EIF2α phosphorylation is regulated in intracellular amastigotes for the generation of infective Trypanosoma cruzi trypomastigote forms

FIGURE S1. Characterisation of parasites overexpressing control and mutated eIF2α. (a) Western blotting of total extracts of non-transfected (NT) or epimastigotes overexpressing the indicated forms of eIF2α was probed for eIF2α and Hsp70 just after selection or after 21 days maintained without Geneticin G418. (b) The graph indicates the mean levels of eIF2α ± standard deviations in relation to Hsp70 from three independent experiments comparing the values of NT. (c) Differential interference contrast (DIC) images of the indicated lines. Bars = 5 μm. (d) Quantification of the projected surface area of the indicated epimastigotes lines was made using Cell ̂ M software. The graph shows mean (white lines), 25 and 75% percentiles (boxes) and the standard deviation as max and min values (n = 100). (e) Western blotting of metacyclic-trypomastigote extract (Meta) or from tissue culture-derived trypomastigotes of non-transfected parasites (NT), trypomastigotes transfected with GFP, or overexpressing a wild-type version of T. cruzi eIF2α (WT) or eIF2α mutations. The blots were probed for eIF2α, Hsp70 and aldolase. (f) Band intensity (n = 3) of eIF2α relative to Hsp70 normalised to NT parasites. The values are means ± standard deviations. The asterisk indicates p .05 and ** indicates p < .01 when comparing the results with NT epimastigotes or with WT overexpressors using one-way ANOVA and Tukey's multiple comparisons.

FIGURE S2. T. cruzi epimastigotes with eIF2α mutated at T169 do not fully arrest translation. Polysome profile at 254 nm of epimastigotes from eIF2α overexpressor lines (a) or Cas9 lines (b) harvested at exponential phase and then maintained in LIT (black trace), or after TAU medium for 2 hr (red trace). The graphs indicate the migrating position of monosomes (80S) and polysomes (P) in each panel. The P/80S ratios were obtained by measuring the area bellow both fractions on each graph. Similar results were obtained in triplicate experiments.

FIGURE S3. Abnormal expression of metacyclic-specific proteins occurred only in eIF2α overexpressing epimastigotes. Extracts of epimastigotes (5 × 10⁶) cultured in LIT medium for 1, 4, or 7 days and from non-transfected and purified metacyclic-trypomastigotes (2 × 10⁷) were loaded in a gel and processed for western blotting. The membranes were probed for Hsp70, aldolase, or metacyclic-specific glycoproteins gp90 and gp82 (a). (b) Graph representing the mean and standard deviations of the levels of gp90 relative to Hsp70 (n = 3). (c) Similar extracts we prepared from exponentially growing overexpressors or Cas9 lines, loaded in a gel and probed with antibodies to gp90. The size markers of each gel are indicated on the right side.

FIGURE S4. Relation betwen the 5′UTR and the presence of μORFs with the difference in protein expression. (a) Shows the relation between the UTR length and the difference in protein levels as found by MS/MS analysis for the hits with –log₁₀ p > 3 as generated by Orange3 data analysis program. (b) Scatter plot showing the presence of hits with μORFs and the corresponding differences in expression generated by Orange3 data analysis program. Non-repressive are in blue and show genes whiteout μORF. The coloured symbols correspond to μORF non-overlapping (no), overlapping (o) and both (o/no) each ORF. Size of each symbol is proportional to the log of MSMS detection. Undetermined ORFs are not shown. (c) Frequency of μORF type among major class of ORFs corresponding to the indicated list. The classes shown in read are those with high proportion of μORF to non-repressive ORFs.

FIGURE S5. Relation between codon adaptation index and proteomic data. (a) The graph shows the relation between the Log (MS/MS counts), which correspond to the total number of identified proteins in both populations and the codon adaptation index (CAI). (b) Volcano plot of the CAI and Difference detected in protein expression. Each hit was coloured according to the Log (MS/MS counts) as indicated in the legend of the figure. Both plots were generated by Orange3 software.

