Multiple domains of the integral KREPA3 protein are critical for the structure and precise functions of RNA Editing Catalytic Complexes in Trypanosoma brucei

The gRNA directed U-insertion and deletion editing of mitochondrial mRNAs that is essential in different life cycle stages for the protozoan parasite Trypanosoma brucei is performed by three similar multi-protein catalytic complexes (CCs) that contain the requisite enzymes. These CCs also contain a common set of eight proteins that have no apparent direct catalytic function, including six that have an OB-fold domain. We show here that one of these OB-fold proteins, KREPA3 (A3), has structural homology to other editing proteins, is essential for editing and is multifunctional. We investigated A3 function by analyzing the effects of single amino acid loss of function mutations most of which were identified by screening bloodstream form (BF) parasites for loss of growth following random mutagenesis. Mutations in the ZFs, an intrinsically disordered region (IDR) and several within or near the C-terminal OB-fold domain variably impacted CC structural integrity and editing. Some mutations resulted in almost complete loss of CCs and its proteins and editing whereas others retained CCs but had aberrant editing. All but a mutation which is near the OB-fold affected growth and editing in BF but not procyclic form (PF) parasites. These data indicate that multiple positions within A3 have essential functions that contribute to the structural integrity of CCs, the precision of editing and the developmental differences in editing between BF and PF stages.


Mitochondrial mRNAs in Trypanosoma brucei, the causal agent of Human African
Trypanosomiasis (aka sleeping sickness), the related T. cruzi and Leishmania spp. pathogens and other kinetoplastids undergo post-transcriptional maturation by RNA editing (Read et al. 2016). The editing generates mature functional mRNAs from transcripts by the insertion and deletion of U-nucleotides as specified by small guide RNAs (gRNAs) (Koslowsky et al. 1990;Pollard et al. 1990;Riley et al. 1994). Some mRNAs undergo quite limited editing, e.g. the insertion of 4 or 34 Us in COII and CYb mRNAs, respectively, whereas editing essentially recodes the sequences of other mRNAs by extensive U insertion/deletion e.g. of +447/-28 and +547/-41 Us respectively in mRNAs for ATPase 6 and COIII oxidation-phosphoryation complex proteins and +132/-28 in RPS12 mitoribosomal protein mRNA (Benne et al. 1986;Feagin et al. 1987;Feagin et al. 1988;Bhat et al. 1990;Read et al. 1992). In addition, the editing of several transcripts differs between the life cycle stages of T. brucei in parallel with its metabolic and developmental differences. Edited cytochrome subunit mRNAs are abundant in the insect midgut stage procyclic form (PF) parasites which generate energy via oxidative phosphorylation whereas these edited mRNAs are dramatically reduced in the mammalian bloodstream form (BF) parasites that produce energy via glycolysis (Panigrahi et al. 2008). Thus, this differential D a v i d g e 3 editing adapts this pathogen to the disparate environments of the midgut of its tsetse fly vector vs. the mammalian bloodstream (Feagin et al. 1987 (A3) is one of the non-catalytic proteins that is common to all three CCs. It contains two C2H2 zinc fingers motifs (ZFs), a C-terminal OB-fold motif, several intrinsically disordered regions (IDRs) but no other motifs that are related to known protein domains or predicted structures (Panigrahi et al. 2001;Schnaufer et al. 2010).
A3 is essential for cell growth and functions in RNA editing in both BF and PF life cycle stages as has been shown by the loss of parasite viability and RNA editing upon A3 expression knockdown by RNAi or in conditional null (CN) cell lines (Guo et al. 2008;Law et al. 2008;Guo et al. 2010;McDermott et al. 2015b). Loss of A3 expression also affects CC structural integrity, albeit to a greater extent in BFs than in PFs as shown by the retention of CCs in PF but not BF A3 CN cells (Brecht et al. 2005;Guo et al. 2008;McDermott et al. 2015b). Mutation analyses showed that the N-terminal ZF (NTZF) and the more C-terminal ZF (CTZF) domains are both required for parasite viability and RNA editing in BFs whereas the latter ZF is not for required in PF (Guo et al. 2008;Guo et al. 2010;McDermott et al. 2015b). Thus, the two A3 ZFs impact CC structure and function somewhat differently between these two life-cycle stages. We report here the identification of multiple single amino acid mutations that result in loss of function (LOF) in BFs and in one case in PFs. Most of these substitutions mapped to the two ZFs, an IDR in a region with no predicted structure or in or adjacent to the OB-fold. Exclusive expression of some of these mutant alleles resulted in loss of all three CCs and editing in BFs but not in PFs except in one case where CCs and editing were reduced but not eliminated in D a v i d g e 5 PFs. Other mutations had lesser effects on CC structure and altered but did not eliminate RNA editing and some of these mutations had differential effects on CCs and editing that were generally more impactful in BFs than in PFs. These results indicate that A3 interacts with multiple proteins in the three CCs and contributes to their structural organization and integrity.
The results suggest that various A3 domains identified by these mutations have specific functions among the numerous events that occur during the complicated process of RNA editing. These functions likely occur via specific molecular interactions that may differ among the three CCs which possess different editing site specificities. The differential effects of the mutations between BF and PF cells imply that some of these A3 domains function differently between these stages despite the indistinguishable protein compositions of CCs.

