Summary drug assays performed in vivo with P. berghei and in vitro with P. falciparum.

A.) Peters’ four-day suppressive test in vivo with P. berghei. Percent suppression of a variable dosing range used to determine both the appropriate clearance dose and potential resistance-generating dose. Parasitaemias of individual mice in each dose cohort were normalised to the mean vehicle control-treated parasitaemia at each time point, and the mean percent-suppression value plotted (error bars represent ±SD). A non-linear regression curve with a variable slope was then fitted to the data points for each dose cohort; B.) End-point (day 4) data for the Peters’ four-day suppressive test. Parasitaemias of individual mice in each dose cohort were normalised to the mean vehicle control-treated parasitaemia at day 4 and each individual value plotted. A non-linear regression curve with a variable slope was then fitted to the data points. Where n<4, mice were euthanised prior to the experimental end point for humane reasons; C.) In vivo drug assay of PbAZMR_G95D_1 (teal) and PbAZMR_G95D_2 (light blue) grown with a full four-day 70 mg/kg q.d. (solid lines, n=2 per genotype) or 140 mg/kg q.d. (dashed lines, n=2 per genotype) drug course versus WT with the vehicle (no drug; black; n=2) and a full four-day 70 mg/kg q.d. drug course (grey; n=2). Treatment was started on day 1. (Values indicate mean parasitaemia, and error bars represent the SEM); D.) Dose response curves of Pf_AZMR_C4 and Pf_AZMR_D2 clones (both with G76V mutation in P. falciparum 50S Rpl4) compared to Pf_WT. Data presented are the mean of three technical replicates and representative of three independent drug trials (error bars represent ±SD); E.) Comparison plot of IC50s of P. falciparum azithromycin resistant clones versus WT. Plotted are the mean IC50 values from three independent drug trials. (error bars represent the SEM, **** = P<0.0001, ordinary one-way ANOVA, Dunnett’s multiple comparisons test); F.) Alignment of region of Rpl4 associated with azithromycin of erythromycin resistance in bacteria and plastids of parasites, algae and plants. Mutations identified in this study are depicted in red and mutations selected from the literature in blue. Included organisms: P. berghei (Plasmodium berghei); P. falciparum (Plasmodium falciparum); T. gondii (Toxoplasma gondii); B. microti (Babesia microti); C. velia (Chromera velia), E.coli (Escherichia coli); D. radiodurans (Deinococcus radiodurans); N. gonorrhoeae (Neisseria gonorrhoeae); S. pneumoniae (Streptococcus pneumoniae); C. reinhardtii (Chlamydomonas reinhardtii);A.thaliana (Arabidopsis thaliana). Asterisks denote invariant residues, colons depect conserved residues groups with strongly similar properties, and periods denote semi-conserved residues groups with similar properties. (AZM = azithromycin; ED = effective dose)

Generation of azithromycin resistant P. berghei parasites in vivo.

PbAZMR_G95D_1 and PbAZMR_G95D_2 were both generated off a 70 mg/kg q.d. dosing regimen, and PbAZMR_S89L was generated off a 60 mg/kg q.d. dosing regimen. The initial four-day dose is represented by the shaded rectangle, and the start of each dose thereafter is indicated by a shaded box to avoid rectangle overlaps. Parasites were deemed resistant when they failed to reduce in parasitaemia on the third day post treatment start (with the exception of PbAZMR_S89L, which grew slowly so was monitored differently).

Blood stage growth of P. berghei azithromycin parasites is impaired compared to WT without drug pressure

Growth assay of PbAZMR_G95D_1 (n=3) and PbAZMR_S89L (n=3) compared to WT (n=3). Parasitaemia was monitored by Giemsa-stained thin blood smears. Dot points represent mean parasitaemia, and error bars indicate ±SD. Growth of both azithromycin resistant parasite lines is delayed compared to WT, and eventually fails to increase.

Mosquito infection data for both P. berghei and P. falciparum azithromycin resistant parasites.

