Mitochondrial proteins and complexes in Leishmania and Trypanosoma involved in U-insertion/deletion RNA editing

Abstract

A number of mitochondrial proteins have been identified in Leishmania sp. and Trypanosoma brucei that may be involved in U-insertion/deletion RNA editing. Only a few of these have yet been characterized sufficiently to be able to assign functional names for the proteins in both species, and most have been denoted by a variety of species-specific and laboratory-specific operational names, leading to a terminology confusion both within and outside of this field. In this review, we summarize the present status of our knowledge of the orthologous and unique putative editing proteins in both species and the functional motifs identified by sequence analysis and by experimentation. An online Supplemental sequence database (http://164.67.60.200/proteins/protsmini1.asp) is also provided as a research resource.

INTRODUCTION

The identification of proteins putatively involved in U-insertion/deletion RNA editing in trypanosome mitochondria has progressed very rapidly recently, owing to the application of mass spectrometry gene identification methods to components of macromolecular editing complexes isolated by a variety of methods, combined with the availability of partial genome databases for both Leishmania major and Trypanosoma brucei (http://www.genedb.org/). Our 2003 review (Simpson et al. 2003) is already considerably out of date.

Putative editing proteins have been identified both by direct biochemical isolations targeted for known enzymatic or RNA binding activities, and by sequencing of protein components of macromolecular complexes active in in vitro editing. Possibly functional motifs have been identified by sequence analysis but only confirmed by analysis of recombinant proteins in a few cases. Evidence for an involvement in editing has been obtained for several of these proteins by down-regulation of expression using conditional gene disruptions and RNA interference, but even this evidence is frequently difficult to interpret.

THE 20S “EDITOSOME”

A full cycle, although fairly inefficient, of gRNA-dependent in vitro editing activity that sedimented at approximately 20S in glycerol gradients was detected in mitochondrial extracts from procyclic T. brucei (Corell et al. 1996; Kable et al. 1996; Seiwert et al. 1996). Enrichment of this editing activity in the Sollner-Webb laboratory using sequential Q Sepharose and DNA cellulose column fractionations of mitochondrial extract led to a complex reported to contain seven major polypeptides; this sedimented at approximately 20S in glycerol gradients and migrated as a single band in a native gel (Rusché et al. 1997). This preparation was capable of both full round U-insertion and U-deletion editing at one or two sites using synthetic pre-edited A6 mRNA and the cognate gRNA, and the efficiency of in vitro editing could be substantially increased by modifications of the mRNA and gRNA sequences (Cruz-Reyes and Sollner-Webb 1996; Cruz-Reyes et al. 1998, 2001, 2002). The polypeptides were operationally labeled Bands I–VII. Bands IV and V proved to be the REL1 and REL2 RNA ligases (Rusché et al. 2001), which had been functionally identified previously as adenylatable proteins (Sabatini and Hajduk 1995; Peris et al. 1997) and Band III, a zinc finger-containing protein (Huang et al. 2002). Down-regulation of Band III by RNAi was lethal and affected in vivo endonucleolytic cleavage and RNA ligation but not addition or removal of uridines. Recognizable nuclease motifs are absent in the Band III protein, indicating these are indirect effects, consistent with the decrease of the sedimentation value of the entire complex (Huang et al. 2002). The notion that these seven proteins represented the complete set of proteins required for editing (Rusché et al. 1997) has evolved into the realization that a variety of additional components, which are most likely present in substoichiometric amounts or interact transiently, are also required (Aphasizhev et al. 2003a; Panigrahi et al. 2003a). For example, the RET2 TUTase required for U-insertion editing (see below) is not one of these seven polypeptides (B. Sollner-Webb, pers. comm.).

A similar fractionation procedure of in vitro editing activity in the Stuart laboratory involving sequential SP Sepharose, Q Sepharose, and Superose 6 chromatography of mitochondrial extract led to a product that contained more than 20 major polypeptides (Corell et al. 1996; Panigrahi et al. 2001a,b, 2003b). A product of a similar complexity was isolated by immunoprecipitation of the 20S region from a glycerol gradient of mitochondrial extract using an antibody against TbMP63, one of the protein components of the core complex (Sollner-Webb lab Band III). Several nonmitochondrial and mitochondrial metabolic or structural proteins were, however, also identified by mass spectrometric analysis of tryptic peptides derived from the biochemically purified material and the monoclonal antibody affinity-purified material (Panigrahi et al. 2003a), some of which represented major stained bands in SDS-PAGE, most likely caused by high abundance of these proteins in the mitochondrial cell fractions used as starting material.

