In Trypanosoma brucei RNA Editing, Band II Enables Recognition Specifically at Each Step of the U Insertion Cycle

Cell lines and extract preparation.The T. brucei procyclic cell lines used here carry an ectopically integrated plasmid that directs band II RNAi, and they were cloned from single cells by extreme dilution, as described previously (28). This transfected plasmid carries a 700-nucleotide internal fragment of the band II coding region flanked by T7 promoters and tetracycline (TET) operators (51, 53) from the pZJM vector (51). Because the recipient 29-13 cells synthesize T7 RNA polymerase and TET repressor (52), TET (1 μg/ml) induces expression of double-stranded band II RNA (28, 51). Before growth phenotypes become evident (5 days postinduction for the band II RNAi cells [28]), extracts were prepared from several independent, clonal band II RNAi cell lines (RNAi +). Extract preparation followed either the traditional protocol for trypanosome mitochondrial extract (18, 38) or a rapid extract protocol used previously for small-scale preparations (20, 28, 36). The traditional mitochondrial extract starts with ∼1 to 16 liters of cells (∼10⁷/ml) and separates the lysate on a Percoll gradient; with the multiple washing steps, this takes ∼6 to 8 h following cell lysis and yields extract at ∼2 × 10⁷ cell equivalents/μl. The rapid, small-scale extract streamlines this protocol to <1.5 h. This protocol starts with ∼2 × 10⁸ cells, which allows shorter centrifugations in a microcentrifuge tube following the initial pelleting, lyses the cells quickly by vortexing rather than by Dounce homogenization plus passage through a syringe needle, and uses half the time of DNase treatment. Additionally, it omits the Percoll gradient and all but two of the STE (250 mM sucrose, 10 mM Tris [pH 8], 2 mM EDTA) washes. Finally, for the 0.5% Triton X-100 treatment (0°C, 5 min), the sample is suspended in 10 mM KCl-MRB (25 mM Tris-HCl [pH 8.0], 10 mM Mg acetate, 1 mM EDTA, 0.5 mM dithiothreitol, 5% glycerol, and 10 mM KCl), 20-fold more dilute than in the traditional mitochondrial extract, and then the supernatant is aliquoted without reextraction, generating extract at 10⁶ cell equivalents/μl. In parallel, extract was also prepared from control cultures (RNAi −), using 29-13 cells for traditional extracts and uninduced band II RNAi cells for rapid extracts. Such rapid extracts of control cells support all the steps of editing (see Fig. 3 to 9 and reference 20), including cleavage reactions (Fig. 1C; see also Fig. 3 and 4) that exhibit substantially less extraneous cutting than is frequently shown.

Band II was similarly depleted in all the RNAi extracts used in this study (see Fig. 1B for the preparations that were used for most of the gels in Fig. 3 to 9, as indicated in their legends). Additional analysis of these extracts, including Western blotting of bands I and III through VII and Northern blotting of bands II and V (IREL) mRNA, was reported previously (28). Comparable amounts of extract were determined by total protein concentration (using Bradford analysis) and were confirmed by Western blotting of specific proteins, including lipoamide dehydrogenase (a control trypanosome mitochondrial protein) and unaffected editing proteins, such as DREL; these assessments gave comparable results.

Activity assays.All assays were performed using extracts from several different clonal band II RNAi cell lines (each with its parallel control extract), and they gave confirmatory results. The particular extracts used for the figures are indicated in the legends. Each assay was shown to be in the approximately linear range for each extract by their titration. Substrate RNAs, transcribed from PCR-amplified DNA or ordered from Oligos Etc., were gel isolated after 3′- or 5′-end labeling (11, 20). The 20-μl reaction mixtures used 10 mM KCl-MRB buffer with 5 mM dithiothreitol and 0.8 U of RNasin (Promega)/ml and were incubated at 27°C for 30 or 45 min unless otherwise noted. RNAs were resolved on 40-cm-long, 9% polyacrylamide-7.5 M urea gels (15% polyacrylamide-7.5 M urea was used for Fig. 6A and 7A), sized to the nucleotide using RNase sequencing products from the same labeled input RNA (7, 44), and visualized by autoradiography. In all experiments, the comparable RNAi − and + lanes use equal amounts of extract, determined as described above.

