Leishmania mexicana pathogenicity requires flagellar assembly but not motility

Abstract

Protists of the order Trypanosomatida possess a single multifunctional flagellum, which powers cellular displacement and mediates attachment to tissues of the arthropod vector. The kinetoplastid flagellar cytoskeleton consists of a nine-microtubule doublet axoneme; further structural elaborations, which can vary between species and life cycle stages, include the assembly of axonemal dynein complexes, a pair of singlet microtubules and the extra-axonemal paraflagellar rod. The intracellular amastigote forms of Leishmania spp. build a short, non-motile cilium whose function has remained enigmatic. Here, we used a panel of 25 barcoded promastigote cell lines, including mutants lacking genes encoding flagellar assembly proteins, axonemal proteins required for normal motility, or flagellar membrane proteins to examine how these defects impact on their virulence in macrophages and mice. Mutants lacking the intraflagellar transport (IFT) protein 88 were avirulent indicating that assembly of a flagellum is necessary to allow for Leishmania survival in a mammalian host. A similarly severe loss of virulence was observed upon deletion of BBS2, a core component of the BBSome complex, which may act as a cargo adapter for IFT. By contrast, promastigotes that were unable to beat their flagella due to loss of core axonemal proteins could establish and sustain an infection and only showed a small reduction of parasite burden in vivo compared to the parental cell lines. These results confirm that flagellar motility is not necessary for mammalian infection, but flagellum assembly and the integrity of the BBSome are essential for pathogenicity.

Keywords: CRISPR screen; Leishmania; flagella; motility; virulence.

Figure 1.

Pooled screen of L. mexicana mutants. (a) Average speed [µm/s] plotted against directionality [velocity/speed] for the flagellar mutants and the parental controls (SBL1-5, L. mex Cas9 T7 marked with unique 17-nt barcodes in the 18S SSU rRNA locus) included in the bar-seq screens in BMDMs and mice. The motility data is taken from [12]. The motility of mutants ΔGDP, ΔPMI and ΔPMM, which were included in all bar-seq screens as controls, was not measured. (b) Schematic of screen design and timeline. At time point 0, pooled stationary phase promastigotes were used to infect mice either with a high parasite dose [HD] or low dose [LD]. (c) BMDM infections were started from pools as in (b). Time points of genomic DNA isolation are indicated. (d) Bar-seq screen in mice, Y-axis: Change in barcode proportions calculated by dividing the normalized read counts at the start of the assay by the normalized read counts at each DNA isolation time point. X-axis: DNA isolation time points. Each value represents the average from six infected mice. Error bars show standard deviations between replicates. Red dotted lines indicate two standard deviations above and below the average of “stable control” cell lines. Mutants included in the bar-seq screen have been grouped as follows: mutants with flagellum-associated defects, mutants previously reported to show attenuated virulence (“drop-out controls”) and barcoded parental control cell lines (SBL1-5; “stable controls”). (e) Bar-seq screen in BMDMs. Axes and mutant groupings as for (d). Each value represents the average from 3 BMDM samples.

Figure 2.

Mouse infections with ΔPF16, ΔIFT88 and ΔBBS2. Lesion progression and parasite burden were measured for Cas9 T7 parental (+/+), knockout (-/-) and addback (-/-/+) cell lines. (a, b) ΔPF16, (c, d) ΔIFT88 and (e, f) ΔBBS2. (a, c, e) Measurement of footpad size over time of infection. The average from five mice is shown with error bars indicating standard deviation. p-values are indicated for the 8 weeks post-infection time point and were calculated using a two-tailed unpaired (unequal variance) Students’ t test. (b, d, f) Parasite burden measured from serial dilutions from footpad lesions and lymph nodes. The average is indicated. p-values were calculated using a Kruskal-Wallis-test with Dunn’s post-hoc test and Bonferroni correction, comparing knockout and addback against each other (asterisks on line) and against the parental (individual asterisks). (p-value: *<0.05, **<0.0001; ns: non-significant; raw values and p-values are provided in the Supplementary Table S4).