The Trypanosomatid Signal Recognition Particle Consists of Two RNA Molecules, a 7SL RNA Homologue and a Novel tRNA-like Molecule*

Trypanosomatids are ancient eukaryotic parasites affecting humans and livestock. Here we report that the trypanosomatid signal recognition particle (SRP), unlike all other known SRPs in nature, contains, in addition to the 7SL RNA homologue, a short RNA molecule, termed sRNA-85. Using conventional chromatography, we discovered a small RNA molecule of 85 nucleotides co-migrating with theLeptomonas collosoma 7SL RNA. This RNA molecule was isolated, sequenced, and used to clone the corresponding gene. sRNA-85 was identified as a tRNA-like molecule that deviates from the canonical tRNA structure. The co-existence of these RNAs in a single complex was confirmed by affinity selection using an antisense oligonucleotide to sRNA-85. The two RNA molecules exist in a particle of ∼14 S that binds transiently to ribosomes. Mutations were introduced in sRNA-85 that disrupted its putative potential to interact with 7SL RNA by base pairing; such mutants were unable to bind to 7SL RNA and to ribosomes and were aberrantly distributed within the cell. We postulate that sRNA-85 may functionally replace the truncated Alu domain of 7SL RNA. The discovery of sRNA-85 raises the intriguing possibility that sRNA-85 functional homologues may exist in other lower eukaryotes and eubacteria that lack the Alu domain.

The signal recognition particle (SRP) 1 is a cytoplasmic ribonucleoprotein (RNP) that couples the synthesis and targeting of secretory and membrane proteins to translocation across the endoplasmic reticulum (ER) membrane (1,2). SRP binds to the signal sequence of a pre-secretory protein as it emerges from the ribosome and delays further translation, a process known as the elongation arrest. After interaction of the SRP with the membrane-bound SRP receptor (SR␣), this pause in translation is relieved and the protein is co-translationally translocated into the ER through the translocon (1,2). This translational pause is thought to prevent misfolding or aggregation of the non-targeted proteins.
SRP homologues have been identified in organisms of all three kingdoms, but they differ considerably. The archeal 7SL RNA resembles the higher eukaryotic RNA, whereas the SRP RNA in most eubacteria is shorter and carries only domain IV, which is the binding site of the signal peptide-binding protein (3). It was proposed that the ancestral SRP molecule was reduced in size during the evolution of bacteria (3).
The mammalian SRP is composed of six proteins and a single RNA molecule, the 7SL RNA (4). Based on in vitro studies using canine microsomes, it has been demonstrated that SRP54 binds the signal peptide as it emerges from the ribosome. SRP54 also binds to the 7SL RNA, and it is a GTPase. SRP19 is thought to facilitate the association of SRP54 with the RNA (5). SRP9/14 binds to domain I, also known as the Alu domain, and functions in elongation arrest. SRP68/72 mediates the interaction of SRP with SRP receptor (6). In yeast, the SRP is also composed of six proteins, Srp72p, Srp68p, Sec65p (SRP19 homologue), Srp54, Srp21, and Srp14p. No homologue of yeast SRP9 exists. Four proteins (14,21,68, and 72) are required for the stable expression of SRP, suggesting that, together with SRP RNA, they build a stable core particle to which Sec65p and Srp54p can bind (7).
Little is known about the biogenesis of SRP. SRP RNA is transcribed by RNA polymerase III, and, in the case of mammalian SRP RNA, it undergoes limited processing at the 3Јend; three uridylates are removed and a single adenylate is added (8). Endogenous SRP RNA was detected in the nucleolus (9). Further studies in yeast showed that the nuclear export of SRP RNA is a carrier-mediated process that depends on the presence of the Alu domain. Assembly of the nuclear exportcompetent SRP takes place in the nucleolus and requires four core SRP proteins that are actively imported into the nucleus by the ribosomal import pathway. Transport into the cytoplasm involves the nuclear export factor Xpo1p (10,11).