(i) Specific editing assays.The three steps that make up each kind of editing cycle were individually assessed, using the standard editing substrate, the 3′ portion of T. brucei ATPase 6 (A6) pre-mRNA (8, 23, 41). (i) The U-deletional cleavage step (see Fig. 2A and 3) was assayed using 3′-end-labeled A6 pre-mRNA m[0,4], which contains four U's at editing site 1 (ES1), and gRNA D33′ (11), which directs efficient removal of three of those U's. This reaction is supplemented with 0.25 mM sodium pyrophosphate (PPi), 4.5 ng of yeast (torula) RNA/μl, and 3 mM AMP-CP (8, 12, 20). (ii) The U-insertional cleavage step (see Fig. 2B and 4) was assayed using 3′-end-labeled m[0,4], which has no U at ES2, and gRNA I47G (12), which directs insertion of two U's at that site. This reaction mixture was supplemented with 0.5 mM PPi and 45 pg of torula RNA/μl but no AMP-CP (8, 20). (iii) The 3′-U-exo step (Fig. 6A) was assayed by the precleaved approach (22), mimicking the cleaved U deletion site noted above by using gRNA D33′ plus the 16-nucleotide upstream segment (5′-end labeled) and 15-nucleotide downstream segment of m[0,4] following its cleavage and including 0.25 mM PPi to inhibit ligation. (iv) The TUTase step (see Fig. 7) was assayed using the upstream (5′-end-labeled) and downstream precleaved mRNA segments and gPCA6-2A gRNA of Igo et al. (21) that mimic a U insertion site following cleavage. This assay included 0.15 mM UTP and 2 mM PPi (see also reference 21) and was incubated for 60 min. (v) The ligation step of U deletion (see Fig. 8A) also used the precleaved approach, mimicking a U deletion site following its cleavage and 3′ U removal with the RNAs of Fig. 7A by adding 0.15 mM ATP to activate the ligase and omitting UTP and PPi (19). This ligation occurs across a 2-nucleotide gap, ensuring it is a U-deletional ligation by DREL, which joins gapped as well as ungapped substrates (12), while U-insertional ligation is specific for ungapped substrates (12, 21). To specifically assay the ligation step of U insertion, it is coupled to the TUTase step (21), but that is not possible for band II RNAi extracts, where the TUTase step is inactive. (vi) Full-round U deletion (see Fig. 9A) was assayed using the same RNAs and reaction conditions as those used for U-deletional cleavage, except that they were supplemented with 0.15 mM ATP and 5 mM CaCl₂, omitting PPi (8, 19). (vii) Coupling between the cleavage step and the 3′-U-exo step (Fig. 9B) was assayed using the same RNAs and conditions as those for the U-deletional cleavage reaction, except the mRNA was 5′-end labeled to visualize U removal from the upstream fragment. For all these reactions, the mRNA (25 to 50 fmol) and gRNA (1.25 pmol) or the precleaved RNAs (0.1 to 0.2 pmol of upstream mRNA, 1 pmol of downstream mRNA, and 0.5 pmol of gRNA) were preannealed as described previously (19, 21).

• Open in new tab
• Download powerpoint

FIG. 2.

All cleavage activities are diminished in traditional extracts of band II RNAi cells. Analysis of editing cleavage reactions, using traditional extracts of control (RNAi −) and induced (RNAi +) band II RNAi cells. (A) The U-deletional cleavage step was assayed using 0.4 and 0.2 μg of extract. (B) The U-insertional cleavage step was assayed using 4 and 0.8 μg of extract. (C) Basic U-insertion-like cleavage was assayed using 0.4 and 0.2 μg of extract. In all experiments, the comparable RNAi − and + lanes used equal amounts of extract. [ext], extract titration. The sizes of the cleavage products are indicated in nucleotides (nt).

• Open in new tab
• Download powerpoint

FIG. 3.

U-deletional cleavage is not diminished upon band II RNAi. Analysis used rapid extracts of induced band II RNAi cells (RNAi +) and control cells (RNAi −), PPi, and AMP-CP unless otherwise indicated. −, reaction without extract. (A) U-deletional cleavage using 2.4 and 0.8 μg of the extracts shown in Fig. 1B, lanes 3 and 4. (B) AMP-CP effect on U-deletional cleavage using 2.4 μg of these same extracts. (C) A larger segment of the first lane of the gel shown in panel B showing specificity of cleavage. In this and subsequent figures, the reaction is diagrammed underneath the gel. [ext], extract titration; *, ³²P label; nt, nucleotide.

