{"paper_id":"a1d897ca-46b7-42f8-bb0b-c1b33a2f1ba5","body_text":"1 of 43 \n \nProteins with proximal-distal asymmetries in axoneme localisation control \nflagellum beat frequency \nAuthors \nCecile Fort1, Benjamin J. Walker2,3, Lore Baert1,4, & Richard John Wheeler1 \nAffiliations \n1 Medawar Building for Pathogen Research, Nuffield Department of Medicine, University of Oxford, \nOxford, UK \n2 Department of Mathematical Sciences, University of Bath, Claverton Down, Bath, BA2 7AY, United \nKingdom \n3 Department of Mathematics, University College London, Gordon Street, London, WC1H 0AY, United \nKingdom \n4 Current acress: Swiss Tropical and Public Health Institute, University of Basel, Basel, Switzerland \n* To whom correspondence should be addressed: richard.wheeler@ndm.ox.ac.uk \nAbstract \nThe 9+2 microtubule-based axoneme within motile flagella is well known for its symmetry. However, \nexamples of asymmetric structures and proteins asymmetrically positioned within the 9+2 axoneme \narchitecture have been identified in multiple different organisms, particularly involving the inner or \nouter dynein arms, with a range of functions. Here, mapped, genome-wide, conserved proximal-distal \nasymmetries in the uniflagellate trypanosomatid eukaryotic parasites. Building on the genome-wide \nlocalisation screen in Trypanosoma brucei we identified conserved proteins with an analogous \nasymmetric localisation in the related parasite Leishmania mexicana. Using deletion mutants, we map \nwhich are necessary for normal cell swimming, flagellum beat parameters and axoneme \nultrastructure, and using combinatorial endogenous fluorescent tagging and deletion, map co-\ndependencies for assembly into their normal asymmetric localisation. This revealed 15 proteins, 8 \nknown and 7 novels, with a conserved proximal or distal axoneme-specific localisation. Most were \nouter dynein arm associated, and showed that there are at least two distinct classes of proximal-distal \nasymmetry – one dependent on the docking complex, and one independent. Many were necessary \nfor normal frequency of the tip-to-base symmetric flagellar waveform, and our comprehensive \nmapping reveals unexpected contribution of proximal-specific axoneme components to frequency of \ndistal waveform initiation. \nKeywords: cilia, flagella, asymmetry, outer dynein arms, docking complex, flagellar beat \nIntroduction \nFlagella and motile cilia are microtubule-based organelles found across diverse eukaryotic lineages \nused for motility. They share a highly symmetric core architecture, known as the axoneme, based on \nnine doublet microtubules around a central pair of singlet microtubules. However, asymmetries in the \ndistribution of axoneme-decorating proteins are emerging as being important for the function of cilia \nand flagella1. \nFlagella undergo different waveforms specific to their function: typically, a planar symmetric near-\nsinusoid flagellar-type beat (e.g. human sperm) or a planar asymmetric ciliary-type beat (e.g. ciliated \nepithelia). Some flagella can switch waveform type, with Chlamydomonas adopting an asymmetric \nbeat for normal swimming and a symmetric beat during a bright-light phototactic response2–4, in this \ncase maintaining a base-to-tip direction of waveform propagation. Other examples can also switch \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n2 of 43 \n \nwaveform direction between base-to-tip and tip-to-base (e.g. Leishmania)5,6. Leishmania are one of \nthe trypanosomatid parasites, a family which also includes Trypanosoma brucei. Due to tractable \nreverse genetics, these uniflagellate human pathogens are powerful model systems for analysing \nflagellar biology7,8. Trypanosomatid parasite flagellum-driven motility is necessary for normal life cycle \nprogression9,10. \nThe flagellar beat is generated by the action of dynein complexes attached to the axoneme. The outer \ndynein arms (ODAs) attached to the nine doublets every 24 nm and are canonically viewed as being \nthe primary driver of the flagellar beat11. Contrastingly, there are multiple different inner dynein arm \n(IDA) complexes attached in a larger 96 nm12 repeating unit, viewed as necessary for controlling the \nflagellar beat waveform rather than generating it11. However, recently, we showed that the preferred \ndirection of waveform propagation in trypanosomatids is associated with a linear proximal-distal \nasymmetry in the ODAs1. We previously identified two paralogues of the ODA-docking complex (DC) \nheterodimer13, one heterodimer specific to the proximal and one to the distal axoneme1. Deletion of \ndDCs lead to loss of the distal ODAs, which caused a switch from the normal tip-to-base symmetric \nbeat to the rarer base-to-tip asymmetric beat. While this precise DC-dependent asymmetry appears \nspecific to the trypanosomatids, analogous DC or ODA asymmetries are found in diverse eukaryotes, \nincluding the unicellular parasite Giardia, the green alga Chlamydomonas and Humans1 \nThis is not the only potentially conserved proximal-distal asymmetry. ARL13B is a well-characterised \nconserved marker of cilia and flagella necessary for normal ciliary / flagellar length14–21. However, in \nsome tissues, including mouse oviduct and tracheal tissue, ARL13B is enriched in the proximal \naxoneme22. A similar proximal localisation is seen in T. brucei21. Phosphodiesterases (PDEs) have \nalso been identified as enriched in the distal flagellum: PDEA in L. mexicana9 and PDEB in T. brucei \n23,24. This suggests complexity in proximal-distal axoneme composition beyond the ODAs. \nCyclic AMP (cAMP) and calcium ion (Ca2+) signalling are often implicated in the control of beat type to \ncontrol cell motility25–31. For example, Ca2+ signalling is involved in the phototactic response of \nChlamydomonas32–34. Trypanosomatid parasites are no exception: Ca2+ and cAMP alter beat type of \ndemembranated reactivated Leishmania axonemes35 and there are likely links to asymmetrically \npositioned proteins. FLAM6 in T. brucei36, has a proximal DC-like localisation and has a predicted \ncAMP binding domain, flagellar cAMP signalling by PDEB is necessary for productive infection of the \nT. brucei insect vector37, and we previously identified LC4-like which has a predicted Ca2+ binding \ndomain as a distal axoneme specific ODA-associated protein necessary for normal beat frequency in \nT. brucei and Leishmania1. Overall, this suggests interplay of signalling and asymmetric protein \ndistribution. \nIt is becoming clear that complex proximal-distal asymmetries exist in the axoneme, and there are \nhints that proteins with a proximal-distal asymmetry in axoneme localisation may function in flagellum \nbeat control. We therefore sought to map, genome-wide, all conserved proximal-distal asymmetries in \na model flagellum. Using genome-wide subcellular protein localisations in T. brucei, we identified all \nproximal- and distal-specific proteins and identified those with a L. mexicana ortholog and an \nanalogous sub-axonemal localisation. We show that L. mexicana have at least two proximal and distal \nasymmetries. A cohort of proximal-distal proteins are DC dependent or/and ODA heavy chain \nassociated, including a new paralagous pair of DC-associated proteins where one is proximal and the \nother distal. Finally, we demonstrated that a subset of proximal and distal-specific proteins are \nnecessary for normal control of flagellum beating, including, surprisingly, that proximal-specific \nproteins can influence frequency of waveforms starting at the distal tip of the flagellum. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n3 of 43 \n \nResults \nFifteen proteins have proximal-distal asymmetry conserved between T. brucei and L. mexicana \nTo comprehensively identify proximal- and distal-specific proteins that may be responsible for \nflagellum beat control, we used the TrypTag (genome-wide subcellular protein localisation in \nTrypanosoma brucei) dataset38. Through a manual survey of proteins annotated with a flagellum or \naxoneme localisation, we identified 55 proximal- or distal-specific flagellar proteins. We reasoned that \nthose with an L. mexicana ortholog that also has a proximal or distal-specific localisation are most \nlikely to be functionally important. Of these 55, we therefore retained those with an L. mexicana \northolog (Table S1, S2, Figure S1, Figure 1) and did not further analyse those that lack an ortholog \n(Table S3, Figure S2, S3). To determine if proximal- or distal-specific localisation was conserved \nbetween T. brucei and L. mexicana, we tagged the L. mexicana proteins with mNeonGreen (mNG) at \nthe C or N terminus at their endogenous loci (Figure 1). Overall, ~50% (25/55) of T. brucei proximal or \ndistal-specific proteins have a L. mexicana ortholog, of which ~60% (15/25) have a comparable \nasymmetric localisation, leaving 10 proteins for detailed analysis. \nBased on the sub-flagella localisation in these two species, we asked whether there are additional \ncharacteristic asymmetries in protein distribution distinct from the previously-characterised pDC/dDC \nasymmetry1. For each T. brucei and L. mexicana cell line, we measured the fluorescence signal \ndistribution along the axoneme (Figure 1). pDC1 and pDC2 and dDC1 and dDC2 regions have a \ncharacteristic ~50%-50% proximal-distal distribution in T. brucei and ~20%-80% proximal-distal in \nL. mexicana1 (Figure 1A). We identified 3 proteins with a localisation similar to pDCs and 2 similar to \ndDCs in both species. We named these “pDC-like” and “dDC-like” localisations (Figure 1A, E and \nFigure 1C,G). This included a paralogous pair, one with a pDC-like and one with a dDC-like \nlocalisation. In L. mexicana LmxM.29.0240 (pDC-like) and LmxM.30.0090 (dDC-like). In T. brucei \nTb927.6.1660 (pDC-like) and Tb927.8.8000/Tb927.4.4370 (dDC-like), where the latter is a more \nrecent gene duplication due to the partial chromosome 4/8 duplication39. \nA subset of proximal-specific proteins had a short ~20% proximal signal in both T. brucei and \nL. mexicana. We named this the “short proximal” localisation (Figure 1B, F). This group included \nARL13B which also had a weak fluorescent signal along entire the flagellum, as previously described \nin T. brucei procyclic forms21. Two proteins localised along the length of the flagellum with stronger \nsignal towards the distal tip in T. brucei. These were PDEB1 and 2, replicating the previously \ndescribed localisation in T. brucei bloodstream forms24, and we observed a corresponding enrichment \nin the distal ~30% of the L. mexicana flagellum. We named this localisation group “distal enriched \nPDEs” (Figure 1D, H).This meta-analysis of two related species with differing morphologies suggests \nat least two types (DC-dependent and independent) of proximal and distal asymmetries. \nDeletion of novel asymmetrically distributed proteins has little impact on axonemal structure \nDeletion of axoneme proteins can cause disruption of axonemal structure, therefore we generated \ndeletions of each protein with proximal or distal-specific localisations in L. mexicana. Complete loss of \nthe respective open reading frames (ORFs) was confirmed using diagnostic PCR from purified \ngenomic DNA from each cell line (Figure S4). We were able to generate all deletion mutants by \nreplacement of both alleles with drug-selectable markers, indicating that none of these proteins are \nvital. Flagellum length was near normal in all except one deletion mutant: ∆ARL13B had shorter \nflagella (Figure 2A,B) as previously observed in T. brucei14 along with many other species14,40. This \nconfirms ARL13B is functioning as expected in Leishmania, consistent with ARL13B mutations in \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n4 of 43 \n \nhumans causing Joubert syndrome ciliopathy 20,41,42 All but one of the proteins are therefore axoneme \ncytoskeleton components rather than factors necessary for axoneme assembly. \nTo identify any critical role in forming the axoneme ultrastructure, we analysed the flagellum structure \nof the deletion mutants by thin section transmission electron microscopy. In trypanosomatids, the \nflagellum protrudes from an invagination at the flagellar base – the flagellar pocket. The short pDC \naxoneme region of L. mexicana means that all cross-sections of flagella protruding through a flagellar \npocket will be within the pDC-like and proximal-short axoneme regions (Figure 2C-E), while cross-\nsections through free flagella will largely (~90%) be in the dDC-like axoneme region and ~40% will be \nin the distal enriched PDEs region (Figure 2C, 2F,G). In both flagellar pocket and free flagellum cross-\nsections, no large defects in the axoneme organisation were visible: the ninefold symmetry of the \nouter doublets and the central pair was unchanged (Figure 2C-G, first column). As might be expected \nfor proteins restricted to a small axoneme sub-domain, no proteins restricted to the pDC or dDC \nregions were necessary for normal overall axoneme organisation. \nNext, we analysed changes in electron density in the flagellar-pocket and free-flagellum axoneme \ncross sections in these deletion mutants to identify subtle changes to axoneme structure. Making use \nof the ninefold radial symmetry of the outer dynein arms, we first perspective-corrected then ninefold \nrotationally averaged the axoneme cross-section images43,44. Subsequently, we used 25 rotationally \naveraged axoneme images to generate average flagellar pocket and free flagellum axoneme electron \ndensity maps and perform statistical comparisons to detect changes in electron density from the \nparental cell line (Figure 2). We confirmed the validity of this methodology by applying it to ∆pDC1, \n∆pDC2, ∆dDC1 and ∆dDC2. Both dDCs are necessary for distal (but not proximal) ODA assembly. \nContrastingly, neither pDC are necessary for proximal ODA assembly, as dDCs relocalise to fill the \nentire axoneme1. Concordantly, ∆pDC1 and ∆pDC2 had changes in electron density in the ODAs and \nat the point of attachment of the ODAs to the outer doublets in flagellar-pocket flagellar cross-sections \n(Figure 2D), while ∆dDC1 and ∆dDC2 had a large loss of electron density in the ODAs in free-\nflagellum flagellar cross-sections (Figure 2F). \nWe predicted that proteins with a pDC or dDC-like localisation will be associated with the DC/ODAs, \nand any electron density change will be in that axoneme region. In flagellar-pocket axoneme cross-\nsections, ∆LmxM.29.0240 had a small loss of electron density at the base of the ODAs (Figure 2D) \nand both ∆FLAM6 36,45 and ∆LmxM.31.2530, are both predicted to have a cyclic nucleotide binding \ndomain, had a clear loss of electron density within the ODAs.  \nContrastingly, deletion mutants of proteins with a short proximal localisation (∆LmxM.32.0390, \n∆LmxM.36.5300, ∆ARL13B) had small or undetectable change in ODA electron density. These data \nwere, however, relatively noisy, with changes in electron density around the central microtubule \ndoublet and loss of electron density at the radial spoke head in several deletion mutants. As these \nincluded ∆pDC2, we believe these noisy observations are spurious, perhaps due to small variability in \ndoublet position (Figure 2D, 2E). For proteins with a dDC-like localisation, in free-flagellum axoneme \ncross-sections, ∆LC4-like had a loss of electron density in the ODAs while ∆LmxM.30.0090 had no \ndetectable change (Figure 2F). For the distal enriched PDEs, neither showed any detectable change \n(Figure 2G). Overall, the most significant changes in axoneme electron density were ODA-associated \nand for proteins with a pDC or dDC-like localisation. This suggests ODAs possess the primary \nstructural asymmetries, although our method is only sensitive to proximal-distal, rather than doublet-\nto-doublet, asymmetries. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n5 of 43 \n \nThere are two mechanisms for proximal and two for distal-specific localisations \nTo build direct evidence for whether proximal or distal-specific proteins are dependent on the DC \nasymmetry, we tested whether their localisation is dependent on DC proteins. First, for proteins with a \npDC-like or short proximal localisation, we deleted pDC1 in the respective cell lines expressing the \ntagged protein and the successful deletion were confirmed by diagnostic PCRs (Figure S5). As \npreviously observed1, this causes loss of pDC2::mNG proximal flagellum signal, as pDC1 and 2 are \nco-dependent for their normal localisation (Figure 3A). ∆pDC1 also caused loss of proximal axoneme \nsignal for all proteins normally with a pDC-like localisation (FLAM6::mNG, LmxM.29.0240::mNG and \nLmxM.31.2530::mNG) (Figure 3A). None of the proteins normally with a short proximal localisation \n(LmxM.32.0390::mNG, LmxM.36.5300::mNG and ARL13B) lost proximal axoneme signal on pDC1 \ndeletion (Figure 3B). To confirm this result, we deleted dDC2 in these tagged lines. Again, as \npreviously observed, this causes extension of pDC1::mNG and pDC2::mNG signal to fill ~30% of the \nflagellum (Figure 3A)1. All three proteins with a normally pDC-like localisation behaved similarly on \ndDC2 deletion (Figure 3A), while proteins normally with a short proximal localisation were unaffected \n(Figure 3B). Our classification of proteins by localisation into pDC-like and short proximal therefore \nrepresents a dependency on the DC asymmetry for their localisation. \nWe next carried out the corresponding experiment for proteins with a distal-specific localisation. Here, \ndDC2 deletion causes loss of distal axoneme signal for dDC1::mNG and LC4-like::mNG (Figure 3C), \nas would be predicted from our previous work1, and LmxM.30.0090::mNG behaved similarly (Figure \n3C). However, the localisation of PDEB2::mNG, a representative distal enriched PDEs, was \nunaffected (Figure 3D). pDC1 deletion causes expansion of signal dDC1::mNG, dDC2::mNG and \nLC4-like::mNG to fill the entire axoneme, again as expected1, and LmxM.30.0090::mNG behaves \nsimilarly (Figure 3C). PDEB2::mNG localisation was unaffected (Figure 3D). We have therefore \naccurately also identified the cohort of distal proteins dependent on the DC asymmetry. Overall, this \nconclusively shows that there are two mechanisms for proximal and distal-specific localisation \noccurring in trypanosomatid parasites. \nA subset of DC-dependent asymmetrically positioned proteins are ODA heavy chain \nassociated \nTo dissect where proteins with pDC-like and dDC-like localisations sit within the ODA-DC complex \nproteins, we tested whether their localisation was dependent on ODA dynein heavy chains. However, \nfirst, we tested the co-dependency of ODA dynein heavy chains on their localisation. Similar to \nhumans, trypanosomatids have two ODA dynein heavy chains, ODAα and ODAβ. These are \northologous to C. reinhardtii ODAγ and ODAβ respectively. In C. renhardtii, the ODA complex \nassembles in the cytoplasm prior to transport into the axoneme46, and mutation of either C. reinhardtii \nODAγ or ODAβ prevents this process47 and ODA448 have disrupted ODAγ or ODAβ respectively and \ncompletely lack ODAs). Thus, we expect L. mexicana ODAα and ODAβ to be co-dependent for \nnormal localisation. Confirmation of ODAα and ODAβ deletion is shown in Figure S6A. As expected, \ndeletion of ODAα in a cell line expressing ODAβ::mNG showed complete loss of ODAβ::mNG signal \nfrom the axoneme. However, surprisingly, deletion of ODAβ in a cell line expressing ODAα::mNG \ncaused loss of ODAα::mNG signal from only the distal axoneme (Figure 4A). \nWe confirmed this result by investigating the ∆ODAα and ∆ODAβ cell lines by thin section electron \nmicroscopy, investigating proximal (flagellar-pocket) and distal (free-flagellum) cross-sections in both \ndeletion mutants (deletion validation is shown in Figure S6B). In ∆ODAα, ODAs were completely \nabsent, while ODAs were only completely absent in the distal flagellum in ∆ODAβ. In the proximal \nflagellum (sections in the flagellar pocket), part of the ODA was present – it appeared that the \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n6 of 43 \n \ninnermost half of the ODA remained (Figure 4B). This is consistent with the expected position of \nODAα as the innermost ODA dynein heavy chain, but not with ODA assembly requiring full \ncytoplasmic pre-assembly prior to flagellum import. Therefore, ODAs require ODAα but, unlike in \nChlamydomonas, ODAs can incompletely assemble preferentially in the proximal axoneme in the \nabsence of ∆ODAβ. \nOn the basis of this result, we analysed whether proximal or distal-specific proteins bind to the ODAs \nrather than the DC. For proximal proteins, this required deletion of ODAα (Figure 5A, validation of \ndeletion Figure S7A), while for distal proteins we deleted ODAβ (Figure 5B, validation of deletion \nFigure S7B). As expected, pDC1::mNG, mNG::dDC1 and dDC2::mNG are not dependent on an ODA \nheavy chain for their normal localisation (Figure 4C, D). Among proximal proteins, FLAM6::mNG, \nLmxM.36.5300::mNG, ARL13B::mNG signal was lost on ODA heavy chain deletion, indicating that \nthese are likely associated with the ODAs rather than the DC directly (Figure 5A). Among distal \nproteins, LC4-like::mNG and LmxM.30.0090::mNG localisation was altered on ODA heavy chain \ndeletion, both with reduced signal and failure to localise all the way to the flagellar tip (Figure 5B). \nNormal distal PDEB::mNG signal was ODA-independent. This reveals