Cotranscriptional Cap 4 Formation on the Trypanosoma brucei Spliced Leader RNA*

mRNA cap formation in trypanosomatid protozoa is mediated through trans-splicing of the capped spliced leader (SL) sequence of the SL RNA onto the 5 * end of all mRNAs. The SL RNA cap structure in Trypanosoma brucei is unique among eukaryotes and consists of 7-meth-ylguanosine (m 7 G) followed by four methylated nucleotides (cap 4): m 7 Gpppm 26 AmpAmpCmpm 3 Um. Using transcriptional arrest in permeable T. brucei cells, we have analyzed the temporal progression of cap 4 formation on the 140-nucleotide-long SL RNA. m 7 G capping of the SL RNA could be detected on prematurely terminated SL RNA transcripts of 56 nucleotides in length and longer. Subsequent modifications characteristic of the SL RNA cap 4 were added successively in a 5 * to 3 * direction and appeared to be independent of core ribonucleoprotein formation. Transcripts between 56 and 67 nucleotides in length were partially modified and carried methyl groups on the first two adenosine residues, whereas a fully modified cap 4 structure was present on transcripts arrested at position 117 and beyond. Taken together, our results are consistent with a cotranscriptional mechanism for generating the cap 4 structure on the SL RNA. The m 7 G 1 cap at the 5 9 end of eukaryotic mRNAs is an essential modification that directs mRNAs to the processing and transport pathways in the cell nucleus and regulates mRNA turnover and translation initiation In all eukaryotes examined, capping is the first bead m g of Molecular glycogen Molecular Biochemicals) °C for h, 3 Hybrids incubated pre- treated streptavidin-agarose beads m l bead

direction and appeared to be independent of core ribonucleoprotein formation. Transcripts between 56 and 67 nucleotides in length were partially modified and carried methyl groups on the first two adenosine residues, whereas a fully modified cap 4 structure was present on transcripts arrested at position 117 and beyond. Taken together, our results are consistent with a cotranscriptional mechanism for generating the cap 4 structure on the SL RNA.
The m 7 G 1 cap at the 5Ј end of eukaryotic mRNAs is an essential modification that directs mRNAs to the processing and transport pathways in the cell nucleus and regulates mRNA turnover and translation initiation (1). In all eukaryotes examined, capping is the first detectable RNA-processing event, occurring cotranscriptionally by the time the transcript is 25 to 30 nucleotides long (2)(3)(4)(5). Capping takes place by a series of three consecutive enzymatic reactions: 5Ј RNA triphosphatase generates a diphosphate terminus, which becomes a substrate for the addition of a guanosine residue through a 5Ј-5Ј triphosphate linkage by RNA guanylyltransferase, and finally, the cap is methylated at the N7 position of guanine by RNA (guanine-7) methyltransferase (6,7). This results in the structure m 7 GpppN, or cap 0, which often is further modified by the addition of 2Ј-O-methyl groups to the first and second transcribed nucleotide, leading to cap 1 and cap 2 structures. Whereas the mechanism of cap 0 formation has been studied in great detail, the biogenesis of the more complex cap structures remains largely unexplored.
In contrast to other eukaryotic organisms, mRNA cap formation in trypanosomatid protozoa occurs by a post-transcriptional RNA processing event. During RNA maturation, trans-splicing transfers the spliced leader (SL) sequence and its cap from the SL RNA to the 5Ј end of all mRNAs. Direct analysis of the SL RNA 5Ј end in Trypanosoma brucei and Crithidia fasciculata revealed that the 5Ј-terminal m 7 G residue is followed by four methylated nucleotides, forming an unusual cap 4 structure: 7-methylguanosine-ppp-N 6 ,N 6 We have previously shown that proper modification of the SL cap 4 structure is essential for utilization of the SL RNA in transsplicing and, therefore, essential for the synthesis of mature mRNAs (9). However, these modifications are not required for the assembly of the SL RNA into a core ribonucleoprotein particle (RNP) nor for the stability of the SL RNA or for the proper folding of the SL RNA in vivo (9).