• Open in new tab
• Download powerpoint

FIG. 4.

U-insertional cleavage is diminished upon band II RNAi. Analysis used rapid extracts, PPi, and no AMP-CP, unless indicated. (A) U-insertional cleavage using 1.15, 0.4, and 0.1 μg of the extracts used for Fig. 3A. (B) AMP-CP effect on U-insertional cleavage using 1.8 μg of another pair of extracts. (C) A larger segment of the second lane of the gel depicted in panel B showing specificity of cleavage. The faint bands seen in the reaction without extract do not represent editing cleavages. nt, nucleotide; [ext], extract titration; *, ³²P label.

(ii) Basic enzymatic assays.The enzymes that catalyze the steps of the editing cycles were assessed independently of their ability to act at an editing site, used minimal substrates. (i) Basic U-insertion-like cleavage activity (see Fig. 2C and 5) was assayed using reaction conditions for U-insertional cleavage and 3′-end-labeled m[0,4] annealed to D33′ (45). In this reaction, cleavage is directed by the 3′ end of the gRNA, which forms an anchor-like duplex with a region of the upstream mRNA that is entirely devoid of U's (see Fig. 5). (ii) Basic 3′-U-exo activity (Fig. 6B) (35) was assayed using ∼0.05 pmol of a gel-isolated 5′-end-labeled RNA band from Anc+U16 DNA (11) that has a 14-nucleotide 3′ oligo(U) tail. (iii) Basic TUTase activity (see Fig. 7B) (4) was assayed using 0.1 pmol of unlabeled m[0,4] and 1.5 pmol of [α-³²P]UTP (3,000 Ci/mmol). (iv) Basic U-deletional ligase activity (see Fig. 8B) was assayed using ∼0.15 pmol of 5′-end-labeled polylinker transcript from EcoRI-cleaved pIBI30 (Pharmacia) (35) and 0.15 mM ATP (20, 35). The last three reaction mixtures were incubated at 22°C.

• Open in new tab
• Download powerpoint

FIG. 5.

Basic U-insertion-like cleavage is not diminished upon band II RNAi. Analysis used rapid extracts, PPi, and no AMP-CP, unless otherwise indicated. (A) U-insertion-like cleavage using 2.4 and 0.8 μg of the extracts used for Fig. 3A and 4A. (B) AMP-CP effect on U-insertion-like cleavage using 2.4 μg of the extracts shown in Fig. 1B, lanes 1 and 2. (C) A larger segment of the second lane of the gel depicted in panel B showing specificity of cleavage. This basic U-insertion-like cleavage assay uses the same U-deletional substrate as that in Fig. 3 but without AMP-CP, so cleavage is opposite the gRNA 3′ end, which is base paired to the pre-mRNA within a polypurine region. Because the gRNA is synthesized using T7 RNA polymerase, an appreciable fraction of the molecules acquire an extra 3′ U. When annealed with the pre-mRNA, that U fortuitously aligns opposite an A and extends the base pairing by 1 residue. Thus, this assay generates doublet bands of 44 and 45 nucleotides (nt) in length. [ext], extract titration; *, ³²P label.

• Open in new tab
• Download powerpoint

FIG. 6.

3′-U-exo is not diminished upon band II RNAi. Analysis using rapid extracts and PPi. (A) The 3′-U-exo step of U deletion, scored using a precleaved substrate and 2.4 and 0.8 μg of the extracts used for Fig. 3A to 5A. The input RNA and products lacking 1, 2, or all 3 terminal U residues are indicated. (B) The basic enzymatic activity of the 3′-U-exo scored on an RNA with a 14-nucleotide (nt) 3′-terminal oligo(U) track, using 0.5 and 0.25 μg of the extracts used for Fig. 4B. The input RNA and the extent of the oligo(U) track that gets removed by the 3′-U-exo are indicated. [ext], extract titration; *, ³²P label.

• Open in new tab
• Download powerpoint

FIG. 7.

The TUTase activity of U insertion is diminished upon band II RNAi, but the enzyme is not lost. Analysis using rapid extracts, UTP, and PPi. (A) The TUTase step of U insertion, scored on a precleaved substrate using 2.4 and 0.8 μg of the extracts used for Fig. 4B and 6B. (B) The basic enzymatic activity of the TUTase scored by U transfer to a single-stranded RNA using 0.4, 0.12, and 0.06 μg of the extracts used for Fig. 3A to 6A. The RNA used in this panel is a mixture of two size classes, so two bands represent the molecules that have acquired one U residue. [ext], extract titration; *, ³²P label; nt, nucleotide.