Mutations that affect growth of BFs are in multiple protein domains:
To elucidate the role of A3 in editing we tested 634 BF CN cell lines that had been cloned in wells by dilution after transfection with a library that contained ~10,000 unique full length randomly mutated A3 alleles and identified 136 wells with growth defects by replica plating in the presence or absence of tet as previously described (Table S1) (McDermott et al. 2015a).
PCR and sequencing of A3 from 59 wells that had strong growth defects identified 14 mutations of interest from sequences that 1) had a single amino acid substitution, the same substitution in multiple clones that had multiple mutations, or a substitution that likely would have a functional consequence based on structural analyses, and 2) which reproduced the growth defect upon exclusive expression in cells in which the mutation was independently reconstructed by sitedirected mutagenesis (Fig. 1A). Three single substitutions (Q299H, L262P and T315A) did not reproduce the growth defects upon reconstruction and thus these residues are not critical for A3 function in BFs and may have originated from wells in which the screened cells were not clonal.
This number of confirmed single amino acid substitutions that affect A3 function is similar to that D a v i d g e 6 identified in B5 (9), B6 (12), B7 (10), and B8 (9) by similar methods (McDermott et al. 2015a;Carnes et al. 2022). We also constructed cell lines with double amino acid substitutions in the more N and C-terminal ZFs (NTZF and CTZF) with an N-terminal V5 tag for comparison with previous constructs carrying the same mutations (Guo et al. 2008;Guo et al. 2010) (Table S2).
The confirmed LOF mutations localized to the two ZF domains and to various positions near or within the OB-fold (Fig. 1A). They were mapped to the partial A3 structure that had been determined by x-ray crystallography and a full-length A3 structure that was predicted using β -strands that contains many charged lysine and glutamate residues. The features of these ZF domains imply that they participate in RNA and/or protein interactions. The other single mutations mapped to a region adjacent to the N-terminal side of the OB-fold which includes one of the predicted intrinsically disordered (IDR) inter-domains and a small helical region and to multiple positions within the highly structured OB fold (Fig. 1B). The single and double mutations in the ZFs had substantial impacts on growth of the BFs (Fig 1 C); the single ZF mutations were not tested in PFs but double mutations in the NTZF but not the CTZF substantially inhibit growth in PFs (McDermott et al. 2015b). The mutations that mapped to the region that is adjacent to or in the first strand of the OB-fold (V282G, C288Y and N290K) or the nearby disordered region (L270R) that was previously predicted (Park and Hol 2012) also had substantial impacts on growth (Fig. 1C). The mutations that mapped to the core of the OB-fold β -barrel structure or in or near loops L 12 and L 45 (F305L, F307S, V311F, Q349P, H377Y and P379T) had lesser effects on growth. All mutations throughout the OB-fold and the nearby region except C288Y and C53A/C56A . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made These qPCR analyses provide an overview of targeted pre-edited and edited mRNA amplicon abundances but do not measure partial or anomalous editing resulting from the mutations. We therefore generated profiles of all of these RNAs from BF and PF cells by performing RT-PCR using primers that anneal to the mRNA 5' and 3' terminal regions of the mRNAs that do not get edited, resulting in amplification of all products regardless of how much editing occurred. We analyzed ATPase 6 (A6), RPS12, and MURF2 mRNAs which are edited in both life cycle stages and used never-edited ND4 mRNA as a control (Fig. 4B). We resolved the resultant products on an agarose gel to visualize the editing products associated with each exclusive-expression cell line (Fig. 4B). Because edited mRNAs contain more insertion sites than deletion sites, larger products result from more editing, and thus more editing occurred in cell lines with larger products. The A3 CN cell line in both the presence and absence of tet was included as a control. The different mutations resulted in distinct RT-PCR product profiles (Fig. 4B). In BFs, the single L66S, H69R, H199Y and H204Y mutations in the helical regions of either ZF and the double C183A/C186A mutation in the CTZF resulted in A6 profiles that were distinct from the conditional null in the presence or absence of tet, which suggests that some editing occurred but was likely anomalous or incomplete, similar to what has been previously described for double mutation in the ZFs (Guo et al. 2010). A similar result was observed for RPS12 except the single mutants had more pre-edited size product in comparison to the CN +tet. The double C53A/C56A mutation of the NTZF alone or together with the double mutation of the CTZF resulted in greater amounts of pre-edited sized product, which is indicative of reduced editing.