A.) Exflagellations were counted per 1×104 red blood cells (RBCs) and are shown as the mean ±95%CI, with no significant difference between strains (ordinary one-way ANOVA with Dunnett’s multiple comparisons; WT [n=17], PbAZMR_G95D_1 [n=20], PbAZMR_G95D_2 [n=10], PbAZMR_S89L [n=2], where ‘n’ is the number of independent biological replicates). At 9-to-14 days post blood feed, a minimum of 10 midguts were dissected for each ‘n’ independent infection (WT [n=13], PbAZMR_G95D_1 [n=11], PbAZMR_G95D_2 [n=5], PbAZMR_S89L [n=2]) and assessed for infection prevalence (pie charts below, percentage indicates the mean; ns= not significant by way of ordinary one-way ANOVA). Oocysts of infected midguts were counted and pooled, and non-zero values were plotted (error bars represent the mean ±95%CI). All azithromycin resistant strains had significantly fewer oocysts than WT (*=P<0.05; ****=P<0.0001; ordinary one-way ANOVA, Dunnett’s multiple comparisons test); B.) Mean number of salivary gland sporozoites per mosquito. Between days 20- and 26-post blood feed, a minimum of 12 mosquitoes (WT) or 20 mosquitoes (PbAZMR_G95D_1, PbAZMR_G95D_2, and PbAZMR_S89L) were dissected. All azithromycin resistant strains produce fewer salivary gland sporozoites than WT (****=P<0.0001, Welch’s t-test; error bar represents +95% CI). Results are from ‘n’ independent mosquito infections: WT (n=21), PbAZMR_G95D_1 (n=14), PbAZMR_G95D_2 (n=10), PbAZMR_S89L (n=2); C.) Quantitative PCR on genomic DNA isolated from midguts infected with both P. berghei azithromycin resistant parasites and WT strains. Copy numbers of genes encoded in each organelle were first normalised to nuclear genes, then expressed as a ratio to WT parasites. There is a clear reduction in apicoplast genome copy number in PbAZMR_G95D_1 (n=4), PbAZMR_G95D_2 (n=3) and PbAZMR_S89L (n=1), where n=number of biological replicates. Error bars represent the SEM; D.) Immunofluorescence assay on midgut oocysts, 14 days post blood feed. Green represents labelling of the P. berghei apicoplast, with anti-ACP antibodies. Hoechst is staining DNA. PbAZMR_G95D_1 oocysts consistently show dispersed ACP staining (bottom panel), while WT oocysts show a branched ACP staining (top panel). Images are maximum projections (except brightfield, which shows the largest single slice). Scale bar=10μm; E.) Quantification of development of blood-feed-ready stage V gametocytes produced by Pf_WT and Pf_AZMR parasites. No significant difference is seen (Unpaired t-test, two-tailed P value); F.) At 7 days post blood feed, a minimum of 10 midguts were dissected for each ‘n’ independent infection (Pf_WT [n=3], Pf_AZMR [n=3]) and assessed for infection prevalence (pie charts below, percentage indicates the mean; ns= not significant by way of Unpaired t-test, two-tailed P value). No significant difference was observed between the number of midgut oocysts in mosquitoes infected with either Pf_WT or Pf_AZMR parasites (Unpaired t-test, two-tailed P value); G.) At day 17 post blood feed, a minimum of 12 mosquitoes were dissected per genotype. No significant difference was observed between the number of salivary gland sporozoites in mosquitoes infected with either Pf_WT or Pf_AZMR parasites (Unpaired t-test, two-tailed P value).

Immunofluorescence assay on PbAZMR_G95D_1, PbAZMR_G95D_2 and WT sporozoites reveals that azithromycin resistant sporozoites lose their apicoplast.

