Protein subcellular relocalization and function of duplicated flagellar calcium binding protein genes in honey bee trypanosomatid parasite

The honey bee trypanosomatid parasite, Lotmaria passim, contains two genes that encode the flagellar calcium binding protein (FCaBP) through tandem duplication in its genome. FCaBPs localize in the flagellum and entire body membrane of L. passim through specific N-terminal sorting sequences. This finding suggests that this is an example of protein subcellular relocalization resulting from gene duplication, altering the intracellular localization of FCaBP. However, this phenomenon may not have occurred in Leishmania, as one or both of the duplicated genes have become pseudogenes. Multiple copies of the FCaBP gene are present in several Trypanosoma species and Leptomonas pyrrhocoris, indicating rapid evolution of this gene in trypanosomatid parasites. The N-terminal flagellar sorting sequence of L. passim FCaBP1 is in close proximity to the BBSome complex, while that of Trypanosoma brucei FCaBP does not direct GFP to the flagellum in L. passim. Deletion of the two FCaBP genes in L. passim affected growth and impaired flagellar morphogenesis and motility, but it did not impact host infection. Therefore, FCaBP represents a duplicated gene with a rapid evolutionary history that is essential for flagellar structure and function in a trypanosomatid parasite.


Introduction
Gene duplication is considered the main source of new genes [1][2][3][4].However, to maintain duplicate genes in a genome, functional divergence is usually necessary, with a few exceptions.Functional divergence can occur through neofunctionalization, where one of the duplicates develops a new function while the other retains the original function of the ancestral gene [1].Alternatively, subfunctionalization can take place, where the ancestral functions are divided between the duplicate genes.For example, their combined levels or patterns of activity could be equivalent to the original single gene [5][6][7].These types of functional divergence can modify the function of the encoded protein or the gene's expression pattern.Subcellular localization is also crucial for a protein's function within a cell.Protein subcellular relocalization (PSR) has been proposed as a mechanism for the functional divergence and retention of duplicate genes [8,9].PSR allows for the rapid subdivision of ancestral localization or acquisition of new localization following gene duplication.After neolocalization or sublocalization, the expression pattern and function of the genes can further diverge.Although this idea was questioned by comparing the frequency of PSR between singletons and duplicates [10], there are numerous examples of PSR for duplicated genes [11][12][13].
The flagellar calcium binding protein (FCaBP or Calflagin) was discovered as one of the abundant proteins in the flagellar membrane of Trypanosoma brucei and Trypanosoma cruzi.It contains four EF-hand calcium binding domains and is considered a conserved signaling protein among trypanosomatid parasites [14,15].FCaBP is acylated with myristate and palmitate at the N-terminus, which is crucial for its localization in the flagellum and association with lipid raft microdomains [16][17][18].Additionally, the N-terminal amino acid sequence of FCaBP is necessary for its targeting to the flagellum in T. cruzi [19].Although the phenotypes of knocking down FCaBPs in T. brucei have been reported [20], the physiological functions of FCaBPs are not well understood.Proteins in the flagellum (and cilium) are initially synthesized in the cell body and then transported to the flagellum through intraflagellar transport (IFT).IFT requires IFT complexes (IFT-A and IFT-B) and motor proteins, including kinesin-2 for anterograde transport and IFT-dynein for retrograde transport, along the axonemal microtubules [21][22][23].The IFT complexes act as carriers, and BBSome and tubby-like protein (TULP) function as adapters by binding to specific sets of flagellar membrane proteins [24][25][26][27][28][29].For example, BBSome is necessary for the exit of a dually acylated phospholipase D from cilia in Chlamydomonas [26].Therefore, FCaBP likely requires IFT complexes and either BBSome or TULP for its import and export to the flagellum.
Lotmaria passim is the most prevalent trypanosomatid parasite that infects honey bees worldwide [30][31][32][33].It specifically colonizes the hindgut of honey bees, affecting the host's physiology, and it may be associated with winter colony loss [34,35].L. passim is a monoxenous parasite that only infects bees and can serve as a model to study the roles of flagellar formation and function for infecting the various hosts of trypanosomatid parasites.
In this study, we identified two FCaBP genes in the genomes of L. passim and closely related species, resulting from tandem duplication.By analyzing the gene in other trypanosomatid species' genomes, the protein subcellular localizations, and N-terminal sorting sequences, we have uncovered their evolutionary history in relation to the flagellar protein transport funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.machinery.Deletion of these genes in L. passim has also revealed novel physiological functions of FCaBPs.Therefore, FCaBP serves as an example to understand how duplicated genes have diversified during the evolution of trypanosomatid parasites.