These results and those with the double mutation in the NTZF and/or CTZF in BF and PF CN cells that exclusively express a C-terminal TAP-tagged (BF) or untagged (PF) A3 reinforce the role for the A3 ZFs in editing (Guo et al. 2008;Guo et al. 2010;McDermott et al. 2015b).
Mutations in or near the OB-fold had more variable effects on editing which were parallel to their effects of CC structure. The V282G, C288Yand N290K substitutions which resulted in . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. substantial loss of CCs and their proteins (Fig. 2) had predominant RT-PCR products consistent with pre-edited mRNAs or short partially edited products (Fig. 4B). The F305L mutant that had reduced, but not eliminated CC abundance had larger RT-PCR products consistent with partially edited sizes. Similar partially edited products were also observed with F307S, V311F, Q349P, H377Y and P379T mutants that broadly retained CC proteins. Interestingly, the L270R mutant that retained CCs had similar A6 profiles to those in these five mutants but had a prominent RPS12 product that was slightly larger the pre-edited rather than larger products seen in the others (Fig. 4B).
Analysis of the effects of a subset of the OB-fold mutations on editing in PFs by gel analysis of the RT-PCR products of the A6, RPS12 and MURF2 mRNA showed more products that are larger than pre-edited than in BFs which reflect the presence of CCs in PFs. The sizes of the RNAs are similar to those in the eWT with the exception of the C288Y mutants where smaller products are seen (Fig. 4B). Thus, mutations in the OB-fold that were assayed in PFs resulted in different effects on editing than in BFs and indicating functional differences of the A3 OB-fold between these developmental stages.
We transfected a Mutant Gamma ATP synthase (MGA) allele into the A3 CN cell line and the reproduced some of the mutant cell lines and eWT to enable the survival and continuous growth of BFs in the absence of RNA editing (Figs S1 and 4B) (Dean et al. 2013;Carnes et al. 2017;Carnes et al. 2023). This allowed us to assess possible incomplete development of the effects of the mutations or secondary effects of the mutations, e.g. physiological consequences due to loss of products encoded by edited mRNAs. This resulted in a shift to a greater proportion of products with sizes at or near that of pre-edited mRNA (Fig. 4B). This was especially evident for the V282G and C288Y mutations that lack most CC material (Figs. 2 and S1) and with the longer A6 and RPS12 mRNAs. The other mutations had products that were larger than preedited but generally smaller than the largest product in wild type. The L270R mutant retained D a v i d g e 1 3 ample ~ 1MDa CCs but the RT-PCR product appeared slightly longer than pre-edited (Fig. 2).
Overall, these results suggest that editing occurred in mutants that retained CCs albeit insufficient in amount or accuracy to support growth.
To determine the characteristics of the RT-PCR products we sequenced multiple RPS12 PCR products that range in size from ~221bp (pre-edited) to 325bp (fully edited) that we cloned from cell lines that retained various proportions of ~1MDa and ~800 kDa CCs. As shown in the diagram (Fig. 5) these sequences indicated that editing had occurred in these cells and is incomplete compared to fully edited RNA including in WT as is typical. However, in the L270R mutant the extent of editing and the proportion of fully edited insertion ESs were reduced and the proportions of deletion ESs that were not fully edited and of editing in non-edited ESs were increased. In the F307S and P379T mutants the extent of editing was greater than in L270R and there were proportional differences in the various types of ESs from those in WT and L270R but not statistically significant in this survey less than in the WT (Fig. S2). These results show that these mutants perform editing but do so with an accuracy and/or efficiency that is insufficient for cell viability.
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. Additionally, MPSS5 has a C2H2 ZF similar to A3, while MPSS5 and MPSS6 have predicted helical domains and regions with no predicted strictures that may contain IDRs, all features that imply functional interactions with RNA and/or protein (Fig. 6A).
We developed a structural model to explore how A3 might bind RNA as many ZF and OB-fold containing proteins bind RNA or DNA and A3 has been suggested to bind RNA in vitro (

