Table S1. Sources of seed sequences used in Fig. S1. The ‘*’ indicate genes that were removed during the procedure of finding clusters of homologues, because they either had no homologues or were too similar to other gene families part of the list.

Table S2. Source of databases used to search for actin and tubulin associated genes.

Table S3. Observed migration speeds of Trichomonas vaginalis T016 exposed to the vaginal epithelial cell line MS74.

Fig. S1. Alignment of TvFim1 and its homologues from a range of eukaryotes is shown in (A). Known alpha-helical structures of the corresponding calponin homology domain (CH) are indicated above the alignment, whereas amino acids thought to be conserved residues for F-actin or suppressor residues are marked with a dot. Characterized actin binding sites are marked as a line below the alignment (based on Klein et al., 2004). (B) TvFim1 includes both, the EF-hand and four calponin homology domains (CH). TvFim2 is N-terminally truncated. For comparison, the human fimbrin protein (HsFim1, L-plastin) is only 16 amino acids longer and shares an overall sequence identity of 43% with TvFim1. The alignment was generated using the CLC Workbench and default settings and the following protein accessions: O.c., Oryctolagus cuniculus (XP_002720283.1); A.c., Acanthamoeba castellanii (ELR12888.1); S.p., Schizosaccharomyces pombe (NP596289.1); M.m., Mus musculus (AAH05459.1); H.s., Homo sapiens (AAB02845.1); G.g., Gallus gallus (P19179); D.m., Drosophila melanogaster (AAF48722.1); D.d., Dictyostelium discoideum (P54680); S.m., Schistosoma mansoni (AAC14025.1); G.p., Gibberella pulicaris (CAA10667.1); A.t., Arabidopsis thaliana (AAC39359.1).

Fig. S2. FPLC-purification of heterologously expressed TvFim1 from Escherichia coli strain C41. In (A) the FPLC run showing the loading, washing and elution phase with the elution peak of the HIS-tagged TvFim1 at 100 mM imidazol. (B) Coomassie-stained SDS-PAGE of 1: non-induced E. coli cell lysate; 2: induced E. coli cell lysate, 3: FPLC purified TvFim1 of 68 kDa (indicated by arrow). Marker (M) in kilodalton.

Fig. S3. Antibody controls.

A. Western blot of crude pre-immune (pIS) and immune rat sera (IS) from TvFim1 peptide immunization on protein extract from Trichomonas vaginalis and the E. coli strain expression recombinant TvFim1.

B. Western blot using purified TvFim1, HA-antibody and GFP-antibody on protein extract of the T. vaginalis strains expressing the corresponding constructs. Cropped bands of this blot were used for Fig. 2. Molecular marker (M) in kDa.

C. Immunofluorescence controls of the T. vaginalis strain T016 on fibronectin-coated slides using TvFim-pre-immune sera and anti-actin (Sigma). Scale bar: 10 μm.

D. shows the gradient-like colocalization of the HA-tagged TvFim1 with that of the endogenous copy towards the periphery of the cells. Scale bar: 2 μm.

Fig. S4. Additional TvFim1 distributions in Trichomonas vaginalis T016 cells. (A–D) TvFim1 concentrates at the protruding edge of a parasite, while gliding across host tissue, contour highlighted by a dashed line in (D). The protein furthermore associates with structures reminiscent of actin cables (E–H). Many parasites proliferate and become multinuclear (arrow heads) during infection (I–L). Arrow heads point to exemplary multi-nuclear cells. (M–R) Further localizations of fimbrin together with the hydrogenosomal marker enzyme ASCT (acetate : succinate CoA-transferase). Scale bar: 10 μm.

Fig. S5. Live cell imaging of Trichomonas vaginalis on human vaginal epithelial cells. Live imaging of the parasite exposed to human epithelial cells reveals T. vaginalis actively migrates across tissue, while they appear to use their flagella and apical tip as the guiding end. A clear trailing end and pseudopodia are also visible. While some rapidly divide on VECs, from what appears to be di-nuclear cells, others massively increase their size (juggernauting) and develop more than two nuclei and flagellar pockets before division. Areas of interest are marked with an asterisk in the first image of every series. For details please also refer to Movies S7–S10. Scale: 10 μm.