One of the most important functions of SRP is elongation arrest. This function is confined to the Alu domain, because SRP sub-particles lacking the Alu domain have lost the elongation arrest capacity (12). It is currently unknown how SRP elicits the pause of translation. It was proposed that, because of the structural similarity of the Alu domain to tRNA, it may compete with incoming tRNA to the A site of the ribosome (13). Experimental data demonstrated that SRP actually interacts with ribosome after the transpeptidylation reaction and before translocation of the peptidyl-tRNA from the A to the P site (14). The role of the elongation arrest in protein translocation in vivo remains to be clarified. Recent studies with Saccharomyces cerevisiae have demonstrated that SRPs carrying a truncated Srp14p did not maintain elongation arrest both in vitro and in vivo, suggesting the coupling of protein translation and translocation (15). Surprisingly, most of the bacterial SRP RNAs lack the Alu domain (3).
Recent studies on the structure of the Alu domain of mammalian SRP have indicated that the Alu domain consists of two helical stacks connected by a central U-turn, a structural motif that is termed the -junction due to its shape. This structure resembles other RNAs such as the hammerhead ribozyme and the part of 23 S RNA that binds to L11 (16). In yeast, both Yarrowia lipolytica and Schizosaccharomyces pombe, the Alu domain is severely reduced and only contains the U-turn loop. The size and shape of the Alu domain are very variable; the only part that is conserved is the U-turn, being the part recognized by SRP9/14. The protein component of the Alu domain also varies across the three kingdoms, because S. cerevisiae possesses a homodimer of SRP14 (17), whereas in Bacillus subtilis the dimeric histone-like HBsu protein binds to the -junction (18).
Trypanosomatids constitute a diverse family of parasitic protozoans that possess exotic RNA processing mechanisms such as trans-splicing (19) and RNA editing (20). These processes may reflect the evolutionary position of trypanosomatids being descendants of one of the deepest branches of the eukaryotic lineage (21).
Very little is known about protein translocation in trypanosomatids. Previous studies have demonstrated the existence of a 7SL RNA homologue in trypanosomatids. Structurally, the trypanosome 7SL RNA homologue resembles the mammalian 7SL RNA except for its Alu-like domain, which is truncated, lacking one of the loops and the potential to form the tRNA-like pseudoknot (22,23). The trypanosomatid Leptomonas collosoma 7SL RNA exists in the cell in two stable conformations that change during the translocation cycle. This conformational change is associated with a novel RNA editing of C to U in position 133 located in domain III (24). Our previous studies demonstrated that a small RNA (sRNA-76) was co-purified with Trypanosoma brucei 7SL RNP. Importantly, sRNA-76 was identified as a tRNA-like molecule (25). This finding led us to hypothesize that trypanosome SRP may differ from all other known SRPs and is composed of two small RNAs (25).
Recently, progress was made in elucidating the trypanosome SRP complex. More specifically, homologues to SRP54, -72, -19, and -68 were identified in the T. brucei genome project, but no homologues to the Alu domain proteins SRP9 and -14 were found. 2 RNA interference was used to silence the SRP54 in T. brucei. SRP54 depletion inhibited growth and cytokinesis, suggesting that the SRP pathway is essential. The translocation of signal peptide-containing proteins was investigated. Surprisingly, these tested proteins were translocated into the ER and properly processed. However, several of them were mislocalized and accumulated in mega vesicles, most likely due to a secondary effect on protein sorting. The translocation of these proteins to the ER upon SRP54 depletion suggests that an alternative pathway for protein translocation must exist in trypanosomes (26). It is therefore conceivable that, as in Escherichia coli, the substrates of the trypanosome SRP pathway consist mainly of polytopic membrane proteins and not the signal peptide-containing proteins like in higher eukaryotes (27,28).
