylation profile of the virus-encoded proteins were obtained. Transmembrane glycoproteins and significantly regulated glycoproteins upon virus infection were marked, which provides clues for future investigation of the role of glycosylation during viral infection in invertebrate systems.

The silkworm were reared on fresh mulberry leaves at 25±1 °C and 70-90% relative humidity. On the first day of 5th instar, the silkworm was fed with 5 µL of BmCPV virion at 10 8 polyhedra/mL in phosphate buffered saline (PBS), and the control larvae were fed with PBS. After 72 hours, silkworms were harvested and residual mulberry leaves were removed from the gut and then preserved at -80 °C. Biological triplicates in each group were prepared for the following sample processing and analysis.

To the silkworm sample, 1.5 mL SDT lysis buffer containing 4% SDS, 100 mM Tris/HCl, and 0.1 M DTT at pH 7.6 was added, and the sample was homogenized in a tissue lyser (Shanghai Jingxin Co., Shanghai, China) using three beating cycles at 120 Hz, 60 s at 4 °C. The solution was then sonicated in an ice bath for 10 minutes, boiled for 15 minutes, and then centrifuged at 14000 g for 40 minutes. The supernatant was collected and the protein concentration was determined by Bradford analysis (ThermoFisher scientific) and stored at -80 ° C for further use.

The protein extract was aliquoted to a tube containing 400 µg of total protein and was boiled for 5 minutes, cooled to room temperature, and then mixed with 200 µL of UA buffer (8 M urea, 150 mM Tris HCl, pH 8.0). The mixture was transferred to a 10 kD ultrafiltration tube (Merck Millipore) and centrifuged at 14000 g at 4 °C for 15 minutes.

The filtrate was discarded and this step was repeated before 100 µL of 100 mM iodoacetamide (IAA) in UA buffer was added. After mixing at 600 rpm for 1 minute, the samples were incubated for 20 min in darkness. The filters were washed three times with UA buffer at 14000 g, 4 °C for 15 minutes, then 100 µL of 25 m M NH 4 HCO 3 were added to the filter and the sample was centrifuged for 15 minutes under the same conditions. This step was repeated twice. Trypsin was added at a ratio of trypsin: protein = 1:50 to the samples and gently mixed for 1 min, and then incubated at 37 °C for 16 J o u r n a l P r e -p r o o f hours. The filtrate was then collected by centrifugation. The filter was washed with 40 µL 25 mM NH 4 HCO 3 and the filtrate was combined.

The digested peptides was transferred to a new 10 kD filter and mixed with 100 μL lectin mixture containing 2.5 mg/mL Con A, 2.5 mg/mL WGA, and 0.8 mg/mL RCA in buffer A (Table S1 ) were set as the variable modifications to the threonine or serine residues. A maximum of 5 PTMs were allowed per peptide. The results were further filtrated using the following parameters: de novo average localized score ≥ 50%, unique peptide ≥ 2, PTM Ascore ≥ 20, and false discover rate ≤ 1%. For quantitative analysis, Perseus software [15] was used for further data processing and visualization. Briefly, the raw intensities obtained by Peaks software were transformed using natural log, and those had at least two valid intensities among the triplicate in at least one group (control or infected) were retained, and the remaining missing values were replaced with normal distribution (width = 0.3; downshift = 2.6).

Student's T-test was performed to obtain statistically different (p<0.05) (glyco)peptides between the groups. Gene ontology enrichment was performed using the ShinyGo webserver [16] . Transmembrane topology was predicted using the Phobius web server [17] . and serve as a protective shield from the host immune systems [18, 19] . Alteration of the glycans on the virus particle can alter the virulence and change the host range [20] [21] [22] .

Therefore (Table 1) . These viral glycoproteins were not found in the control P50 group.

Polyhedrin, encoded by the genome segment 10 of the BmCPV virus and organized in octameric or dodecameric forms [23] , is the major matrix protein surrounding the virus envelope and protects the virions against long term harsh environments. Polyhedrin has typical molecular weight of 27-31 kDa, but no significant sequence similarities were found between the polyhedrins derived from the CPVs of different host species [9, 24, 25] . While baculovirus polyhedrin has been found with glycosylation potential [23] , there has been controversy on whether the cypovirus polyhedrin is glycosylated or not. It has been previously suggested to be glycosylated based on positive color-reactions with acid-Schiff reagent [11] , however, molecular cloning of this protein showed that polydedrin does not contain a transmembrane signal peptides which guide peptide glycosylation [9] . In this study, glycosite mapping showed that BmCPV polyhedrin is Nglycosylated at the single site N77, confirming that it is a glycoprotein, and it is speculated that the sugar coating could further enhance the stability and protection of the virion. Another glycoproteins P101, which is a 101 kDa protein encoded by the J o u r n a l P r e -p r o o f genome segment 5, was identified with a single glycosylation site at N684, although this protein was predicted with three possible N-glycosylation sites based on motif analysis [26] . The other two viral glycoproteins were VP3, a capping protein [8] , which was glycosylated at N801 and the non-structural (NS) protein, which was mapped with three glycosylation sites at N69, N48, and N138. The functions of these viral proteins have not been clarified or experimentally validated so far, and even less is known about what their glycosylation does for these proteins during virus infections. Therefore, it would be of great interest to investigate these glycosylated viral proteins and their roles during virus infection.