To further our understanding of cap 4 formation on the T. brucei SL RNA, we have investigated the temporal progression of cap synthesis in vivo. Using transcriptional arrest induced by 3Ј-O-methyl-GTP, we analyzed the 5Ј ends of prematurely terminated SL RNA transcripts in permeable cells. Our data are consistent with a cotranscriptional event for both initial m 7 G cap addition and subsequent methylation of base and sugar moieties of the first four transcribed nucleotides resulting in a cap 4 structure.

Transcription in T. brucei Permeable Cells-Procyclic T. brucei
YTat1.1 cells were permeabilized with lysolecithin as described previously (9,10). Prematurely terminated SL RNA species were generated by adding 3Ј-O-methyl-GTP instead of GTP to the transcription reaction at a final concentration of 200 M. After incubation at 28°C for 10 min, total RNA was extracted with Trizol reagent (Life Technologies, Inc.) and stored as an ethanol precipitate for further analysis.
Affinity Selection of SL RNA Transcripts-Total RNA was subjected to hybrid selection with the RNA oligonucleotide GM01 (5Ј-GGAGCUUCUCAUAC5555A-3Ј, 5 ϭ biotinylated 2Ј-deoxythymidine; all ribonucleotides are 2Ј-O-methylated), complementary to nucleotides 40 to 54 of the SL RNA. Total RNA from approximately 2 ϫ 10 7 cells and 200 ng of GM01 were mixed in 1ϫ PIPES (40 mM PIPES (pH 6.8), 400 mM NaCl, 1 mM EDTA) and 50% formamide in a total volume of 10 l and incubated for 10 min each at 85, 75, 65, 55, 45, 35, and 25°C. Before use, streptavidin-agarose beads (Sigma) were washed 10 times with WB400 (20 mM HEPES, 0.01% Nonidet P-40, 400 mM KCl, 1 mM EDTA), adjusted to 100-l packed bead volume/1 ml of buffer, blocked with 10 g of tRNA (Roche Molecular Biochemicals) and 20 g of glycogen (Roche Molecular Biochemicals) at 4°C for 1 h, then washed 3 times with WB400, and stored at 4°C. Hybrids were incubated with pretreated streptavidin-agarose beads (10-l packed bead volume in 100 l of WB400) on ice for 60 min with occasional mixing. After 5 washes with 500 l of WB400 each, bound RNAs were eluted with 200 l of 2 mM EDTA by incubation at 85°C for 10 min with occasional mixing. If residual counts were present in the bead pellet, the elution procedure was repeated with 100 l of 2 mM EDTA. Eluted RNA was precipitated by adding 10 l of 20ϫ SET (3 M NaCl, 20 mM EDTA, 0.6 M Tris-HCl), 10 g of glycogen, and 2.5 volumes of ethanol.
Immunoprecipitations-Cell extracts for immunoprecipitations were prepared from permeable cells by lysis in 100 l of NET-2 (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.05% Nonidet P-40) containing 20 units of RNase inhibitor. The lysate was spun at 14,000 ϫ g for 10 min, and the supernatant was then incubated with antibodies against the T. brucei common proteins (11) bound to protein A-Sepharose. Total RNA was subjected to immunoprecipitations with rabbit polyclonal antibodies against the m 7 G cap (a kind gift of Dr. F. Richards). After incubation at 4°C for 2 h, the beads were washed 10 times with NET-2, and RNA was prepared with Trizol reagent (Life Technologies, Inc.).
