(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



The Trypanosoma brucei MISP family of invariant proteins is co-expressed with BARP as triple helical bundle structures on the surface of salivary gland forms, but is dispensable for parasite developme [1]

['Aitor Casas-Sanchez', 'Department Of Vector Biology', 'Liverpool School Of Tropical Medicine', 'Liverpool', 'United Kingdom', 'Department Of Tropical Disease Biology', 'Raghavendran Ramaswamy', 'Department Of Biochemistry', 'Microbiology', 'University Of Victoria']

Date: 2023-04

Trypanosoma brucei spp. develop into mammalian-infectious metacyclic trypomastigotes inside tsetse salivary glands. Besides acquiring a variant surface glycoprotein (VSG) coat, little is known about the metacyclic expression of invariant surface antigens. Proteomic analyses of saliva from T. brucei-infected tsetse flies identified, in addition to VSG and Brucei Alanine-Rich Protein (BARP) peptides, a family of glycosylphosphatidylinositol (GPI)-anchored surface proteins herein named as Metacyclic Invariant Surface Proteins (MISP) because of its predominant expression on the surface of metacyclic trypomastigotes. The MISP family is encoded by five paralog genes with >80% protein identity, which are exclusively expressed by salivary gland stages of the parasite and peak in metacyclic stage, as shown by confocal microscopy and immuno-high resolution scanning electron microscopy. Crystallographic analysis of a MISP isoform (MISP360) and a high confidence model of BARP revealed a triple helical bundle architecture commonly found in other trypanosome surface proteins. Molecular modelling combined with live fluorescent microscopy suggests that MISP N-termini are potentially extended above the metacyclic VSG coat, and thus could be tested as a transmission-blocking vaccine target. However, vaccination with recombinant MISP360 isoform did not protect mice against a T. brucei infectious tsetse bite. Lastly, both CRISPR-Cas9-driven knock out and RNAi knock down of all MISP paralogues suggest they are not essential for parasite development in the tsetse vector. We suggest MISP may be relevant during trypanosome transmission or establishment in the vertebrate’s skin.

The Trypanosoma brucei group of parasites are exclusively transmitted to the vertebrate host by the tsetse vector alongside insect saliva. To better understand trypanosome transmission, we investigated the protein composition of T. brucei-infected tsetse saliva using a mass spectrometry proteomics approach. We found that, in addition to proteins from tsetse saliva and Sodalis glossinidius (a bacterial tsetse symbiont), trypanosome-infected saliva contains several parasite surface glycoproteins, including a partially characterized family of invariant proteins herein named Metacyclic Invariant Surface Proteins (MISP). We show that MISP is primarily expressed, together with metacyclic Variant Surface Glycoprotein (mVSG) and Brucei Acidic Repetitive Protein (BARP), on the surface of the infectious metacyclic stage of T. brucei. Its triple helix bundle architecture appears tethered to the outer membrane by an extended glycosylphosphatidylinositol-anchored C-terminal tail that putatively projects MISP above the VSG coat. Our findings provide new insights into the surface architecture of the T. brucei metacyclic stage and describe the challenges associated with developing transmission-blocking vaccines against tsetse-transmitted trypanosomes.

During a blood meal, the infected tsetse inoculates MCFs into the host along with saliva [ 26 , 32 ] to inhibit the host hemostatic response to the cutaneous trauma [ 33 , 34 ]. However, during parasite development in the tsetse SGs, there is a drastic (~80%) transcriptional down regulation of genes encoding for salivary proteins, which induces a feeding phenotype that may amplify disease transmission [ 32 , 35 , 36 ]. Although the proteome of tsetse saliva from T. brucei-infected flies has been determined [ 34 ], soluble parasite factors have not been characterized in detail and thus, the molecular mechanism by which these parasites manipulate the vector to promote transmission is largely unknown. Here, we identified trypanosome proteins from infected tsetse saliva using a high-resolution mass spectrometry-based proteomics approach. We found that T. brucei-infected saliva is particularly enriched with peptides from trypanosome surface proteins such as BARP, mVSG and a partially characterized family of glycoproteins herein named Metacyclic Invariant Surface Proteins (MISP). We show evidence that MISP, previously named as salivary gland epimastigote 1 (SGE1) [ 37 , 38 ], are a small family of invariant glycoproteins primarily expressed on the surface of the metacyclic stage of T. brucei. Structural data reveal a triple helix bundle architecture for MISP that appears to be tethered to the outer membrane by an extended, unstructured C-terminal tail, potentially projecting MISP above the mVSG coat and thereby allowing antibody recognition. Our findings provide a new insight into the surface architecture of MCFs and are discussed with respect to the molecular crosstalk between parasite and vector, and the challenges associated with developing transmission-blocking vaccines against tsetse-transmitted trypanosomes.

In contrast to BSFs that possess a large repertoire of vsg genes [ 27 , 28 ], MCFs express a single mVSG protein (also called metacyclic variant antigen type; mVAT) from a much-reduced gene repertoire. Based on antibody recognition, it is estimated that a population of tsetse-derived MCFs express different mVSGs (as many as 27) to ensure infection of hosts that may have been pre-exposed to any of these antigenic types [ 29 , 30 ]. Given that MCFs develop in the tsetse SGs and their persistence in the host after transmission is relatively short [ 26 ], they do not seem pressured to undergo antigenic variation. Despite the differences in the expression mechanisms between VSGs and mVSGs, it is assumed that their function and structural organization is equivalent. BSF VSGs form a protective outer layer that covers the parasite surface (~5 million homodimers/cell) [ 31 ] to mask the invariant surface proteins from the host immune system. Many of these invariant proteins are predicted to be shorter in height than the VSG homodimers and to be attached to the plasma membrane via transmembrane domains [ 8 , 11 ]. The exceptions to this are the haptoglobin-hemoglobin (HpHbR) and transferrin receptors, which are GPI-anchored and exclusively localize in the flagellar pocket [ 9 , 10 ].