L. passim contains two FCaBPs with flagellar and entire body localizations
The genome of L. passim encodes two FCaBP genes (LpFCaBP1 and LpFCaBP2) that are separated by 1063 bp, suggesting tandem gene duplication.To determine their localizations, we expressed GFP-tagged versions of LpFCaBP1 and LpFCaBP2 (at C-terminus) in L. passim.LpFCaBP1-GFP was highly enriched in the flagellum, similar to FCaBPs in T. cruzi and T. brucei [16,17].On the other hand, LpFCaBP2-GFP was present at the flagellum and the cell body membrane (Fig 1).We also observed that GFP alone is uniformly distributed in the entire body, while the paraflagellar rod protein 5 (LpPFRP5)-GFP fusion protein is specific to the flagellum (Fig 1).These findings indicate that the two duplicated FCaBP genes encode proteins enriched in the different part of parasite body with common localization in the flagellum.

The N-terminal 16 amino acids of LpFCaBP function as either flagellum or cell body membrane sorting sequence
By aligning the amino acid sequences of LpFCaBP1 and LpFCaBP2, we found that they are almost identical except for their N-terminal ends (S1 Data).To determine the flagellar and entire body membrane sorting sequences, we swapped the N-terminal 16 or 28 amino acids between LpFCaBP1 and LpFCaBP2 in GFP fusion proteins.We also determined the ratio of mean fluorescence of flagellum to that of cell body using Image-J.There are no statistical differences between the ratios for LpFCaBP1-GFP (2.4

FCaBP genes have rapidly evolved in trypanosomatid parasites
We examined FCaBP genes in various trypanosomatid parasites' genomes.T. brucei has four FCaBP and the related genes as previously reported [20].T. cruzi appears to contain the large number of FCaBP genes due to its repetitive genome [36].Trypanosoma theileri genome has three FCaBP genes (S2 Data).These results suggest that FCaBP gene duplicated in the common ancestor of Trypanosoma and Leishmaniinae and has undergone further duplication in each Trypanosoma species.
In the genomes of various Leishmaniinae species, microsynteny around FCaBP genes is well conserved with the same orders and orientations of genes encoding three novel proteins, dynein light chain, and sucrose-phosphate synthase-like protein (Fig 3).In Leishmania infantum, Leishmania braziliensis, Leishmania mexicana, and Leishmania donovani, two FCaBP genes are present but FCaBP2 appears to have become a pseudogene by generating a stop codon in 5' end of the ORF (open reading frame).If this gene is translated, the truncated protein lacking the N-terminal sorting sequence is synthesized and is likely to be present in the cytosol rather than the membrane, as it lacks the N-terminal glycine and cysteine residues necessary for myristoylation and palmitoylation.Although the protein should be able to bind calcium, its interactions with other membrane associated proteins as a signaling factor would be lost.Thus, above four Leishmania species may contain only FCaBP1.Intriguingly, two FCaBP genes in Leishmania major have become pseudogenes by accumulating multiple internal stop codons in the ORFs (Fig 3).These characteristics are shared with three independently assembled genome sequences derived from different strains of L. major (Friedlin, SD75.1, and LV39c5), suggesting that this is not the result of genome assembly error.
In species related to L. passim (34), Crithidia fasciculata, Crithidia expoeki, and Crithidia bombi have two FCaBP genes, while Leptomonas seymouri has only FCaBP1 gene (Fig 3).At downstream of LsFCaBP1, there is 173 bp genomic sequence capable of encoding a protein with low similarity to C-terminal 56 amino acid of FCaBP (e-value: 0.013) if translated.This may represent the remnant of LsFCaBP2 pseudogene.We primarily characterized L. seymouri assembled genome sequence derived from the strain ATCC 30220 Lsey_0068 (Accession number: LJSK01000068); however, FCaBP genomic sequence is the same in another independently assembled genome sequence derived from the different strain, BHU1095 (Accession number: ANAF02000031).Thus, the loss of LsFCaBP2 is unlikely to be caused by the error of genome assembly.Leptomonas pyrrhocoris contains FCaBP1 and two FCaBP2 genes based on the comparison of N-terminal sorting sequences (S1 Data).These results demonstrate that FCaBP gene has rapidly evolved in trypanosomatid parasites.fasciculata, L. seymouri, and L. passim.Glycine and cysteine residues targeted for myristoylation and palmitoylation are well conserved; however, there are more variations in the sorting sequences of FCaBPs compared to the main sequences with EF-hand calcium binding domains.To elucidate the pivotal amino acids within the sorting sequence governing the membrane localization of LpFCaBP in the flagellum or entire body, we performed targeted mutagenesis on LpFCaBP1 and LpFCaBP2.Specifically, lysine at position 11 (K11) in LpFCaBP1 was substituted with isoleucine, and vice versa, generating LpFCaBP1N16(K11I)-GFP and LpFCaBP2N16(I11K)-GFP.Additionally, serines at positions 9 (S9) and 10 (S10) in   and 4C).Thus, it is likely that S9, S10, T9, and Q10 are required for efficient acylation at glycine (G2) and cysteine (C3) in the N-termini of FCaBP1 and FCaBP2.These lipid modifications appear to play a crucial role in determining FCaBP localization, either in the flagellum as previously reported [17,19] or in the entire body membrane.However, LpFCaBP1N16-TQ-GFP and LpFCaBP2N16-SS-GFP may fail to associate with the cell body membrane due to a mechanism other than defective acylation.