Approximately 16 polypeptides with mitochondrial targeting signals were identified that appeared by one or more criteria to represent possible components of the editing machinery (see Table 1 ▶). These proteins have been given the operational designation of TbMPx in the Stuart laboratory, where x represents the calculated molecular weight of the mitochondrial pre-protein. The criteria included either cosedimentation in the 20S region in glycerol gradients, coimmunoprecipitation by antibodies against specific proteins demonstrated to be essential for editing, double affinity purification using TAP-tagged proteins, or demonstration of an inhibition of RNA editing by conditional down-regulation of expression in vivo. The genetic method is the most definitive especially because major nonediting protein contaminants were present in the biochemically purified and in the antibody-purified fractions. TbMP123, MP103, MP100, MP99, MP90, MP67, MP61, and MP24 represent proteins with mitochondrial signal sequences that were identified by mass spectrometry analysis of purified editing complexes but for which no definite evidence for inclusion in the editing complex was reported (Table 1 ▶).

TABLE 1.

Putative editing-related proteins in Leishmania and T. brucei

Operational names
Leishmania	T. brucei				Possible function					Effect of down-regulation on editing activity
Simpson lab	Stuart lab	Sollner-Webb lab	Functional names	Motifs	Putative	Exptl. verified	L-complex	GP-complex	MRP-complex	In vivo	In vitro	% ID of Lm and Tb	Ref
LC-1	MP81	Band II		ZnF_C2H2	Structural scaffold	Interacts with RET2 and REL2	TAP, IP, CC			Lethal, A6, ND7, Cyb	U-insertion	27	Panigrahi et al. 2001b; Aphasizhev et al. 2003a
LC-2	MP100			Exo_endo_phos	3′–5′ exonucl. of mRNA		TAP, IP, CC					39	Aphasizhev et al. 2003a; Panigrahi et al. 2003b
LC-3	MP99			Exo_endo_phos	3′–5′ exonucl. of mRNA		TAP, IP, CC					25	Aphasizhev et al. 2003a; Panigrahi et al. 2003b
LC-4	MP63	Band III		ZnF_C2H2	Structural scaffold		TAP, IP, CC			Lethal	mRNA cleavage and ligation	36	Panigrahi et al. 2001b; Aphasizhev et al. 2003a
LC-5	MP46			ZnF_NFX	RNA binding		TAP, IP, CC					19	Aphasizhev et al. 2003a; Panigrahi et al. 2003b
LC-6a	MP61			RNase III	Cleavage of mRNA		TAP, IP, CC					38	Aphasizhev et al. 2003a; Panigrahi et al. 2003b
LC-6b	MP57		RET2	3′-TUTase	Addition of Us at editing site	Recomb. protein TUTase activity	TAP, IP, CC			Lethal, ND8, COIII Cyb, COII	U-insertion	67	Aphasizhev et al. 2003a; Ernst et al. 2003
LC-7a	MP52	Band IV	REL1	Ligase	Ligation of mRNA fragments	Recomb. protein ligase activity	TAP, IP, CC			Lethal, A6, ND7, RPS12		65	McManus et al. 2001; Panigrahi et al. 2001a; Rusché et al. 2001; Schnaufer et al. 2001; Aphasizhev et al. 2003a
LC-7b	MP42	Band VI		ZnF_C2H2	Structural scaffold		TAP, IP, CC					42	Panigrahi et al. 2001b; Aphasizhev et al. 2003a
LC-7c	—						TAP, CS					11	Aphasizhev et al. 2003a
LC-8	MP44			RNase III	Cleavage of mRNA		TAP, IP, CC			Lethal, ND7, A6, ND4, RPS12	Disrupts L-complex	41	Aphasizhev et al. 2003a; Panigrahi et al. 2003b
LC-9	MP48	Band V	REL2	Ligase	Ligation of mRNA fragments	Recomb. protein ligase activity	TAP, IP, CC			No phenotype	No effect	64	McManus et al. 2001; Panigrahi et al. 2001a; Aphasizhev et al. 2003a
LC-10	—						TAP						Aphasizhev et al. 2003a
LC-11	MP18	Band VII		SSB	Binds RNA		TAP, IP, CC					44	Panigrahi et al. 2001b; Aphasizhev et al. 2003a
	MP121						IP, CC						Panigrahi et al. 2003b
Lm MP90	MP90			RNase III	Cleavage of mRNA		IP, CC					38	Panigrahi et al. 2003b
	MP103						IP, CC						Panigrahi et al. 2003b
Lm MP67	MP67			RNase III	Cleavage of mRNA		IP, CC					24	Panigrahi et al. 2003b
	MP24						IP, CC						Panigrahi et al. 2003b
Lm RGGm	TbRGGm-			RRM	RNA binding		IP					57	Vanhamme et al. 1998; Panigrahi et al. 2003a
	REAP-1				mRNA binding								Madison-Antenucci et al. 1998
	RBP16			Cold shock domain	RNA binding					Letha, Cyb, COI, ND4		72	Pelletier and Read 2003
Lm mHel61	mHEl62			DEAD box	Helicase?					Slow growth, inhibited	No effect	48	Missel et al. 1997
LtTUTase	TbTUTase		RET1	3′-TUTase	3′-U addition to gRNAs	Recomb. protein TUTase activity		CC	IA, TAP	Lethal, ND7, ND8, COIII, Cyb, MURF2	U-insertion, length of 3′-oligo(U) tail	38	Aphasizhev et al. 2002
Ltp26	gBP21		MRP1		gRNA-mRNA binding	Recomb. protein RNA annealing			TAP, CC			45	Koller et al. 1997; Aphasizhev et al. 2003b
Ltp28	gBP25		MRP2		gRNA-mRNA binding	Recomb. protein RNA annealing			TAP, CC			43	Blom et al. 2001; Aphasizhev et al. 2003b
AP1			MRP-AP1						TAP			48	Aphasizhev et al. 2003b
AP2			MRP-AP2	mre11_DNA_bind	RNA binding				TAP			59	Aphasizhev et al. 2003b
AP3			MRP-AP3	NUDIX	Hydrolase				TAP			42	Aphasizhev et al. 2003b