• Open in new tab
• Download powerpoint

FIG. 8.

RNA ligation by DREL is not diminished upon band II RNAi, but the IREL protein is lost. Analysis using rapid extracts and no PPi. (A) Assay of ligation on a precleaved substrate that mimic a U deletion site following U removal (20) using 0.4 and 0.12 μg of extract. (B) Assay of the basic DREL enzymatic activity, scored by joining free RNA ends, using 0.4 and 0.12 μg of the extracts used for Fig. 5B and 7B. Ligated dimers are shown. (C) SDS-PAGE analyzing adenylylation reaction, following deadenylylation, with the extracts used for panel A. [ext], extract titration; *, ³²P label; nt, nucleotide.

Utility of nontraditional extract preparation for RNAi cells.Using the traditional extract protocol (18, 38) (see Materials and Methods) that has been routinely used to study editing (6, 11, 23, 33, 35, 41), we prepared extracts from four clonal lines of band II RNAi cells induced for 5 days (before they exhibited appreciable growth effects) (28) and parallel control cultures. In contrast to the active control extracts, these band II RNAi extracts were markedly diminished in numerous editing activities, including U-deletional cleavage (Fig. 2A), U-insertional cleavage (Fig. 2B), basic U-insertion-like cleavage (Fig. 2C), and basic 3′-U-exo activities (data not shown; see Materials and Methods for descriptions of these assays). This could suggest that band II is important for all the above activities. However, we also considered that band II-depleted complex might not withstand the traditional extract protocol, which entails ∼6 to 8 h of handling following cell lysis (18, 38). We therefore also examined extract prepared more rapidly, using a protocol that reduces the handling time postlysis by ∼80% (20, 28, 36) yet is compatible with all the steps of editing (Fig. 1C; see also Fig. 3 to 9). As will be shown below, such rapid extracts, prepared in parallel from uninduced and 5-day-induced clonal band II RNAi cell lines, reproducibly retained U-deletional cleavage, basic U-insertion-like cleavage, and basic 3′-U-exo activities (see Fig. 3, 5, and 6) that were depleted in the traditional extracts (Fig. 2). This indicates that these activities were present in the RNAi cells and were lost during the longer traditional extract preparation, evidently due to instability of the band II-depleted complex. Thus, all subsequent experiments used these rapid extracts.

• Open in new tab
• Download powerpoint

FIG. 9.

Full-round U deletion occurs despite band II RNAi. (A) Full-round U deletion assay using 1.2, 0.4, and 0.12 μg of the rapid extracts used for Fig. 5B, 7B, and 8B, as well as AMP-CP and no PPi. The upper panel shows the gel region containing the input and fully edited (−3) RNA; the lower panel shows the region of the same gel that contains the cleaved 3′ fragment. The bands representing the product ligated with incomplete U removal (−2) and the cleaved but unligated fragment (3′ frag) are no more intense, relative to the complete −3 U deletion product, than with control extract, even though this particular experiment used an RNAi extract where U-deletional cleavage is slightly reduced. (B) The 3′-U-exo step of U deletion assayed in combination with gRNA-directed cleavage (as in Fig. 3A, except with 5′-end-labeled mRNA) using 1.2 and 0.4 μg of the rapid extracts used for panel A. The 3′-U-exo efficiency is the fraction of the upstream cleavage fragments (the sum of the cleavage, −1U, −2U, and −3U products) from which all three U's (−3U) have been removed. The various fragments are distinguished relative to sequence-specific markers (data not shown). This experiment used the same RNAi extract as that used for panel A, and although it generates a slightly reduced amount of cleaved RNA, the 3′-U-exo reaction appears as complete as that with control extract. [ext], extract titration; *, ³²P label; nt, nucleotide.