complexity: proteins can be \nproximal-specific and associated with the ODAs, with their proximal localisation either dependent \n(FLAM6, LmxM.36.5300) or independent (LmxM.36.0830) of the asymmetry of the DCs attaching the \nODAs to the doublet. There are, therefore, two different mechanisms involved for proximal-distal \nasymmetry of ODA-associated proteins. The pDC-dependent and ODA heavy chain-independent \nlocalisation of LmxM.29.0240 and LmxM.31.2530 is also potentially indicative of a function \nassembling pDCs. \nLmxM.29.0240 is necessary for pDC assembly \nTo dissect which proteins are necessary for the DC assembly, we first deleted proteins with a pDC-\nlike localisation in cells expressing pDC1::mNG localisation or deleting proteins with a dDC-like \nlocalisation in cells expressing dDC2::mNG (validation of cell lines in Figure S8). LC4-like and \nLmxM.30.0090 deletion did not affect the distal localisation of dDC2::mNG, although some \naccumulation of dDC2::mNG around the transition zone occurred following LmxM.30.0090 deletion \n(Figure 6B, validation of cell lines in Figure S8B). Therefore, these proteins are not necessary for dDC \nassembly. Contrastingly, LmxM.29.0240 was necessary for normal pDC1::mNG localisation, while \nFLAM6 and LmxM.31.2530 were not (Figure 6A, validation of cell lines in Figure S8A). LmxM.29.0240 \nis, therefore, acting as a pDC component necessary for pDC assembly; thus we named it pDC3. \nFollowing our naming scheme for the proximal and distal DC1 and DC2 paralogous pairs, we named \nLmxM.30.0090 dDC3 – although it was not necessary for dDC assembly. \nTo comprehensively test whether pDC3 is truly necessary for pDC assembly and map the co-\ndependencies of pDC-associated components, we used combinatorial tagging and deletion. pDC3 \nwas necessary for the normal localisation of FLAM6::mNG and LmxM.31.2530::mNG in addition to \npDC1::mNG (Figure 6C, validation of cell lines in Figure S8C). Normal LmxM.31.2530::mNG \nlocalisation was also dependent on FLAM6. pDC1, 2 and 3 therefore form the pDC, with FLAM6 and \nLmxM.31.2530 binding to the structure, LmxM.31.2530 in a FLAM6-dependent manner. \nProximal and distal-specific proteins contribute to the control of the tip-to-base flagellum beat. \nProteins that are not vital for flagellum assembly or normal ultrastructure but are well-conserved \nbetween related species are good candidates for flagellar beat regulators. To identify any correlation \nof beat regulation function with proximal or distal asymmetry, we carried out detailed cell swimming \nand flagellum beat analysis of all proximal and distal proteins not necessary for normal flagellum \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n7 of 43 \n \nassembly. We carried out three analyses (Figure 7, S10): 1) Average cell swimming speed, in a deep \nvolume and analysed away from the slide and coverslip to avoid surface interaction effects16; 2) \nProportion of cells undergoing high frequency tip-to-base symmetric (flagellar-type), low frequency \nbase-to-tip asymmetric (ciliary-type), low frequency aperiodic movement (static/uncoordinated) beats \nwhen in a thin volume between a slide and coverslip; 3) Beat waveform properties (amplitude, \nfrequency and number of waves per flagellum) for cells in a thin volume between a slide and coverslip \nundergoing a tip-to-base symmetric waveform. For the latter, we noted that changes in mutants were \ndominated by change to beat frequency rather than amplitude, waves per flagellum, or the additional \ncontrol measures (Figure S10). \nFirst, we consider proteins with a dDC-like localisation, where previous analyses are more \ncomprehensive. As previously described, dDC1 and dDC2 deletion caused reduced swimming speed \nand reduced ability to carry out the normal tip-to-base symmetric beat (Figure 7A, B). When they did \nundergo a tip-to-base symmetric beat, it was much lower frequency. Again, as previously described, \nLC4-like deletion caused a small but significant increase in swimming speed, arising from a higher tip-\nto-base symmetric beat frequency. Deletion of the novel dDC-associated protein, LmxM.30.0090, had \nno significant effect on beat frequency, proportion of time spent undergoing a tip-to-base symmetric \nbeat or beat frequency (Figure 7C). This confirmed the expected phenotypes but revealed little of \nnovelty. \nDeletion of either distal enriched PDEBs had little effect on swimming speed (Figure 7A) and reduced \nthe number of cells able to undergo a tip-to-base symmetric beat (Figure 7B); however, those that did \nhad a significantly increased beat frequency (Figure 7C). Distal enriched PDEs are therefore potential \npositive regulators of tip-to-base symmetric beats, but negative regulators of beat frequency. \nNext, we considered proteins with a proximal localisation, seeing more surprising changes for both \nproteins with a pDC-like or short proximal localisation. In our previous work1, no large effect of pDC1 \ndeletion on cell swimming speed was observed, thus was not analysed in detail. Here, consistent with \nthis, pDC1, pDC2 and pDC3 deletion had little effect on swimming speed (Figure 7D). However, \npDC2 and pDC3 deletion caused less tip-to-base symmetric beating, with a prominent increase in \nbase-to-tip asymmetric beats upon pDC3 deletion (Figure 7E). pDC1 and pDC2 deletion changed tip-\nto-base symmetric beat frequency, giving a bimodal distribution of frequencies, i.e. cells tended to \nhave either faster or slower beat frequency (Figure 7F). As pDC1, pDC2 and pDC3 are co-dependent \nfor assembly of the pDC, we might expect these deletion mutants to have similar phenotypes. There \nwas, however, mutant-to-mutant variability, but overall pDC proteins were necessary for a normal \nbeat profile. \nFLAM6, LmxM.31.2530 (pDC-like localisation), LmxM.31.0240 or LmxM.32.0390 (short proximal \nlocalisation) deletion all had little effect on swimming speed (Figure 7D) or proportion of time spent \nexhibiting different beat types (Figure 7E). However, all four had a large increase in beat frequency of \ntip-to-base symmetric beats (Figure 7F). Potentially, a proximal or distal-specific protein may have an \neffect on only the proximal or distal waveform. However, analysing frequency, amplitude and \nwavelength for the proximal and distal flagellum separately showed no significant difference between \nthe two for any mutant with altered overall beat frequency (Figure S11). Therefore, surprisingly, \nproximal-specific axoneme proteins can be negative regulators of the flagellum-wide beat frequency \nof beats starting at the far end of the flagellum. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n8 of 43 \n \nDiscussion \nWe have shown that proximal-distal asymmetry of axonemal organisation is a complex phenomenon. \nAnalysis of genome-wide T. brucei data identified 55 examples of flagellar proximal- or distal-specific \nprotein localisations. ~50% (25/55) have orthologs among the trypanosomatid parasites and ~30% \n(15/55) have an ortholog in L. mexicana with a comparable asymmetric localisation, summarised in \nFigure 8. Many of these are necessary for a normal flagellar beat, summarised in Table S4. \nPerhaps surprisingly, some aspects of proximal-distal asymmetry were evolutionarily well-conserved, \nwhile others appear to be recent innovations. This study identified several novel asymmetrically \nlocalised proteins whose localisation is conserved between L. mexicana and T. brucei: a paralogous \npair associated with the proximal and distal docking complex, LmxM.29.0240 and LmxM.30.0090 \nwhich we name pDC3 and dDC3 respectively, a proximal DC-associated protein, LmxM.31.2530 \nwhich we name pDAP1, and two proteins which localise to a short proximal axoneme region, \nLmxM.36.5300 and LmxM.32.0390 which we name SPA1 and SPA2 respectively. We previously \nnoted that ODA proximal-distal asymmetries occur across diverse eukaryotes49–54 although are not \nnecessarily orthologous (i.e. not all involving a pDC and dDC pair of paralogous heterodimers) to \nT. brucei and L. mexicana1, and the majority of proteins we analysed in detail were ODA-associated \n(Figure 5), hinting at a tendency for proximal-distal asymmetries to generally arise in ODAs. \nImportantly, uncharacterised orthologs of pDC3/dDC3 and SPA2 are found in the genomes of several \nflagellate/motile ciliate species, including Chlamydomonas reinhardii, and Alphafold2 structure \npredictions of pDC3/dDC3, SPA1 and SPA2 predicted globular domains not identified by sequence-\nbased protein domain detection (PFAM, Superfamily, etc.)55–57. \nThe proximal and distal flagellum appear, overall, similarly complex in terms of asymmetrically \npositioned molecular machinery. This applies to both genome-wide analysis of T. brucei and the set of \nproteins with an ortholog in L. mexicana. However, there was a large degree of apparently recent \nadaptation. Many asymmetrically positioned T. brucei proteins lacked an L. mexicana ortholog, and a \nfurther set had a whole flagellum rather than asymmetric localisation (Figure S1) – indicating either \nloss of asymmetry in L. mexicana or gain of asymmetry in T. brucei. Either way, this indicates notable \nadaptation within the different highly motile trypanosomatid parasite lineages on the time scale of \nhundreds of millions of years. It is an intriguing possibility that evolvability of flagellum beat control \ninvolves evolvability of asymmetric position within the flagellum. \nWe have confidently identified at least two distinct types of proximal and two types of distal \nasymmetry, indicating that there must be multiple mechanisms underlying their formation. Previously, \nwe saw that the pDC/dDC asymmetry is likely achieved by competition for axoneme binding on top of \nIFT transport to form a proximal-distal concentration gradient, leading to mutually exclusive axoneme \nregions1. Interestingly, our data showed that OADα and OADβ were not fully co-dependent for \nassembly into the axoneme, unlike in C. reinhardtii, perhaps related to differences in DC biology \ncompared to C. reinhardtii. Here, we comprehensively mapped the conserved proteins whose \nasymmetric localisation is dependent on the dDC/pDC asymmetry, finding pDC3, pDAP1 and FLAM6 \nas pDC-dependent and dDC3 and LC4-like as dDC-dependent. Interestingly, pDC3 is necessary for \nthe correct assembly of the pDC1/2 heterodimer into the proximal axoneme while its paralog, dDC3, \nwas not necessary for normal dDC1/2 distal axoneme localisation. Perhaps this is a key distinguishing \nfeature of the proximal and distal docking complex. \nThe other axoneme asymmetries we identified do not seem to have the same mutual exclusivity as \nthe pDC/dDC asymmetry: we identified novel short-proximal (SPA1 and SPA2) and short-distal \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n9 of 43 \n \nregions (PDEB1 and PDEB2), but no corresponding conserved long-proximal and long-distal regions, \nrespectively. A simple explanation for short-distal and short-proximal regions is limiting protein \nabundance, with their asymmetric localisation maintained by anterograde or retrograde transport by \nintraflagellar transport (IFT), respectively. Short-proximal localisation could also be explained by no \nactive transport by IFT, only diffusion up the flagellum, and high binding affinity. The latter was \npreviously proposed for ARL13B21 and we confirmed the proximal ARL13B localisation was indeed \nnot linked with DC asymmetry, but its axonemal localisation was ODA-dependent. Contrastingly, \ndistal PDEB localisation was not DC-dependent and not ODA-dependent. \nThere are, however, other potential mechanisms for generating asymmetry. One is tubulin post-\ntranslational modification, which could define specialised microtubule regions to which proteins could \nbind. Detyrosination is a well-characterised example in trypanosomatids58–61 and many other \nmodifications could be infolved62. Alternatively, some axonemal regions may be defined by other \nintraflagellar or extraflagellar structures, like the intraflagellar para-axonemal structure called the \nparaflagellar rod (PFR)63 or the lateral attachment of the flagellum to the flagellar pocket neck by the \nflagellum attachment zone (FAZ)64. The PFR-free and FAZ structures broadly line up with the short-\nproximal region, but dependence of PFR or FAZ positioning on axoneme asymmetries is unlikely as \nwe saw no obvious change to the PFR or FAZ by electron microscopy in the SPA1 and SPA2 deletion \nmutants. \nHaving mapped the conserved asymmetries in L. mexicana flagella, the critical question is the \nfunction of these asymmetrically positioned proteins. We showed that many asymmetrically positioned \nproteins are necessary for normal beating. However, we did not perturb asymmetry, thus do not have \nevidence that the asymmetric positioning is also necessary for normal beating. In general, deletion \nmutants had complex changes to the preferred mode of flagellum movement but, by focusing on only \nthe tip-to-base waveform, we carried out systematic and comparable analysis of all deletion cell lines. \nThis showed that deletion of dDC1/2 is an outlier phenotype. Total removal of distal ODAs by dDC1/2 \ndeletion were the only mutants with a strong preference for a base-to-tip beat and, furthermore, on the \nrare occasion when these cells manage a tip-to-base beat, it was at low frequency. Other deletion \nmutants of asymmetrically positioned proteins either had little change to beat frequency or gave a \nsignificant population with increased tip-to-base beat frequency. Therefore, distal ODAs are needed \nfor efficient distal waveform initiation, and asymmetrically positioned flagellar proteins are often \nnegative regulators of beat frequency. \nInterestingly PDEB deletion was previously shown to have no effect on T. brucei swimming speed24 \nand we saw no statistically significant effect of PDEB deletion on swimming speed in Leishmania, \nalthough PDEB2 deletion gave a small speed increase and PDEA deletion was previously shown to \nincrease Leishmania swimming speed. While the effect on swimming speed may be marginal, \nvariable9, or contingent on deleting both PDEBs, beat frequency was clearly and strongly affected by \ndeletion of either PDEB. The T. brucei PDEB deletion mutant has defects in chemotaxis-like \nbehaviour termed “social motility” and inability to progress through the tse-tse fly vector23,24. This likely \nis a result of changes to beat frequency control, and emphasises the importance of cAMP signalling to \nregulate flagella. We suspect the tendency for bimodal beat frequency, some increased and some \ndecreased, is the result of an increase in beat frequency to the point that it cannot be stably \nmaintained. \nRoles in regulation of tip-to-base waveform frequency control are intuitive for distal-specific proteins, \nlike PDEB. But deletion of the either FLAM6 or pDAP1, both pDC-associated, caused a significant \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n10 of 43 \n \nincrease in the tip-to-base beat frequency. How can removal of protein present only in a short \nproximal region of the flagellum increase beat frequency by promoting initiation of waveforms at the \ndistal end of the flagellum?. We can consider a two key hypotheses: Their deletion may lead to a \nconstant signal that increases the rate of tip-to-base waveform initiation thus increasesing frequency, \nor their deletion may allow faster transmission of individual signal that initiates individual tip-to-base \nwaveforms in more rapid succession. The former seems unlikely – this requires proximal proteins to \nact as sinks for a molecule that acts as a constant negative signal. FLAM6 and pDAP1 are predicted \nto have cAMP binding domains, and cAMP can be a positive regulator of beating26–28. However, the \npredicted binding rather than enzymatic degradation represents a limited sequestration rather than a \nsink, making their deletion unlikely to have a large effect on flagellar cAMP concentration. The latter is \npossible, but only through limited mechanisms. Signal from the flagellar base would have to reach the \ntip in ~25 ms to initiate successive tip-to-base waveforms at 40 Hz (a speed of ~1,000 μm/s). \nDiffusing or actively transported (cf. IFT train speed in trypanosomatids65) signalling molecules are far \ntoo slow; however, a mechanical signal like shear force on the outer doublet microtubules would be \ntransmitted almost instantaneously. We suggest that these proximal-specific proteins act to modulate \nshear force generation as waveforms approach the base of the flagellum. \nOverall, we have shown that there are multiple mechanisms for generating proximal-distal asymmetry \nin axoneme organisation, often involving the ODAs. Surprisingly, proximal-specific proteins can be \nnecessary for normal beat frequency for flagellar waveforms starting at the distal end of the flagellum. \nMethods \nParasite cell lines, maintenance and culturing \nCas9T7 L. mexicana derived from WHO strain MNYC/BZ/62/M379, expressing Cas9 and T7 RNA \npolymerase8 were grown in M199 (Life Technologies) supplemented with 2.2 g/L NaHCO3, 0.005% \nhemin, 40 mM HEPES HCl (pH 7.4) and 10% FCS. L. mexicana cultures were grown at 28°C. Culture \ndensity was maintained between 1 × 105 and 1 × 107 cells/mL for continued exponential population \ngrowth. Culture density was measured using a haemocytometer. Identity was confirmed by recent \nmRNA and genomic sequencing, and lack of mycoplasma contamination was confirmed by \nfluorescent microscopy with a Hoechst 33342 DNA stain. \nProtein sequence analysis \nTagging was carried out in L. mexicana and we considered genes for tagging if they had a syntenic \northolog in T. brucei. Ortholog proteins were identified by reciprocal best protein sequence search hits \ncarried out using  The BLAST Sequence Analysis Tool66.Genes were selected for tagging at N or C \nterminal tagging using TrypTag PCF protein localisation data available up to 12th March 201867 and \nTriTrypDB version 36. AlphaFold protein structure from genome in T. brucei and L. mexicana \nsequencing are done by the methods published by RJ Wheeler 57.  \nDomain identification, ortholog identification, structure prediction, etc. Cite tritrypdb, cite my alphafold. \nGenetic modifications  \nConstructs and sgRNA templates for endogenous mNG-tagging templates were generated by PCR \nas previously described8 and were transfected as previously described68. The pLrPOT series of \nvectors was used as PCR templates for generating tagging constructs1, specifically pLrPOT mNG \nNeo. Constructs and sgRNA templates for ORF deletion were generated by PCR and transfected as \npreviously described, using pT Blast, pT Puro and pT Neo as templates8. Primers were designed \nusing LeishGEdit (www.leishgedit.net/)8. Transfectants were selected with the necessary combination \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n11 of 43 \n \nof 20 μg/mL puromycin dihydrochloride, 5 μg/mL blasticidin S hydrochloride and 40 μg/mL G418 \ndisulfate. \nDiagnostic PCR for gene knockout validation. \nTo verify loss of the target ORF in drug-resistant transfectants, a diagnostic PCR was performed by \namplifying a short PCR product (100–300 bp) within the ORF of the target gene. We used a positive \ncontrol of with genomic DNA from the parental cell line to confirm successful detection of the target \nORF. Primers to amplify a short fragment of the PF16 (LmxM.20.1400) ORF was amplified as a \ntechnical control to confirm presence of genomic DNA from the knockout cell line. The PCR mixture \nfor each reaction was: ≤100 ng of gDNA (as required) and 10 μM (1.25μL) each of the forward and \nreverse primers mixed with the FastGene Optima HotStart Ready Mix with dye (12.5 μL) (Nippon \nGenetics, [P8-0082]) up to 25 μL of PCR-grade water. The thermocycle: Step Initial denaturation at \n95°C for 3 mins, then 30 cycles of 95°C for 15 seconds, 58°C for 15 seconds, 72°C for 30 seconds, \nthen final elongation at 72°C for 1 min. \nMicroscopy \nL. mexicana expressing fluorescent fusion proteins were imaged live. Cells were washed three times \nby centrifugation at 800 g followed by resuspension in PBS. DNA was stained by including 10 μg/mL \nHoechst 33342 in the second washing. Washed cells were settled on glass slides and were observed \nimmediately. Widefield epifluorescence and phase-contrast images were captured using a Zeiss \nAxioimager.Z2 microscope with a 63×/1.40 numerical aperture (NA) oil immersion objective and a \nHamamatsu ORCA-Flash4.0 camera. Cell morphology measurements were made in ImageJ 69. \nMotility Assays \nFor motility analysis in L. mexicana, swimming behaviours are analysed for cells in the exponential \ngrowth phase in normal culture medium essentially as previously described (Wheeler, 2017). For cell \nswimming analysis, a 25.6 s video at five frames/s under darkfield illumination was captured from 5 μL \nof cell culture in a 250 μm deep chamber using a Zeiss Axioimager.Z2 microscope with a 10×/0.3 NA \nobjective and a Hamamatsu ORCA-Flash4.0 camera. Particle tracks were traced automatically, and \nmean cell speed, mean cell velocity and cell directionality (the ratio of velocity to speed) were \ncalculated as previously described 9. \nFlagellum beat type analysis \nTo determine the proportion of cells in a population undergoing different beat types, 1 mL of \nexponential growth cells (between 1 × 106 and 1 × 107 cells/mL) was centrifuged for 5 min at 800 g. \nBetween 700 and 950 μL (depending cell density) of supernatant was removed and cells were \nresuspended in M199 (300 to 500 μL). 