During its life cycle, T. brucei undergoes many developmental changes to adapt to the different environments within the mammalian host and the tsetse vector [ 3 ]. All T. brucei life stages have in common the expression of major surface glycosylphosphatidylinositol (GPI)-anchored glycoproteins known to be important for parasite survival. In bloodstream form (BSF) stages, the variant surface glycoprotein (VSG) coat is responsible for antigenic variation [ 4 – 6 ], clearance of host immunoglobulins bound to VSGs [ 7 ] and preventing antibodies from binding to conserved VSG epitopes and invariant surface proteins (ISG), such as transporters and receptors [ 8 – 11 ]. When ‘stumpy’ BSF trypanosomes are ingested by tsetse, they quickly (24 to 48 hours) differentiate into the procyclic trypomastigote stage within the fly’s midgut [ 3 , 12 ]. During differentiation, VSGs are no longer expressed and the parasites lose the VSG coat via the concerted action of major surface metalloproteases and the GPI-phospholipase C (GPI-PLC) [ 13 , 14 ]—a process that appears to modulate the tsetse’s immunity in favor of parasite infection [ 15 ]. Concomitant with VSG disappearance, the parasites express a new coat comprised of GPEET- and EP-procyclin isoforms once the infection establishes in the midgut [ 16 – 18 ] and proventriculus (PV) [ 19 ]. Within the tsetse midgut, procyclic trypomastigotes first colonize the ectoperitrophic space of the anterior midgut, from which they further migrate to the tsetse cardia or PV [ 20 ]; alternatively, both the anterior midgut and PV can be simultaneously colonized as trypanosomes reach the ectoperitrophic space after crossing an immature peritrophic matrix at the PV [ 19 ]. After 10–15 days post-infection, depending on parasite strain, trypanosomes differentiate into epimastigote forms (EMFs) within the PV, which then migrate to the salivary glands (SGs), attach to the epithelial cell microvilli and switch on the expression of Brucei Alanine-Rich Proteins (BARP) [ 21 ]. The attached EMFs further differentiate into pre-metacyclic trypomastigotes [ 22 , 23 ], during which time BARP expression is lost, a new coat of metacyclic VSG (mVSG) develops [ 24 , 25 ], and cells detach from the SG epithelium. Finally, mature metacyclic trypomastigote forms (MCFs) [ 23 , 26 ] are inoculated along with saliva into the skin of a vertebrate host in a subsequent feed. Unlike in BSFs, the surface composition of T. brucei MCFs is less well characterized primarily due to the lack of an efficient in vitro culture system of this life stage, the low percentage of flies with salivary gland infections, and the low yields of MCFs from infected salivary glands.

African trypanosomes are the causative agents of human African trypanosomiasis (HAT) or sleeping sickness and animal African trypanosomiasis (AAT or Nagana disease) in most sub-Saharan countries. Multiple trypanosome species, including the group of Trypanosoma brucei parasites (i.e. T. b. gambiense, T. b. rhodesiense, and T. b. brucei), are transmitted by several tsetse species (Glossina spp.). Despite decreasing HAT cases since 2010, due largely to a program of active screening and treatment of the human population together with an efficient vector control program, AAT continues to cause significant economic losses in the agricultural sector with >1 million cattle dying each year [ 1 ]. Since these parasites undergo antigenic variation to effectively evade the host’s immune system, preventive vaccines have been difficult to develop, and only one has recently shown protection against T. vivax [ 2 ].

To gain more insight into the essentiality of MISP, we created a T. b. brucei transgenic cell line that allows the RNAi knockdown of all misp paralogs at once through a tetracycline-inducible RNAi stem loop system [ 55 ] targeting a conserved gene region. Although misp knockdown was successfully induced, both transcript and protein downregulation were only partial and did not lead to any conclusive phenotype either in PCF in vitro or within the tsetse salivary glands ( S19 Fig ). We then used a CRISPR-Cas9 strategy to knock out (KO) misp by targeting DNA breaks at both 5’ and 3’ sequences conserved in all paralogs, and replacement with drug selection markers ( Fig 5A ). An addback (AB) rescue cell line was created by inserting a tetracycline-inducible ectopic misp360 gene for overexpression. As expected, due to the native expression pattern of misp, neither PCF cell line presented any growth or morphology phenotype in vitro. We then infected tsetse with either misp KO or AB cell lines and assessed infection rates at 30 dpi. All cell lines (i.e. KO, uninduced (AB-) and induced (AB+) addback) yielded normal midgut and proventricular infection rates, comparable to that of wild type cells. However, salivary gland infection rates of KO and AB- experienced a slight (although not statistically significant) reduction ( Fig 5B ) despite the observation that morphology and motility were apparently not affected, suggesting that MISP may have a minor but non-essential role in salivary gland infection establishment and/or maintenance. Lastly, misp KO cells re-expressing MISP360 (AB+ cells) failed to restore normal salivary gland infection rates. It remains to be determined whether this is due to the differential expression level compared to the endogenous gene and/or whether more than one isoform is necessary for complementation. Lack of MISP expression in KO cells was confirmed by immunostaining with anti-MISP antibody ( Fig 5C ), while expression of BARP and CRD remain comparable to the wild type and AB+ cells ( S20 Fig ).

Representation of MISP ( A ) and BARP ( B) recombinant proteins produced for mouse vaccination; sections in gray were not included; signal peptide (SP), C-terminus domain (CTD), GPI anchor peptide (GPI); numbers indicate position from Methionine 1. Size exclusion chromatography chromatograms of recombinant MISP ( C ) and BARP ( D ), indicating the purified peaks including the SDS-PAGE of the peak fraction as inset. Western blotting of BSA as control, recombinant MISP and BARP, probed with either anti-MISP polyclonal antibodies ( E ) or anti-BARP ( F ). ( G ) T. brucei parasitaemia development (parasites/mL blood) in mice either vaccinated with PBS only (gray), PBS and adjuvant (dark gray), recombinant MISP and adjuvant (blue), recombinant BARP and adjuvant (yellow) or dead metacyclic cells (purple) from 5 to 32 days post-challenge; dots indicate the mean of each group (n = 10 per group). ( H) , parasitemia levels from 5 to 8 days post-challenge highlighting the significant (*) reduction of parasitemia at day 7 post-challenge in mice vaccinated with dead metacyclic cells.