DISCUSSION
We show here that multiple functional domains of A3 are essential for structural integrity and accurate functioning of the CCs that perform U insertion and deletion editing. Multiple single amino acid (saa)-LOF mutations were identified in the two ZFs, C-terminal OB-fold, adjacent region domains and in a nearby intrinsically disordered region. Saa-LOF mutations that disrupt CC integrity indicate that A3 is critical to CC structure (Figs 2-5) and show that its interaction with A6 is especially important for this ( Fig 7A). . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023.
Multiple domains of A3 are critical for the structure and functions of the three different

CCs:
Each domain of A3 contributes to its overall function as is evident from the diverse effects of the mutations which range from the alteration but not loss of editing to the complete loss of CCs, their proteins and editing (summarized in Fig. 7). The C2H2 ZFs contribute to CC structural organization via their interactions with other CC proteins as shown by the effects of their mutation on CC integrity, i.e. less than ~1 MDa CC material and reduced amounts of A1 protein was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023.

RNA interactions and effects on editing:
Our structural modeling suggests that the A3 OB-fold can interact with RNA, i.e. substrate RNA (Fig. 6). The characteristics of the A3 OB-fold and its structural similarities to ssDNA and RNAbinding OB-fold proteins, e.g. P. falciparum SSB and mammalian LIN28A proteins (Theobald et al. 2003;Horvath 2011;Wang et al. 2017) suggests that these interactions could occur via a combination of base stacking interactions with planar residues and electrostatic interactions with . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ; specific steps in the editing process. Structural modeling suggests how A3 might interact with RNA but further detailed analyses will be necessary to elucidate the specific effects of mutations in A3 on each of the three distinct CCs and on other components of the editing machinery, e.g. SCs. The structural homology between A3 and other proteins involved with mt RNA processing . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

Library screening
The mutated library was screened as previously described ( in HMI-9 with penicillin-streptomycin, 10% FBS, and 5 ng/mL tetracycline, which is the minimum amount that allows normal cell growth. Cells were plated in 24 well plates at a density of ~ one transfected cell per well which resulted in 634 puromycin-resistant cell lines. These potential cell clones were consolidated into seven 96 well plates, replica plated at a 1:100 dilution into media plus or minus tetracycline. The cells were grown for three days then passaged into another set of 96-well plates with plus or minus tet media for another three days. Following a total of six days of growth + or -tet, 20 μL of Alamar Blue cell viability reagent (Thermo Fisher) was added to each well and the plates were incubated for 4 h. Fluorescence was measured using a SpectraMax M2 microplate reader (Molecular Devices) with an excitation wavelength of 544nm . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made and an emission wavelength of 590nm to identify wells with viable (pink) and non-viable cells (blue). Comparison of the tet plus and minus replica plates was used to identify 117 cell lines with a strong growth defect (blue in minus tetracycline) and 19 with a lesser growth defect (purple in minus tetracycline). These clones were consolidated into new 96 well plates and 10µL of confluent cells were lysed for gDNA extraction as previously described (McDermott et al. 2015a).

PCR and sequencing of mutants
59 full growth defect clones were sequenced using primers 9571 and 5356 or 10150 (Tables S1   and S2). Forward and reverse sequence pairs were assembled using Geneious and aligned to the wild-type T. brucei 427 Lister A3 sequence to allow for identification of mutations and resulting amino acid substitutions (Table S1).

