An extensive endoplasmic reticulum-localised glycoprotein family in trypanosomatids

Abstract

African trypanosomes are evolutionarily highly divergent parasitic protozoa, and as a consequence the vast majority of trypanosome membrane proteins remain uncharacterised in terms of location, trafficking or function. Here we describe a novel family of type I membrane proteins which we designate 'invariant glycoproteins' (IGPs). IGPs are trypanosome-restricted, with extensive, lineage-specific paralogous expansions in related taxa. In T. brucei three IGP subfamilies, IGP34, IGP40 and IGP48 are recognised; all possess a putative C-type lectin ectodomain and are ER-localised, despite lacking a classical ER-retention motif. IGPs exhibit highest expression in stumpy stage cells, suggesting roles in developmental progression, but gene silencing in mammalian infective forms suggests that each IGP subfamily is also required for normal proliferation. Detailed analysis of the IGP48 subfamily indicates a role in maintaining ER morphology, while the ER lumenal domain is necessary and sufficient for formation of both oligomeric complexes and ER retention. IGP48 is detected by antibodies from T. b. rhodesiense infected humans. We propose that the IGPs represent a trypanosomatid-specific family of ER-localised glycoproteins, with potential contributions to life cycle progression and immunity, and utilise oligomerisation as an ER retention mechanism.

Keywords: Trypanosoma brucei; endoplasmic reticulum; evolution; exocytosis; protein sorting; trypanosoma; variant surface glycoprotein.

Figure 1. FIGURE 1: The invariant glycoprotein (IGP) family.

(A) Phylogenetic reconstruction of the IGP family. The tree shown is the best Bayesian topology with branch support for important nodes indicated from both Bayesian and PhyML calculations. Clades are indicated by vertical bars. (B) Schematic diagram showing the secondary structure prediction and domain architecture of the IGP family. Lumen indicates the portion of the molecule predicted to be within the ER lumen and Cyt designates the predicted cytoplasmic domain. CLECT indicates C-type lectin domain. (C) Schematic structures of epitope-tagged and domain-swap IGP constructs used in this study. All constructs contain an HA tag as indicated by small triangle below the bar. Abbreviations: IGP48, full length IGP48; IGP48-ISG65, IGP48 ectodomain fused to ISG65 TMD and cytoplasmic domain; BiPN-IGP48, IGP48 ectodomain replaced with BiPN, retaining IGP48 TMD and cytoplasmic domain; IGP40, full length IGP40; IGP40-ISG65, IGP40 ectodomain fused to ISG65 TMD and cytoplasmic domain; BiPN-IGP40, IGP40 ectodomain replaced with BiPN, retaining IGP40 TMD and cytoplasmic domain; dTRIM, IGP48 with predicted trimerisation domain deleted.

Figure 2. FIGURE 2: The IGP family are developmentally regulated ER proteins.

(A) Copy numbers of IGP48 , IGP40 and IGP34 mRNAs measured by qRT-PCR, in different life cycle stages, normalised to long slender BSF mRNA levels at 1.0. Error bars denote standard errors of the mean from triplicate measurements on independent RNA samples. Western blot of trypanosome whole cell lysates using anti-IGP48 affinity-purified antisera raised against E. coli-expressed recombinant protein at 1:100 dilution. The blot was re-probed for PAD1 (protein associated with differentiation 1), and which is specifically unregulated in the stumpy bloodstream form, to validate the short stumpy lysate. Rightmost; whole cell lysate probed with anti-IGP48 antisera to validate specificity. Abbreviations: BSF (LS), long slender bloodstream form; BSF (SS), short stumpy bloodstream form and PCF, procyclic culture form. (B) Intracellular localisation of IGP48-HA and IGP40-HA in BSF cells under permeabilised conditions, and detected with anti-HA antibody (red). Top panel: co-staining with anti-TbBiP (green). Lower panels: co-straining with anti-TbRabX2 and anti-p67 (green) using confocal microscopy. Bar = 2µm. Inset: expression of IGP48-HA and IGP40-HA in 427 BSF cells detected by Western blotting using anti-HA antibody. (C) Digestion of IGP48 with PNGase F or Endo H, or treatment with tunicamycin results in a large molecular weight shift. IGP48 was detected in fractionated lysates using anti-HA antibody. (D) Turnover of IGP48 and IGP40. Quantification of anti-HA reactivity in lysates of cells expressing IGP48-HA and IGP40-HA following inhibition of protein synthesis with cycloheximide. Error bars represent the standard deviation and values were normalised against a loading control, BiP (n = 3).

Figure 3. FIGURE 3: IGP40 and IGP48 are required for normal cell proliferation.