Since we found MISP isoforms to be highly conserved, expressed on the metacyclic surface and susceptible to antibody binding in native conditions, we considered MISP as transmission-blocking vaccine candidates against T. brucei. To explore the experimental immunoprotective properties of MISP, the N-terminal ectodomain of MISP360 ( Fig 4A ) was recombinantly produced in E. coli, purified by nickel affinity and size exclusion chromatography ( Fig 4B–4C ), depleted from endotoxins, and used to formulate a vaccine with glucopyranosyl lipid adjuvant aqueous formulation (GLA-AF) as adjuvant. In parallel, we also immunized mice with recombinant BARP ( Fig 4D–4F ) and dead whole T. b. brucei metacyclic cells. Ten BALB/c mice per group were immunized by subcutaneous injection and received two more boost doses after 10 and 20 days ( Fig 4G–4H and S2 File ). Ten days after the last booster, each mouse was bitten once by a single tsetse with a confirmed SG infection. Infection was followed up to 31 days post-challenge and parasitemia was regularly measured for each individual ( Fig 4G ). Animals vaccinated with either MISP or BARP developed an infection and did not show a significant reduction in parasitemia compared with the control groups (PBS and adjuvant only), at any point. Only mice immunized with dead metacyclic cells showed a significant reduction of parasitemia at 7 days post-challenge ( Fig 4H ), but this was restored to levels comparable to the controls the day after. Despite the immunized animals successfully developing an immune response against either antigen by generating high titers of specific antibodies as shown by immunoblotting ( S2 File ), vaccination did not prevent the development of a T. b. brucei infection.

A structural comparison between the VSG MiTat 1.2 N-terminus (PDB: 1VSG), MISP360, and the BARP model revealed they all share a similar height (~99.9 Å, ~93.0 Å and ~89.2 Å, respectively) ( S18A Fig ). However, when comparing the sequence length of their unstructured C-terminal domains ( Fig 3G ) with that of VSGs, it was clear that MISP360 and MISP380 have ~71% and ~34% longer C-termini; MISP440 has an equivalent length and MISP400/420 have a 60% shorter C-terminus. All BARP isoforms (except Tb927.9.15590) lack a C-terminal domain. By combining the crystal structure of MISP360 and the MISP C-terminal sequences, we generated structural models for all T. brucei MISP isoforms using I-TASSER [ 51 – 53 ] and IntFOLD [ 54 ] ( S18B Fig ). Similar to TcHpHbR and VSG monomers, we predict that the MISP monomeric triple helical bundle is vertically oriented and, within the MCF surface context, is in close contact with neighboring mVSG homodimers. Since MISP360/380 have longer C-terminal domains than VSGs, and despite their lack of secondary structure, we predict that the MISP N-terminal domains (except for MISP400/420) could extend above the mVSG coat (Figs 3H and S18B ). Consistent with our live immunostaining results ( S14 Fig ), these data suggest that despite being covered by a dense coat of mVSG homodimers, metacyclic trypomastigotes still allow room for the co-exposure of invariant GPI-anchored glycoproteins throughout the cell surface like BARP and MISP.

Surface analysis of the MISP360 and GARP [ 48 ] crystal structures, and the BARP high confidence homology model revealed a similar distribution of acidic and basic patches along the entire length of the structure with no clear localized charge densities that would indicate a molecular recognition site ( Fig 3E ). However, a surface pocket of similar dimensions was identified at the membrane distal end near the region where the core helices bend in each protein ( Fig 3E ), but intriguingly, the residues forming the pockets are quite divergent. While the pocket in MISP360 is mainly formed by hydrophobic residues, BARP’s contain both hydrophobic and basic residues, and GARP’s are formed by acidic, polar and hydrophobic amino acids. This variability in pocket residue composition prompted us to map the residues conserved between MISP360, BARP and GARP onto the MISP360 crystal structure using ConSurf [ 49 , 50 ] ( Fig 3F , left), which showed no clear conservation among the pocket-forming residues. Moreover, mapping conservation between the five T. brucei MISP isoforms on MISP360, revealed that most of the conserved residues are found forming the core of the helices and pocket. However, a striking 51% of the non-conserved residues (representing 22.3% of the full protein sequence) are in close proximity to the pocket ( Fig 3F , right), suggesting these residues may mediate differential affinity or recognition of putative ligands.

The lack of significant sequence identity between MISP360 and any protein with known function led us to perform a DALI search [ 44 ] to identify structural homologs. T. congolense’s GARP, of unknown function, was identified as the top hit with a Z-score of 21.4. A least squares superposition between the two structures resulted in a root-mean-square deviation of 1.7 Å over 183 C α atoms ( Fig 3E , left). The DALI search also revealed structural homology to the previously characterized haptoglobin-hemoglobin receptor (HpHbR) from T. brucei (PDB: 4X0J) and T. congolense (PDB: 4E40) [ 45 , 46 ], and to a T. brucei VSG monomer (PDB: 1VSG) with Z-scores of 18.1, 17.7 and 6.5, respectively. All these proteins exhibit a complementary core of twisted three helical bundles ( S17 Fig ). Structural comparison, however, indicates a closer architectural similarity with HpHbR compared to VSG, due in part to the breakdown of the third helix into loops and extensions enabling substantial conformational diversity in VSG [ 47 ]. Moreover, MISP360 is monomeric in contrast to dimeric VSGs [ 47 ]. Despite the general structural similarity with GARP, BARP, HpHbR and the VSG monomer, the overall low sequence identity and the lack of key conserved residues suggest a different biological role for MISP.