Reconstruction of mutations and generation of exclusive expression cell lines
Mutations of interest, either single substitutions that disrupted function, substitutions found in more than one cell line, or substitutions at residues mentioned in the literature, were reconstructed in new cell lines as single amino acid substitutions using site-directed mutagenesis. The pHD1344tub(PAC)-Nterm3V5-A3 plasmid was mutagenized using the QuickChange Lightning site-directed mutagenesis kit (Agilent, product #210518) and primers listed in Table S2. All mutations were confirmed by Sanger sequencing using primers 9571 and 5356 or 10150 (Table S2). Substitutions of the C residues in both ZF domains were made using site-directed mutagenesis and the primers listed in were used for our study as no protein could be detected from a previous construct, NTZF-myc, . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ; D a v i d g e 2 3 as previously described (Fig. S4)

Cell culture and transfections
BF cells were grown in HMI-9 with 10% FBS at 37 °C, 5% CO2. PF cells were grown in SDM-79 with 10% FBS at 27 °C. The PF A3 CN cell line used contained an untagged tet-regulated A3 construct expressed from the rDNA locus and was produced as previously described (McDermott et al. 2015b). Unless otherwise stated, the concentrations of drugs used for . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

Growth rate analyses
BF cells were seeded at an initial density of 2x10 5 cells/mL and PF cells at 2x10 6 cells/mL unless noted otherwise. Cells were counted using a Z1 Coulter particle counter each day. BF were reseeded at 2 x 10 5 cells/mL in 10 mL every day, while PF were reseeded at 2 x 10 6 cells/mL in 10 mL every two days. The ratio of cumulative cell numbers in -tet to + tet cultures for each cell line was calculated (Carnes et al. 2022). Log2 of the ratio is displayed on a blueorange heat map created using Microsoft Excel.

Glycerol gradient fractionation
Glycerol gradients were run as previously described (McDermott et al. 2015a). Briefly, 3x10 9 BFs were grown in the absence of tet for 72 hours, collected by centrifugation and washed with PBS plus 6mM glucose. Cell pellets were flash frozen in liquid nitrogen and stored at -80°C until use. Pellets were lysed in 500-1000µL of lysis buffer containing 10mM Tris pH 7.2, 10mM MgCl 2 , and 100mM KCl supplemented with cOmplete mini protease inhibitor tablets (Roche), . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ; D a v i d g e 2 5 pepstatin A, leupeptin, Pefabloc (Roche), and 1 mM DTT. Triton-X100 was then added to a final concentration of 1%. Cells were incubated on ice for 10 min followed by centrifugation at 13,000 rpm in a microcentrifuge to clear. 50 to 100µL of the cleared lysate was taken for a western blot and the remainder was loaded onto 11 mL 10-30% glycerol gradients and centrifuged in a Beckman Optima XPN-80 ultracentrifuge for 9 hours at 38,000 rpm using an SW40 rotor. 500µL fractions were collected from the top and 100µL samples were taken for analysis by both native and denaturing western blots. Remaining fractions were flash frozen in liquid nitrogen and stored at -80°C. Western blots from gradients were developed using either ECL (Pierce) or SuperSignal West PICO Plus ECL (Pierce).

SDS-PAGE, BN-PAGE, and Western blotting
Western blots were conducted as previously described (Carnes et al. 2022). Briefly, cell pellets containing 5x10 7 -2x10 8 cells were resuspended on ice in IPP150 buffer (10mM Tris-HCL pH 8.0, 150mM NaCl, and 0.1% NP-40) with cOmplete protease-inhibitor cocktail mini tablets (Roche). Triton-X100 was added to a final concentration of 1% and the samples were incubated on ice for 10-20 mins to lyse the cells. Lysates were cleared by centrifugation at 13,000 rpm in a microcentrifuge for 10 mins, diluted 1:1 in Tricine sample buffer (Bio Rad) with β mercaptoethanol and up to ~6x10 6 cell equivalents were loaded onto each lane of 10% TGXacrylamide Criterion gels (Bio Rad). Gels were run at 100V for ~ 2 h or until the dye front ran off the gel. Protein was transferred to PVDF membrane (Millipore) using a Tris/glycine transfer buffer with 20% methanol. Membranes were blocked in PBS-Tween with 5% milk for 15 min before the addition of primary antibody. Antibodies and dilutions are listed in Table S4.
Western blots were exposed to ECL (Pierce) and visualized using X-ray film (McKesson). For BN-PAGE, cleared lysates were prepared the same as for the denaturing western blots. 5x10 6 cell equivalents were loaded into wells on a 3-12% Bis-Tris gradient gel with unstained NativePage protein standards in the first well (ThermoFisher) and run according to . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made manufacturer's instructions. Protein was then transferred to a PVDF membrane (Millipore) overnight with NuPAGE transfer buffer ThermoFisher) without methanol at 22-24V at 4°C.
Following transfer, gels were stained with Ponceau SP to visualize the protein standards, fixed in 10% acetic acid, and probed with antibodies (Table S6).