Flagellar localization of FCaBP depends on both the sorting sequence and trans-acting factors in L. passim
To test the conservation and plasticity of the flagellar and cell body membrane sorting sequences of FCaBPs, we expressed GFP fusion proteins containing N-terminal sorting sequences from various trypanosomatid parasites in L. passim.Because TcFCaBP and TbFCaBP (Calflagin, Tb24) are flagellar proteins, their N-terminal amino acids should function as flagellar sorting sequences in T. cruzi and T. brucei, respectively [17,19].Indeed, the Nterminal 24 but not 12 amino acids of the TcFCaBP were sufficient to direct GFP to the flagellum of T. cruzi [19].However, we found that GFP fused with the sorting sequence of TbFCaBP is localized in the cell body membrane of L. passim.GFP with the sorting sequence of either TcFCaBP or LdFCaBP is present in both cell body membrane and the proximal part of flagellum (Fig 4C).The sorting sequences of CfFCaBP1 and 2 directed GFP to flagellum (most) as well as cell body membrane and cell body membrane, respectively.GFP with the sorting sequence of LsFCaBP1 is primarily localized in flagellum (Fig 4C).These results may suggest that flagellar localization of FCaBP requires the recognition of N-terminal sorting sequence by IFT-associated proteins (trans-acting factors).

The N-terminal sorting sequence of LpFCaBP1 is in close proximity to the BBSome complex but not TULP
Since the import and export of flagellar membrane proteins depend on either the BBSome complex or TULP [37], we first tested the interaction between LpFCaBP1N16-GFP and Flagtagged LpBBS1 or LpTULP in the parasites expressing both proteins by immunoprecipitation.However, we were unable to detect binding with either protein.This would be due to the transient interaction between a flagellar membrane protein and BBSome or TULP as the adaptor [37].We thus used an engineered biotin ligase, ultraID [38], instead of GFP.This fusion protein allowed us to identify proteins in the close proximity through biotinylation.In L. passim expressing both LpFCaBP1N16-UltraID and Flag-tagged LpBBS1 or LpTULP, the ultraID fusion protein was detected in the proximal part of the flagellum (

LpFCaBPs are essential for flagellar morphogenesis and motility but not the host infection of L. passim
To determine the functions of LpFCaBPs, we deleted both LpFCaBP1 and LpFCaBP2 genes in L. passim by replacing them with hygromycin resistance gene by CRISPR-mediated homology-directed repair (Fig 6A).Although LpFCaBP1 and 2 primarily localize in the flagellum and entire body membrane, they could be still functionally redundant.As we previously reported, expressing Cas9 and gRNA does not modify the target gene in L. passim.Thus, the off-target effects are basically absent (39).By genomic PCR, we identified both heterozygous (+/-) and homozygous (-/-) deletion clones of LpFCaBPs (Fig 6B).To confirm the disruption of genes, we tested LpFCaBP1 and 2 mRNA expression in wild type, heterozygous and homozygous deletion parasites by RT-PCR using the gene-specific reverse primer.As shown in Fig 6C , both mRNAs are absent in the homozygous mutant (LpFCaBP1/2 -/-) parasites.
We also established LpFCaBP1/2 -/-parasites, wherein either LpFCaBP1 or LpFCaBP2 was introduced on a plasmid DNA vector containing a bleomycin resistance gene.Comparative phenotypic analyses were performed with wild-type and the parent mutant parasites at 29˚C.LpFCaBP1/2 -/-parasites exhibited reduced growth rates during the logarithmic phase, reaching stationary phase at a density comparable to wild-type parasites.The growth defect in .These results suggest that the molecular functions of LpFCaBP1 and LpFCaBP2 remain the same as calcium-binding proteins.However, in L. passim, both LpFCaBP1 concentrated in the flagellum and LpFCaBP2 present throughout the entire parasite are necessary.We then compared the infection of wildtype and LpFCaBP1/2 -/-parasites in honey bee hindgut and found that they infect the host at the comparable level (Fig 7H).Thus, less motile L. passim with short flagellum by the loss-of function of LpFCaBP genes is still capable of infecting honey bee hindgut effectively.