( A ) Schematic representation of MISP360 and BARP protein sequences, highlighting in color the region recombinantly expressed in E. coli and crystallized; numbers indicate residue positions, signal peptide (SP), C-terminal domain (C-term), GPI-anchor peptide (GPI). ( B ) FPLC SEC elution chromatogram of recombinant MISP360; SDS-PAGE of monomeric recombinant MISP360 (inset). ( C ) MISP360 N-terminus crystal structure with molecular surface representation. The structure was found to be 83.3 Å tall, 25.5 Å in width with a 30° helix bend at the top (left); B-factor putty model with ordered regions displayed in blue and thin tubes and flexible regions in red and thicker tubes (middle); secondary structure depiction highlighting the organization of the helices (right); bend-forming residues are in parentheses; coiled structures are represented in white. (D) BARP high confidence homology model generated with Modeller 9v18 [ 43 ] shown with a semi-transparent surface representation. The BARP model reveals an overall architecture that is 82.7 Å tall, 25.0 Å in width with a 34° helix bend consistent with the molecular envelope calculated from the 3.2 Å X-ray crystallography diffraction data set. ( E ) Structural superimposition of the MISP360 N-terminus crystal structure (cyan) with GARP (purple, PDB: 2Y44) and the high confidence model of BARP (orange) (left). A black dashed line highlights the area where the apical surface pocket is found. Top view of an electrostatic surface representation of MISP360, BARP and GARP where the surface pocket is found (middle), and its orthogonal view indicating the residues forming the pocket and size (right). ( F ) Conservation of the pocket-forming (top) and overall (bottom) residues across MISP360, BARP and GARP depicted on the MISP360 structure (left). Residue conservation between the five T. brucei MISP isoforms on the MISP360 structure (right); from most variable (dark blue) to most conserved (dark purple). ( G ) Schematic comparison of the C-terminal domain lengths of BARP, VSG and MISP; scale bar = 10 residues; dashed line sets length baseline (VSG). ( H ) Structural model of the proposed T. brucei metacyclic surface glycocalyx displaying the GPI-anchored metacyclic VSG homodimers (mVSG; light and deep purple), MISP-A (cyan), MISP-B (dark blue) and remains of BARP (yellow). The mVSG structure is represented with a model of mVAT4 based on the crystal structure of VSG 221 N- (PDB: 1VSG) and C-terminus (PDB: 1XU6). MISP is represented with the crystal structure of MISP360 N-terminus (PDB: 5VTL) and a model of its C-terminus; BARP is represented with our confidence model.

To complement the expression and localization analyses, we next determined the crystal structure of the conserved N-terminal ectodomain (Gly 24 -Ala 234 ) of MISP360 (Tb427.07.360; T. b. brucei MISP360 homolog in reference strain Lister 427) to 1.82 Å resolution ( Fig 3 ). Crystals of purified, monomeric MISP360 (Figs 3A–3B and S16A–S16B ) were obtained using the hanging drop method and grew in space group P2 1 2 1 2 1 with a single molecule in the asymmetric unit. The structure of MISP360 was determined by molecular replacement using a truncated form of T. congolense GARP (PDB entry: 2Y44) as the search model. The identification of GARP as a suitable model was based on secondary structure predictions as the amino acid sequence identity of the N-terminal ectodomain is only 15%. The overall structure of MISP360 was well defined with only two residues from the N-terminus remaining unmodelled (Gly 24 and Ser 25 ). The core of the N-terminal ectodomain adopts an elongated structure measuring approximately 83 Å in height and spanning approximately 26 Å in width ( Fig 3C , left). The ectodomain is well ordered with low B-factors throughout the protein ( Fig 3C , middle). MISP360 adopts an overall triple helical bundle structure composed of a core of extended twisted helices capped by a smaller helical bundle at the N-terminal end. The helical bundle that dominates the structure consists principally of three helices ( Fig 3C , right): helix I is comprised of residues Val 29 -Ser 83 , helix II of Glu 88 -Thr 130 and helix VI of Phe 182 -Ala 233 . The three helices adopt a bend of approximately 30 degrees at Gly 44 (helix I), Ala 127 (helix II) and Leu 195 (helix VI) and collectively form the helical bundle cap. In addition to the terminal portions of the three major helices, the helical bundle includes three shorter helices: helix III (Asp 134 -Glu 142 ), which is connected by a 5-residue loop to helix IV (Ser 148 -Gly 156 ), and helix V (Ser 166 -Phe 179 ) connected to helix IV by a 9-residue loop. Helices IV and V lie at either side of helix I and form the broadest face of the helical bundle cap. Moreover, the head structure of MISP360 is anchored by two disulphide bonds; one between Cys 36 and Cys 157 , and the other between Cys 177 and Cys 185 ( Fig 1D ).

As an additional approach to investigate MISP localization, we engineered a T. b. brucei transgenic cell line to ectopically express HA/eGFP-tagged MISP360 under a tetracycline-inducible promoter (Figs 2E and S13 ). Upon induction, ectopic MISP localized on the metacyclic cell body surface in a pattern resembling that of wild type cells ( S13C Fig ). However, in BSF, ectopic MISP was only detected in the flagellar pocket and, upon differentiation to PCF, it was then found redistributed to the flagellum and co-localised with the paraflagellar rod (PFR) ( S13C Fig ). We also used high-resolution scanning electron microscopy (HRSEM) to precisely locate anti-MISP antibodies on the surface of stained cells ( Fig 2F ). While uninduced MISP360 overexpressing PCF cells did not show antibody binding, induction of the transgene led to detection of the ectopic MISP mainly along the flagellum as in Fig 2E . In contrast, native MISP was equally detected in wild type metacyclic cells on both flagellum and cell body, thus validating the observations from confocal microscopy ( Fig 2B and 2D ). To confirm exposure of MISP epitopes in native conditions and free of potential fixative-derived artifacts, we performed immunostaining on live wild type metacyclic cells, which still displayed positive binding of anti-MISP antibodies on the cell surface ( S14 Fig ). Lastly, metacyclics extracted from flies infected with T. brucei TSW196 strain are also recognized by the anti-MISP antibody ( S15 Fig ), confirming that staining is not specific for AnTat 1.1 metacyclic trypanosomes.