RNA extraction, RT-PCR and qPCR
All procedures for RNA extraction, quantification, cDNA synthesis, and qPCR were designed to follow the MIQE guidelines. Briefly, 1X10 8 cells were harvested by centrifugation, washed with PBS and 6mM glucose, resuspended in Trizol, and stored at -80°C until use. Total RNA was extracted from the Trizol samples, quality was checked using a BioAnalyzer (Agilent) and quantified using the NanoDrop One spectrophotometer (Thermo Scientific). RNAs were treated with DNase I and cDNA was generated using random hexamer primers with TaqMan reverse transcription reagents and MultiScribe reverse transcriptase (ThermoFisher). Prior to qPCR, cDNAs were pre-amplified using Taqman PreAmp master mix (ThermoFisher), treated with Exonuclease I to remove excess primers, and diluted as previously described (McDermott et al. 2015a). High-throughput BioMark qPCR was performed as previously described using SsoFast EvaGreen with low ROX (Bio-Rad) and the Fluidigm BioMark HD system (McDermott et al.

2015b). Data was analyzed using the Fluidigm real-time PCR analysis software and Microsoft
Excel. In most cases, heat maps were generated using the average of two or three biological replicates with two technical replicates each. However, there is only one ND7 3' ed and ND8 ed replicate for V311F, H377Y, and P379T. Primers used for this assay are listed in Table S2. For RT-PCR (Fig. 4), cDNA was synthesized from 1µg of RNA using Multiscribe reverse transcriptase (Invitrogen) and a gene specific reverse primer (4394 (MURF2), 5380 (A6), 3620 (RPS12), or 3707 (ND4)) ( Table S2). The entire amount of cDNA was used in a PCR reaction to amplify MURF2, RPS12, A6, or ND4 transcripts using primer pairs 6204/4934 (MURF2), . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made PCR products were resolved on a 3% agarose gel containing ethidium bromide and visualized.

Cloning and Sanger sequencing of RT-PCR products
Remaining RT-PCR products from Figure 4 were resolved on an 1.5% agarose gel and all RPS12 products spanning pre-edited through edited sizes (~200-350 bp) were excised for each sample (BF L270R, BF eWT, and BF CN), purified, and cloned into pGEM-T-Easy (Promega) according to manufacturer's instructions. Ligations were transformed into 5α E. coli (NEB) and plated on Amp/IPTG/X-gal plates for blue/white screening. White colonies were selected for Sanger sequenced using T7 and M13R primers and assembled using Geneious. 11 clones for eWT, 24 clones for L270R, 13 clones for F307S, and 16 clones for P379T were analyzed. 4 clones from the CN were also obtained. Editing events were determined and sequences aligned by hand. Editing sites were defined as each position between two non-U nucleotides, beginning with the most 3' deletion site in RPS12, as this region was not spanned by primer 3620 used for RT-PCR. Editing at each site was then tabulated for each set of samples and recorded as follows: : I-Ed.= expected insertion sites that are edited (the number of Us matches the fully edited (F.E.) RPS12 sequence), I-Part. Ed.= partially-edited insertion sites (do not match F.E. or Pre-edited (P.E.) RPS12), D-Ed.= edited deletion sites, D-Part. Ed.= partially-edited deletion sites, N-Part. Ed.= partial editing at a site that is non edited in F.E. RPS12, not edited, and N/A= a clone contained an A, C, or G mismatch so it could not be analyzed at one or more ESs.
Graphs, which show the type of editing at each site as a percent of total transcripts, were generated using Microsoft Excel. Graphs include all potential ESs (positions between non-U nucleotides). The type of editing (insertion, deletion, or no editing) expected at each site based on the fully-edited RPS12 sequence is indicated beneath each bar by a yellow (insertion), blue (deletion), or black (no editing) dot. Aligned sequences from each set of samples are shown in Figure S2.
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  (Table S4) were visually identified using the AlphaFold structures of each protein in comparison to A3.