The cleavage steps.Using 3′-end-labeled A6 pre-mRNA and a gRNA to target the U-deletional (first) editing site, plus pyrophosphate (PPi) to inhibit ligation, we specifically assessed the U-deletional cleavage step of editing (8, 10, 19). The rapid extracts of band II RNAi and control cells are similarly active for this editing step (Fig. 3A). Titration shows that some RNAi extracts retain full U-deletional cleavage activity while others exhibit slightly less (within twofold), likely reflecting a small extent of the lability that this activity shows in traditional extracts of band II RNAi cells (Fig. 2A). Verifying its U-deletional nature (8), this cleavage by the rapid RNAi extracts is quite inactive without adenosine nucleotides, and it is strongly stimulated by AMP-CP (Fig. 3B). Furthermore, these RNAi extracts cleave at the same site as control extracts (Fig. 3A). Thus, by all established criteria, U-deletional cleavage appears normal upon band II RNAi.

When instead using a gRNA to target the U-insertional (second) editing site of 3′-end-labeled A6 pre-mRNA, and again including PPi to inhibit ligation, we specifically assessed the U-insertional cleavage step of editing (8, 10, 19). In contrast to U-deletional cleavage, U-insertional cleavage is markedly diminished in rapid extracts of band II RNAi cell lines (Fig. 4A). Extract titration shows that the amount of U-insertional cleavage activity remaining upon band II RNAi is obtained with only ∼1/10th as much control extract. Nonetheless, this residual cleavage activity appears normal, since it is inhibited by AMP-CP (Fig. 4B) (8) and shows the same site specificity as the control extract. Because induction of the RNAi leaves some residual band II and therefore some wild-type editing complex (28), the actual band II-depleted complexes must be even more deficient in U-insertional cleavage than it appears from Fig. 4. Thus, band II is critical for endonucleolytic cleavage in U insertion but not in U deletion.

Since U-insertional cleavage is absent from rapid extracts of band II RNAi cells, we assayed for the basic enzymatic activity of the U-insertional endonuclease independent of its ability to act at an editing site. The editing endonuclease activities have been found to cleave at a point abutting an anchor duplex, not only at natural editing site but also when only one strand extends (5′) from a duplex (Fig. 5) (45). Reactions using such minimal substrates mimic U-deletional or U-insertional cleavage depending on whether there are U's or a non-U, respectively, just 5′ of the duplex (see the diagrams in Fig. 1), and this specificity is verified by its AMP-CP stimulation or inhibition, respectively (45; A. Zhelonkina, J. Cruz-Reyes, and B. Sollner-Webb, unpublished data). Remarkably, cleavage of such a basic U-insertion-like substrate remains as active with rapid extracts of the RNAi cells as with control extracts (Fig. 5A). Furthermore, this activity was verified as U-insertion-like by its AMP-CP inhibition (Fig. 5B). As expected because U-deletional cleavage remains active (Fig. 3), band II RNAi extract also cleaves a basic U-deletion-like substrate as efficiently as the control extract (data not shown). The data of Fig. 4 and 5 lead to the interesting conclusion that band II is needed for the U-insertional endonuclease to productively recognize an actual U insertion site but not for this enzyme to be present or active on a nonediting substrate.

The 3′-U-exo step of U deletion.The second step in the U-deletional cycle is removal of U's by the 3′-U-exo at the 3′ end generated in the cleavage step. To individually assess this editing step, we bypassed the endonuclease step by using a precleaved substrate (three annealed short RNAs that mimic an editing site following cleavage [21, 22]) and again inhibited the ligation step by adding PPi. Our precleaved U deletion substrate (see Materials and Methods and the diagram in Fig. 6A) was designed to reproduce the cleaved U deletion site studied above (Fig. 3), with the upstream fragment 5′-end labeled to reveal the U removal. Using this substrate, we found that the 3′-U-exo step is as efficient with band II RNAi extracts as with control extracts (Fig. 6A).

As above, the basic enzymatic activity of the 3′-U-exo can be assayed using a minimal substrate, in this case a single-stranded RNA with a 3′ oligo(U) tail that the enzyme preferentially removes (35). Rapid extracts of band II RNAi cells are active in this assay (Fig. 6B). In fact, they reproducibly remove U's from a larger fraction of the substrate molecules and, on average, remove a larger number of U's than do their control extracts (Fig. 6B; see also Discussion). The combined data in Fig. 3 and 6 demonstrate that band II is not important for either of the first two steps of U deletion.

The TUTase step of U insertion.In the second step of U insertion, TUTase adds U's to the mRNA 3′ end generated in the cleavage step. This TUTase step was assessed using a precleaved U-insertional substrate and PPi (21), analogous to the above-described assay of the 3′-U-exo step. However, reactions with rapid band II RNAi extracts support only minimal U addition (Fig. 7A), indicating that the TUTase step of editing is inactive without band II. Thus, band II appears critical for both the first and second steps of U insertion.