Rapid evolution of FCaBP genes in trypanosomatid parasites
We characterized FCaBP genes in the genomes of trypanosomatid parasites and observed that each species, except L. seymouri, has more than two FCaBP genes in the same genomic contigs (Fig 3).This suggests that gene duplication occurred in the common ancestor of Trypanosoma and Leishmaniinae.While T. brucei, T. theileri, and L. pyrrhocoris have further duplicated FCaBP genes, it appears that four Leishmania species and L. seymouri have lost FCaBP2 by pseudogenization and deletion, respectively.When we expressed GFP fused with the sorting sequence of TbFCaBP in L. passim, we observed the cell body membrane localization.Meanwhile, GFPs fused with the sorting sequences of TcFCaBP, LdFCaBP, and CfFCaBP1 were present in both the flagellum and cell body membrane (Fig 4C).These findings indicate that the BBSome complex in L. passim may not recognize the flagellar sorting sequence of TbFCaBP, but partially recognizes those of TcFCaBP, LdFCaBP, and CfFCaBP1.Therefore, it appears that dually acylated FCaBP is naturally targeted to the plasma membrane of the entire body, and specific N-terminal sorting sequences and the BBSome complex are required for flagellar enrichment.Additionally, the inefficient localization of GFP at the distal part of the flagellum using the TcFCaBP and LdFCaBP sorting sequences suggests a functional or structural separation in the flagellum (See also Fig 5A).Meanwhile, GFP fusion proteins may also exist in the flagellar pocket and pocket neck membranes in these parasites.The lack of functional FCaBP in L. major demonstrates that it is not essential for completing the life cycle in sand fly and mammalian hosts, suggesting the presence of other proteins that complement the absence of FCaBP in L. major.
In trypanosomatid parasites with two FCaBP genes, mutations have accumulated in the 5' end of the ORF, modifying the N-terminal amino acid sequences after gene duplication.As a result, FCaBP1 and FCaBP2 have different subcellular localizations at the flagellum and entire body membrane, respectively.Thus, this can be considered as one example of PSR to retain duplicated genes.However, this process may not have occurred in the five Leishmania species and L. seymouri.Without an outgroup paralogous gene or a species with a single FCaBP gene, it is difficult to conclude whether this represents neolocalization or sublocalization.However, sublocalization is unlikely due to the absence of FCaBP2 in the aforementioned species.Since neolocalization to the flagellum requires co-evolution with the BBSome complex, the chance of this occurring is quite low.Therefore, we like to propose that neolocalization to the entire body membrane has occurred, and FCaBP2 in the plasma membrane of the cell body must provide some advantages for C. fasciculata, L. passim, C. bombi, C. expoeki, and L. pyrrhocoris.
We showed that S9 and S10 conserved among FCaBP1s across four trypanosomatid species are not only important for the efficient flagellar localization but also myristoylation and palmitoylation at G2 and C3.Similarly, T9 and Q10 of FCaBP2 appear to be essential for the acylation at the same amino acid residues (Fig 4B).Thus, myristoylation and palmitoylation at G2 and C3 in FCaBPs are likely to be context dependent influenced by the surrounding amino acid residues.