FIGURE S6. Intracellular replication of amastigotes overexpressing different forms of eIF2α. (a) Panels show images obtained from high-content analysis (HCA) for the indicated lineages 3 or 4 days after infection with 40:1 multiplicity of infection. Bar = 20 μm. (b–d) The graphs show the number of amastigotes per infected cell 2, 3 and 4 days after infection of U-2 OS cells with the indicated multiplicity of infection. The values are means ± standard deviation of three independent experiments. The asterisk indicates p < .05, ** < .01 and *** p < .001 calculated using the two-way ANOVA, at the fourth day of replication. (e and f) Intracellular replication of amastigotes from non-transfected (NT), GFP expressing (GFP) and overexpressing eIF2α (WT, S43A, T169A and S43A/T169A) after 2 and 3 days in LLC-MK2-infected cells. The number of amastigotes per cell was counted in 100 infected cells using optical microscopy after Giemsa staining, as described in Methods. The values show means, min and max values and the standard deviation of triplicate (2 days) and duplicate (3 days) experiments. The asterisk indicates p < .05 and ** indicates p < .01 both calculated using the Student's t test in relation to NT parasites. # indicates p < .05, ## p < .01 and ### p < .001, calculated using the Student's t test in relation to WT overexpressors.

FIGURE S7. Effect of immunosuppression on mice infected with eIF2α overexpressors. Mice were previously infected with trypomastigotes as shown in Figure 5. Forty days after infection, the mice started to be immunosuppressed with cyclophosphamide as described in Methods and blood parasitemia was followed again for the indicated period starting at the initial day of immunosuppression. The graphics displayed the percentage of animals with detectable blood parasites (a) and the respective parasitemia levels (b).

TABLE S1. Primes used in this work. The bases in italic in the sgeIF2 primer correspond to the T7 promoter for the in vitro transcription, in bold, the homology region to eIF2α sequence and the underlined sequence corresponds to the bases of the pUC Sg vector. The double underlined bases in eIF2T169don are modified residues to replace the Thr 169 for Ala and insert the BssHII diagnostic site.

TABLE S2. Protein identification by MS/MS analysis. (Table S1.xlsx). Summary of proteomic data (n = 5) for controls and eIF2α mutants. Sheet 1 is the LFQ output for the 4,301 protein groups detected. For each protein group, a LFQ ratio of T169A mutant versus parental cells, the respective t-test difference and a corresponding log p-value are indicated. For ratio formation, a constant was added to each LFQ value, to avoid division by zero (such ‘infinite ratios’ are clearly distinguishable from genuine ratios by being significantly larger, smaller or exactly 1.0). GO annotations were downloaded from TriTrypDB. Sheets 2 and 3 are reduced to protein groups with significant abundance shift (–log₁₀ p-value >1). Sheets 4, 5 and 6 are the output tables of the Maxquant analysis (summary, parameters and protein groups).

TABLE S3. Function attribution of the hits with -log₁₀P-value > 1. (Table S2.xlsx). The most significant abundance shifts upon eIF2α T169A mutation detected by proteomic analysis with functional annotation.

TABLE S4. UTRs and μORFs calculated for the T. cruzi YC6 CDS. (Table S3.xlsx). In the first sheet, the predicted best 5′UTRs using the UTRme software are indicated and in the second, the identified μORFs. The third sheet summarises the enrichment analysis of the tree μORF categories (Fo = overlapping, Fno = non-overlapping, no-repression) and of predicted trans-sialidases, within both protein cohorts, increased and decreased as a result of the T169A mutation (–log p > .5) respectively. Categories with enrichment over 1.5-fold and Benjamin–Hochberg false discovery rate of 0.02 are highlighted. Sheet 4 lists the 15 predicted trans-sialidases used in the enrichment analysis. Sheet 5 lists the predicted best 3′UTR also using the UTRme software. The corresponding GFF files for the 5′UTR and 3′UTR are provided as separate supplementary files.

TABLE S5. CAI calculated for the T. cruzi YC6 CDs. (Table S4.xlsx). The values were calculated considering codon usage table for all CDSs, or using the codon usage table for the 100 top hits detected by MS/MS counts, both using the YC6 genome data.

Data S1. Additional supplementary files: (YUTR-5 best-score.gff) and (YUTR-3 best-score.gff) correspond to the annotation of the best hits for the 5′ and 3′UTR detected by using the UTRme software.