A total of 762 N-glycopeptides belonging to 389 N-glycoproteins were identified in the P50 control and BmCPV-infected groups (Table S2) , revealing a high level of glycosylation in this lepidopteran insect. More than 95% of the mapped glycoproteins were within 200 kDa, and a significant portion of them were closely resided within 10-100 kDa (Figure 1a) . The N-glycoproteins were predominantly singly or doubly glycosylated, although 14 of them were mapped with more than 5 glycosites per protein ( Figure 1b) . Figure 1c showed that all the 762 identified glycosylation site has the consensus sequence N-X-T/S, demonstrating the highly conserved motif required for Nglycosylation. Among the 389 identified N-glycoproteins, 118 were predicted to be transmembrane glycoproteins (Table S3) , which can be interesting candidates to explore as potential cell surface receptors to outside pathogens. The total glycoproteins identified in the silkworm were analyzed by gene ontology (GO) enrichment in the category of biological process, indicating that the silkworm glycoproteome mainly involve in metabolic processes, cell adhesion, and proteolysis processes (Figure 1d ).

Journal Pre-proof [27, 28] . Previous study indicated that it is a glycoprotein [29] , but the specific sites were not known. In this study, the storage protein is mapped with three N-glycosites at N208, N519, and N598. The second most abundant silkworm N-glycoprotein is a putative cuticle protein glycosylated at N66. Although many insect cuticle proteins were found to be glycosylated [30] , the functions of these modifications

are not yet clear. The protein containing the highest number of glycosite was H9JRT0, with 10 N-glycosites, and was the eighth most abundant N-glycoprotein. However, it has not been characterized at the protein level. As Table 2 shows, a lot of the glycoproteins are "uncharacterized", thus continued effort are needed to investigate their structures and functions, including their glycosylation status.

To understand how protein glycosylation status is affected by viral infection, the total glycoproteomes in the P50 control and BmCPV-infected groups were analyzed and compared. For reliable quantitative comparison between the glycoproteomes, the data were further filtered retaining those detected at least twice in the triplicate in at least one group. This yields 489 glycopeptides belonging to 287 glycoproteins (Table S4) , which were then used for quantitative comparison between the control and BmCPV-infected group. The glycopeptide intensities within each group are highly replicable (Figure 2a) , as demonstrated by low discrepancy in intensity between the replicates. Majority of the glycopeptide intensities in the replicates were within 10% from each other.

A heat map was constructed to compare the identified glycopeptides in the two groups ( Figure S1 ), which showed that majority of the N-glycopeptides had a similar profile between the control and the virus-infected groups. Statistically significant (p<0.05) glycopeptides between the two groups were shown in Figure 2c and 2d. A total of 14 glycopeptides were significantly downregulated in the virus-infected group, and another 27 glycopeptides significantly upregulated. Among them, 9 were predicted transmembrane proteins (Table S3) physiological and pathological processes [32] . Previous studies showed that silkworm integrin beta interacts with the BmCPV virion, and that silencing integrin beta gene could inhibit viral infection to the silkworm [33] . Our data showed that integrin beta was upregulated during BmCPV infection, indicating its potential role as a cell adhesion receptor facilitating viral propagation. Sepin, a glycosylated serin protease inhibitor, was previously found significantly upregulated in the bacteria-infected fat body of the silkworm [34] , similar to what we observed in the silkworm upon viral infection in this study.

It is noted that the ecdysone oxidase and another glycoprotein (H9IUI0) both contain two glycopeptides that were downregulated. However, another interesting observation is that while the glycopeptide (N181) of SSR was downregulated, its glycopeptide at N126 wass upregulated. The results indicate that while viral infection affects the level of protein glycosylation, it may or may not influence in the same way. Therefore, to evaluate or compare the profiles of multiply glycosylated proteins, it is necessary to locate individual glycopeptides instead of the glycoprotein as a whole, because each glycosylation site might change independently from the rest of the glycosylation sites upon perturbation or stimuli. The O-glycosylation mapping results for the P50 control and virus-infected groups were shown in Table 3 . Only those identified at least twice in the triplicate were retained for comparison purposes. A total of 13 O-glycoproteins were mapped in the P50 and its

BmCPV-infected strains. The most frequently occurred O-glycans were fucose (F) and

HexNAc (N), followed by Hex1HexNAc2 (H1N2), HexNAc2 (N2), HexNAc3 (N3) and

HexHexNAcFuc (H1N1F1). A lot of the O-glycoproteins are also N-glycosylated (Table   3 ). It was noticed that the uncharacterized protein H9IXN5 contain a surprisingly large number of N-glycosite: a total of 20, in addition to the O-glycosite identified (Table 3) . A protein coverage map with its supporting peptides for this unknown but interesting protein was shown in Figure S2 , demonstrating a reliable glycosite mapping of this protein.