Interaction between LpFCaBP1 and the BBSome complex in L. passim
The flagellar localization of LpFCaBP1 may depend on the interaction between the N-terminal sorting sequence and the BBSome complex, rather than TULP (Fig 5).These results are consistent with the requirement of the BBSome complex for the exit of dually acylated phospholipase D to the cilia in Chlamydomonas [40].Although we did not detect interaction between LpFCaBP2N16-UltraID and Flag-LpBBS1, this could be due to the mislocalization of LpFCaBP2N16-UltraID at the anterior end of L. passim.It remains to be determined whether the entry/exit of LpFCaBP1 to the flagellum requires the BBSome and IFT complexes.

Roles of FCaBPs in flagellar morphogenesis and motility of L. passim
We found that LpFCaBPs are not essential for the viability of L. passim.However, the loss of LpFCaBPs altered the growth characteristics of L. passim in the culture medium and impaired flagellar morphogenesis and motility.The formation of rosettes, which requires the clustering of parasites through flagellar movement [41], was also defective in LpFCaBPs-deleted parasites ( Fig 7).In contrast, knocking down TbFCaBPs (Calflagin Tb17, Tb24, and Tb44) in T. brucei did not affect parasite growth and motility [20].The effects of FCaBP loss-of-function may differ between monoxenous parasites like L. passim (infecting only honey bees) and dixenous parasites like T. brucei (infecting both tsetse flies and mammals).It is also possible that the small amount of TbFCaBPs remaining after gene knockdown [20] is sufficient to support the normal growth and motility of T. brucei in the culture medium.We also found that neither FCaBP1 nor FCaBP2 alone can fully rescue the phenotypes of LpFCaBPs-deleted parasites ( Fig 7).Assuming that the amount of ectopic LpFCaBP1 or LpFCaBP2 protein is comparable to that of the endogenous protein, L. passim relies on both proteins with flagellar and entire body localizations.This is consistent with a PSR hypothesis [8,12].FCaBP likely functions as a calcium signaling protein, and it will be important to uncover how downstream effectors are involved in the growth and flagellar formation of L. passim.

Roles of the flagellum in the host infection of L. passim
We observed that LpFCaBPs-deleted parasites can infect honey bee as effectively as wild-type parasites in our assay (Fig 7H).Although the mutant parasites are less active, they can reach the honey bee hindgut through normal gut flow.L. passim appears to attach to the hindgut wall through the structurally modified flagellum [42], and the short flagella of the mutant parasites would be sufficient for attachment.In contrast, TbFCaBPs-depleted T. brucei showed attenuated parasitemia in infected mice [20], but the underlying mechanism and the infection in the tsetse fly gut were not clarified.The precise in vivo roles of FCaBP in infecting insects and mammalian hosts remain to be answered.
We collected actively growing L. passim (4 × 10 7 ), washed twice with 5 mL PBS, and then resuspended in 0.4 mL of Cytomix buffer without EDTA (20 mM KCl, 0.15 mM CaCl 2 , 10 mM K 2 HPO 4 , 25 mM HEPES and 5 mM MgCl 2 , pH 7.6) [44,45].The parasites were electroporated twice (1 min interval) with 10 μg of each plasmid DNA constructed above using a Gene Pulser X cell electroporator (Bio-Rad) and cuvette (2 mm gap).We set the voltage, capacitance, and resistance at 1.5 kV, 25 μF, and infinity, respectively.The electroporated parasites were cultured in 4 mL of modified FP-FB medium [46], and then G418 (200 μg/mL, Sigma-Aldrich) was added after 24 h to select the G418-resistant clones.We washed live L. passim expressing the GFP fusion protein three times with 1 mL PBS, and then placed them on poly-L-lysinecoated slide glass for imaging using a NIKON eclipse Ni-U fluorescence microscope.The exposure time remained constant at 200 msec.Using Image-J, we measured the mean fluorescence of GFP in both the entire flagellum and cell body of individual parasites, as well as the background.Then, we calculated the ratio between the fluorescence in the flagellum and the cell body after subtracting the background fluorescence.Statistical comparison was performed using the Steel test.

Analysis of microsynteny with FCaBP genes in trypanosomatid parasites
We used TBLASTN and LpFCaBP1 as a query to identify the genomic contigs containing FCaBP in trypanosomatid parasites.The order and orientation of genes in the microsynteny with FCaBP of five Leishmania species, L. seymouri, and L. pyrrhocoris were determined based on annotated gene lists.For L. passim, C. fasciculata, C. bombi, and C. expoeki, we performed the analysis using TBLASTN.The amino acid sequences of 11 FCaBPs (S1 Data) were aligned using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/).