(A) qRT-PCR of control and tetracycline-induced RNAi lines for IGP34, IGP40 and IGP48 24 hours post-induction (n = 3). Right panel shows a Western blot of IGP48 following RNAi induction after 24 hours with β-tubulin as loading control. (B) Growth curves of control and tetra-cycline-induced RNAi lines for IGP34, IGP40 and IGP48, showing a growth defect within 24 hours post induction. Representative results for one of two clonal cell lines for each construct studied are shown (n = 2) (all subsequent experiments were performed using this cell line). Cultures were diluted daily to maintain cell densities between 10⁵ and 2 × 10⁶ cells/ml, and a cumulative pseudo-growth curve is shown. Counts were carried out in triplicate, error bars represent standard error of the mean.

Figure 4. FIGURE 4: Localisation of IGP sorting signals.

(A) Immunofluorescence demonstrating the locations of IGP constructs in BSF trypanosomes. HA-tagged constructs were detected in permeabilised cells with anti-HA antibody (red). The ER was stained with anti-TbBiP (green) and DNA stained with DAPI (blue). BiPN constructs, in which the lumenal/ecto-domain is replaced by the BiP ATPase domain, no longer co-localise with TbBiP and so are not retained within the ER. (B) Expression of IGP40-HA and IGP48-HA in 427 BSF cells detected by Western blotting using anti-HA antibody. Note that the BiPN chimeras migrate slower than predicted from their molecular weight, an observation that is consistent with the behaviour of BiPN-ISG65 chimeras reported previously . (C) Location of BiPN constructs was determined by co-localisation with p67, Rab5A or Rab11, markers for the lysosome, early or recycling endosomes respectively (green). BiPN-IGP40 and BiPN-IGP48 both demonstrate significant overlap with all three intracellular markers. Scale bar 2 µm.

Figure 5. FIGURE 5: The IGP ectodomain is required for retention.

(A) Western blot analysis of surface-biotinylated (Pellet) and non-biotinylated (Supernatant) cells to detect surface-exposed IGP proteins and chimeras. Blots were also probed for an intracellular (TbRabX1) and surface (ISG75) control, to demonstrate that cells are intact and that surface components are successfully biotinylated. Note that the Pellet lanes have been moved in Photoshop simply for clarity and no other manipulation has taken place. Surface presence of BiPN-IGP48 and BiPN-IGP40 was further demonstrated by confocal microscopy. Non-permeabilised cells were stained with anti-HA antibodies (red) and for DNA (blue). Scale bar = 1 μm. (B) Kinetics of protein secretion. BSF trypanosome cells expressing IGP48-HA, BiPN-IGP48 and BiPN were pulse-labeled with ³⁵S-Met/Cys for 15 minutes and then chased for 3 hours. At 0 and 3 hours cultures were separated into cell and medium fractions. Labeled proteins were immunoprecipitated with anti-HA and separated by SDS-PAGE. OE, OverExposure to reveal a 70 kDa proteolytic fragment (P) cleaved from BiPN-IGP48. Lower panel: Detailed kinetics of BiPN-IGP48 secretion. Cells were pulse-labeled with ³⁵S-Met/Cys for 15 minutes and at the indicated chase times, aliquots were treated as described above. The proteolytic BiPN-IGP48 50 kDa fragment (P) appears in the medium after 1 hour. (C) Left panel: Turnover kinetics of BiPN-IGP48 (open symbols) and BiPN-IGP40 (closed symbols) was determined by blocking protein synthesis with cycloheximide and detecting residual protein with anti-HA antibodies. Right panel: Turnover is sensitive to inhibition by 10 µM MG-132 (open symbols) or 20 mM NH₄Cl (closed symbols). Results were normalised to 100% at t = 0. The graph represents the mean of two independent experiments, with the standard error of the mean indicated. Numbers to the left of some panels indicate the positions of co-migrated molecular weight standards and are in kDa.

Figure 6. FIGURE 6: Subcellular localisation of BiPN-IGP48 chimeras is not dependent on the lumenal domain.

(A) The IGP48 CLEC (dCLEC) or trimerisation lumenal domain (dTRIM) was replaced with the BiP ATPase domain and the location determined by immunofluorescence. In each case the BiPN chimera is in red and a marker protein visualised using polyclonal antibodies is in green. Scale bar = 1 μm. Verification of protein expression was carried out by Western blot (inset at right). Due to low expression, blots have been deliberately overexposed and relevant reactivity is indicated with arrows. (B) Turnover of dCLEC and dTRIM constructs. Protein degradation following cycloheximide treatment was monitored as described in Figure 2. Experiments were done in duplicate and error bars indicate standard error of the mean.

Figure 7. FIGURE 7: IGP48 forms a complex in vivo.

Lysates from cells expressing various IGP48 chimeras were subjected to native PAGE, followed by detection by Western blotting. Left: Replacement of the IGP48 ectodomain with the BiP ATPase domain (BiPN-IGP48) results in loss of the high molecular weight (450 - 500 kDa) complexes. Right: Cells expressing IGP48-HA or IGP48-HA (48-HA) plus IGP48-FLAG (48-FLAG) were immunoprecipitated with anti-FLAG antibody, followed by Western blotting with anti-HA antibody. Whole cell lysates are shown to the left, and wild-type (427) and single transfected cells, as well as a bead plus lysate with no antibody IP control (Beads).