1uL of 5 μm polysyrene beads diluted 1:100 in M199 (Sigma \n79633) was added, which ensure a 5 μm sample depth. 1 μL of cell sample was added to the center \nof a 2 by 5 cm area marked with a hydrophobic pen on a slide, and a glass coverslip (1.0 thickness) \nadded. Videomicrographs of swimming cells under phase contrast illumination were captured with an \nAndor Neo 5.5 camera at 200 frames/s for 0.5 sec, using a x20 NA 0.3 objective lens on a Zeiss \nAxioimager.Z2 inverted microscope. Cells with one flagellum (non-dividing) were manually classified \ninto symmetrical tip-to-base (continuous or interrupted), asymmetric base-to-tip, wave type switch and \nstatic and uncoordinated. \nFlagellar beating analysis \nParental cell line and deletion cell lines were analysed by high-speed video microscopy. A 5 s video at \n200 frames/s under phase-contrast illumination was captured from a thin film of cell culture between a \nslide and coverslip using ZeissAxioimager.Z2 microscope with a 100x/1.4 NA objective and an Andor \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n12 of 43 \n \nNeo 5.5 camera. Flagellar beat behaviours for each cell lines were classified manually and only \nsymmetrical tip-to-base waveforms were analysed for this study. Automated image analysis and \nflagellum tracking70 in ImageJ [version 1.52a] was used to digitise the flagellar waveforms of a target \n25 cells (all at least 19), with a target of 950 (all at least 30 frames) per cell, manually excluding cells \nwhich swam out of focus or out of the frame during video capture. Digitised waveforms were screened \nbased on the variation in measured flagellum length over each video (a proxy for consistency of \ndigitisation), then smoothed in space and time with smoothing splines in MATLAB. Finally, waveforms \nwere excluded if they were poorly approximated by a sinusoidal beat (wild-type L. mexicana has a \nsinusoidal tip-to-base beat), as measured by a least-squares fit. A range of beating characteristics \nwere computed for each cell and all deletion cell lines were compared to the parental cell line. Full \ncode for the analysis pipeline (including thresholds for exclusion) is available on request. \n \nTransmission electron microscopy \nFor transmission electron microscopy, L. mexicana were fixed directly in medium for 10 min at room \ntemperature in 2.5% glutaraldehyde (glutaraldehyde 25% stock solution, EM grade, Electron \nMicroscopy Sciences). Centrifugation was carried out at room temperature for 5 min at 16,000 g. The \nsupernatant was discarded and the pellet was fixed in 2.5% glutaraldehyde and 4% PFA (16% stock \nsolution, EM grade, Electron Microscopy Sciences in 0.1 M PIPES (pH 7.2)) for minimum 2h. Cells \nwere embedded in 3% agarose and contrasted with OsO4 (1%) (osmium tetroxide 4% aqueous \nsolution, Taab Laboratories Equipment) during 2 hours at 4°C. Cells are stained with 2% of Uranyl \nAcetate for 2h 4°C. After serial dehydration with ethanol solutions, samples were embedded in low-\nviscosity resin Agar 100 (Agar Scientific, UK) and left to polymerise at 60°C for 24 h. Ultrathin \nsections (90 nm thick) were collected on nickel grids using a Leica EM UC7 ultra microtome and \nstained with uranyl acetate (1%, w/v) (uranyl acetate dihydrate, Electron Microscopy Sciences) and \nReynolds lead citrate71 (Lead nitrate (Thermofisher, L/1450), sodium citrate (Sigma, 71405) and \nsodium hydroxide (SLS, CHE3422). Observations were made on a Thermo Fisher Scientific Tecnai12 \nor JEOL 2100 Plus 200kV transmission electron microscope with a Gatan OneView camera. \nNinefold rotational averaging of L. mexicana axonemes \nFor generation of averaged axoneme views, axoneme images were first perspective corrected to \nensure circularity, followed by nine-fold rotational averaging as previously described72. 25 rotationally \naveraged axonemes were then aligned and averaged. Difference maps were generated by \ncomparison with the average 25 rotationally averaged parental cell line, and per-pixel statistical \nsignificance of electron density changes calculated by Mann Whitney U test (with multiple-comparison \ncorrection for the number of pixels within the axoneme cross-section). \nAcknowledgments \nWe thank Dr Errin Johnson and Dr Charlotte Melia for technical assistance for transmission electron \nmicroscopy (Electron microscopy facility at Sir William Dunn School of Pathology, Oxford University, \nUnited Kingdom). This work was supported by a Wellcome Trust Sir Henry Dale Fellowship \n[211075/Z/18/Z] awarded to RJW. BJW is supported by the Royal Commission for the Exhibition of \n1851. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n13 of 43 \n \nReferences \n1. Edwards, B. F. L. et al. Direction of flagellum beat propagation is controlled by proximal/distal \nouter dynein arm asymmetry. Proc Natl Acad Sci U S A 115, E7341–E7350 (2018). \n2. Gadelha, C., Wickstead, B. & Gull, K. Flagellar and ciliary beating in trypanosome motility. Cell \nMotility 64, 629–643 (2007). \n3. Shiba, K., Shibata, D. & Inaba, K. Autonomous changes in the swimming direction of sperm in the \ngastropod Strombus luhuanus. Journal of Experimental Biology jeb.095398 (2013) \ndoi:10.1242/jeb.095398. \n4. Yang, Y. & Lu, X. Drosophila Sperm Motility in the Reproductive Tract. Biol Reprod 84, 1005–\n1015 (2011). \n5. Baron, D. M., Kabututu, Z. P. & Hill, K. L. Stuck in reverse: loss of LC1 in Trypanosoma brucei \ndisrupts outer dynein arms and leads to reverse flagellar beat and backward movement. Journal \nof Cell Science 120, 1513–1520 (2007). \n6. Wheeler, R. J. Use of chiral cell shape to ensure highly directional swimming in trypanosomes. \nPLOS Computational Biology 13, e1005353 (2017). \n7. Santi-Rocca, J. et al. Imaging intraflagellar transport in trypanosomes. Methods Cell Biol 127, \n487–508 (2015). \n8. Beneke, T. et al. A CRISPR Cas9 high-throughput genome editing toolkit for kinetoplastids. R \nSoc Open Sci 4, 170095 (2017). \n9. Beneke, T. et al. Genetic dissection of a Leishmania flagellar proteome demonstrates \nrequirement for directional motility in sand fly infections. PLoS Pathog 15, e1007828 (2019). \n10. Rotureau, B., Ooi, C.-P., Huet, D., Perrot, S. & Bastin, P. Forward motility is essential for \ntrypanosome infection in the tsetse fly. Cell Microbiol 16, 425–433 (2014). \n11. Brokaw, C. J. & Kamiya, R. Bending patterns of Chlamydomonas flagella: IV. Mutants with \ndefects in inner and outer dynein arms indicate differences in dynein arm function. Cell Motil. \nCytoskeleton 8, 68–75 (1987). \n12. Bui, K. H., Sakakibara, H., Movassagh, T., Oiwa, K. & Ishikawa, T. Molecular architecture of inner \ndynein arms in situ in Chlamydomonas reinhardtii flagella. J Cell Biol 183, 923–932 (2008). \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n14 of 43 \n \n13. Takada, S., Wilkerson, C. G., Wakabayashi, K., Kamiya, R. & Witman, G. B. The Outer Dynein \nArm-Docking Complex: Composition and Characterization of a Subunit (Oda1) Necessary for \nOuter Arm Assembly. Mol Biol Cell 13, 1015–1029 (2002). \n14. Caspary, T., Larkins, C. E. & Anderson, K. V. The Graded Response to Sonic Hedgehog \nDepends on Cilia Architecture. Developmental Cell 12, 767–778 (2007). \n15. Cevik, S. et al. Joubert syndrome Arl13b functions at ciliary membranes and stabilizes protein \ntransport in Caenorhabditis elegans. J Cell Biol 188, 953–969 (2010). \n16. Duldulao, N. A., Lee, S. & Sun, Z. Cilia localization is essential for in vivo functions of the Joubert \nsyndrome protein Arl13b/Scorpion. Development 136, 4033–4042 (2009). \n17. Li, Y., Wei, Q., Zhang, Y., Ling, K. & Hu, J. The small GTPases ARL-13 and ARL-3 coordinate \nintraflagellar transport and ciliogenesis. J Cell Biol 189, 1039–1051 (2010). \n18. Lu, H. et al. A function for the Joubert syndrome protein Arl13b in ciliary membrane extension and \nciliary length regulation. Dev Biol 397, 225–236 (2015). \n19. Ishikawa, H. & Marshall, W. F. Ciliogenesis: building the cell’s antenna. Nat Rev Mol Cell Biol 12, \n222–234 (2011). \n20. Cantagrel, V. et al. Mutations in the cilia gene ARL13B lead to the classical form of Joubert \nsyndrome. Am J Hum Genet 83, 170–179 (2008). \n21. Zhang, Y. et al. The unusual flagellar-targeting mechanism and functions of the trypanosome \northolog of the ciliary GTPase Arl13b. J Cell Sci 131, jcs219071 (2018). \n22. Cevik, S. et al. Active Transport and Diffusion Barriers Restrict Joubert Syndrome-Associated \nARL13B/ARL-13 to an Inv-like Ciliary Membrane Subdomain. PLoS Genet 9, e1003977 (2013). \n23. Oberholzer, M. et al. The Trypanosoma brucei cAMP phosphodiesterases TbrPDEBl and \nTbrPDEB2: flagellar enzymes that are essential for parasite virulence. The FASEB Journal 21, \n720–731 (2007). \n24. Oberholzer, M., Saada, E. A. & Hill, K. L. Cyclic AMP Regulates Social Behavior in African \nTrypanosomes. mBio 6, e01954-14 (2015). \n25. Nlend, M.-C. et al. Calcium-mediated, purinergic stimulation and polarized localization of calcium-\nsensitive adenylyl cyclase isoforms in human airway epithelia. FEBS Lett 581, 3241–3246 (2007). \n26. Balbach, M., Beckert, V., Hansen, J. N. & Wachten, D. Shedding light on the role of cAMP in \nmammalian sperm physiology. Mol Cell Endocrinol 468, 111–120 (2018). \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n15 of 43 \n \n27. Aoki, F., Sakai, S. & Kohmoto, K. Regulation of flagellar bending by cAMP and Ca2+ in hamster \nsperm. Mol Reprod Dev 53, 77–83 (1999). \n28. Beltrán, C., Zapata, O. & Darszon, A. Membrane Potential Regulates Sea Urchin Sperm \nAdenylylcyclase. Biochemistry 35, 7591–7598 (1996). \n29. Hyams, J. S. & Borisy, G. G. Isolated flagellar apparatus of Chlamydomonas: characterization of \nforward swimming and alteration of waveform and reversal of motion by calcium ions in vitro. \nJournal of Cell Science 33, 235–253 (1978). \n30. Differential regulation of Paramecium ciliary motility by cAMP and cGMP. J Cell Biol 106, 1615–\n1623 (1988). \n31. Mukhopadhyay, A. G. & Dey, C. S. Role of calmodulin and calcineurin in regulating flagellar \nmotility and wave polarity in Leishmania. Parasitol Res 116, 3221–3228 (2017). \n32. Pivato, M. & Ballottari, M. Chlamydomonas reinhardtii cellular compartments and their \ncontribution to intracellular calcium signalling. J Exp Bot 72, 5312–5335 (2021). \n33. Harz, H. & Hegemann, P. Rhodopsin-regulated calcium currents in Chlamydomonas. Nature 351, \n489–491 (1991). \n34. Pazour, G. J., Sineshchekov, O. A. & Witman, G. B. Mutational analysis of the phototransduction \npathway of Chlamydomonas reinhardtii. J Cell Biol 131, 427–440 (1995). \n35. Mukhopadhyay, A. G. & Dey, C. S. Effect