In this study we have provided direct evidence that the trypanosomatid L. collosoma SRP is composed of two separate RNA molecules: 7SL RNA and sRNA-85, a novel tRNA-like molecule. sRNA-85 interacts with 7SL RNA by base pairing, and together these RNAs comprise a small RNP of ϳ14 S that interacts with ribosomes. Disruption of the potential base pairing between sRNA-85 and 7SL RNA completely abolished sRNA-85 interaction with 7SL RNA; the RNA was found at the top of the gradient with the bulk of the cellular tRNAs. In situ hybridization suggests that both 7SL RNA and sRNA-85 are mainly confined to the cytoplasm but also exist in the nucleolus. Being a tRNA-like molecule, sRNA-85 may assist in the arrest function of the SRP. The trypanosomatid SRP may represent the first attempt to separate the arrest function of the SRP from the signal recognition and translocation promotion entities of the particle. This novel tRNA-like molecule, like the Alu domain, may also function in the transport of the SRP particle from the nucleolus to the cytoplasm. Functional homologues of sRNA-85 may be present in other lower eukaryotes and prokaryotes that either carry a truncated Alu domain or lack this domain entirely. PCR to generate the probes for in situ hybridization; 8721, 5Ј-CGG-GATCCAGCCGGAGCCTTGCTC-3Ј, sense, covering positions 1-16 of the 7SL RNA with a BamHI site, used for PCR to generate the probes for in situ hybridization; 12407, 5Ј-AGCTATATCTCTCGAA-3Ј, antisense, complementary to positions 83-98 of U6 snRNA, used for Northern analysis; 5303, 5Ј-GCAGAGCACCACGTCAACGC-3Ј, antisense, complementary to positions 82-101 of 7SL RNA, used for Northern analysis; and SM-04, 5Ј-GGCTTAGAGGCGTTCAGCCGAGAACCAC-CGTGCGA-3Ј, antisense, complementary to the 3Ј-end of rRNA1, used for Northern analysis.

Oligonucleotides
Extract Preparation and Purification of the SRP Complex-L. collosoma growth and extract preparation were as previously described (29). Cells were disrupted using nitrogen cavitation, and RNPs were extracted at 0.4 M KCl. DEAE-Sephacel column chromatography was as previously described (30). The flowthrough fraction from the DEAE-Sephacel was concentrated on a 0.5-ml DEAE-Sephacel column by binding at 50 mM KCl and eluting the particles at 0.4 M KCl. The KCl eluate was layered on continuous 10 -30% (w/v) sucrose gradients. Gradients were centrifuged at 4°C for either 3 or 22 h at 35,000 rpm using a Beckman SW41 rotor.
Sequencing of sRNA-85 and Cloning of Its Coding Gene-RNA was extracted from the sucrose peak fraction carrying sRNA-85, and samples (0.1 g) were 3Ј-end-labeled with [5Ј-32 P]pCp (3000 Ci/mmol) (31). The labeled RNA was fractionated on an 8% polyacrylamide, 7 M urea denaturing gel. The eluted small RNA was subjected to sequencing using base-specific nucleases as previously described (29). The alkaline partial hydrolysis ladder was produced by boiling the RNA in 50 mM NaOH and 1 mM EDTA for 2 min. The labeled RNA (250,000 cpm) was also used to screen a genomic library (32). Two 700-bp fragments derived from digesting the DNA with RsaI and Sau3A were subcloned into pBluescript and were sequenced from both strands using T3 and T7 primers.
Affinity Purification of the SRP Complex Using 2Ј-O-Methyl Biotinylated Oligonucleotide-The extracts or column fractions for affinity selection were in buffer A containing 400 mM KCl. Oligonucleotide 85-R-1, 5Ј-GATCCAGGACTCGAACCXXXXA-3Ј, complementary to positions 58 -74 of sRNA-85, is composed of 2Ј-O-methyl nucleotides except X, which is biotinylated 2Ј-deoxythymidine. Affinity selection was performed essentially as previously described (33) except that hybridization with the RNPs was performed in 200 -250 mM KCl.