Having identified SG stages as the major expressors of misp and barp transcripts, we next determined which life stages translated these transcripts into proteins by immunostaining with polyclonal antibodies that recognize epitopes conserved across all MISP or BARP isoforms. Antibody binding was only to SG parasites, both within the glands ( S6 Fig ) and ex vivo ( Fig 2B and 2C ), and no recognition was detected on BSFs, PCFs, midgut procyclics or PV parasites by either antibody ( S7 Fig ). Surprisingly, despite PV-associated trypanosomes showing high levels of misp and barp transcripts, the corresponding gene products are not produced until they reach the next developmental stage in the salivary glands. Instead, we confirmed that both mesocyclic and epimastigote forms in the proventriculus expressed EP-procyclin as the major surface GPI-anchored protein ( S8 Fig ). Furthermore, both polyclonal anti-MISP and anti-BARP antibodies recognized all parasite cells and stages in infected SGs, including attached epimastigotes, pre-metacyclics and mature metacyclic forms (Figs 2B, 2D and S6 ). The relative mean fluorescence intensity of MISP was higher in metacyclics compared to epimastigotes (2.3-fold); in contrast, BARP intensity was found to be higher in epimastigotes than in metacyclics (3.3-fold) ( S9 Fig ). Moreover, MISP appeared to localize evenly across the cell surface of non-permeabilized fixed epimastigote and metacyclic cells ( Fig 2B ). In contrast, BARP appears homogeneously expressed on the surface of all epimastigotes, yet on metacyclics, BARP is detected on the surface of few (~10.4% of the cells; n = 481), while the rest show exclusively flagellar expression (Figs 2C and S10 ). To confirm the identity of SG-specific trypanosomes, we measured the length of parasites and the relative position of kDNA and nucleus ( S11 Fig ), and performed immunostainings with anti-CRD polyclonal antibodies, which recognize the GPI cross-reacting determinant (CRD) epitope [ 23 , 42 ]. While epimastigotes and pre-metacyclics were not detected by the antibody, mature metacyclics showed surface staining as indirect detection of mVSGs that are partially cleaved by GPI-PLC ( S12 Fig ). To confirm the localization of MISP and BARP in metacyclics, immuno-stained cells were 3D-rendered using confocal laser scanning microscopy and orthogonal views from z-stacks confirmed the fluorescence only localized to the edges of the cell body ( Fig 2D ).

( A ) Relative mRNA expression levels (arbitrary units, normalized to the expression of tert) of T. b. brucei misp paralogs (blue) and barp (yellow), in cultured bloodstream forms (BSF), procyclic cultured forms (PCF), midgut procyclics (MG), proventricular forms (PV), salivary gland forms (SG) and purified metacyclic forms (MCF). Bars represent mean values; error bars represent ±S.D from two technical replicates. ( B ) Representative images of non-permeabilized fixed parasites extracted from infected tsetse salivary glands, immunostained with either anti-MISP ( B ) or anti-BARP ( C ) polyclonal antibodies; epimastigote (EMF), pre-metacyclic (P-MCF) and metacyclic forms (MCF); shown as phase, anti-MISP (cyan) or anti-BARP (yellow), and DAPI (magenta); kinetoplasts (K) and nuclei (N) are annotated in the DAPI channel; anterior (A) and posterior (P) ends of the parasite are annotated in the Phase channel. ( D ) Maximum intensity projection of representative metacyclic cells immunostained with anti-MISP (top) or anti-BARP (bottom) polyclonal antibodies and 3D-reconstructed from z-stacks; y-z orthogonal views show surface localization of the signal and graphs on the right plot fluorescence intensity over the white line crossing the cell section. ( E ) Localization of anti-HA monoclonal antibodies on metacyclic (MCF), bloodstream form (BSF) and procyclic cultured form (PCF) transgenic T. b. brucei cells overexpressing ectopic HA-tagged MISP360 upon doxycycline induction; shown as phase, anti-HA (cyan), DAPI (magenta) and merge; kinetoplasts (K) and nuclei (N) are annotated in the DAPI channel; anterior (A) and posterior (P) ends of the parasite are annotated in the Phase channel. ( F ) Representative HRSEM images of wild type metacyclic (MCF), procyclic MISP360 overexpressor uninduced (PCF MISP -) and induced (PCF MISP+) cell surfaces immunostained with anti-MISP polyclonal antibodies; secondary electron detection (SE), 50% inverted backscatter electron detection (50% BSD); scale bars annotated in SE.

To understand the expression pattern of individual misp paralogs throughout the T. b. brucei’s life cycle, we used semi-quantitative RT-PCR to detect transcript levels of individual misp isoforms and barp, which are markers for salivary gland epimastigotes [ 21 ] ( Fig 2A ). Total RNA was purified from all isolated T. b. brucei life stages obtained from both in vitro cultures (i.e., bloodstream forms and procyclic cultured forms) and tsetse-derived parasites at 30 dpi. (i.e., midgut procyclics, proventricular parasites (mixture of 95% mesocyclics and 5% epimastigotes), whole infected SGs, and metacyclics isolated from infected saliva (mixture of 95% metacyclics and 5% epimastigotes). Thirty day old uninfected flies were used as an age-matched control. Due to the high sequence identity among misp paralogs, assessing the expression of individual misp transcripts by real-time RT-PCR was not feasible and hence, we used semi-quantitative RT-PCR to target the different C-terminal misp repetitive regions, which are unique amongst four of the paralogs with differences of 78 bp ( S4 Fig ). While all control samples from uninfected flies were negative, whole infected SGs and purified metacyclics from saliva showed the highest levels of misp transcripts compared to midgut procyclics (11.8-fold) and BSFs (25.1-fold) ( Fig 2A ). Expression of misp transcripts was also found to be high in trypanosomes colonizing the PV, suggesting it is in those stages where gene up regulation starts. Although there seems to be an overall preferential selection for paralogs with the shortest C-termini (misp400/420), the longer paralogs misp380 and misp360 were up regulated and significantly higher in SG stages. In parallel, the transcriptional profile of misp was validated by measuring that from a transgenic cell line overexpressing misp360 under a tetracycline-inducible promoter ( S4 Fig ). Interestingly, all barp transcripts (originated from 14 paralogs sharing >60% identity, S5 Fig ) followed a similar expression pattern, although PV-associated trypanosomes were the highest expressors, suggesting that the expression of barp may precede that of misp genes during development.