The basic TUTase enzymatic activity was scored independently of editing using a minimal substrate by following transfer of UMP (from [α-³²P]UTP) to the 3′ end of a single-stranded RNA (4, 35). It has been shown that the TUTase that adds the U's in U insertion (RET2) (2), can be distinguished from the TUTase that adds the U's onto the gRNA (RET1) (2) when using such a minimal substrate, because RET2 (TbMP57) preferentially adds a single U while RET1 (TbMP108) adds very long poly(U) tracts (2, 3, 14). Scoring RET2 basic activity by this assay, we found that addition of a single U was similarly robust with rapid band II RNAi extracts and control extracts (Fig. 7B). (The RNAi extract additionally generates some heterogeneous size addition products, possibly due to increased availability of enzyme in the absence of band II [see Discussion].) Therefore, without band II the TUTase of U insertion does not utilize a U-insertional editing substrate but remains present and active. We conclude that band II enables the second step of U insertion by facilitating action on an editing substrate, much like that at the first step of this editing cycle.

The ligation steps.The ligation step of U deletion is catalyzed by the DREL ligase (12, 19, 39), which remains present upon band II RNAi (see Fig. 8C and reference 28). To assay the ligation step of U deletion, we again used the precleaved approach (21) with RNAs (see the diagram in Fig. 8A) that mimic a U deletion site following endonucleolytic cleavage and 3′ U removal (19). Such U-deletional ligation requires addition of ATP (21), which is needed to activate the DREL ligase (12, 20, 35) that seals this form of editing (12, 19). Using this assay, rapid band II RNAi extracts were found to retain full activity relative to control extracts (Fig. 8A). Thus, the third step of U deletion also remains active upon band II RNAi.

To assay the basic enzymatic activity of this U-deletional ligase, we scored dimer formation of an unstructured RNA (Fig. 8B). Previous studies have shown that this ligation is catalyzed by DREL (36), not IREL (12, 20). This assay shows that the basic DREL enzymatic activity is as robust in band II RNAi extracts as in control extracts.

The second ligase of the editing complex, IREL (TbREL2), is specific for U insertion (12, 39). Adenylylation assays, which include an initial deadenylylation step to detect all ligase molecules (35), show that the IREL protein is lost from band II RNAi cells (Fig. 8C; see also reference 28). When absent, IREL surely cannot serve its function in ligating U insertion. Western and Northern blotting plus sedimentation analysis (28) affirmed that IREL loss occurs at the protein level, not the mRNA level, evidently due to degradation of ligase that is not in the editing complex (28, 36). Thus, band II appears critical for the IREL protein to be recognized and retained in the editing complex and hence to ligate U insertion. Taken together, these analyses reveal that band II is important at each step of U insertion, yet it provides none of these basic enzymatic activities.

Coordination of the individual U deletion steps within the editing cycle.Although we have shown that each step of U deletion remains active upon band II RNAi, this does not ensure that those individual steps will function in a coordinated manner to catalyze a full cycle of editing. This is especially a concern because the depleted complex may be more extended (see Discussion) (28) than the wild-type editing complex. If coordination between the cleavage and 3′-U-exo or between the 3′-U-exo and ligase was impaired, the full-round U deletion reaction (Fig. 9A) should generate more molecules that are cleaved but not ligated (designated 3′ frag) and/or that exhibit partial U removal (−2 edited product). However, there is no accumulation of these molecules with band II RNAi extracts (Fig. 9A), indicating that the cleaved fragments are converted to fully edited product (−3) as efficiently as with control extracts. To verify coordination between the cleavage and 3′-U-exo steps (Fig. 9B), we used 5′-end-labeled pre-mRNA in a U deletion reaction to directly visualize 3′-U-exo action on the terminal U's of the upstream cleavage fragment (10, 19), with PPi added to prevent ligation (Fig. 9B). The lack of discernible partial U removal products (the position indicated by cleavage; −1U and −2U) affirms that the 3′-U-exo acts efficiently on the cleavage product in both the RNAi and control extracts (Fig. 9B). A repeat of this assay using a gRNA that is less efficient for the 3′-U-exo step (D32 [11]) reinforced this conclusion (data not shown). The results shown in Fig. 9 thus demonstrate that band II is also unnecessary for coordination of the steps of U deletion into a complete editing cycle.