Fig. S6. A TvFim1::HaloTag fusion construct was generated using the identical primers as used for TvFim::HA construct (see Material & Methods of manuscript). The HaloTag plasmid (Martincová et al., 2012, PLoS ONE 7: e36314) was kindly provided by Pavel Dolezal (Charles University, Prague) and the product cloned into the plasmid through NdeI and BamHI and correct insertion verified through sequencing. RT-PCR and Western blot analysis were carried out according to the methods described in the manuscripts main text. For live imaging of TvFim1 we fused the gene to the recently reported HaloTag (Martincová et al., 2012). We obtained several clones, three of which we analysed and shown above. (A) Reverse transcriptase PCR on RNA isolated from the transfected T016 cells and using specific primers, shows the fusion gene TvFim1::HaloTag is expressed in clones 2 and 3. (B) No fusion-protein is detected by Western blot analysis using an anti-haemagglutinin antibody (an HA-tag is included in the fusion protein generated by pHaloTag). This correlates with the lack of fluorescence during microscopic analysis of the clones.

Fig. S7. A. We tried to knockdown TvFim1 using anti-sense RNA as described previously by Ong et al. (2007, JBC 282: 6716–25). The anti-sense region was amplified from genomic DNA using the primers TvFim1-as_BamHI_F: 5′-GGTGGTGGATCCCGTATGAGCCTTCTCAAGAC-3′ and TvFim1-as_NdeI_R: 5′-GGTGGTCATATGGCCGTCAGCGATGCCG-3′. After sequence verification the fragment was cloned into pTagvag2 for standard expression under the control of the SCS promoter. T. vaginalis was then transfected with 30 μg of plasmid DNA and selected through G418 (see ‘Gene cloning’ section).

B. RNA from recombinant and wild-type trichomonads was transcribed into cDNA and used as a template for real-time quantitative PCR using TvFim1 specific primers (see ‘Quantitative PCR’ section) and biological and technical triplicates. The expression level of TvFim1 in the wild type was used as the reference (100%) against TvFim1-as, which revealed no significant upregulation or downregulation of the gene.


Movie S1. Total internal reflection microscopy movie of actin self-assembly.


Movie S2. Total internal reflection microscopy movie of actin self-assembly in the presence of 3 μm TvFim1.


Movie S3. Total internal reflection microscopy movie of actin self-assembly in the presence of 5 μm TvFim1.


Movie S4. Live imaging of TvFim1::GFP. A daughter cell can be seen budding from a multinuclear mother cell that is attached to host tissue. Fimbrin::GFP is found to cluster around contractile ring-like structure of the daughter cell. One image was taken every 3 s and run at 7 frames per second.


Movie S5. Live imaging of TvFim1::GFP. During the migration of the parasite across tissue, waves of TvFim1::GFP can clearly be seen in close proximity to the migration front and moving away from it. One image was taken every 3 s and run at 7 frames per second.


Movie S6. Live imaging of TvFim1::GFP. Clustering of TvFim1::GFP was predominately observed to occur in areas of the parasite attaching to the vaginal epithelial cells. One image was taken every 5 s and run at 7 frames per second.


Movie S7. Time lapse movie of T. vaginalis on vaginal epithelial cells showing concerted movement and the parasite using the apical tip as the guiding end and for flagellar sensing. 40 min into infection and one image taken every second and run at 7 frames per second.


Movie S8. Time lapse movie of a T. vaginalis nicely showing the pseudopods and the trailing end of the parasite while gliding along host tissue 20 min after infection. One image was taken ever 0.2 s and run at 10 frames per second.


Movie S9. Time lapse movie of a T. vaginalis cell dividing on a vaginal epithelial cell just 5 min after infection. One image taken every second and run at 7 frames per second.


Movie S10. Time lapse movie of T. vaginalis on vaginal epithelial cells 70 min into infection, demonstrating some adherent cells to massively increase their overall cell mass, a process we refer to as juggernauting. One image taken every second and run at 28 frames per second.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.