MISP proteins are part of the large Fam50 family of trypanosome surface proteins [ 40 ]. T. b. brucei contains five misp paralog genes (Tb927.7.360, Tb927.7.380, Tb927.7.400, Tb927.7.420, Tb927.7.440 in reference strain TREU927), which encode for proteins that are unique in sequence and have no predicted function. All five T. b. brucei paralogs are localized to chromosome 7 ( Fig 1C ) and are separated by unrelated genes encoding hypothetical proteins (Tb927.7.370, Tb927.7.390, Tb927.7.410 and Tb927.7.430). Notably, the same misp homologs are found in T. b. gambiense (Tbg972.7.270 and Tbg972.7.290). Furthermore, slightly divergent versions of misp genes are found in the animal trypanosome species T. evansi (TevSTIB805.7.300, TevSTIB805.7.320 and TevSTIB805.7.380) and T. congolense (TcIL3000.0.02370), and no homologs were found in the T. vivax genome ( S3 Fig ). Using a phylogenetic analysis (PhyML 3.1) [ 41 ], we determined that MISP are comprised of two sub-families (MISP-A and MISP-B) ( Fig 1B ). MISP-A members are mainly characterized by containing one to three 26 residue repetitive motives at the C-terminus (see below) and MISP-B isoforms, in contrast, present no C-terminal repeats. The T. congolense homolog (TcIL3000.0.02370) shares only 34.2% identity with the others and was not included into either sub-family ( S3B Fig ). A comparison of all T. b. brucei MISP paralogs revealed an overall amino acid sequence identity of ~79% ( Fig 1D ), with the N-terminal domain (Met 1 –Glu 237 ) showing greater conservation (88.7% identity) compared to the divergent C-terminal (Lys 238 -Ser 357 ) (49.0% identity). Importantly, sequence divergence is not random but primarily localized to the charged 26 residue repetitive motives (EVEEVPKKDPEGNVEVPEDKEKTERT or DVQEISREEFEGNVEVPEDKEKTERT). We identified predicted signal peptides for all MISP isoforms (Met 1 –Ala/Thr 17 ). While no transmembrane domains were identified, all MISP isoforms contain a predicted GPI omega site at the C-terminus ( Fig 1D and S5 Table ). Furthermore, all MISP isoforms contain at least one potential N-glycosylation site (Asn 59 ), while Tb927.7.400 and Tb927.7.420 contain an additional site on Asn 121 ( S5 Table ). Overall, sequence-based predictions suggest that MISP is a family of surface GPI-anchored glycoproteins presenting little sequence divergence and differential C-terminal lengths.

(A) Summary table of trypanosome GPI-anchored proteins identified in T. brucei-infected tsetse saliva by nLC-MS/MS. All identified proteins and peptides have >95% confidence and were manually curated. Protein accession code (Protein ID), protein description (Annotation), number of identified unique peptides (Peptides), and number of spectra (Spectra) are indicated; see peptides and spectra in S1 File . Accession codes either from TriTrypDB (BARP and MISP) or UniProt (VSG). (I) representative ID for Tb927.9.15520, Tb927.9.15530, Tb927.9.15550, Tb927.9.15590, Tb927.9.15600, Tb927.9.15620 and Tb927.9.15660; (II) representative ID for VSGs 646, 769, 3613, 1125.408, 1125.474, 1125.1142, 1125.4207 and 1125.4707; (III) representative ID for Tb927.7.360, Tb927.7.380, Tb927.7.400, Tb927.7.420, Tb927.7.440. (B) Phylogenetic tree generated from a multiple sequence alignment of all MISP predicted proteins. Note the demarcation of two sub-families MISP-A (light blue) and MISP-B (dark blue) and the least-conserved T. congolense isoform (red). The T. b. brucei MISP are boxed. Tree created using PhyML 3.1 and the WAG substitution model of maximum likelihood; bootstrap of 500; numbers show branch support values (%); scale bar: 0.1 substitutions/site. (C) Schematic representation of the T. b. brucei chromosome 7 region containing the misp gene array. Genes represented by arrows; misp-A (light blue), misp-B (light purple), unrelated genes (light grey); C-terminal repeats represented by purple rectangles on the arrows. (D) Multiple protein alignment of T. b. brucei MISP isoforms. Residues colored from grey (full conservation) to bright red (highest divergence). Protein domains highlighted: N- signal peptide (green), predicted N-glycosylation sites (yellow asparagine), disulphide bonds identified in crystal structure (cysteines in black boxes), C-terminus repetitive motives of 26 residues (blue) and GPI-anchor signal peptide (red). The conserved peptide ‘SVAEDNSAASTAR’ identified by nLC-MS/MS is highlighted in orange. The top bar numbers indicate residue position starting from Met 1 . Top bar colors indicate the main protein domains N-terminus (purple) and C-terminus (grey).

Discussion

Transmission of vector-borne pathogens usually requires insect saliva as a vehicle, but it is also accompanied by a series of soluble components from both the parasite and the insect. For instance, in the sand fly, Leishmania spp. parasites secrete abundant promastigote secretory gel and exosomes that are important virulence factors for transmission and establishment of the parasite infection in the vertebrate host [56–58]. Here, we performed semi-quantitative proteomics to analyze the composition of T. b. brucei-infected tsetse (Glossina m. morsitans) saliva with the aim of identifying potential soluble factors involved in parasite transmission. Alongside hits from Glossina, we found proteins from the bacterial symbiont S. glossinidius, which compared to naïve flies seem to be more abundant in saliva during a trypanosome infection, as shown by immunoblotting. This could be explained in part by a higher cell permeabilization or microbial dysbiosis led by the trypanosome infection, as Sodalis is capable of intracellular and extracellular growth within tsetse, including in the salivary glands [39]. An overall reduction in the number of salivary proteins was confirmed in infected saliva as previously observed [32,34]. In addition, we identified 27 different T. b. brucei proteins in infected saliva, of which 62.9% were cytosolic and cytoskeletal. It remains unclear why trypanosome proteins are found in tsetse saliva and how they are released by the parasite. Their source could be dying parasites, and/or a more specific mechanism of protein secretion such as shedding of extracellular vesicles (EV) [59]. Indeed, it has been demonstrated that T. brucei BSFs secrete EVs that contain several virulence factors and proteins involved in parasite persistence within the host, including serum resistance-associated protein, VSGs, adenylate cyclase (GRESAG4), and GPI-PLC [60]. Notably, we showed that T. brucei-infected saliva is particularly enriched with trypanosome GPI-anchored surface proteins (37.1%); i.e. several VSGs, BARPs, and one (Tb927.7.360) belonging to a small family of hypothetical proteins, we herein named as metacyclic invariant surface proteins (MISP). Nevertheless, it remains to be determined whether MISP, VSGs, BARP, and other proteins here found in tsetse saliva are part of EVs, secreted in soluble form, or both.