Extract preparation.Although traditional mitochondrial extract (see Materials and Methods and references 18 and 38) is frequently used for the study of editing complex from wild-type cells (12, 23, 34, 35, 41), we considered that band II-depleted complex might be less stable and thus might not withstand this lengthy preparation. We therefore compared extract from band II RNAi cells prepared by this protocol and that prepared by a protocol that takes <1/5 the time (20, 28, 36) yet is compatible with all steps of editing (for an example, see Fig. 1C). Interestingly, the U-deletional cleavage, basic U-insertion-like cleavage, and basic 3′-U-exo activities that are retained in rapid extracts of band II RNAi cells (Fig. 3, 5, and 6), and thus surely were in the original cells, are lost in traditional extracts of those cells (Fig. 2 and data not shown). This indicates that band II-depleted editing complex is less stable than wild-type complex, and the apparent lack of these activities in traditional extracts is not a direct effect of band II loss but rather an indirect effect due to lability of the band II-depleted complex. This lability presumably explains why occasional rapid extracts of band II RNAi cells exhibit slightly reduced U-deletional cleavage activity (Fig. 9). In contrast, all extracts of the RNAi cells show greatly reduced U-insertional editing activities, evidently reflecting a direct effect of band II depletion.

Cleavage steps.While band II loss affects the U-insertional cleavage step, it does not affect the U-deletional cleavage step (Fig. 3 and 4) or the basic U-insertion-like cleavage activity (Fig. 5), indicating that the band II protein is not the endonuclease. This challenges an earlier proposal that band II may be the editing endonuclease, based on a TbMP81 (band II) RNAi extract exhibiting reduced cleavage of a CYb substrate (13). However, that cleavage activity appeared to sediment somewhat faster than the editing proteins, minimally overlapping the DREL (TbMP52) ligase (13). Indeed, there is another endonuclease activity that sediments slightly faster than the editing proteins and is known to efficiently cleave this CYb substrate adjacent to its editing site (33), suggesting that study (13) may not have scored an editing endonuclease. In contrast, our data demonstrate that band II enables the U-insertional endonuclease to productively recognize its editing substrate, a novel activity for an editing protein.

3′-U-exo and TUTase steps.Our assays for the individual 3′-U-exo and TUTase steps suggest that band II plays a recognition role at the second step of editing, specifically in U insertion, much like that at the first step. The initial study of band II (TbMP81) RNAi (13) suggested that both the 3′-U-exo activity in precleaved U deletion and the TUTase activity in precleaved U insertion were diminished, the latter approximately two- to threefold more than the former, as quantitated from the residual unligated reaction intermediate. Their reported decrease in the TUTase intermediate is consistent with our data (Fig. 7A), but we suspect their reported decrease in the 3′-U-exo intermediate (13) is not a direct effect of band II RNAi but rather an indirect effect due to lability of the complex. Supporting this contention, the 3′-U-exo remains fully active in rapid extracts of band II RNAi cells (Fig. 6) yet is diminished during a lengthy extract preparation. Our additional assessment of the basic TUTase enzymatic activity (Fig. 7B) by RET2 (2, 14) demonstrates that its inactivity in U insertion upon band II RNAi is not due to loss or inactivation of the TUTase enzyme (Fig. 7B). Instead, all data support band II facilitating specific recognition by the TUTase.

Product profiles in the basic 3′-U-exo and TUTase enzymatic assays of band II RNAi extracts are also noteworthy. The basic 3′-U-exo activity is more robust in band II RNAi extracts than in control extracts (Fig. 6B), suggesting that exogenous RNA has greater access to the 3′-U-exo in the band II-depleted complex. That could result if band II-depleted complex is more extended than the wild type, as we earlier suggested (28). In the basic TUTase assay, the profile of products is altered upon band II RNAi, with many additional molecules acquiring short U tracts (Fig. 7B). If this U addition is by the RET2 TUTase, which frequently adds short U tracts in U insertion (2, 14), it could also suggest that exogenous RNA has greater access to enzymes upon band II depletion. Alternatively, it has been proposed that U insertion arises from the TUTase generating longer U tracts which are then trimmed by 3′-U-exo (6, 21, 26). If the final U length is determined by this proposed additional step in the U insertion cycle, then the longer U tracts produced in the basic TUTase assay by the RNAi extract (Fig. 7B) would likely reflect impaired coupling of this TUTase step to the 3′-U-exo step, so all four steps of such a U insertion cycle would depend on band II. Although the RET1 TUTase appears separate from the editing complex (2, 14) and generally adds much longer U tracks when assayed for basic TUTase activity (3, 14), we cannot rule out that the short U tracts that arise upon band II RNAi could be due to this TUTase.