Assessing the interaction between the N-terminal sorting sequence of LpFCaBP1 or LpFCaBP2 and either LpBBS1 or LpTULP
We constructed plasmid DNA for expressing ultraID with the N-terminal sorting sequence of either LpFCaBP1 or LpFCaBP2.The DNA fragment encoding the sorting sequence was PCR amplified using 1444F and Ultra-Rev-LpFCaBP1 or 2 primers.We also PCR amplified the DNA fragment encoding ultraID using Ultra-For-LpFCaBP1 or 2 and Ultra-3-XhoI primers and used pSF3-ultraID [38] (ADDGENE: #172878) as a template.The two DNA fragments were ligated through fusion PCR followed by BamHI and XhoI digestion and cloned into the pTrex-n-eGFP plasmid at the same restriction enzyme sites.We first constructed plasmid DNA carrying triple Flag tags by inserting the annealed complementary oligonucleotides (3Flag-5 and 3Flag-3) into the XbaI and HindIII sites of tdTomato/pTREX-b [47] (ADDGENE: #68709).Then, we PCR amplified the entire ORFs of LpBBS1 and LpTULP (S2 Data) using LpBBS1-5-HindIII and LpBBS1-3-ClaI stop primers as well as LpTULP-5-HindIII and LpTULP-3-ClaI stop primers.After digesting with HindIII and ClaI, the PCR products were cloned in the same sites of the above plasmid DNA with Flag tags.
L. passim was electroporated as described above with 10 μg each of LpFCaBP1N16-UltraID or LpFCaBP2N16-UltraID and either Flag-LpBBS1 or Flag-LpTULP.The parasites expressing both proteins were selected using G418 and blasticidin (50 μg/mL, Macklin).We detected the expressed proteins by immunofluorescence.The parasites were washed and mounted on a poly-L-lysine-coated 8-well chamber slide as above, fixed with 4% paraformaldehyde, permeabilized with PBS containing 0.1% TX-100 (PT), and blocked with PT containing 5% normal goat serum (PTG).The samples were incubated overnight at 4˚C with rabbit anti-Myc (500-fold dilution for the ultraID fusion proteins) or rabbit anti-Flag epitope (500-fold dilution for the Flag-tagged proteins) polyclonal antibodies (Proteintech) in PTG.The samples were washed five times with PT (5 minutes for each wash) and then incubated with Alexa Fluor 488 (for the ultraID fusion proteins) or Alexa Fluor 555 (for the Flag-tagged proteins) anti-rabbit IgG antibody (ThermoFisher) for 2 hours at room temperature.After another round of washing, the samples were observed under a microscope.
We incubated the parasites (4 × 10 8 ) in 20 mL culture medium with 50 μM biotin for 3 hours at 28˚C, followed by washing with 5 mL PBS twice.The samples were then suspended in 2 mL lysis buffer (10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5% NP-40, 0.5 mM EDTA) containing a protease inhibitor cocktail (Beyotime) and sonicated three times (5 seconds each) at amplitude 10 using a Q700 sonicator (Qsonica).After centrifugation, the supernatants were collected, and 50 μL of streptavidin agarose (Yeasen) and 12.5 μL of DYKDDDDK (Flag) Fab-Trap Agarose (Chromotek) were added to the half volume of each sample, respectively.The beads were washed five times (5 minutes each) with washing buffer (10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% NP-40, 0.5 mM EDTA) and then suspended in 60 μL of SDS-PAGE sample buffer (2% SDS, 10% glycerol, 10% β-mercaptoethanol, 0.25% bromophenol blue, 50 mM Tris-HCl, pH 6.8).The samples were heated at 95˚C for 5 minutes, and then 7.5 μL and 30 μL of each sample precipitated by Flag Fab beads and streptoavidin beads, respectively was applied to a 10% SDS-PAGE gel.The proteins were transferred to a nitrocellulose membrane (Pall Life Sciences) and the membrane was blocked with PBST (PBS with 0.1% Tween-20) containing 5% BSA at room temperature for 30 minutes.The membrane was incubated with a 1000-fold diluted anti-Flag epitope polyclonal antibody overnight at 4˚C.After washing five times with PBST (5 minutes each), the membrane was incubated with a 10,000-fold diluted IRDye 680RD donkey anti-rabbit IgG (H+L) secondary antibody (LI-COR Biosciences) in PBST containing 5% skim milk at room temperature for 2 hours.Following another round of washing, the membrane was visualized using ChemiDoc MP (BioRad).