T2 Assay and Thin Layer Chromatography Analysis of the cap 4 Structure-Full-length SL RNA as well as prematurely terminated SL RNA species were excised from denaturing polyacrylamide gels and eluted at 37°C overnight with 400 l of water in the presence of 20 g of tRNA. Eluted RNAs were purified through SPINX centrifuge tube filters (Costar), precipitated with ethanol, and digested with 4 l each of RNase A (800 g/ml) and RNase T2 (80 units/ml) overnight at 37°C in a total volume of 50 l of 50 mM ammonium acetate (pH 4.5) and 2 mM EDTA. Reactions were lyophilized and resuspended in 50 l of water, and 500 l of 10 mM ammonium formate (pH 8) was added. Digestion products were applied to a 100-l DEAE-Sepharose CL-6B column equilibrated with 10 mM ammonium formate. RNA fragments were step-eluted with 3 times 250 l each of 0.2, 0.3, 0.35, 0.4, and 0.5 M ammonium formate. An aliquot of the eluate was fractionated on a 20% polyacrylamide, 7 M urea gel, and fractions containing T2-resistant fragments were pooled and lyophilized. The samples were digested in 50 l of 10 mM ammonium acetate, pH 5.3, with 5 units of nuclease P1 for 3 h at 45°C, and after the addition of 5 l of 1 M NH 4 HCO 3, pH 7.8, and 0.002 units of nucleotide pyrophosphatase, the incubation was continued for 2 h at 37°C. Digestion products were lyophilized and analyzed by two-dimensional thin layer chromatography as described (12).

Transcriptional Arrest with 3Ј-O-Methyl-GTP Produces Distinct, Prematurely Terminated SL RNA Transcripts in Perme-
able Cells-Terminating ribonucleotide analogs have been used by several groups to examine RNA polymerase pausing and 5Ј cap formation in vitro and in nuclear run-on assays (2,3). We decided to apply this approach to permeable trypanosome cells and investigate the progression of cap 4 formation on SL RNA transcripts. To test whether we could prematurely terminate SL RNA transcription, we incubated procyclic trypanosomes permeabilized with lysolecithin (10) in the presence of 3Ј-O-methyl-GTP. Fig. 1A shows a profile of total ␣-32 Plabeled RNA synthesized in a 10-min incubation period (lane 1). The most abundant transcript is the SL RNA of 140 nucleotides, for which we have previously shown that 50 to 70% carries the modifications characteristic of the cap 4 structure. We found that the addition of 200 M 3Ј-O-methyl-GTP to permeable cells significantly altered the profile of newly synthesized RNAs (Fig. 1A, lane 3). In particular, we observed a marked decrease in the accumulation of tRNA-size molecules and the appearance of a series of discrete transcripts in the size range of 50 to 140 nucleotides. To analyze the origin of these RNA species, total RNA samples were subjected to affinity selection with a biotinylated RNA oligonucleotide complementary to nucleotides 40 to 54 of the SL RNA, which are located in the SL intron just downstream of the 5Ј splice site. As shown for control RNA, this procedure specifically and with almost 100% efficiency selected the SL RNA and the free SL intron, which is diagnostic of active trans-splicing (Fig. 1A, lane 2). In addition, high molecular weight RNA was selected, which most likely represents pre-mRNA Y-branched structures that are intermediates in the trans-splicing reaction. Using this method on RNA synthesized in 3Ј-O-methyl-GTP-treated cells, we selected full-length SL RNA, as expected, and most of the newly synthesized transcripts visible in unselected RNA (compare lanes 3 and 4 in Fig. 1A). Electrophoresis of the selected transcripts alongside a dideoxy-GTP sequencing ladder of the cloned SL RNA gene revealed essentially the same pattern of bands and, thus, allowed us to assign each RNA species to a termination event at a particular G residue (Fig. 1B). The difference in mobility between the RNA and DNA ladder most likely reflects the presence of modified nucleotides in the RNA fragments (see below). Taken together, our results are consistent with the hybrid-selected transcripts being prematurely terminated SL RNAs ending at Gs. To further confirm the identity of the selected RNAs, we employed site-directed cleavage with RNase H and a DNA oligonucleotide complementary to the very 5Ј end of the SL RNA (nucleotides [1][2][3][4][5][6][7][8][9][10][11]. This resulted in the specific cleavage of all selected RNA species, except the one corresponding to the SL RNA intron of 100 nucleotides (data not shown).