Abbreviations: IP, protein detected in immunoaffinity-purified material; TAP, protein detected in material isolated by tandem affinity purification; CC, protein detected in material purified by column chromatography and/or sedimentation; ID, identical amino acids.

The TbMP52 and TbMP48 proteins (the Sollner-Webb laboratory Bands IV and V) were identified as the two mitochondrial RNA ligases (Sabatini and Hajduk 1995; McManus et al. 2001) and were designated REL1 (RNA Editing Ligase 1) and REL2 (McManus et al. 2001; Panigrahi et al. 2001a; Rusché et al. 2001; Schnaufer et al. 2001). Down-regulation of REL1 was lethal and inhibited editing, but down-regulation of REL2 (Drozdz et al. 2002; Stuart et al. 2002; Gao and Simpson 2003) had no effect on cell growth. Down-regulation of TbMP81 and Band III (= TbMP63) expression by conditional RNAi inhibited in vivo editing but not transcription or mRNA stability.

The TbMP57 protein showed a nucleotidyltransferase motif and poly(A) polymerase-associated motifs, and recombinant protein immunoprecipitated from a reticulocyte transcription–translation system (Ernst et al. 2003) or expressed in Leishmania tarentolae as a TAP-tagged cytosolic protein by removal of the mitochondrial signal peptide (Aphasizhev et al. 2003c) both exhibited a 3′-TUTase activity that added mainly a single U residue to the 3′-OH of a single-stranded RNA substrate. This protein has been designated RET2 (RNA Editing TUTase 2; Aphasizhev et al. 2003a; Ernst et al. 2003), as suggested by K. Stuart (pers. comm.). Down-regulation of RET2 in procyclic T. brucei was lethal and led to a complete inhibition of U-insertion precleaved in vitro editing and no effect on U-deletion editing or on the length of the 3′-oligo[U] tails of the gRNAs, whereas down-regulation of RET1 expression produced a small effect on in vitro U-insertion editing and a large effect on the length distribution of the 3′-oligo[U] tails of gRNAs (Aphasizhev et al. 2003c). It was concluded that RET2 is an integral part of a core editing complex and is responsible for the addition of Us to the editing site and RET1 is responsible for the 3′ addition of Us to the gRNAs (Aphasizhev et al. 2003c).