RT-PCR
Total RNA was extracted from wild-type, LpFCaBPs heterozygous, and homozygous mutant parasites using TRIzol reagent (Sigma-Aldrich).Reverse transcription of 0.2 μg of total RNA was performed using ReverTra Ace (TOYOBO) and random primers, followed by PCR with KOD-FX DNA polymerase.To specifically detect LpFCaBP1 and 2 mRNAs, we designed reverse primers corresponding to the N-terminal sorting sequences (LpFCaBP1-36R and LpFCaBP2-36R).L. passim splice leader sequence (LpSL-F) was used as the forward primer.To detect LpGAPDH mRNA, we used LpGAPDH-F and LpGAPDH-R primers.

Culture, flagellar and cell body length measurement, and motility measurement of L. passim
To construct plasmid DNA expressing LpFCaBP1 or LpFCaBP2, we first PCR amplified the ORF using LpFCaBP1-5 and LpFCaBP1-3-XhoI or LpFCaBP2-5 and LpFCaBP2-3-XhoI primers.The plasmid DNA to express the corresponding GFP fusion protein was used as a template.We digested the resulting PCR products with XbaI and XhoI followed by subcloning into the same restriction enzyme sites of pTrex-n-eGFP plasmid DNA wherein neomycin resistance gene was replaced by bleomycin resistance gene.We electroporated LpFCaBPs homozygous mutant (LpFCaBP1/2 -/-) parasites with above plasmid DNA and the parasites expressing LpFCaBP1 or LpFCaBP2 were selected using hygromycin and zeocin (50 μg/mL, InvivoGen).
For culture, flagellar and cell body length measurement, and motility measurement, we inoculated wild-type, LpFCaBP1/2 -/-, and LpFCaBP1/2 -/-expressing either LpFCaBP1 or LpFCaBP2 parasites in the culture medium at 10 5 /mL at 29˚C.The number of parasites was counted every 12 hours using a hemocytometer, and images of the cultured parasites were captured simultaneously for a week.We measured the length of both the flagellum and cell body of individual parasites at 60 hours after culture, utilizing phase images with Image-J.To record the movement of parasites, videos were created by taking images for 1 minute every second.The movement of parasites was tracked using TrackMate v7.10.2 [51] as a Fiji [52] plugin.The videos were imported to Fiji, converted to 8-bit grayscale, and the brightness and contrast were adjusted for better tracking.The Laplacian of Gaussian detector was used with an estimated object diameter of 30.0-36.0 pixels and a quality threshold of 0.2-0.5 for the detection of individual parasites.We used the Simple Linear Assignment Problem tracker to track the parasites by adjusting the linking maximum distance and gap-closing maximum distance to 35.0-200.0pixels and gap-closing maximum frame gap to 1.We characterized above phenotypes with two independent clones of LpFCaBP1/2 -/-parasites (D1 and F6) and their phenotypes were the same.We thus used the clone D1 for further experiments.

Honey bee infection
To infect honey bees with L. passim, parasites collected during the logarithmic growth phase (5 × 10 5 /mL) were washed with PBS and suspended in sterile 10% sucrose/PBS at 5 × 10 4 /μL.Newly emerged honey bee workers were collected by placing the frames with late pupae in 33˚C incubator and starved for 2-3 hours.Twenty individual honey bees were fed with 2 μL of the sucrose/PBS solution containing either wild-type or LpFCaBPs homozygous mutant (10 5  parasites in total).The infected honey bees were maintained in metal cages at 33˚C for 14 days and then frozen at -80˚C.This experiment was repeated three times.We sampled eight honey bees from each of the above three experiments, and thus analyzed 24 honey bees in total infected with either wild-type or LpFCaBPs homozygous mutant.Genomic DNAs were extracted from the whole abdomens of individual bees using DNAzol reagent (Thermo Fisher).We quantified L. passim in the infected honey bee by qPCR using LpITS2-F and LpITS2-R primers which correspond to the part of internal transcript spacer region 2 (ITS2) in the ribosomal RNA gene.Honey bee AmHsTRPA was used as the internal reference using AmHsTRPA-F and AmHsTRPA-R primers [35].The relative abundances of L. passim in the individual honey bees (24 each infected by wild-type or mutant L. passim) were calculated by the ΔC t method and we set one sample infected by wild-type as 1.The statistical analysis was performed using the Brunner-Munzel test.All of the above primers are listed in S3 Data.