Cotranscriptional m 7 G Capping of the SL RNA-The 5Ј end of the SL RNA is capped with a 7-methyl GMP moiety, which is linked to the first transcribed nucleotide by a 5Ј-5Ј pyrophosphate bond. To directly look at cap formation in 3Ј-O-methyl-GTP-treated trypanosome cells, oligonucleotide-selected transcripts were digested with tobacco acid pyrophosphatase, an enzyme capable of releasing GMP by cleaving the pyrophosphate bond ( Fig. 2A). As a result, pyrophosphate-treated transcripts should move slightly faster then untreated control transcripts, if properly capped. All selected SL RNA transcripts,  4). RNA samples were separated on a 6% denaturing polyacrylamide gel. The position of the SL RNA and the SL intron is indicated. M, 32 P-labeled MspI digest of pBR322. B, oligonucleotide-selected transcripts (lane 1) were electrophoresed alongside an 35 S-labeled dideoxy-GTP DNA sequencing ladder of the SL leader gene generated with an oligonucleotide primer complementary to the very 5Ј end (lane 2). Products were resolved on a 8% denaturing polyacrylamide gel. Asterisks indicate RNA bands not corresponding to transcript termination sites at G residues; they were not seen consistently. except the linear SL intron, were sensitive to this treatment, resulting in a slight but noticeable increased electrophoretic mobility (compare lanes 1 and 2). Next, we analyzed the methylation status of the cap nucleotide by performing immunoprecipitations with an antibody specific for the m 7 G residue. As shown in Fig. 2B, immunoprecipitations with anti-m 7 G antibodies (lane 2), but neither control precipitation (lanes 3 and 4), displayed the same transcripts that were shown by our oligonucleotide selection procedure to represent prematurely terminated SL RNA transcripts (lane 5). Thus, according to these two assays transcripts of 56 nucleotides in length or greater were capped at the 5Ј end with an m 7 G residue.
cap 4 Formation Takes Place on Nascent Transcripts-The trypanosomatid SL RNA 5Ј end is unusual in that, in addition to the m 7 G cap nucleotide, it contains four consecutive modified nucleotides, forming a cap 4 structure: m 7 G(5Ј)ppp(5Ј)-N 6 ,N 6 ,2Ј- (8). Having established that formation of the m 7 G cap nucleotide occurs on transcripts as short as 56 nucleotides, we next asked whether hypermethylation of the SL RNA 5Ј end also takes place cotranscriptionally. The presence of 2Ј-O modifications at positions 1 to 4 of the SL RNA 5Ј end is revealed by digestion with ribonuclease T2, which cleaves RNA at every position, except pyrophosphate bonds or 5Ј bonds adjacent to 2Ј-O-modified nucleotides. Thus, digestion of fully modified SL RNA with ribonuclease T2 gives rise to a characteristic T2-resistant fragment with the sequence m 7 GpppAACUA, which can be displayed on a denaturing 20% polyacrylamide gel (8,12). Indeed, when full-length SL RNA, synthesized in the presence of 3Ј-O-methyl-GTP, was gel-purified and digested to completion with T2, a T2-resistant fragment characteristic of the cap 4 structure was obtained (Fig. 3, lane 2). Surprisingly, a similar analysis with SL transcripts terminated at position 111 also gave rise to a T2-resistant fragment, albeit with an increased electrophoretic mobility (Fig. 3, lane 3). This result suggested that the 111-nt-long transcript was partially modified at the 5Ј end and raised questions about the timing and progression of cap 4 formation. To address this issue, RNA synthesis in permeable cells was carried out in the presence of three ␣-32 P-labeled nucleotide triphosphates ([␣-32 P]ATP, [␣-32 P]CTP, and [␣-32 P]UTP) and 3Ј-O-methyl-GTP, and the precise nucleotide composition at the 5Ј end of three different size classes (indicated in Fig. 1B) of prematurely terminated SL RNA transcripts was determined. The transcripts are named according to the last G residue they incorporated: G67 refers to a pool of transcripts terminating between positions 56 and 67, G111 includes transcripts terminating between positions 109 and 111, and G127 represents a pool of transcripts terminating between position 117 and 127. Following gel purification and T2 ribonuclease digestion, T2-resistant fragments were separated from mononucleotides by chromatography on DEAE-Sepharose columns (see "Experimental Procedures").