One protein was identified that was homologous to the mitochondrial RNA-binding protein, TcRGGm, and was therefore designated TbRGG1, but there is no evidence that this is a component of an isolated editing complex (Vanhamme et al. 1998; Panigrahi et al. 2003a).

The difference in the number of polypeptides obtained in these two laboratories is most likely due to technical differences in isolation and detection methodologies and to differential stability of the complex components. In both cases, the final preparations showed both U-insertion and U-deletion in vitro editing activity. There is clearly a fairly stable macromolecular complex composed of ~16 proteins sedimenting at approximately 20S in glycerol gradients, but it may be premature to designate this as the “editosome” (Panigrahi et al. 2003a) because work in the L. tarentolae system discussed below has indicated a higher complexity of interacting components of the editing machinery required for functional editing (Aphasizhev et al. 2003a,b).

THE L-COMPLEX

The ligase-containing complex (L-complex) was first identified as a high-molecular-weight band in native gels of L. tarentolae and T. brucei mitochondrial extracts that could be labeled with [α-³²P]ATP (Sabatini and Hajduk 1995; Peris et al. 1997; Rusché et al. 1997) because of the presence of the two adenylatable RNA ligases, REL1 and REL2 (Sabatini and Hajduk 1995). This complex sedimented in glycerol gradients in a broad region from 20S to 25S but as a single band in a native gel and is almost certainly identical to the “editosome” complexes isolated from T. brucei discussed above. The L-complex was isolated from L. tarentolae by the TAP procedure using a REL1 fusion protein (Aphasizhev et al. 2003a). The purified L-complex sedimented as a single homogeneous adenylatable 20S particle both in gradients and in native gels. The isolated material showed precleaved U-insertion and U-deletion editing activity. Substoichiometric amounts of RET1 and MRP1 and MRP2 were also present in the final preparation, indicating a low-affinity interaction of the L-complex with the MRP-complex, the GP-complex, and gRNA (see below).

The gradient-isolated TAP-tagged 20S L-complex was composed of 13–16 bands in an SDS gel in approximate stoichiometric amounts (Aphasizhev et al. 2003a). The absence of contamination with abundant mitochondrial metabolic enzymes was shown by probing with antiserum for glutamate dehydrogenase. Substoichiometric amounts of RET1 and the MRP1,2 proteins could be identified by Western analysis. So far, 15 polypeptides have been identified by MS/MS analysis of tryptic peptides from gel-isolated stained bands. All except LC-10 have orthologs in the T. brucei database, and orthologs to MP90 and MP67 were identified in the L. major database. LC-6b (MP57) was shown to be a 3′-TUTase and was labeled RET2, as mentioned above.

Functional motifs in the L-complex proteins

All identified Leishmania editing proteins and the orthologs in T. brucei are shown in Table 1 ▶, together with the operational designations and the functional gene names if assigned. The apparent absences of LC-10 in T. brucei and MP90, MP103, and MP121 in Leishmania are probably due to database gaps rather than species-specific differences. All sequences in FastA format can be obtained from the online spreadsheet at http://164.67.60.200/proteins/protsmini1. asp. Alignments of orthologs from Leishmania and Trypanosoma with identified motifs indicated are also presented in the online spreadsheet as pdf files.

The presence of the Leishmania LC-proteins in the TAP-affinity isolated complex in approximately stoichiometric yields (Aphasizhev et al. 2003a) indicates that these are, indeed, integral components of the L-complex, and that the orthologs in T. brucei are also L-complex components in that species.

RNA and/or protein-binding motifs

LC-1 (MP81), LC-4 (MP63), LC-7b (MP42), and LC-5 (MP46) have zinc finger motifs, which could be nucleic acid or protein-binding. LC-11 (MP18) also has an SSB-like single-strand-binding motif (Fig. 1 ▶). SSB-like motifs can be recognized in LC-7b (MP42) and LC-10 by alignment of the LC-11 (MP18) SSB motif sequence, as indicated in Figure 1 ▶. Only the zinc finger motifs in LC-4 (MP63) have so far been experimentally shown to have a biochemical function in protein–protein interactions (X. Kang and L. Simpson, unpubl.). Possible RNA-binding functions of these motifs remain to be experimentally demonstrated.

FIGURE 1.