A.) Sporozoites were stained with anti-circumsporozoite protein (CSP; red) to mark the outer surface of the sporozoite, anti-ACP (green) to mark the apicoplast, and Hoechst (blue) to mark the nucleus. WT sporozoites have a clear, punctate apicoplast (top panel), while azithromycin resistant sporozoites have diffuse apicoplast marker staining (bottom two panels). Images are representative of the calculated majority; B.) Quantification of apicoplast presence (API+, green) and absence (API-, red) in WT (n=2068), PbAZMR_G95D_1 (n=445), and PbAZMR_G95D_2 (n=855) sporozoites, where n=number of total sporozoites counted. A minimum of 110 sporozoites were counted for each parasite strain, for each of three independent mosquito infections. Green bars indicate the mean percentage apicoplast positive apicoplast, error bar represents +95% CI. Both azithromycin resistant strains produce sporozoites with significantly fewer intact apicoplasts than WT (P<0.0001, Fisher’s exact test); C.) Quantification of sporozoite motility trails from PbAZMR_G95D_1 (n=159), PbAZMR_G95D_2 (169), and WT (n=175) sporozoites, from two independent mosquito infections. There was no significant difference in types of trails observed between each parasite line (chi-square test).

Transmission of azithromycin resistant P. berghei parasites to a naïve host is severely impaired, and in vitro liver stage development appears stalled.

A.) Kaplan-Meier curves of P0 infections by biting. At least 15 infected mosquitoes were allowed to feed on naïve mice for at least 15 minutes and mice were monitored for 14 days by Giemsa smear (one mouse infected with PbAZMR_G95D_1 was only monitored until day 7, indicated by the initial triangle marker). Number of transmission attempts per strain were WT, n=8; PbAZMR_G95D_1, n=4; PbAZMR_G95D_2, n=4. Naïve mice infected with azithromycin strains by biting failed to establish an infection, and this was significantly different than WT (P=0.0007, Mantel-Cox test) B.) Kaplan-Meier curves of P0 infections by 1×103 IV of sporozoites. Mice were monitored for 14 days by Giemsa smear. Number of transmission attempts per strain were WT, n=6; PbAZMR_G95D_1, n=1; PbAZMR_G95D_2, n=4. Naïve mice infected with azithromycin strains by 1×103 IV failed to establish an infection, and this was significantly different than WT (P=0.01, Mantel-Cox test). C.) Kaplan-Meier curves of P0 infections by 1×104 IV of sporozoites. Mice were monitored for 14 days by Giemsa smear. Number of transmission attempts per strain were WT, n=10; PbAZMR_G95D_1, n=5; PbAZMR_G95D_2, n=6. Only naïve mice infected with either WT or PbAZMR_G95D_2 by 1×104 IV were able to establish an infection, but infection with PbAZMR_G95D_2 sporozoites showed delayed patency (mean 8.5 days ± 3.8 days [SD]) and impaired growth. This was significantly different than WT (P<0.0001, Mantel-Cox test). D.) Growth curve comparison of P0 infections established by WT and PbAZMR_G95D_2 by 1×104 IV of sporozoites. Mice infected with PbAZMR_G95D_2 sporozoites grow slowly. Points shown indicate mean parasitaemia (±SEM), WT, n=9; PbAZMR_G95D_2, n=6, where ‘n’ represents the number of infected mice monitored for the period; E.) Immunofluorescence assay of liver cells infected with WT, PbAZMR_G95D_1 and PbAZMR_G95D_2 sporozoites in vitro and fixed at 48 hours post infection. The top two panels show HC-04 cells infected with WT sporozoites, the middle two panels show HC-04 cells infected with PbAZMR_G95D_1 sporozoites and the bottom two panels show HC-04 cells infected with PbAZMR_G95D_2 sporozoites. Azithromycin resistant liver stage schizonts only show dispersed ACP (apicoplast) staining. Nuclei (blue) in azithromycin resistant liver stage schizonts appear clustered and donut-shaped. Red staining indicates anti-Pb, marking the outside of the liver stage parasite. All images are maximum projections and scale bars represent 10 μm; F.) Quantification of parasite size (μm2) at 48-hours post sporozoite inoculation of HC-04 cells in vitro pooled from two independent experiments, with WT (n=60), PbAZMR_G95D_1 (n=31), and PbAZMR_G95D_2 (n=47), where ‘n’= number of individual liver stage schizonts measured. PbAZMR_G95D_1 and PbAZMR_G95D_2 liver stages are significantly smaller than WT (P<0.0001, Ordinary one-way ANOVA with Dunnett’s multiple comparisons test; bars represent the mean ±SD); G.) Pooled quantification of nuclei counts from the same liver stage schizonts quantified in F. PbAZMR_G95D_1 and PbAZMR_G95D_2 liver stages have significantly fewer nuclei than WT (P<0.0001, Ordinary one-way ANOVA with Dunnett’s multiple comparisons test; error bars represent the mean ±SD); H.) Nuclei counts, normalised to the size of each individual liver stage parasite described in F (number of nuclei per μm2). PbAZMR_G95D_1 and PbAZMR_G95D_2 liver stages have significantly fewer nuclei than WT even when normalised to liver stage parasite size (P<0.0001, Ordinary one-way ANOVA with Dunnett’s multiple comparisons test; error bars represent the mean ±SD).