Ligation step.Band II is also specific for U insertion at the ligation step. DREL (band IV/TbMP52/TbREL1), which is needed to seal U deletion (12, 19, 39), remains present (Fig. 8C and reference 28) and active in ligating U deletion (Fig. 8A and 9A) upon band II RNAi. In contrast, IREL (band V/TbMP48/TbREL2), which only seals U insertion (12, 39), is lost (as revealed by adenylylation assays that include the deadenylylation step and Western analyses [Fig. 8C and reference 28]). Although U insertion can be sealed by DREL in editing complexes lacking only IREL (13, 16, 28), it is unknown whether DREL appreciably serves this function in the wild-type editing complex. Furthermore, DREL activity in the U-insertional cycle cannot be assessed in the editing complex lacking band II, because assaying for U-insertional ligation requires the TUTase step (13, 16, 28) (which is inactive upon band II RNAi) (Fig. 7A). An earlier study (13), which scored adenylylation of ligases without prior deadenylylation, showed that the IREL (TbMP48) protein was either depleted or preferentially preadenylylated in band II (TbMP81) RNAi extract. However, in their wild-type extract, most of the IREL appeared preadenylylated (13) and the preadenylylated fraction of this ligase should increase upon band II RNAi (28), suggesting that these adenylylation differences (13) may not reflect altered protein abundance. The initial band II (TbMP81) RNAi study (13) additionally showed ligation in precleaved editing assays. Those authors reported somewhat inhibited ligation in precleaved U deletion, which we did not observe and which may again reflect instability of band II-depleted complex in their extract preparation, and twofold more inhibited ligation in precleaved U insertion, which presumably reflects both TUTase inactivity and IREL loss. Thus, all data support band II being important for retention of IREL but not DREL in the editing complex.

As understanding about U deletion and U insertion has grown, models for the editing complex have evolved, oscillating between extremes. When these two forms of processing were originally found to involve three parallel reaction steps (Fig. 1A) (10, 23, 35, 41), it seemed that these parallel steps could share enzymatic activities (17, 46). However, finding that the U-deletional and U-insertional endonuclease activities have different features (8, 11), that the 3′-U-exo is not a reverse TUTase reaction (10, 35), and that the two ligases (35, 37) serve the two forms of editing differently (12, 19, 20, 39) suggested that U deletion and U insertion use separate activities, likely present in different domains of a common editing complex (8, 9, 39). While the two forms of editing could be catalyzed by separate protein complexes (46), such complexes would need to exchange many times during editing of an mRNA, and extensive purification (1, 9, 31, 35, 39) has not separated natural complexes that catalyze only one form of editing. Interestingly, overexpressed tagged ligase proteins have enhanced the abundance and facilitated the isolation of ∼10S subcomplexes, one containing DREL (TbREL1), band III (TbMP63), and band I (TbMP99) that catalyzes precleaved U insertion, and another containing IREL (TbREL2), band II (TbMP81), and TbMP57 (RET2) that catalyzes precleaved U deletion (39). Although neither subcomplex catalyzes cleavage (39), these data attest that at least the later two steps of U deletion and U insertion are catalyzed by different domains of the editing complex. Indicating that all three steps of U insertion utilize a distinct subcomplex, we find that band II is critical for the cleavage as well as the later two steps of U insertion but not for the steps of U deletion (Fig. 3 to 8). Additionally reinforcing the notion of functionally separate subcomplexes for the two forms of editing, all the steps of U deletion remain coordinated in catalyzing the full editing cycle, despite the band II-depleted complex being less stable (Fig. 2) and likely less compact (28) and having defects in all steps of U insertion (Fig. 4, 7, and 8). However, band II depletion causing the U-deletional activities to be more labile (Fig. 2A) and the U-deletional 3′-U-exo to be more accessible (Fig. 6C) implies that the subcomplexes also exhibit significant interdependence within the intact editing complex.