L-complex proteins with protein- and nucleic acid-binding motifs. The locations of the tryptic peptide sequences used for the L. major gene identifications are indicated by triangles in this and the following figures. All motif identifications in this and the following figures were performed using SMART (http://smart.embl-heidelberg.de/) and PFAM (http://www.sanger.ac.uk/Software/Pfam/index.shtml). The e-values are indicated in parentheses. Mitochondrial target signals were identified by MitoProt (http://www.mips.biochem.mpg.de/cgi-bin/proj/medgen/mitofilter), SMART, or SignalP (http://www.cbs.dtu.dk/services/SignalP/). The probabilities from MitoProt are given above the figure. Note that the motifs in LC-10 and LC-7b were identified by similarity with the SSB motif in LC-11. All sequences were obtained from the GeneDB Leishmania major and Trypanosoma brucei databases.

RNase III motifs

Three proteins with RNase III motifs were identified from Leishmania major and T. brucei (Fig. 2 ▶). RNase III enzymes degrade double-stranded RNA and are involved in rRNA processing and degradation of some mRNAs. The presence of an RNase III-like motif in LC-8 (MP44) was only detected by alignment of the LC-6a (MP61) RNase III motif sequence (Fig. 2 ▶). There is as yet no biochemical evidence for enzymatic activity of these proteins. A possible function is the gRNA-mediated cleavage of mRNA at editing sites.

FIGURE 2.

L-complex proteins with RNase III motifs. Note that the motif in LC-8 was identified by similarity with the RNase III motif in LC-6a. The L. major homologs to the T. brucei MP90 and MP67 were identified by database mining.

Exonuclease–endonuclease–phosphatase motifs

There are two L-complex proteins, LC-2 (MP100) and Tb MP99, with exonuclease–endonuclease–phosphatase motifs (Fig. 3 ▶). The L. major LC-3 protein is clearly an ortholog to MP99 but, interestingly, lacks most of the motif sequence.

FIGURE 3.

L-complex proteins with nuclease motifs. The LC-3 protein identified by mass spectrometry lacks most of the Exo_endo_phos motif present in MP99, but otherwise is very similar.

3′-TUTase motif

The L-complex LC-6b (MP57; Aphasizhev et al. 2003a; Ernst et al. 2003) has a nucleotidyltransferase motif separated between the second and third catalytic aspartates by a stretch of amino acids that is not found in other nucleotidyltransferase proteins, as exemplified by the RET1 TUTase (Fig. 4 ▶; Aphasizhev et al. 2002). There are also poly(A) polymerase-associated motifs. This is the ortholog of RET2 in T. brucei, which has been expressed and shown to add mainly a single U residue to the 3′-end of RNA substrates (Aphasizhev et al. 2003c; Ernst et al. 2003).

FIGURE 4.

The L. major L-complex RET2 3′-TUTase. The homolog in T. brucei was identified by database searching (Aphasizhev et al. 2002) and has been characterized experimentally (Ernst et al. 2003).

Ligase motifs

The five motifs in the REL1 and REL2 proteins indicated in Figures 5 ▶ and 6 ▶ are conserved among both DNA and RNA ligases (McManus et al. 2001; Ho and Shuman 2002). The localizations of the lysine residues that covalently bind AMP were also determined experimentally (Huang et al. 2001). Ligase activities of recombinant TbREL1 and REL2 expressed in an Escherichia coli system (McManus et al. 2001; Rusché et al. 2001), TbREL1 and TbREL2 expressed in a reticulocyte cell-free coupled transcription–translation system (Palazzo et al. 2003), and LtREL1 and LtREL2 expressed in the Baculovirus system (G. Gao and L. Simpson, unpubl.) have been demonstrated.

FIGURE 5.

The L-complex REL1 and REL2 RNA ligases. Five conserved ligase motifs (McManus et al. 2001) are indicated as is the location of the lysine residues that covalently bind AMP. These motifs can be recognized by multiple alignments with known ligase enzymes.

FIGURE 6.

Multiple alignments of portions of selected genes to illustrate the five conserved motifs in REL1 and RELL2 characteristic of DNA and RNA ligases (Shuman and Schwer 1995; Odell et al. 2000; McManus et al. 2001). (T3) Bacteriophage T3; (T7) bacteriophage T7; (Vac) Vaccinia virus; (At) Arabidopsis thaliana; (Mj) Methanocaldococcus jannaschii; (Cj) Campylobacter jejuni; (Sc) Saccharomyces cerevisiae; (Hu) human; (Ec) E. coli; (Ae) Aquifex aeolicus; (Mg) Mycoplasma genitalium; (T4) bacteriophage T4; (ChlV) Chlorella virus; (Tb) T. brucei; (Ce) Caenorhabditis elegans; (Lt) L. tarentolae; (Lm) L. major; (*) αPO₄-binding residues; ($) metal-binding residues; (@) ribose O-binding residue; (#) residue binding to Arg; (&) adenine-binding residue.