of inhibition of axonemal dynein ATPases on the \nregulation of flagellar and ciliary waveforms in Leishmania parasites. Mol Biochem Parasitol 225, \n27–37 (2018). \n36. Subota, I. et al. Proteomic Analysis of Intact Flagella of Procyclic Trypanosoma brucei Cells \nIdentifies Novel Flagellar Proteins with Unique Sub-localization and Dynamics. Mol Cell \nProteomics 13, 1769–1786 (2014). \n37. Shaw, S. et al. Flagellar cAMP signaling controls trypanosome progression through host tissues. \nNat Commun 10, 803 (2019). \n38. Billington, K. et al. Genome-wide subcellular protein map for the flagellate parasite Trypanosoma \nbrucei. Nat Microbiol 8, 533–547 (2023). \n39. Jackson, A. P. Evolutionary consequences of a large duplication event in Trypanosoma brucei: \nChromosomes 4 and 8 are partial duplicons. BMC Genomics 8, 432 (2007). \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n16 of 43 \n \n40. Song, P., Dudinsky, L., Fogerty, J., Gaivin, R. & Perkins, B. D. Arl13b Interacts With Vangl2 to \nRegulate Cilia and Photoreceptor Outer Segment Length in Zebrafish. Invest Ophthalmol Vis Sci \n57, 4517–4526 (2016). \n41. Parisi, M. A. Clinical and molecular features of Joubert syndrome and related disorders. American \nJournal of Medical Genetics Part C: Seminars in Medical Genetics 151C, 326–340 (2009). \n42. Alkanderi, S. et al. ARL3 Mutations Cause Joubert Syndrome by Disrupting Ciliary Protein \nComposition. The American Journal of Human Genetics 103, 612–620 (2018). \n43. Gadelha, C., Wickstead, B., McKean, P. G. & Gull, K. Basal body and flagellum mutants reveal a \nrotational constraint of the central pair microtubules in the axonemes of trypanosomes. J Cell Sci \n119, 2405–2413 (2006). \n44. Gluenz, E., Wheeler, R. J., Hughes, L. & Vaughan, S. Scanning and three-dimensional electron \nmicroscopy methods for the study of Trypanosoma brucei and Leishmania mexicana flagella. \nMethods Cell Biol 127, 509–542 (2015). \n45. Chiurillo, M. A., Ahmed, M., González, C., Raja, A. & Lander, N. Gene editing of putative cAMP \nand Ca2+-regulated proteins using an efficient cloning-free CRISPR/Cas9 system in \nTrypanosoma cruzi. Journal of Eukaryotic Microbiology 70, e12999 (2023). \n46. Fowkes, M. E. & Mitchell, D. R. The Role of Preassembled Cytoplasmic Complexes in Assembly \nof Flagellar Dynein Subunits. Mol Biol Cell 9, 2337–2347 (1998). \n47. Mitchell, D. R. & Rosenbaum, J. L. A motile Chlamydomonas flagellar mutant that lacks outer \ndynein arms. Journal of Cell Biology 100, 1228–1234 (1985). \n48. Kamiya, R. Mutations at twelve independent loci result in absence of outer dynein arms in \nChylamydomonas reinhardtii. Journal of Cell Biology 107, 2253–2258 (1988). \n49. Bui, K. H., Yagi, T., Yamamoto, R., Kamiya, R. & Ishikawa, T. Polarity and asymmetry in the \narrangement of dynein and related structures in the Chlamydomonas axoneme. J Cell Biol 198, \n913–925 (2012). \n50. Fliegauf, M. et al. Mislocalization of DNAH5 and DNAH9 in respiratory cells from patients with \nprimary ciliary dyskinesia. Am J Respir Crit Care Med 171, 1343–1349 (2005). \n51. Panizzi, J. R. et al. CCDC103 mutations cause primary ciliary dyskinesia by disrupting assembly \nof ciliary dynein arms. Nat Genet 44, 714–719 (2012). \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n17 of 43 \n \n52. Hoops, H. J. & Witman, G. B. Outer doublet heterogeneity reveals structural polarity related to \nbeat direction in Chlamydomonas flagella. J Cell Biol 97, 902–908 (1983). \n53. Dean, A. B. & Mitchell, D. R. Late steps in cytoplasmic maturation of assembly-competent \naxonemal outer arm dynein in Chlamydomonas require interaction of ODA5 and ODA10 in a \ncomplex. Mol Biol Cell 26, 3596–3605 (2015). \n54. Hagen, K. D. et al. Novel structural components of the ventral disc and lateral crest in Giardia \nintestinalis. PLoS Negl Trop Dis 5, e1442 (2011). \n55. Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 \n(2021). \n56. Mirdita, M. et al. ColabFold: making protein folding accessible to all. Nat Methods 19, 679–682 \n(2022). \n57. Wheeler, R. J. A resource for improved predictions of Trypanosoma and Leishmania protein \nthree-dimensional structure. PLOS ONE 16, e0259871 (2021). \n58. Pugacheva, E. N., Jablonski, S. A., Hartman, T. R., Henske, E. P. & Golemis, E. A. HEF1-\ndependent Aurora A activation induces disassembly of the primary cilium. Cell 129, 1351–1363 \n(2007). \n59. Ran, J., Yang, Y., Li, D., Liu, M. & Zhou, J. Deacetylation of α-tubulin and cortactin is required for \nHDAC6 to trigger ciliary disassembly. Sci Rep 5, 12917 (2015). \n60. Sherwin, T., Schneider, A., Sasse, R., Seebeck, T. & Gull, K. Distinct localization and cell cycle \ndependence of COOH terminally tyrosinolated alpha-tubulin in the microtubules of Trypanosoma \nbrucei brucei. Journal of Cell Biology 104, 439–446 (1987). \n61. Sasse, R. & Gull, K. Tubulin post-translational modifications and the construction of microtubular \norganelles in Trypanosoma brucei. J Cell Sci 90 ( Pt 4), 577–589 (1988). \n62. Sloboda, R. D. Posttranslational Protein Modifications in Cilia and Flagella. in Methods in Cell \nBiology vol. 94 347–363 (Elsevier, 2009). \n63. Bastin, P., Ellis, K., Kohl, L. & Gull, K. Flagellum ontogeny in trypanosomes studied via an \ninherited and regulated RNA interference system. J Cell Sci 113 ( Pt 18), 3321–3328 (2000). \n64. Wheeler, R. J., Sunter, J. D. & Gull, K. Flagellar pocket restructuring through the Leishmania life \ncycle involves a discrete flagellum attachment zone. Journal of Cell Science jcs.183152 (2016) \ndoi:10.1242/jcs.183152. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n18 of 43 \n \n65. Kohl, L., Robinson, D. & Bastin, P. Novel roles for the flagellum in cell morphogenesis and \ncytokinesis of trypanosomes. EMBO J 22, 5336–5346 (2003). \n66. Madden, T. The BLAST Sequence Analysis Tool. in The NCBI Handbook [Internet] (National \nCenter for Biotechnology Information (US), 2003). \n67. Billington, K. et al. TrypTag: Genome-wide subcellular protein localisation in Trypanosoma brucei. \nZenodo https://doi.org/10.5281/zenodo.7719295 (2022). \n68. Dean, S. et al. A toolkit enabling efficient, scalable and reproducible gene tagging in \ntrypanosomatids. Open Biol 5, 140197 (2015). \n69. Collins, T. J. ImageJ for microscopy. Biotechniques 43, 25–30 (2007). \n70. Automated identification of flagella from videomicroscopy via the medial axis transform | Scientific \nReports. https://www.nature.com/articles/s41598-019-41459-9. \n71. Reynolds, E. S. THE USE OF LEAD CITRATE AT HIGH pH AS AN ELECTRON-OPAQUE \nSTAIN IN ELECTRON MICROSCOPY. J Cell Biol 17, 208–212 (1963). \n72. Gadelha, A. P. R., Cunha-e-Silva, N. L. & de Souza, W. Assembly of the Leishmania \namazonensis flagellum during cell differentiation. Journal of Structural Biology 184, 280–292 \n(2013). \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n19 of 43 \n \nFigures \n \nFigure 1. Comparative analysis of T. brucei proximal and distal-specific axoneme proteins in L. mexicana \nrevealed 5 localisation groups. Quantitative analysis of fluorescence signal distribution along the axoneme from \nproximal and distal specific axoneme proteins endogenously tagged with mNG at the N and/or C terminus. Each \nrow corresponds to a T. brucei protein and its L. mexicana ortholog. In the first and third columns, an example of \nT. brucei and L. mexicana cells, respectively. Phase contrast (grey), DNA (Hoechst 33342, magenta) and mNG \n(green) overlay (left) and mNG fluorescence (right) are shown. In the second and the fourth columns, graphs \nrepresenting the mNG fluorescence signal intensity along the axoneme, from the base to tip. Data points represent \nthe mean of n = 15 axonemes in 1K1N cells, normalised by maximum signal intensity per cell. From this analysis in \nT. brucei and L. mexicana, we infer 4 protein localisation groups: A,E. pDC-like. B,F. short proximal. C,G. dDC-like. \nD-H. Distally enriched proteins, only observed for PDEs. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n20 of 43 \n \n∆pDC1∆pDC2\n∆FLAM6∆LmxM.29.0240\npDC3\n∆LmxM.31.2530\npDAP1\n∆LmxM.32.0390\nSPP1\n∆LmxM.36.5300\nSPP2\n∆ARL13B\nExample Average\nParental \npDC-likeShort proximal\nZoom C\nD\nE\n∆dDC1\n∆dDC2\n∆LC4-like\n∆LmxM.30.0090\ndDC3\ndDC-like\n∆PDEB2∆PDEB1\nDistal enriched PDEs\nF\nG\nAverageExample\nn=25\nn=25\nn=25\nn=25\nn=25\nn=7\nn=25\nn=25\nn=25\nn=25\nn=25\nn=25\nn=25\nn=26\nDifference map Difference map \nn=43\nZoom \nProximal (Flagellar pocket) Distal (Free flagellum)\n100 nm\nA B\nn=55 n=65 n=52 n=54 n=72 n=57\n0.0099\nn=53 n=58 n=56\n<0.0001\n0\n5\n10\n15\n20\n25\n30\nParental\n∆pDC1\n∆pDC2\nΔFLAM6\nΔLmxM.29.0240\nΔLmxM.31.2530\nΔLmxM.32.0390\nΔLmxM.36.5300\nΔArl13B\nFlagellum length (µm)\nn=55 n=56 n=65 n=53 n=70\n0.0247\nn=51 n=51\n0\n5\n10\n15\n20\n25\n30\nParental\n∆dDC1\n∆dDC2\n∆LC4-like\n∆LmxM.30.0090\n∆PDEB2\n∆PDEB1\nFlagellum length (μm)\nChange in electron density\nnot significant\n0.00010.0001 0.05 1 0.05p =\nloss gain dDC3\npDC3\npDAP1\nSPP2\nSPP1\nn=25\nParental \n \nFigure 2. Deletion of proteins with pDC and dDC-like localisations cause minor changes in axonemal \nstructure. A-B. Box and whisker plot of flagellum length in deletion mutants. Points represent the mean; box and \nwhiskers represent the quartile ranges and the 5th and 95th percentile. n indicates the number of cells. Statistically \nsignificant differences (p<0.05, two-tailed T test) are indicated. C-G. Thin section electron microscopy of axoneme \nstructure. The first column of each shows one representative axoneme cross-section in the flagellar pocket or in the \nfree flagellum; the second column shows an averaged axoneme structure, in which axoneme cross-sections have \nhad perspective deviation from circularity corrected (n indicates the number of axonemes used); the third column \nshows the result of ninefold rotational averaging and averaging across multiple axonemes; the fourth column \nshows a zoomed view of one microtubule doublet. Electron micrographs of transverse sections of axonemes in C. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n21 of 43 \n \nParental cell line and D. Proximal pDC-like, E. Short proximal and F. distal dDC-like and G. Distal enriched PDEs \nproteins mutants. There is a specific loss in outer dynein arm (ODA) structure in dDC1/dDC2 images. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n22 of 43 \n \n \n \nFigure 3. Proximal and distal-specific localisations can be either DC asymmetry dependent or independent. \nProtein localisation changes on pDC1 or dDC2 deletion. First column: micrographs of L. mexicana cell line \nexpressing tagged proximal and distal proteins. Second column: after deletion of both alleles of pDC1. Third \ncolumn: after deletion of dDC2. Phase contrast (grey), DNA (Hoeschst 33342, magenta) and mNG (green) overlay \nand mNG fluorescence are shown. Tagged proteins are grouped by localisation. A. pDC-like. B. Short proximal. C. \ndDC-like, D. Distal enriched PDEs. Cell lines where the protein fails to localise to the axoneme are outlined in \norange, and cell lines where the localization is axonemal but changed are outlined in blue. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n23 of 43 \n \nODAα mNG::LmxM.25.0980 ∆ODAβ mNG::ODAα\nA Parental Deletion mutant \nB\nPhase DNA mNG mNG\nODAβ mNG::LmxM.13.1650 ∆ODAα mNG::ODAβ\n10μm\nPhase DNA mNG mNG\nChange in electron density\nnot significant\n0.00010.0001 0.05 1 0.05p =\nloss gain\nn=25\nn=25\nn=33\n n=32\nn=32\nn=23\n 100 nm\nExample Average Zoom AverageExampleDifference map Difference map Zoom \nProximal (Flagellar pocket) Distal (Free flagellum)\n∆ODAβ∆ODAα\n∆ODAβ∆ODAα\nParental \nParental \n \nFigure 4. ODAβ deletion limits ODAα incorporation in the proximal axoneme while ODAβ requires ODAα for \naxonemal incorporation. A. First column: micrographs of L. mexicana cell line expressing ODAα or ODAβ tagged \nwith mNG at the C terminus. Second column: before and after deletion of both alleles of ODAα or ODAβ. Phase \ncontrast (grey), DNA (Hoeschst 33342, magenta) and mNG (green) overlay and mNG fluorescence are shown for. \nB. Ultrastructure changes upon ODA beta and ODA alpha deletion. First and second columns: one representative \naxoneme cross-section in the flagellar pocket and in the free flagella. Third and fourth columns: an averaged \naxoneme structure. (n indicates the number of axonemes used). Fifth and sixth columns: an electron density \ndifference map, resulting from subtraction of deletion mutant average axoneme image from the parental cell line. \nYellow indicates a loss of electron density in the deletion mutant. Cell lines where the protein fails to localise to the \naxoneme are outlined in orange, and cell lines where the localization is axonemal but changed are outlined in blue. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n24 of 43 \n \n \n \nFigure 5. Some non-DC dependent localisations are ODA dynein heavy chain dependent and some DC \ndependent localisations are ODA heavy chain independent. A. Micrographs of L. mexicana cell line expressing \nproximal pDC-like and short proximal proteins and B. Distal dDC-like proteins tagged with mNG at the C terminus, \nbefore and after deletion of both alleles of ODAβ. Cell lines where the protein fails to localise to the axoneme are \noutlined in orange, and cell lines where the localization is axonemal but changed are outlined in blue. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n25 of 43 \n \n \n \nFigure 6. Localisation co-dependency among DCs and proteins with a DC-like localisation identifies an \nadditional pDC component. A. Tagging of pDC1 and deletions of pDC-like proteins localisation. In the first \ncolumn, micrographs of L. mexicana cell line expressing proteins tagged with mNG at the C terminus. In columns 2 \nto 4, micrographs after deletion of FLAM6, LmxM.29.0240 or LmxM.31.2530. B. Micrographs of L. mexicana cell \nline expressing dDC2 protein tagged with mNG at the C terminus, before and after deletion of both alleles of LC4-\nlike and LmxM.30.0090. C. Combinatorial tagging and deletion of pDC1 or proteins with a pDC-like localisation. In \nthe first column, micrographs of L. mexicana cell line expressing proteins tagged with mNG at the C terminus. In \ncolumns 2 to 4, micrographs after deletion of FLAM6, LmxM.29.0240 or LmxM.31.2530. Cell lines where the \nprotein fails to localise to the axoneme are outlined in orange, and cell lines where the localization is axonemal but \nchanged are outlined in blue. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n26 of 43 \n \n \nFigure 7. Distal and proximal proteins contribute to the control of flagellum beat and their frequency. \nA,D. Graphic representing the normalized swimming speed in parental cell line and all knockout mutants. Error \nbars represent the standard deviation of three replicates. p-values from Student’s t-test compared to the parental \ncell line. B,E. The proportion of cells undergoing different types of flagellar movement, comparing deletion mutants \nto the parental C9T7 cell line. C,F. Amplitude, dominant frequency and wavelength per flagellum in parental cell \nline and knockout distal mutants. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n27 of 43 \n \nA\nTb927.5.1900\nLmxM.15.0540\nTb927.7.5660\nLmxM.06.1040\nTb927.6.1660\nLmM.29.0240\nTb927.11.16090\nLmxM.31.2900\nTb927.8.8000/\nTb927.4.4370\nLmxM.30.0090\nTb927.8.4400\nLmxM.10.0960\npDC1\ndDC1\npDC2\ndDC2\npDC3\ndDC3\nTb927.9.4420\nLmxM.01.0620\nTb927.3.5020\nLmxM.08_29.1040\nTb927.11.15730\nLmxM.31.2530\nFLAM6\npDAP1\nLC4-like\nTb927.9.5040\nLmxM.15.1481\nTb927.9.5100\nLmxM.15.1480\nTb927.10.5230\nLmxM.36.0820\nTb927.10.11010\nLmxM.32.0390\nTb927.11.10610\nLmxM.36.5300\nSPA1\nSPA2\nARL13B\nPDEB1\nPDEB2\nDocking complex Docking complex asymmetry-dependent\nNon-docking complex asymmetry-dependent\nB\npDC pDC\nPredicted secondary structure\n α helix\n β sheet\nAlphafold2 predicted local distance\ndifference test (pLDDT) score\n 50 > pLDDT\n 50 > pLDDT > 70\n 70 > pLDDT > 90\n pLDDT > 90\n100 aa\nSSF51206\nSSF51206\nSSF47473\nSSF56112\nSSF52540SSF47391\nSSF55781 SSF55781\nSSF55781 SSF55781\nODAα and ODAβ\n \nFigure 8. Summary of conserved proximal-distal asymmetry in the Leishmania and \nTrypanosoma flagellum. A. Cartoon summary of asymmetrically localised axonemal proteins, \nwhere proteins drawn overlapping indicate broadly summarises dependency for assembly.  B. \nSummary of Alphafold2-predicted protein structure and predicted protein domains for the \nasymmetrically localised proteins in A.\n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n28 of 43 \n \nSupplemental Figures and Tables \n \nTable S1. All flagellum proximal and distal proteins in T. brucei whose L. mexicana ortholog has \ncomparable asymmetric distribution. \n  \ngene ID in T. brucei PFam ID PFam Description gene ID in                     \nL. mexicana\nTb927.8.4400 N/A N/A N/A N/A LmxM.10.0960 86699 806 7.57\nTb927.7.5660 N/A N/A N/A N/A LmxM.06.1040 79107 722 5.39\nTb927.3.5020 SSF51206 Cyclic nucleotide-binding-like Cyclic nucleotide-binding domain PF00027 LmxM.08_29.1040 203554 1898 8.87\nTb927.6.1660 N/A N/A N/A N/A LmxM.29.0240 65899 583 6.46\nTb927.11.15730 SSF51206 Cyclic nucleotide-binding-like N/A N/A LmxM.31.2530 102051 958 8.49\nTb927.10.11010 N/A N/A N/A N/A LmxM.32.0390 40816 369 9.82\nTb927.11.10610 SSF56112 Protein kinase-like domain \nsuperfamily N/A N/A LmxM.36.5300 32959 301 5.84\nTb927.10.5230 SSF47391; \nSSF52540\nP-loop containing nucleoside \ntriphosphate hydrolase Small GTPase superfamily, ARF/SAR type PF00025 LmxM.36.0820 119520 1110 4.65\nTb927.5.1900 N/A N/A N/A N/A LmxM.15.0540 75593 678 8.15\nTb927.11.16090 N/A N/A N/A N/A LmxM.31.2900 70535 618 4.76\nTb927.9.4420 SSF47473 EF-hand domain pair N/A N/A LmxM.01.0620 59566 566 9.51\nTb927.8.8000 N/A N/A N/A N/A\nTb927.4.4370 N/A N/A N/A N/A\nSSF109604; \nSSF55781 N/A 3'5'-cyclic nucleotide phosphodiesterase, \ncatalytic domain;GAF domain PF00233;PF01590 LmxM.15.1480 102700 930 5.53\nSSF109604; \nSSF55781 N/A 3'5'-cyclic nucleotide phosphodiesterase, \ncatalytic domain;GAF domain PF00233;PF01590 LmxM.15.1481 103765 940 4.92\nTb927.9.5040 SSF109604; \nSSF55781 N/A 3'5'-cyclic nucleotide phosphodiesterase, \ncatalytic domain;GAF domain PF00233;PF01590 LmxM.15.1480 102700 930 5.53\nIdentified groups in function of their \nrespective fluorescent localisation\nMolecular \nWeight\nProtein \nLength\nIsoelectric \nPointSuperfamily DescriptionSuperfamily ID\n6.15\nShort proximal\npDC-like\nLmxM.30.0090\nTb927.9.5100\ndDC-like\ndistal enriched PDEs\n49256116\n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n29 of 43 \n \n \n \nTable S2. All flagellum proximal and distal axoneme-specific proteins in T. brucei whose L. mexicana \northolog was not asymmetrically distributed within the flagellum. \n  \nFluorescent localisation                                                \nin T. brucei gene ID in T. brucei Superfamily ID Superfamily Description PFam ID PFam Description gene ID in L. \nmexicana\nFluorescent \nlocalisation in  \nL. mexicana\nMolecular \nWeigtht\nProtein \nLength\nIsoelectri\nc Point\nTb927.5.2820 SSF51316; \nSSF56112\nMss4-like superfamily;Protein \nkinase-like domain superfamily PF00069 Protein kinase domain LmxM.08.0930 22649 210 5.55\nTb927.7.4600 SSF52540 P-loop containing nucleoside N/A N/A LmxM.14.0100 116353 1087 9.01\nTb927.11.1230 N/A N/A N/A N/A LmxM.27.0610 38217 366 8.56\nTb927.11.10360 N/A N/A N/A N/A LmxM.36.4090 46175 432 7.8\nTb927.10.8800 N/A N/A N/A N/A LmxM.36.6010 73032 670 7\nTb927.10.6570 N/A N/A PF13181;PF13414 Tetratricopeptide repeat LmxM.36.2070 46175 432 7.8\nTb927.11.3920 SSF47576 CH domain superfamily N/A N/A LmxM.13.0910 116434 1029 6.92\nTb927.3.4270 SSF82185 N/A PF02493 MORN motif LmxM.08_29.1700 99497 916 7.28\nTb927.5.4470 SSF55073 Nucleotide cyclase N/A N/A LmxM.05.0050 94987 841 6.38\nTb927.6.2220 N/A N/A N/A N/A LmxM.29.0770\nLysosome, \nlysosome \nassociated \nmicrotubule\n14871 129 5.61\nAxonemal \nEnriched proteins\n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n30 of 43 \n \n \n \n \nTable S3. All flagellum proximal and distal proteins in T. brucei which lack an L. mexicana \northolog. \n  \nFluorescent localisation in                                             \nT. brucei gene ID in T. brucei Superfamily ID Superfamily Description PFam ID PFam Description Molecular Weigtht Protein Length Isoelectric \nPoint\nshort proximal Tb927.1.4340 N/A N/A N/A N/A 77650 716 9.57\nshort proximal Tb927.10.7770 N/A N/A N/A N/A 25760 230 10.63\nshort proximal Tb927.10.4720 SSF48403 Ankyrin repeat-containing domain \nsuperfamily PF13857 N/A 39545 360 4.85\nshort proximal Tb927.11.15610 SSF48452 Tetratricopeptide-like helical \ndomain superfamily N/A N/A 109790 997 9.24\nshort proximal Tb927.11.5770 SSF49265 Fibronectin type III superfamily N/A N/A 222130 2030 6.1\nshort proximal Tb927.11.9650 N/A N/A N/A N/A 53172 484 7.32\npDC-like Tb927.3.5260 N/A N/A N/A N/A 50908 469 10.54\npDC-like Tb927.5.1950 N/A N/A N/A N/A 66973 611 8.82\npDC-like Tb927.6.3410 N/A N/A N/A N/A 34704 311 8.38\npDC-like Tb927.11.15740 N/A N/A N/A N/A 60161 542 11.36\npDC-like Tb927.3.1200 SSF47473 EF-hand domain pair N/A N/A 28693 263 9.85\nlong proximal Tb927.11.5790 N/A N/A N/A N/A 30645 273 10\nlong proximal Tb927.8.1280 N/A N/A N/A N/A 153905 1393 9.01\nlong proximal Tb927.9.2075 N/A N/A N/A N/A 302504 2843 4.2\nlong proximal Tb927.9.7100 N/A N/A N/A N/A 25626 246 6.13\nlong proximal Tb927.9.9300 N/A N/A N/A N/A 54512 482 9.69\nlong proximal Tb927.6.2120 N/A N/A N/A N/A 