Sex-specific glycosylation of secreted immunomodulatory proteins in the filarial nematode Brugia malayi

The extended persistence of filarial nematodes within a host suggests immunomodulatory mechanisms that allow the parasites to resist or evade the host immune response. There is increasing evidence for immunomodulatory glycans expressed by a diversity of parasitic worms. In this study, we integrate multiple layers of the host-parasite interface to investigate the glycome of a model filarial parasite, Brugia malayi. We report a significant overrepresentation of terminal GalNAc moieties in adult female worms coupled with an overall upregulation in O-glycosylation, T-antigen expression, and a bias for galactose containing glycans. Adult males preferentially displayed a bias for terminal GlcNAc containing glycans, and fucosylated epitopes. Subsequent proteomic analysis confirmed sex-biases in protein glycosylation and highlighted the sex-specific glycosylation of well characterized immunomodulators expressed and secreted by B. malayi. We identify sex-specific effectors at that interface and suggest approaches to selectively interfere with the parasitic life cycle and potentially control transmission.


156
As observed for most eukaryotic organisms, helminths glycosylate secreted proteins. These 157 glycoproteins play key roles in immune function, and glycosylation is crucial to their activity 158 (Cvetkovic et al.; Ahmed et al.). It is therefore important to define glycan structures decorating 159 secreted proteins at the host-parasite interface. Determining the diversity of glycan structures 160 expressed by an organism is largely dictated by the nature and number of glycosylation-161 related enzymes and available substrates. To construct a comprehensive list of glycosylation 162 enzymes, we interrogated the B. malayi genome using a series of functional annotations.

163
Gene Ontology (GO) and KEGG annotations for genes and pathways were used to assign 164 glycosylation-related proteins. We identified 116 genes with an associated GO term relevant 165 to protein glycosylation and 143 genes assigned to KEGG glycosylation pathways. A full 166 BLASTp search of the B. malayi proteome against a "human glycosylation proteome" 167 uncovered 136 genes with significant similarities. A full genome Hidden Markov Model scan 168 for protein families (Pfam) was done to complement the analysis, uncovering 112 genes with analysis reveals a high degree of conservation within filarial nematodes (Fig. 2B, VII-XII) with 174 around 56% of the genes conserved at >50% similarity. The data also indicated clear 175 divergence from blood and liver flukes (Fig. 2B, V-VI) with 33% of the genes having no 176 orthologs, and an intermediate profile of conservation when compared to hookworms, 177 whipworms or the free-living nematode C. elegans with only 14% of the genes having no 178 orthologs (Fig. 2B, I, IV, and II, respectively). Overall, the data show a higher conservation 179 and similarity in the glycogenomes of filarial parasites and class III nematodes, and a clear   . Table S2b). Mapping these results onto the biosynthetic pathways 193 suggests higher expression of core 2/6 O-glycans in male worms (Fig. 3A). The results also 194 indicate sex-dependent expression of specific fucose type O-glycans, with males having 195 higher levels of GlcNAc-Fuc-(ser/thr), while females modify this epitope further to form 196 predominantly Gal-GlcNAc-Fuc-(ser/thr) (Fig. 3B).

198
Females also show a higher expression of a protein-O-acetylglucosamine transferase 199 (Bm4815), leading to the uncommon GlcNAc O-linked glycan epitope (Fig. 3C). Similarly, the 200 mannose type O-glycan biosynthetic pathway shows male-biased expression of core M1 and 201 core M2 glycans (Fig. 3D). Sex-dependent expression profiles are also apparent for N-type 202 glycan biosynthetic pathways where males display a higher expression of 17 genes within the 203 N-type glycosylation pathway as compared to 7 genes with higher expression in females; most 204 are N-glycan precursor and trimming enzymes. Both sexes also express genes related to 205 glycan degradation, such as mannosidases and glycosidases, primarily involved in 206 oligomannose and paucimannose N-glycan biosynthetic pathways (Supp.  The study of glycosylation in B. malayi has primarily focused on GPI-anchored glycoproteins 213 and staining of worms with specific lectins (Mersha et al.; Schraermeyer et al.; Kaushal et al.).