THE MRP-COMPLEX

Two mitochondrial RNA-binding proteins (MRPs) from L. tarentolae, MRP-1, were shown to form a stable 100-kD complex (Aphasizhev et al. 2003b). These mitochondrial proteins, which were previously called Ltp26 and Lt28 (Aphasizhev et al. 2003b), are orthologs of TbgBP21 (Koller et al. 1997) and TbgBP25. We suggest the MRP designation instead of the original gBP (guide RNA binding protein) designation (Koller et al. 1997) because it is now known that these proteins bind RNA fairly nonspecifically (Aphasizhev et al. 2003b). TAP-isolation using an MRP-1 fusion protein was used to identify three additional interacting proteins, AP1, AP2, and AP3 (Fig. 7 ▶; Aphasizhev et al. 2003b). The TAP-isolated ~30S complex contained bound gRNA, and the native and recombinant MRP proteins from both species and the reconstituted 100-kD complex exhibited binding to both double- and single-stranded RNA and catalyzed RNA annealing (Muller et al. 2001; Aphasizhev et al. 2003b). We have designated this as the MRP-complex (Aphasizhev et al. 2003c). There is evidence that down-regulation of both MRP1 and MRP2 together affects viability of procyclic T. brucei (A. Simpson, R. Aphasizhev, I. Aphasizheva, and L. Simpson, unpubl.), but it remains to be demonstrated that the complex is involved in vivo with catalyzing the annealing of the gRNA and mRNA.

FIGURE 7.

MRP-complex proteins. The L. major MRP RNA-binding proteins and the three associated proteins (AP1, AP2, AP3) identified by double affinity isolation (Aphasizhev et al. 2003b) are indicated.

THE RET1-COMPLEX

A biochemical isolation of the major processive 3′-TUTase activity from L. tarentolae mitochondria yielded a 121-kD protein belonging to the nucleotidyltransferase superfamily (Fig. 8 ▶; Aphasizhev et al. 2002). The recombinant protein expressed in E. coli exhibited a U-specific 3′ transferase activity that added multiple Us to the 3′-OH of any single-stranded RNA. An involvement with RNA editing was evidenced by the correlation of down-regulation of expression with a decrease in editing in vivo. The homologous protein was identified in T. brucei, and it was shown that down-regulation of expression by RNAi led to a decrease in the average length of the 3′-oligo[U] tail of the gRNAs, indicating that the observed inhibition of editing in vivo was indirect. This protein was designated as RET1 (RNA Editing TUTase 1; Aphasizhev et al. 2003c), as suggested by K. Stuart (pers. comm.).

FIGURE 8.

RET1 proteins. The L. major and T. brucei RET1 TUTases are shown (Aphasizhev et al. 2002).

The majority of the protein was present as an ~500-kD tetramer, but a small portion was in the form of an ~700-kD complex that could be separated by ion exchange chromatography followed by sedimentation in a glycerol gradient or electrophoresis in a native gel. We have designated this as the RET1-complex (Aphasizhev et al. 2003c). It remains to be investigated whether a previously identified U-specific 3′–5′ exonuclease activity (Aphasizhev and Simpson 2001; Igo et al. 2002) is also associated with this complex. It is possible that the 700-kD complex is a breakdown product of a larger complex after ion exchange chromatography, but this must be further examined.

PROTEINS WITH POSSIBLE ROLES IN RNA EDITING

It should be noted that there are several additional mitochondrial RNA-binding proteins that interact with guide RNA and mRNA, such as Tb REAP-1 (Madison-Antenucci et al. 1998; Madison-Antenucci and Hajduk 2001), RPB16 (Hayman and Read 1999; Pelletier et al. 2000; Pelletier and Read 2003) and RGG1 (Fig. 9 ▶; Vanhamme et al. 1998). Down-regulation of Tb RBP16 by conditional RNA interference showed that this protein is required for viability. The effect on maxicircle transcripts was similar to that obtained by down-regulation of both MRP1 and MRP2 (R. Aphasizhev and L. Simpson, unpubl.). There was a 98% decrease in the abundance of edited Cyb transcripts with little effect on other edited RNAs, and a 80% decrease in never edited RNAs, COI and ND4. Immunoprecipitation of RBP16 coprecipitated gRNAs. But no stable interaction of RBP16 with purified editing complexes was observed (Stuart et al. 2002).