Both G67 and G111 transcripts revealed a T2-resistant fragment that migrated faster in a 20% denaturing polyacrylamide gel than the one characteristic of the cap 4 structure (Fig. 3, lane 3, and data not shown). On the other hand, treatment of G127 RNA with RNase T2 resulted in two resistant fragments: one comigrated with the G67 and G111 fragment (G127-short), whereas the other (G127-long) had the same mobility as the one originating from a fully modified cap 4 structure (data not shown). Next, the various T2-resistant fragments were further digested to mononucleotides with nucleotide pyrophosphatase and nuclease P1 and separated by two-dimensional thin layer chromatography ( Fig. 4 and Table I). As a control, the T2resistant fragment characteristic of mature SL RNA was purified from total T. brucei RNA and processed similarly (panel E). Consistent with previously published TLC analysis (12,13), the five spots of the cap 4 structure were identified as pm 2 6 Am, pAm, pCm, pm 3 Um, and pA (note that pm 7 G is not visible here, since we did not include [␣-32 P]GTP in the transcription reaction). A sixth minor spot (pC), which accounts for less than 5% of the total C residues was also noted, but its origin is unclear.
As shown previously (12) the intensity of the labeled nucleotide spots does not correlate with their representation in the SL cap 4 structure, which should theoretically be 1:1:1:1. This is caused by a difference in the pools of endogenous ribonucleotide triphosphates, which leads to different specific activities for each radiolabeled nucleotide. Indeed, in total RNA the ratio of [␣-32 P]CTP: [␣-32 P]UTP: [␣-32 P]ATP was found to be 9.5: 1.8:1. Correcting for the unequal specific activity of radiolabeled nucleotides, we deduced the following structures for the various T2-resistant fragments. As predicted from the size of the T2-resistant fragment, TLC analysis of G127-long gave rise to nucleotides with mobilities and intensities comparable with   2 and 3), and full-length SL RNA (lanes 1 and 2) and SL transcripts terminated at G111 were gel-purified. After digestion with RNase T2, the T2-resistant fragments (T2R) were analyzed by electrophoresis through a 20% denaturing polyacrylamide gel. Nucleoside 3Ј-monophosphates (Nps) are indicated. those of control RNA (compare panels D and E in Fig. 4). Thus, a proportion of transcripts arrested between positions 117 and 127 carried a fully modified cap 4 structure. In contrast, G127short displayed seven nucleotide spots (one more than in control RNA), and the relative intensities of several spots were significantly different when compared with control RNA (compare panels C and E). In particular, 80% of cytidine was present as unmodified pC, and only 20% was modified to pCm. In contrast, in control RNA about 95% of cytidine was present as pCm. Furthermore, small amounts of pm 3 Um and unmodified pU were detected. This result is most consistent with the presence in G127-short of three different T2-resistant fragments: m 7 Gpppm 2 6 AmpAmpC, m 7 Gpppm 2 6 AmpAmpCmpm 3 UmAp, and to a lesser extent, m 7 Gpppm 2 6 AmpAmpCmpUp. Within the limits of our analysis the nucleotide composition of G111 and G67 appeared to be very similar, and both revealed three major differences when compared with that of control RNA (compare panels A and B with panel E in Fig. 4). First, pm 3 Um was not detectable. Second, a novel nucleotide spot was visible just below pAm. According to its relative mobility, we tentatively identified this spot as pm 6 Am (14), which could be an intermediate in the synthesis of pm 2 6 Am. Third, pC contained the majority of the [␣-32 P]CTP label (ϳ97%), and less than 3% was present as pCm. One conclusion from these data was that in addition to the m 7 G cap, only A and C residues were represented in the T2-resistant fragments of G111 and G67 transcripts. Since pU was lacking from the chromatogram, a proportion of these T2-resistant fragments must contain the sequence AAC. Furthermore, as pA in the fully modified cap 4 structure can only be derived from the adenosine residue at position 