Multiple sequence alignments
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ; Sequence alignments in Figure 1 were constructed using the MUSCLE alignment tool in Geneious. The gene IDs for T. brucei, T. cruzi, and T. congolense A3 used for the alignment are Tb427.08.620, TcCLB.509611.110, and TcIL3000_8_100.1, respectively.

ACKNOWLEDGMENTS:
This work was supported by the National Institutes of Health grant R01AI014102 to K.S. We thank Xuemin Guo for providing the A3 CN cell lines.       was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ; deleted Us that match the F.E. sequence are indicated in red or by *, respectively. Inserted and deleted Us that do not match the F.E. sequence are also indicated. Junctions in each transcript are underlined. Highlighted regions contain a non-U nucleotide mismatch. Analysis of this data to produce the graphs in Figure 5 is described in detail in the Materials and Methods.   . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

References
The copyright holder for this preprint (which this version posted April 19, 2023. ; D  a  v  i  d  g  e  3  3   B  r  e  c  h  t  M  ,  N  i  e  m  a  n  n  M  ,  S  c  h  l  u  t  e  r  E  ,  M  u  l  l  e  r  U  F  ,  S  t  u  a  r  t  K  ,  G  o  r  i  n  g  e  r  H  U  .  2  0  0  5  .  T  b  M  P  4  2  ,  a  p  r  o  t  e  i  n  c  o  m  p  o  n  e  n  t  o  f  t  h  e  R  N  A  e  d  i  t  i  n  g  c  o  m  p  l  e  x  i  n  A  f  r  i  c  a  n  t  r  y  p  a  n  o  s  o  m  e  s  ,  h  a  s  e  n  d  o  -e  x  o  r  i  b  o  n  u  c  l  e  a  s  e  a  c  t  i  v  i  t  y  .  M  o  l  e  c  u  l  a  r  c  e  l  l   1  7   :  6  2  1  -6  3  0  .  B  r  e  v  e  t  A  ,  C  h  e  n  J  ,  C  o  m  m  a  n  s  S  ,  L  a  z  e  n  n  e  c  C  ,  B  l  a  n  q  u  e  t  S  ,  P  l  a  t  e  a  u  P  .  2  0  0  3  .  A  n  t  i  c  o  d  o  n  r  e  c  o  g  n  i  t  i  o  n  . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ; D  a  v  i  d  g  e  3  5   R  e  a  d  L  K  ,  L  u  k  e  s  J  ,  H  a  s  h  i  m  i  H  .  2  0  1  6  .  T  r  y  p  a  n  o  s  o  m  e  R  N  A  e  d  i  t  i  n  g  :  t  h  e  c  o  m  p  l  e  x  i  t  y  o  f  g  e  t  t  i  n  g  U  i  n  a  n  d  t  a  k  i  n  g  U  o  u  t  .  W  i  l  e  y  i  n  t  e  r  d  i  s  c  i  p  l  i  n  a  r  y  r  e  v  i  e  w  s  R  N  A   7   :  3  3  -5  1  .  R  e  a  d  L  K  ,  M  y  l  e  r  P  J  ,  S  t  u  a  r  t  K  .  1  9  9  2  .  E  x  t  e  n  s  i  v  e  e  d  i  t  i  n  g  o  f  b  o  t  h  p  r  o  c  e  s  s  e  d  a  n  d  p  r  e  p  r  o  c  e  s  s  e  d  m  a  x  i  c  i  r  c  l  e  C  R  6  t  r  a  n  s  c  r  i  p  t  s  i  n  T  r  y  p  a  n  o  s  o  m  a  b  r  u  c  e  i  .  T  h  e  J  o  u  r  n  a  l  o  f  b  i  o  l  o  g  i  c  a  l  c  h  e  m  i  s  t  r  y   2  6  7   :  1  1  2  3  -1  1  2  . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023