FIGURE 9.

Several proteins that may be involved in RNA editing. The L. major orthologs of the RNA-binding proteins, RBP16 and RGGm, and the putative helicase, mHel61, are shown. The T. brucei REAP-1 is shown because there is no ortholog of this protein in the L. major database.

No ortholog for REAP-1 has been found in Leishmania. REAP-1 was identified in a 20S editing complex by Western analysis and was shown to coimmunoprecipitate with anti-REL1 antiserum, indicating an interaction with the L-complex (Madison-Antenucci et al. 1998). It was also shown that REAP-1 interacted with the pre-edited region of mRNA (Madison-Antenucci and Hajduk 2001), indicating a possible role in the interaction of gRNA and mRNA.

TBRGG1 is a mitochondrial oligo[U]-binding protein that cosedimented with in vitro editing activity in glycerol gradients. Peptides from this protein were identified by mass spectrometry analysis of a partially purified editing complex (Panigrahi et al. 2003a).

Hel61 is a putative helicase because of the presence of a DEAD-box motif, but there is no biochemical evidence for this activity (Missel and Goringer 1994; Missel et al. 1995, 1997). A double allele knockout in procyclic trypanosomes led to a reduced growth rate and significantly lower amounts of edited mRNAs, with little effect on never edited RNAs. Re-expression of the gene from an ectopic copy restored the ability to synthesize edited mRNAs. However, no effect was observed on full round in vitro editing by mitochondrial extracts of the mHel61-null cells (Missel et al. 1997).

MODEL FOR HIGH-LEVEL ORGANIZATION OF EDITING COMPLEXES

Evidence that the MRP-complex, the RET1-complex, and the L-complex interact via RNA linkers

Evidence for interactions of the L. tarentolae L-complex with RET1 and MRP1/2 was obtained by showing that the L-complex could be immunodepleted from the 20S region and coimmunoprecipitated by anti-RET1 antibody and also by anti-MRP1/2 antibody (Aphasizhev et al. 2003a,b). RNase digestion of the L. tarentolae mitochondrial extract prior to glycerol gradient sedimentation eliminated the co-IP of the L-complex with anti-RET1 or anti-MRP1/2 antibodies, but had no effect on the S value or polypeptide composition of the L-complex. In addition, the RNase pretreated L-complex eluted from a native gel lacked any detectable in vitro U-insertion editing activity, implying that the presence of the substoichiometric RET1 and MRP1/2 components was essential for editing activity (R. Aphasizhev and L. Simpson, unpubl.). This, however, must be investigated further.

The MRP1/2 homologs in T. brucei, TbgBP21 and gPB25, also showed RNA-binding and RNA-annealing activities, and anti-gPB21 antibodies coimmunoprecipitated in vitro editing activity, ATP-labeled REL1 and REL2, and TUTase activity (Allen et al. 1998). This co-IP was sensitive to micrococcal nuclease digestion, indicating that it is likely that the GMR–RNP complex is also present in T. brucei and that it also interacts with the L-complex via RNA.

We propose that the core editing complex composed of ~16 proteins be designated the L-complex in both species, both for reasons of historical priority (Peris et al. 1997) and descriptiveness. The term “editosome” should be reserved for the complete functional editing supercomplex yet to be completely defined but consisting of at least the L-complex linked by RNA to the RET1-complex and the MRP-complex. A diagrammatic model for the organization of the three complexes is shown in Figure 10 ▶. It is, of course, also possible that there is no static supercomplex but that there is an dynamic functional interaction of multiple components. The analogy with the growing complexity of the spliceosome is striking (Rappsilber et al. 2002).

FIGURE 10.

Diagrammatic model for higher-order organization of editing complexes. The L-complex is shown as a dimer, for which the evidence is suggestive, and the three complexes are shown as interacting via RNA. The gRNA and mRNA indicated are the most likely candidates, but the existence of other RNA species has not been ruled out. The gRNA structure shown is taken from the model of Hermann et al. (1997).