We identified and characterized the MISP family by studying gene and protein expression during parasite life cycle, cellular localization and accessibility to antibodies, protein tertiary structure, immunoprotective potential, and essentiality for parasite development in the tsetse vector. Most of these studies were conducted alongside BARP, a known family of surface proteins expressed by epimastigotes in SGs [21]. In our proteomic analyses, we detected at least four different BARP isoforms in infected saliva (Tb927.9.15520, Tb927.9.15530 and Tb927.9.15570), in agreement with a previous report [34]. MISP were previously included in the trypanosome surface phylome [40], within the Fam50 family that interestingly also includes BARP, and the T. congolense proteins GARP [61,62] and CESP [63]. Apart from T. vivax, genomes from most of the disease-relevant species of African trypanosomes such as T. brucei (including T. b. gambiense and T. b. rhodesiense), T. congolense, and T. evansi encode misp homolog genes. The preservation of MISP only across trypanosome species with full development in the tsetse vector suggests a conserved function and potential role in development or transmission.

We have confirmed that misp and barp gene expression is developmentally regulated in T. b. brucei SG stages as previously determined [36,37,64]. Our semi-quantitative RT-PCR method has allowed the detailed study of individual misp paralogs by exploiting the different number of repetitive elements at their C-termini. We found that misp up regulation is not homogeneous across isoforms, but instead biased towards a preferential expression for the shortest misp-B paralogs. Except for misp-440 that seems to have a low but stable expression, misp-A transcripts are only detected during the late stages of parasite development in SG. It is unknown how these genes are expressed in such different ways despite being part of a gene array that is polycistronically-transcribed with >99.7% sequence conservation in UTR and intergenic sequences. Differences in homolog transcript levels, however, may not reflect changes in the respective protein abundance, which remain unknown as the polyclonal antibodies used in this study are expected to cross-react with all isoforms due to the overall high amino acid sequence conservation. Although proventricular trypanosomes produce large amounts of misp and barp transcripts [21], these are only translated into proteins in SG stages and localize evenly across the cell surface. While BARP expression peaks in SG epimastigotes and is gradually reduced during metacyclogenesis, MISP expression seems to increase with development and peaks in metacyclic forms, leading us to consider MISP as predominant metacyclic surface proteins (S21 Fig). Our observations disagree with a previous study [38] describing MISP (therein named SGE1) as proteins exclusively expressed by SG epimastigotes and absent in metacyclic forms, based on single-cell RNA-seq and immunostaining analyses. Such divergences may be attributed to several factors, including (1) use of different T. b. brucei strains and tsetse infection timing; (2) RNA-seq data not correlating with actual transcript translation like in PV-associated trypanosomes presenting high levels of MISP and BARP transcripts but not proteins; and (3) use of different polyclonal antibodies, particularly relevant when having to bind to scarcely accessible epitopes within the mVSG coat in metacyclic cells. Instead, we used confocal fluorescence microscopy and HRSEM to confirm the surface localization of MISP in metacyclic cells, and importantly, demonstrated that MISP epitopes are susceptible to antibody binding using live cells. Lastly, two more recent studies also support the expression of MISP in metacyclic cells using scRNA-seq [65], or a combination of transcriptomics and proteomics [66] (S3 File). Of importance, these studies were performed using either EATRO1125 [65] or Lister 427 (29:13 clone) parasite strains [66], which together with our MISP immunodetection in TSW196 (S15 Fig), strongly suggest that MISP expression in AnTat 1.1 metacyclic trypanosomes is not strain-specific. As for BARPs, previous observations pointed at their expression being lost in MCFs [20,23], but we detected significant amounts of transcripts and protein levels in MCFs with variable localization (2C Fig). A similar effect was observed in PCFs where ectopic MISP localized to the flagellum, but then regained surface localization when cells differentiated into MCFs (where MISP are natively expressed).

Metacyclic forms are cells with an arrested cell cycle with low transcription activity. Therefore, detecting high levels of misp transcripts prompted us to hypothesize that these proteins may be abundant and play an important role in parasite development, transmission, or in establishing a mammalian host infection. To test this, we attempted to determine misp essentiality by knocking down transcripts from all paralogs using a single RNAi construct due to the high sequence conservation. The developmental expression of misp only in SG stages limited the selection process of clones in the culturable stages of the parasite (PCF). Although partial downregulation was successfully achieved in salivary gland trypanosomes, MISP could still be detected in MCFs making results inconclusive despite the absence of phenotype. We then knocked out all ten misp alleles using CRISPR-Cas9, which did not result in any apparent phenotype in vitro, and only produced a slight reduction in salivary gland infectivity, partially supporting the findings obtained by RNAi. Altogether, these data suggest that MISP are not essential for the development of T. brucei in the vector and may function during or after transmission to a mammalian host [26]. It remains to be determined whether MISP subfamilies have either distinct or complementary functions, as adding back a single misp did not rescue the slight infection phenotype observed, which may be simply due to loss of virulence from the extended culture time needed to generate the transgenic cell lines.