Fig 1 .
Fig 1. Subcellular localization of LpFCaBP1-and 2-GFP fusion proteins.Lotmaria passim expressing LpFCaBP1-GFP, LpFCaBP2-GFP, LpPFRP5-GFP, or GFP were examined under visible (DIC) or fluorescence (Fluorescence) light.Merged images are also shown.Flagellum at the anterior end and cell body are indicated by arrows.The ratio of the mean fluorescence in the flagellum to that in the cell body (mean value ± SD) along with number of parasites analyzed (n) is presented on the right.Scale bar: 2 μm.https://doi.org/10.1371/journal.pgen.1011195.g001

Fig 2 .
Fig 2. Flagellum and cell body membrane sorting sequences of LpFCaBP1 and 2. The N-terminal 16 or 28 amino acids of LpFCaBP1 (in red) and LpFCaBP2 (in blue) were swapped as illustrated in the diagram.The positions of amino acids in the cores of LpFCaBP1 and 2 proteins are also indicated.GFPs fused with the N-terminal 16 amino acids of LpFCaBP1 and 2 are shown at the bottom.The ratio of the mean fluorescence in the flagellum to that in the cell body (mean value ± SD) along with number of parasites analyzed (n) is presented on the right.Scale bar: 2 μm.https://doi.org/10.1371/journal.pgen.1011195.g002

Fig
Fig 4A shows the N-terminal ends of aligned FCaBPs from T. cruzi, T. brucei, L. donovani, C. fasciculata, L. seymouri, and L. passim.Glycine and cysteine residues targeted for myristoylation and palmitoylation are well conserved; however, there are more variations in the sorting sequences of FCaBPs compared to the main sequences with EF-hand calcium binding domains.To elucidate the pivotal amino acids within the sorting sequence governing the membrane localization of LpFCaBP in the flagellum or entire body, we performed targeted mutagenesis on LpFCaBP1 and LpFCaBP2.Specifically, lysine at position 11 (K11) in LpFCaBP1 was substituted with isoleucine, and vice versa, generating LpFCaBP1N16(K11I)-GFP and LpFCaBP2N16(I11K)-GFP.Additionally, serines at positions 9 (S9) and 10 (S10) in

Fig 4 .
Fig 4. Subcellular localization of GFPs with N-terminal sorting sequences of FCaBPs from various trypanosomatid parasites.(A) N-terminal parts of the aligned sequences of Trypanosoma cruzi (AAA99985.1),Trypanosoma brucei (XP_847374.1),Leishmania donovani (XP_003859779.1),Crithidia fasciculata, Leptomonas seymouri (KPI87939.1),and Lotmaria passim FCaBPs are shown.The amino acids fused to GFP in (C) are indicated in bold letters.Glycine (red) and cysteine (blue) residues are targeted for myristoylation and palmitoylation, respectively.Conserved amino acids are shown by asterisks, and similar amino acids are indicated by either a full stop (.) or a colon (:).(B) Subcellular localization of GFPs fused with the mutated N-terminal sorting sequences of LpFCaBP1 and 2. (C) Subcellular localization of GFPs fused with the N-terminal sorting sequences in (A).The ratio of the mean fluorescence in the flagellum to that in the cell body (mean value ± SD) along with number of parasites analyzed (n) is presented on the right.Scale bar: 2 μm.https://doi.org/10.1371/journal.pgen.1011195.g004 Fig 5A).Surprisingly, LpFCaBP2N16-UltraID concentrated at the anterior end of cell body (Fig 5A) in contrast to the uniform cell body membrane localization of LpFCaBP2N16-GFP (Fig 2).Thus, ultraID tends to anchor at the anterior side of cell body and this may also inhibit the localization of LpFCaBP1N16-UltraID to the distal part of flagellum in L. passim (Fig 5A).We detected biotinylated Flag-LpBBS1 but not Flag-LpTULP in the parasites expressing LpFCaBP1N16-but not LpFCaBP2N16-UltraID (Fig 5B), suggesting that LpFCaBP1N16 is in close proximity to LpBBS1.The apparent size of Flag-LpTULP was larger than that of Flag-LpBBS1 and the expected size (59 kDa) (Fig 5B).LpTULP may have post-translational modification or migrate slowly through 10% SDS-PAGE.These findings suggest that the flagellar localization of LpFCaBP1 may be mediated by the BBSome complex.