27386 250 10.45\nlong proximal Tb927.8.5300 N/A N/A N/A N/A 63847 562 10.59\nlong distal Tb927.10.3870\nSSF101908; \nSSF47473; \nSSF50978\nEF-hand domain pair;WD40-repeat-\ncontaining domain superfamily PF00400 WD40 repeat 237134 2151 7.79\nlong distal Tb927.7.5430 N/A N/A N/A N/A 98743 906 9.22\nlong distal Tb927.8.3700 N/A N/A N/A N/A 161957 1486 8.65\nlong distal Tb927.6.410 SSF52058 N/A N/A N/A 93966 858 4.93\nlong distal Tb927.7.5430 N/A N/A N/A N/A 98743 906 9.22\nshort distal Tb927.6.1100 N/A N/A N/A N/A 53686 492 7.6\nshort distal Tb927.6.3020 N/A N/A N/A N/A 32217 288 10.23\nshort distal Tb927.1.750 N/A N/A N/A N/A 132691 1215 6.86\naxonemal Tb927.10.3960 N/A N/A N/A N/A 17598 155 4.69\naxonemal Tb927.11.14970 N/A N/A N/A N/A 93163 823 5.13\nTb927.4.4400 SSF56104 N/A PF03770 Inositol polyphosphate \nkinase 82630 756 5.63\nTb927.9.13770 N/A N/A PF04683 Proteasomal ubiquitin \nreceptor Rpn13/ADRM1 31304 279 5.38\n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n31 of 43 \n \nGene \nname  \nGene ID \n(L. mexicana)  Localisation  \nElectron density \nchange upon \ndeletion \nDC-\ndependent  \nODA-\ndependent  \nFlagellum \nlength  \nSwim \nspeed  \nTip-to-\nbase beat \nincidence \nBase-to-\ntip beat \nincidence \nTip-to-\nbase beat \nfrequency \npDC1 LmxM.10.0960 pDC-like \nReduced near ODA \n(low confidence) Yes* No No change Slow Normal Normal Bimodal \npDC2 LmxM.06.1040 pDC-like \nReduced near ODA \n(low confidence) Yes* No No change Normal Less More Bimodal \nFLAM6 LmxM.08_29.1700 pDC-like Reduced near ODA Yes Yes No change Normal Normal Normal Increased \npDC3 LmxM.29.0240 pDC-like Reduced near ODA Yes No No change Normal Less More Normal \nPDAP1 LmxM.31.2530 pDC-like Reduced near ODA Yes No No change Normal Normal Normal Increased \nSPA2 LmxM.32.0390 short proximal None No No No change Slow Normal Normal Bimodal \nSPA1 LmxM.36.5300 short proximal None No Yes No change Normal Normal Normal Increased \nARL13B LmxM.36.0820 short proximal None No Yes Short N/D N/D N/D N/D \ndDC1 LmxM.15.0540 dDC-like Reduced ODA Yes No No change Very slow Less More Decreased \ndDC2 LmxM.34.2900 dDC-like Reduced ODA Yes No No change Very slow Less More Decreased \nLC4-like LmxM.01.0620 dDC-like Reduced near ODA Yes Yes No change Fast Less Normal Increased \ndDC3 LmxM.30.0090 dDC-like None Yes Yes No change Normal Less Normal Normal \nPDEB2 LmxM.15.1480 \nDistal enriched \nPDEs None  No No  No change Normal  Less  More  Normal  \nPDEB1 LMXm.15.1481 \nDistal enriched \nPDEs None  No  No  No change Normal  Less  More  Increased  \nTable S4. Qualitative summary of protein localisation and deletion mutant phenotypes. \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n32 of 43 \n \n \n \nFigure S1. Proximal or distal axoneme-specific proteins in T. brucei where the L. mexicana \northolog did not have a proximal or distal-specific localisation. \n \nLocalisation of proteins with a proximal or distal specific localisation in T. brucei but in L. mexicana. The first column \nshows widefield epifluorescence micrographs of endogenous mNG tagging at the N and C termini in T. brucei. Phase \ncontrast (grey), DNA (Hoechst 33342, magenta) and mNG (green) overlay and mNG fluorescence are shown. The \nsecond column is a graph representation of the mNG fluorescent signal intensity along the axoneme from the base \nto the tip. Data points represent the mean of n = 15 axonemes in 1K1N cells, normalised by maximum signal intensity \nper cell. The third column shows widefield epifluor escence of endogenous tagging at the C terminus of the L. \nmexicana ortholog. These were categorised as not specific to the proximal or distal axoneme. \n \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n33 of 43 \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n34 of 43 \n \nFigure S2. Proximal axoneme-specific proteins in T. brucei which lack a detectable L. mexicana \northolog. \n \nQuantitative analysis of fluorescence signal distribution along the axoneme from proximal specific axoneme proteins \nendogenously tagged with mNG at the N and/or C terminus. In the first column epifluorescence microscopy images \nof mNG fluorescence in example T. brucei cell. Phase contrast (grey), DNA (Hoechst 33342, magenta) and mNG \n(green) overlay (left) and mNG fluorescence (right) are shown. In the second column, graphs representing the mNG \nfluorescence signal intensity along the axoneme, from the base to tip. Data points represent the mean of n = 15 \naxonemes in 1K1N cells, normalised to maximum signal intensity per cell. We categorise these as: A. Short proximal \nlocalisation. B. pDC-like localisation. C. Proximal long localisation. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n35 of 43 \n \n \n \nFigure S3. Distal axoneme-specific proteins in T. brucei which lack a detectable L. mexicana \northolog. \n \nIn A to D, the first column shows widefield epifluorescence micrographs of mNG fluorescence at the C terminus in \nthree different groups of proximal proteins in  T. brucei, then the second column shows widefield epifluorescence \nmicrographs of mNG fluorescence at the N terminus in same different groups of proximal proteins in T. brucei. Phase \ncontrast (grey), DNA (Hoechst 33342, magenta) and mNG (green) overlay and  mNG fluorescence are shown. The \nthird column represent the measure proteins signal intensity along flagellum. Localisations are the same as with N \nterminal tagging. A. Long distal proteins in T. brucei. B. Distal short proteins in T. brucei. C. Axonemal proteins in \nT. brucei. D. Low signal to determine a specific localisation of proteins in T. brucei.  \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n36 of 43 \n \n.  \nFigure S4. Validation of proximal and distal mutants. Diagnostic PCRs to confirm deletion of both alleles \nof A. Proximal and B. Distal proteins in the clonal L. mexicana deletion cell lines. For each, gel electrophoresis of \nPCR products from genomic DNA (gDNA) are shown. Control PCR product from an unaffected open reading frame \n(ORF), PF16, are shown to confirm presence of deletion mutant gDNA and PCR products from parental gDNA are \nshown to confirm that the test primers can amplify the dedicated ORF.   \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n37 of 43 \n \n \n \nFigure S5. Validation of pDC1 and dDC2 mutants in different tagged cell lines. Diagnostic PCRs to \nconfirm deletion of both alleles of A. pDC1 and B. dDC2 proteins in the clonal L. mexicana deletion cell lines. For \neach, gel electrophoresis of PCR products from genomic DNA (gDNA) are shown. Control PCR product from an \nunaffected open reading frame (ORF), PF16, are shown to confirm presence of deletion mutant gDNA and PCR \nproducts from parental gDNA are shown to confirm that the test primers can amplify the dedicated ORF.   \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n38 of 43 \n \n \nFigure S6. Validation of ODA mutants. Diagnostic PCRs to confirm deletion of both alleles of A. ODAβ in \ntagged mNG::ODAα and ODAα in tagged mNG::ODAβ and B. ODAα and ODAβ proteins in the clonal L. mexicana \ndeletion cell lines. For each, gel electrophoresis of PCR products from genomic DNA (gDNA) are shown. Control \nPCR product from an unaffected open reading frame (ORF), PF16, are shown to confirm presence of deletion \nmutant gDNA and PCR products from parental gDNA are shown to confirm that the test primers can amplify the \ndedicated ORF.   \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n39 of 43 \n \n \nFigure S7. Validation of ODA mutants. Diagnostic PCRs to confirm deletion of both alleles of A. ODAα in \ntagged pDC-like proteins and B. ODAβ in tagged dDC-like proteins in the clonal L. mexicana deletion cell lines. For \neach, gel electrophoresis of PCR products from genomic DNA (gDNA) are shown. Control PCR product from an \nunaffected open reading frame (ORF), PF16, are shown to confirm presence of deletion mutant gDNA and PCR \nproducts from parental gDNA are shown to confirm that the test primers can amplify the dedicated ORF. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n40 of 43 \n \n \n \nFigure S8. Validation of mutants. Diagnostic PCRs to confirm deletion of both alleles of A. pDC-like proteins \nin tagged pDC1, B. dDC-like proteins in tagged dDC2 and C. pDC-like proteins in tagged FLAM6, pDC3 and \nLmxM.31.2530 in the clonal L. mexicana deletion cell lines. For each, gel electrophoresis of PCR products from \ngenomic DNA (gDNA) are shown. Control PCR product from an unaffected open reading frame (ORF), PF16, are \nshown to confirm presence of deletion mutant gDNA and PCR products from parental gDNA are shown to confirm \nthat the test primers can amplify the dedicated ORF. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n41 of 43 \n \n \n \nFigure S9. Validation of pDC1, dDC2 and ODAβ mutants in tagged cell lines. Diagnostic PCRs to \nconfirm deletion of both alleles of A. pDC1 and dDC2 in tagged of flagellum long tip proteins and B. ODAβ in \ntagged LmxM.14.1410, LmxM.27.0330 and LmxM.31.3200 in the clonal L. mexicana deletion cell lines. For each, \ngel electrophoresis of PCR products from genomic DNA (gDNA) are shown. Control PCR product from an \nunaffected open reading frame (ORF), PF16, are shown to confirm presence of deletion mutant gDNA and PCR \nproducts from parental gDNA are shown to confirm that the test primers can amplify the dedicated ORF. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n42 of 43 \n \n \nFigure S10. Additional control measures for beat waveform properties for distal and proximal \nproteins. Graphics represented the percentage of bad frames that are not traceable, the quality of fit with nice \nfrequency and the flagellum length in pixels in A. dDC-like and distal enriched PDEs and in B. pDC-like and short \nproteins. n indicates the number of analysed cells. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint \n\n43 of 43 \n \n \nFigure S11. Comparison of dominant beat frequency and amplitude measured from the proximal \nor distal region of individual flagella, for deletion mutants of proteins with pDC and dDC or distal \nenriched PDE-like localisations. Every line in each graph represent a single flagellum, with the data points \ncorresponding to beat frequency in the proximal and distal flagella A-C. For deletion mutants of proteins with pDC-\nlike localisation: ∆pDC1, ∆pDC2, ∆FLAM6 and ∆LmxM.31.2530. B-D. For deletion mutants of proteins with distal \nlocalisations: ∆LC4-like and ∆PDEB2 or ∆PDEB1. No difference between the proximal and distal flagellum were \nstatistically significant (p > 0.05, two-tailed T test) \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 10, 2024. ; https://doi.org/10.1101/2024.05.08.593170doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}