To gain more insight into the function of MISP, we solved the crystal structure of the ordered N-terminal domain of the MISP360 (Tb427.07.360). Structural analysis revealed a triple helical bundle structure that resembles other trypanosome surface proteins such as the Tb/TcHpHbR [67], GARP [48] or a VSG monomer [47]. However, it is important to consider that mature MISP360 contains an additional C-terminal domain predicted to be highly disordered. In the absence of definitive structural data, we modelled the C-terminus, which indicated that it might serve as a long tether effectively elevating the N-terminal head group from the surface of the parasite membrane. This may partly explain the recognition of MISP by polyclonal antibodies shown by live immunostaining. It is tempting to speculate that this extension facilitates a biologically relevant interaction between the parasite and its environment such as promoting the acquisition of nutrients, serving as an adhesion or decoy, or modulating a response in the vertebrate host. The previously characterized TbHpHbR lies partially within the VSG layer of the BSF and allows the trypanosome to acquire nutrients from the blood of the mammalian host. The mechanism by which it acquires nutrients is facilitated by a C-terminal extension that increases its relative distance from the membrane, thus making the ligand-binding site accessible by protruding above the VSG layer [45]. Since MCFs express mVSGs [68] alongside MISP, a C-terminal extension could support a similar role although the substantial length of the extension would likely render cellular uptake of a nutrient payload difficult. However, the unstructured extension is also suited to support an adhesion role for MISP, similar to the function described for the Haemophilus Cha adhesins [69]. Intriguingly, the only study that defined an adhesin function of a trypanosome surface protein (T. congolense CESP) is also part of the Fam50 family of proteins [63], thereby suggesting that MISP could also play a role in adhesion. Our structural analyses revealed the existence of a predominantly hydrophobic pocket in the apical end of MISP360, which we suggest may allow the coordination of a molecular partner, consistent with the putative role as an adhesin. Despite high sequence identity between isoforms, the residues forming and surrounding the pocket lacked sequence conservation and resulted in a distorted cavity between the homolog models that may reflect the capacity to coordinate a structurally diverse set of ligands, potentially from different hosts. Furthermore, the homologs differ in the length of the C-terminal region, originating from the variable number of 26-residues motifs. This region could offer the parasite a robust mechanism to engage a variety of cell types or provide a sequential binding effect to enable tighter adhesion. It is noteworthy that these putative unstructured C-terminal regions contain lysine and arginine residues, which are inherently susceptible to proteolysis consistent with predictions using the PROSPER software [70] (S22 Fig). The predicted proteolytic susceptibility may serve a strategic function in enabling the parasite to release the molecular interactions anchoring it to the tsetse salivary glands and thereby promote transmission. It has been assumed that the MCF surface is structurally and functionally equivalent to that of BSFs as they show the same thick VSG (electron-dense) layer in TEM sections [12,24,71]. This implies that mVSGs also form a dense macromolecular barrier that possibly masks MCF invariant surface proteins from the host’s immune system. Most of the characterized BSF invariant surface proteins (e.g., ISG65, ISG75, ESAG4 and hexose transporters), which are found on either the flagellar or cell surface, are attached via transmembrane domains [8,11]. Two exceptions are the transferrin [9] and the HpHbR receptors [10], which are GPI-anchored and exclusively localize in the flagellar pocket. This restricted localization has been proposed to be due partly to the presence of only one GPI anchor molecule, differing from VSGs, which form homodimers (i.e. two GPI anchors) that allow their homogeneous surface expression [72]. In contrast, our structural and immunostaining evidence shows that MISP appear to be displayed on the entire cell surface as monomer. This may suggest that, unlike BSFs, MCFs allow the homogeneous surface expression of monomeric GPI-anchored invariant proteins, although we cannot rule out the possibility of MISP multimerizing in vivo. This clearly differs when MISP is ectopically expressed in BSF or PCF, in which the proteins localize at the flagellar pocket and/or the flagellar membrane. It remains to be determined the mode of association to the metacyclic membrane of another small family invariable surface proteins known as Fam10 [40], as these molecules lack identifiable hydrophobic transmembrane domain or GPI-signal addition site in the C-terminus.

Due to its surface exposure, antigenicity, and conservation, we considered MISP as potential vaccine candidates. Our immunization experiment failed to provide protection against a tsetse challenge, as it also failed against T. b. brucei needle challenge in a previous study [38]. However, in neither study was the MISP immunogen designed for optimal performance; improvements could include: 1) using an alternative expression system (e.g. Leishmania tarantolae) to conserve all post-translational modifications; 2) immunizing with full length protein; or 3) testing different adjuvants that elicit a protective immune response against a trypanosome infection.

Our proteomic analyses identified peptides from several VSG proteins. The identification of mVSG peptides in infected saliva is not surprising given that infected SGs contain thousands of quiescent metacyclic cells, which may release mVSGs into the SG lumen. mVSG release may occur partly by the action of the parasite GPI-PLC, which is highly expressed in MCFs [23,73] and/or by the action of tsetse salivary proteases. Interestingly, among the VSG species detected, only one corresponds to a canonical mVSG (mVAT4), which has not been previously detected as protein, but its metacyclic expression site (MES) has been well characterized [74]. Notably, a similar study detected mVAT5 in tsetse saliva infected, but from a different T. brucei strain [34]. Although it has been reported that BSFs may express mVSGs from bloodstream expression site (BES) at low levels [75], no study has described the expression of BSF VSG proteins in MCFs. This supports the assumption that, although recombination may occur from MES to BES, it does not appear to occur in the opposite direction since MES lack the 70 bp repeats upstream of the mVSG gene. However, we identified peptides found in at least four BSF VSGs, including MiTat 1.2 (identified by three unique peptides). This suggests that either (1) MCF could have active BES (unlikely since MCFs do not appear to express ESAG proteins [73]) or 2) there is recombination at the MES replacing the mvsg by a vsg gene, or 3) there is recombination between mvsg and vsg genes leading to the formation of mosaics. In addition, we cannot rule out that the undetermined VSG repertoire of the TSW196 strain [76] (used in this study) could differ from other reference strains [77]. Importantly, previous studies have found similar results although this aspect has been overlooked [34,36,64,73] (S6 Table). For example, at least two BSF VSGs were identified in saliva from flies infected with T. b. brucei EATRO 1125 [34]; one BSF vsg was found to be highly transcribed in a T. b. brucei RUMP503 SG infection [64]; and 15 BSF vsgs had similar transcription levels than mvsgs (and a few were detected as protein) in T. b. brucei Lister 427 MCF overexpressing the RNA-binding protein 6 [73].

In summary, not only does the T. brucei-infected tsetse saliva contain infective MCFs, but also a cocktail of G. morsitans salivary proteins, bacterial and trypanosome soluble factors. How this complex composition modulates parasite transmission is a fascinating question yet to be addressed. Among the detected parasite proteins in infected saliva, MISP represent a family of immunogenic, invariant surface proteins that are exposed on the cell surface of the mammalian-infectious metacyclic stage of T. brucei. Based on their external exposure and sequence conservation, we considered MISP as potential vaccine candidates against T. brucei spp-caused trypanosomiasis, in a similar way to the malaria vaccine that targets the sporozoite CSP protein [78]. Lastly, it could be advantageous to use MISP as a xenodiagnosis marker in the field for the identification of disease-transmissible tsetse. Having a parasite marker that discerns between a mature tsetse salivary gland infection (disease vector) and trypanosomes in the midgut (potential non-vectors), would provide a key tool for accurately assessing both vector incrimination and the prevalence of trypanosomiasis in the field.

[END]
---
[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1011269

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.