Figure 8. FIGURE 8: IGP48 is retained in the ER in short stumpy-like cells.

(A) BSF cells expressing IGP48-HA were incubated at 37°C, 20°C (cold-shock) or with pCPT-cAMP for 12 hours. IGP48-HA was visualised with anti-HA antibody and co-stained with anti-BiP antibody. IGP48-HA remains in the ER in short stumpy-induced cells. DNA was visualised using DAPI. All images are captured at the same magnification, scale bar 2 μm. (B) Surface biotinylation was performed to determine if IGP48-HA reaches the cell surface in short stumpy-like cells. Cells were cultured in vitro at 37°C or 20°C for 12 hours and the biotinylation assay was carried out as described previously. IGP48-HA was detected by Western blot with anti-HA antibody. Blots were stripped and re-probed for an intracellular marker, RabX1 (localises to the ER) and a surface marker, ISG75, which localises to both the surface and endosomal compartments.

Figure 9. FIGURE 9: Effects of IGP knockdown on major surface protein copy number and exocytosis.

(A) Cells were sampled from either induced or uninduced IGP48 RNAi cell lines at the times indicated. Membranes were probed with anti-VSG221, TbBiP, ISG65, or ISG75 and relative protein abundance was determined by densitometry. Experiments were done in duplicate and bars indicate standard error of the mean. Data were normalised to 100% at t = 0. (B) Export of newly synthesised VSG in induced and uninduced IGP48 RNAi cells. Surface accessible VSG was hydrolysed by GPI-PLC after hypotonic lysis of the cells. Soluble and membrane-form VSG was recovered by incubation with ConA-sepharose. Data represent the kinetics of newly synthesised VSG transported from the endomembrane system to the cell surface, shown as percent of VSG at the cell surface. Data were taken from two independent experiments and standard error of the mean is shown. Student’s t-test showed statistically significant difference between induced and uninduced cells at the time point indicated with an asterisk (p < 0.05). Right panel: Metabolic labelling of newly synthesised VSG following 24 hour RNAi induction. Newly synthesised VSG was labelled with ³⁵S-methionine and detected by autoradiography. (C) Top: Hypotonic lysis, followed by separation of surface (supernatant, S) and intracellular (pellet, P) VSG 221 was visualised by SDS-PAGE and Commassie staining and levels of VSG compared to that in whole cell lysates (L). No significant difference in VSG distribution is seen between induced and uninduced cells. Lower: Levels of the BiPN reporter following labelling of cells with ³⁵S-methionine were detected in induced and uninduced cells, following a 3 hour chase. No significant change is seen between export of BiPN from the cell in induced compared to uninduced trypanosomes.

Figure 10. FIGURE 10: Ultrastructural analysis of IGP48 RNAi cells reveals defects in the ER.Transmission electron micrographs of IGP48 RNAi cell lines.

Top left: Representative uninduced IGP48 RNAi cell with major secretory pathway organelles indicated. Other panels are induced cells after 24 hours induction. Top centre: Distorted ER with apparent lumenal inclusion. Top right: ER tubules with apparent normal morphology. Lower left: Extensive vesicles associated with the Golgi complex. Based on observations that the Golgi complex is concave towards the trans-face (see top left panel), these vesicles are likely ER to Golgi transport intermediates corresponding to a structure similar to the ERGIC, i.e. ER-GIC-like. Lower right: Examples of extensive clusters of vesicles in close association with ER tubules. Scale bars are 500 nm. Abbreviations: ER, endoplasmic reticulum; ERGIC, ER-Golgi intermediate compartment; TGN, trans-Golgi network.

Figure 11. FIGURE 11: IGP48 elicits variable IgG and IgM responses in human T. b. rhodesiense infections.

Recombinant IGP48 was resolved on a 10% SDS-PAGE gel. After Western Blotting, PVDF membranes were probed with plasma from T. b. rhodesiense patients and controls. Lane M: molecular weight markers. Lane 1 (from left): Ponceau red staining of IGP48 (*). Lanes 2, 3 and 4: membrane probed with plasma L2 with anti-IgG, anti-IgM and no secondary antibody control respectively. Lanes 5, 6 and 7: membrane probed with plasma L3 with anti-IgG, anti-IgM and no secondary antibody control respectively. Lanes 8, 9 and 10: membrane probed with plasma L16 with anti-IgG, anti-IgM and no secondary antibody control respectively. Lanes 11, 12 and 13: membrane probed with endemic control plasma LC7 with anti-IgG, anti-IgM and no secondary antibody control respectively. Lanes 15 and 16: anti-IgG and IgM controls with no primary antibody.