214
A more general glycomic analysis of B. malayi has not been performed. To confirm whether 215 the observed sex-biased differential expression of the B. malayi glycogenome translates into 216 sex-biased display of glycan epitopes, we analyzed whole worm lysates on lectin microarrays   (Pilobello, Krishnamoorthy, et al.; Propheter et al.). This technology uses the known glycan 218 binding specificities of lectins to provide an epitope-specific readout of glycosylation. We 219 observed significant differences in glycosylation profiles between male and female worms 220 (Fig. 4). Female worms showed higher levels of terminal Gal/GalNAc epitopes, including core

226
We suspected that biases in glycan epitopes would be associated with differential distribution 227 within tissues in male and female worms, especially in the organs involved in secretion and 228 excretion. To test this, we determined the spatial localization of fucosylated and/or 229 galactosylated proteins in the heads and tails of adult male and female B. malayi. We stained Heatmap summarizing lectin array data comparing male and female glycans and representing significantly differentially displayed glycan epitopes between sexes (p < 0.05). Data shown represent Log2 of the fluorescence ratios between each sample and the control. Lectins are grouped by their corresponding recognition epitopes and representative glycan structure displaying respective epitopes are shown.
in males) and GS-I lectin, which binds to a-galactose residues (enriched in females). We 232 observed a clear localization of fucosylated and galactosylated epitopes within the mouth and 233 cephalic alae of both males and females consistent with the known distribution of sensory 234 organs (Fig. 5A). Females showed a diffuse body staining with galactose-specific lectin GS-I,

235
with higher levels at the cuticular areas and within the ovaries, compared to a much lower 236 intensity of galactose staining across the body of male worms. In contrast, males displayed a 237 very specific and high intensity fluorescence of fucosylated residues in reproductive organs 238 and tissues (Fig. 5B) coupled to higher levels of fucosylated residues at the mouth and in the 239 cephalic regions.

243 244
The differential localization of galactosylated and fucosylated residues suggests such 245 modifications occur on different proteins. To identify the protein partners of the sex-biased 246 glycan epitopes, we carried out proteomic analysis on glycoproteins isolated with AAL and 247 GS-I (Fig. 6A). Consistent with our findings from the lectin arrays, binding to AAL and GS-I 248 showed higher affinity to male and female lysates, respectively (Fig. 4). Total protein extracts 249 from both male and female worms were subjected to affinity chromatography on AAL and GS-

262
To further investigate this subset of differentially glycosylated proteins, we profiled them for 263 GO term enrichments (Supp. Table S3c). We observed a significant enrichment in proteins 264 in the extracellular region (p-value = 0.0014). This finding, coupled to the detection of Bma-265 mif-1, a known immunomodulatory protein at the host-parasite interface, led us to determine 266 whether these differentially glycosylated proteins were part of the B. malayi ES. All 56 detected 267 proteins were observed in the B. malayi secretome defined in this study, suggesting a potential 268 role for differential glycosylation in parasite glycoprotein-mediated immunomodulation.   Table S4).

292
To further shortlist the candidates that are highly expressed and differentially glycosylated by        Table S5a. The genome was filtered for genes 428 associated with any of the glycosylation GO terms. A full list of 135 genes was identified and 429 can be found in Supp. Table S5b. A similar analysis was done with the genome using the 430 KEGG database (Kanehisa et al.) and all protein glycosylation related pathways can be found 431 in Supp. Table S5c. Subsequently old gene identifiers from KEGG were assigned to the new 432 IDs using BioMart for conversion (Smedley et al.). A total of 112 genes were found and are 433 listed in Supp. Table S5d. To complement the functional analysis, B. malayi's genome was 434 analyzed using HMMER (Potter et al.) to annotate all Pfam domains. The full results are shown 435 in Supp. Table S5e. Pfam annotations related to protein glycosylation were extracted from 436 the Pfam database (Finn et al.); annotated genes can be found in Supp. Table S5f. were allowed and filtered for a maximum e-value of 0.05. All hits with at least 50% sequence 442 identity were considered true and Supp.   Table S6a. We note that we are unable to observe a subset of epitopes (e.g.

571
The RAW data files were processed using MaxQuant (version 1.6.1.0) to identify and quantify 572 protein and peptide abundances. The spectra were matched against the Brugia malayi Uniprot 573 database (downloaded August 18, 2018) with standard settings for peptide and protein 574 identification, that allowed for 10 ppm tolerance, a posterior global false discovery rate (FDR) 575 of 1% based on the reverse sequence of the mouse FASTA file, and up to two missed trypsin wash counterparts. An adjusted p-value cutoff of 0.05 and a minimum 2-fold change was used 582 and a list of significantly eluted proteins can be found in Supp. Table S3b.

583
Label-free data relevant to stage and sex-specific secretomes were normalized by total High magnification cut-outs were imaged using the 100X oil objective using 0.29 μm z stack "Median Filter" processing option with x/y kernel size set at 3 voxels.