5 of the T2-resistant fragment m 7 Gpppm- 2 6 AmpAmpCmpm 3 UmpAp, a proportion of the G67 and G111 fragments should have the sequence AA. Taking all of the above into consideration, our data are consistent with the presence of multiple T2-resistant fragments with the following proposed structures: m 7 GpppAmpAmpCp, m 7 Gpppm 6 AmpAmCp, m 7 Gpppm 2 6 AmAmCp, and m 7 Gpppm 2 6 AmpAp. Finally, the presence of pCm, albeit at a very low level (less than 3%, when compared with pC), also indicated that in some cases the cytidine nucleotide at position 4 was also methylated, suggesting an even more complex mixture of different methylation patterns. cap 4 Synthesis Is Independent from Binding of Common Proteins to the SL RNA-We previously showed that for in vitro cap 4 modification only the core SL RNP and not deproteinized SL RNA was used as a substrate (12). This suggested that either one, or more of the SL RNP components is recognized by the cap 4 biosynthetic machinery or that a structural determinant or a specific sequence at the SL RNA 5Ј end is masked in naked RNA. To test whether SL RNA cap 4 formation in permeable cells is dependent on assembly of the core SL RNP proteins with the SL RNA, we performed immunoprecipitation assays on total cell extracts of 3Ј-O-methyl-GTP-treated cells with an antibody specific for the common proteins, which associate with both the SL RNA and the major U small nuclear RNAs of trypanosomes (11). As shown in Fig. 5, full-length SL RNA as well as the linear SL intron was immunoprecipitated. Furthermore, transcripts arrested at position 120 and beyond were present in the precipitate, arguing that they were bound by common proteins. Faint bands comigrating with G100-and G90 -91-terminated transcripts were also observed, but their intensity varied from experiment to experiment. Since modified nucleotides characteristic of the cap 4 structure begin to appear in transcripts as short as 56 nt, these observations indicated that common protein binding is not a prerequisite for cap 4 modification.

DISCUSSION
Here we describe the temporal progression of cap 4 formation on the SL RNA in permeable T. brucei cells by taking advantage of the incorporation of the transcription terminator 3Ј-Omethyl-GTP into the growing RNA chain. This allowed us to arrest transcription at guanosine residues downstream of the transcription start site and then to determine the state of modification at the 5Ј end of these transcripts.
Using oligonucleotide-selected RNAs, we found that SL RNA transcripts of 56 nucleotides or greater contain an m 7 G cap structure (Fig. 2, A and B), and immunoprecipitations with anti-m 7 G antibodies indicated that transcripts as short as 30 nucleotides are capped ( Fig. 2 and data not shown). These results are consistent with previous reports on mRNA cap formation in other systems. In a Drosophila nuclear run-on analysis cap addition was observed on transcripts between 20 and 30 nucleotides in length (3), and similarly, in a vaccinia virus system Hagler and Shuman report capping on nascent RNA chains 31 nucleotides in length or longer (2). More recent studies have explored the implications of these findings and demonstrated that formation of the cap early in transcription is mediated by recruitment of the capping machinery to the phosphorylated carboxyl-terminal domain of the largest subunit of RNA polymerase II (15)(16)(17)(18). At present the RNA polymerase transcribing the SL RNA gene has not been identified unambiguously, although there is a strong suggestion that it might be RNA polymerase II (19,20). It is conceivable that, similar to FIG. 4. Analysis of cap 4 constituents of 3-O-methyl-GTP-terminated SL RNA transcripts. T2-resitant fragments from a control reaction (panel E) or from three prematurely terminated SL RNA transcripts (panels A-D) were purified over DEAE-Sepharose columns and digested with nuclease P1 and nucleotide pyrophosphatase. The digestion products were separated on cellulose TLC plates as described (12). T2 digestion of G127 revealed two T2-resistant fragments (G127short and G127-long).