Myosin regulatory light chain (RLC), an important structural element of myosin, is critical to cell movement and muscle contractions. RLC is a well-known phosphoprotein and its dynamic phosphorylation state is tissue-specific and case-specific [35, 36] . However, it has not been reported to be glycosylated before. In this study, silkworm RLC was found to be O-fucosylated at S104, S121 or S124, depending on whether the silkworm was infected with the virus or not. The silkworm larval cuticle protein LCP-17 was found to be N-glycosylated at N52 and N97 in both groups but only O-glycosylated in the control group at site S104 (Table 3) . On the other hand, the silkworm storage protein (H9JHM9) and the uncharacterized proteins H9JRT0, H9JTY5, H9J9M0, H9JTA0, and H9J609

were (Table S5 ).

The changes in glycosylation states, together with other molecular differences such as genetic mutations, may contribute to the different responses these strains develop to virus invasions.

It has been well established that alteration of glycosylation is closely associated with changes in physiological and pathological states. Our previous study also showed that baculovirus resistant and susceptible silkworm strains were differentially glycosylated [14] . Combining the results from this study, which showed that viral infection altered protein glycosylation in the silkworm, the relationships between host glycosylation and virus/microbe infection can be intertwined: 1) host glycosylation changes lead to infection and inflammation, 2) infection and inflammation lead to host glycosylation changes, which leads to molecular and cellular functional differences.

Glycosylation modification is highly dynamic and can be affected by many cellular factors and processes. During BmCPV infection, the silkworm's cellular machinery was hijacked and many biological pathways were altered, including the glycosylation pathway.

In this study, the total glycoproteome in the silkworm P50, including the N-and O- 

Additional data are provided as supplementary Figures S1-2 

A draft sequence for the genome of the domesticated silkworm (bombyx mori)

Silkworm: A promising model organism in life science

Glycosylation in health and disease

Infection, inflammation and host carbohydrates: A glyco-evasion hypothesis

Biological roles of glycans

Itraq-based quantitative proteomic analysis of midgut in silkworm infected with bombyx mori cytoplasmic polyhedrosis virus

Transcriptome analysis of bombyx mori larval midgut during persistent and pathogenic cytoplasmic polyhedrosis virus infection

The dsrna viruses

Molecular cloning and characterization of cytoplasmic polyhedrosis virus polyhedrin and a viable deletion mutant gene

Identification and characterization of vp7 gene in bombyx mori cytoplasmic polyhedrosis virus

Biochemical properties of polyhedra and virus particles of the cytoplasmic polyhedrosis virus of bombyx mori

The insect viruses

Precision mapping of an in vivo n-glycoproteome reveals rigid topological and sequence constraints

Precision mapping of n-and o-glycoproteins in viral resistant and susceptible strains of bombyx mori

The perseus computational platform for comprehensive analysis of (prote)omics data

Shinygo: A graphical enrichment tool for ani-mals and plants. bioRxiv

Advantages of combined transmembrane topology and signal peptide prediction-the phobius web server

Glycan shield and epitope masking of a coronavirus spike protein observed by cryo-electron microscopy

The epsteinbarr virus glycoprotein gp150 forms an immune-evasive glycan shield at the surface of infected cells

Removal of either n-glycan site from the envelope receptor binding domain of moloney and friend but not akv mouse ecotropic gammaretroviruses alters receptor usage

Role of n-linked glycosylation of the 5-ht2a receptor in jc virus infection

Alteration of the n-linked glycosylation condition in e1 glycoprotein of classical swine fever virus strain brescia alters virulence in swine

How baculovirus polyhedra fit square pegs into round holes to robustly package viruses

Nucleotide sequence of the polyhedrin gene of euxoa scandens cytoplasmic polyhedrosis virus (escpv). Virology

Analysis of the cpolyhedrin genes from different geographical isolates of a type 5 cytoplasmic polyhedrosis virus

Nucleotide sequence of genome segment 5 from bombyx mori cypovirus 1

Vitellogenin receptor selectively endocytoses female-specific and highly-expressed hemolymph proteins in the silkworm, bombyx mori

Expressional analysis of the silkworm storage protein 1 and identification of its interacting proteins

Structural determination of the n-glycans of a lepidopteran arylphorin reveals the presence of a monoglucosylated oligosaccharide in the storage protein

Cuticle proteins of the boll weevil, anthonomus grandis, abdomen: Structural similarities and glycosylation

A signal sequence receptor in the endoplasmic reticulum membrane

The integrins

Integrin beta and receptor for activated protein kinase c are involved in the cell entry of bombyx mori cypovirus

Serpin-5 regulates prophenoloxidase activation and antimicrobial peptide pathways in the silkworm, bombyx mori

Feifei Zhu: Conceptualization, Investigation, Formal analysis, Funding acquisition

Dandan Song: Investigation, Formal analysis

Shangshang Ma: Validation

Peng Lü: Resources, Validation; Xiaoyong Liu & Qin Yao: Review and editing

Keping Chen: Funding acquisition, Supervision, Project

None.J o u r n a l P r e -p r o o f