Liver stage development of P. falciparum azithromycin resistant parasites is impaired.

A.) HC-04 cell traversal by Pf_WT and Pf_AZMR sporozoites at multiplicity of infection (MOI) 0.3, measured by FITC-Dextran uptake and counted by FACS. Cells were fixed at 3-hours post inoculation and show no significant difference between Pf_WT and Pf_AZMR indicating normal sporozoite motility in azithromycin resistant parasites (ns = not significant, Paired t-test two-tailed P value); B.) HC-04 cell invasion by Pf_WT and Pf_AZMR sporozoites measured by FACS after incubation for 18-hours using CSP-positive antibody staining of fixed permeabilized cells. No significant difference was seen in invasion of HC-04 between Pf_WT and Pf_AZMR sporozoites (ns = not significant, Paired t-test two-tailed P value); C.) Quantification of parasite liver load in humanised mice infected by i.v. injection of 8 x 105 Pf_WT sporozoites (n=2 mice) or Pf_AZMR sporozoites (n=3 mice) by qRT-PCR. Livers of infected humanised mice were harvested at day 5 post sporozoite i.v. and tissue homogenised for subsequent loads analysis and determination of the degree of chimerism. Pf_AZMR infected mice showed a significantly lower liver stage parasite load than mice infected with Pf_WT sporozoites (Unpaired t-test, two-tailed P value, P=0.0023, error bars represent mean ±SD); D.) Immunofluorescence assay of liver sections from livers of humanised mice infected with either Pf_WT or Pf_AZMR sporozoites. Liver stage schizonts were stained with anti-HSP70 (cytosolic marker) and anti-PTEX150 (parasitophorous vacuole marker) and DAPI DNA stain. Pf_AZMR liver stage schizonts show little to no staining with either anti-HSP70 or anti-PTEX150 and show aberrant nuclear staining, with individual nuclei unclear. Scale bar 10 μm; E.) Immunofluorescence assay of liver stage schizonts and stained with anti-ACP (apicoplast marker), anti-CSP (parasite marker) and DAPI DNA stain. Pf_AZMR livers stage schizonts show a loss of anti-ACP and anti-CSP stain compared to WT, and again show aberrant nuclear staining, with DAPI appearing to aggregate around holes. Scale bar 10 μm; F.) Quantification of liver stage schizont size (μm2) from liver sections from livers of humanised mice infected with either Pf_WT or Pf_AZMR sporozoites. Sections from independent mice were selected and individual ‘n’ liver stage schizonts measured, Pf_WT (n=10), Pf_AZMR (n=10). No significant difference in liver stage schizont size was seen (ns = not significant, Unpaired t-test, two-tailed P value); G.) Quantification of nuclear centres counted from liver stage schizonts imaged in D and E. Pf_AZMR liver stage schizonts show significantly fewer nuclei than Pf_WT, indicating impaired liver stage development (Unpaired t-test, two-tailed P value, P<0.0001; error bars represent mean ±SD).