other systems, there is a specific interaction between the capping machinery and the SL RNA transcription apparatus, either through the RNA polymerase large subunit or another component of the transcriptional machinery. Following m 7 G addition, the 5Ј end of the SL RNA is further modified by the addition of seven methyl groups, giving rise to a cap 4 structure (8). Our data presented here provide a detailed look at the progression of cap 4 formation in vivo. Modification of transcripts in the size range of 56 to 111 nucleotides was mostly restricted to the first two adenosine residues, whereas in longer transcripts (117 to 127 nucleotides) both partially and fully modified cap 4 structures were present. The progression of cap 4 modification was most readily noticeable through the gradual acquisition of the modified cytidine residue (pCm), barely visible in transcripts arrested between G67 and G111 but clearly present in transcripts longer than 111 nucleotides. Similarly, modification to N 3 ,2Ј-O-dimethyluridine appeared to be restricted to transcripts longer than 111. Thus, one implication from our results is that conversion of the initial m 7 G-capped transcript to the fully modified cap 4 structure (m 7 Gpppm 2 6 Ampm 6 AmpCmpm 3 Um) takes place cotranscriptionally and proceeds sequentially in a 5Ј to 3Ј direction. If, as our data suggest, cap 4 biosynthesis is cotranscriptional, it is tantalizing to speculate that the cap 4 methyltransferases engage in an interaction with the SL RNA transcriptional com-plex. This is an area of investigation that we are actively pursuing.
The vectorial addition of methyl groups onto the SL RNA 5Ј end is likely to reflect substrate requirements of the different methyltransferases that carry out the modifications. The vectorial nature of this process would ensure that only correctly modified SL RNA is produced by requiring that, for instance, the addition of the methyl group at the N3 position of uridine can only take place when the preceding nucleotides are fully modified.
Immunoprecipitation experiments with anti-common protein antibodies revealed that both m 7 G capping as well as methylation of the first four nucleotides is independent of the SL RNA being assembled into a core RNP. In particular, most transcripts in the size range of 56 to 111 nucleotides, which were capped and partially modified, were not associated with common proteins. Only SL RNA transcripts longer than 112 nucleotides were precipitable with anti-common protein antibodies. This result helps to clarify an interesting aspect we observed previously in our in vitro cap 4 modification system (12). It appeared that cap 4 formation required the SL RNA substrate to be in an RNP, although at the time we did not exclude the possibility that a structural determinant or a specific sequence at the 5Ј end becomes masked in deproteinized SL RNA. The observation in the present study that the onset of cap 4 formation precedes binding of common proteins would argue for the latter scenario. Consistent with this conclusion is an in vivo mutational analysis of the Leptomonas SL RNA (21). Sequence elements necessary for cap 4 modification were found to reside exclusively within the modification site itself and stem loop I, whereas substitution or deletion of stem loops II and III had no effect on cap formation. The observation that common protein binding occurs when the SL RNA is at least 112 nt long suggests that the synthesis of the structural determinant for common protein binding is not yet completed in transcripts shorter than 112 nt. Interestingly, transcripts shorter than 112 nt do not contain the entire single-stranded region between nt 110 and 120 of the SL RNA, which has been proposed to be the region analogous to the Sm binding site of spliceosomal U small nuclear RNAs of other eukaryotic organisms. Last, our results have implications for the biogenesis of the SL RNA. Since cap 4 synthesis and assembly with the SL RNP proteins occurs on prematurely terminated transcripts, SL RNP biogenesis most likely occurs in the nucleus, as we have previously suggested.  Fig. 4 were scanned with phosphorimaging, and the relative intensities of the various spots are given in arbitrary units (first number). The numbers in parentheses are corrected for the unequal specific activity of radiolabeled nucleotides (see "Results").