Acidic Gels for Examining the Charging Status of sRNA-85 and tRNA Molecules-RNA was prepared from L. collosoma cells under acidic conditions (35). Briefly, cells (5 ϫ 10 8 ) were pelleted at 4°C, resuspended in 0.3 ml of 0.3 M sodium acetate (pH 4.5) and 10 mM EDTA. The RNA was subjected to two extractions with equal volumes of phenol equilibrated with the same buffer. After ethanol precipitation, the RNA was mixed with sample buffer (0.1 M sodium acetate (pH 5.0), 8 M urea, 0.05% bromphenol blue and 0.05% xylene cyanol) and fractionated on a 15% polyacrylamide gel containing 8 M urea in 0.1 M sodium acetate buffer (pH 5.0). To discharge the tRNA, we treated the RNA sample with a mild alkaline by incubating the RNA in 0.2 M Tris-HCl (pH 9.6) at 37°C for 20 min.
Plasmid Construction and Transformation-sRNA-85 mutants were generated by PCR mutagenesis using primers carrying the mutations and oligonucleotides situated upstream and downstream of the gene. The different oligonucleotides used for the PCR amplification were previously listed. The mutated sRNA-85 genes were subcloned into the BamHI site of the pX-neo expression vector. All mutations were confirmed by DNA sequencing. Stable cell lines carrying the mutated genes were established as described previously (32).
UV Cross-linking of Living Cells Treated with AMT-Psoralen-L. collosoma cells were harvested at 5 ϫ 10 7 cells/ml and washed with PBS. Cells (5 ϫ 10 9 ) were concentrated and incubated on ice. AMT (Sigma) was added to the cells at a concentration of 0.2 mg/ml (36). Cells treated with AMT-psoralen were kept on ice and irradiated using a UV lamp at 365 nm with an intensity of 10 milliwatts/cm 2 for 30 min. RNA was prepared from the cells using TRIzol reagent and subjected to the splint-labeling technique.
In Situ Hybridization-Hybridization probes specific to 7SL RNA and sRNA-85 were synthesized by PCR amplification. For the synthesis of the 7SL RNA probes, two oligonucleotides, 15137 and 8721, were used, and for the production of the sRNA-85, oligonucleotides 16721 and 18488 were used. PCR was performed under standard conditions using the expand high fidelity enzyme (Roche Molecular Biochemicals), but the dNTPs mixture contained 200 M of each of the nucleotides except dTTP (130 M) and digoxigenin-11 (DIG)-dUTP (70 M). For labeling the mutant-specific oligonucleotides, the oligonucleotides were tailed with DIG-dUTP using the enzyme terminal transferase (Roche Molecular Biochemicals) as described by the manufacturer. The conditions for fixation and hybridization were similar to those published previously (37).
Cells in mid-logarithmic phase were washed with PBS (0.15 M sodium chloride, 10 mM sodium phosphate, pH 7.2) and resuspended in PBS before fixation. The cells were then diluted 1:2 in fixation solution (1ϫ fixation solution: 4% formaldehyde and 5% acetic acid in PBS) and incubated at room temperature for 20 min while rotating. The fixed cells were centrifuged for 10 min at 3000 ϫ g, and the fixation medium was replaced by 70% ethanol, followed by two additional washes in 70% ethanol to remove all traces of formaldehyde. Microscopic slides were prepared by dropping 20 l of fixed-cell suspension on glass slides. Slides were allowed to air-dry and were then baked at 80°C for 10 min to improve the adherence. Before hybridization, the cells were subjected to partial hydrolysis and proteolysis by using 0.1% pepsin in 0.01 M HCl for 5 min at 37°C. After treatment, the cells were washed with 2ϫ SSC, dehydrated with ethanol, and then air-dried. The DIG-labeled PCR product was in 60% deionized formamide, 2ϫ SSC, 50 mM sodium phosphate buffer, and 500 g/ml denatured salmon sperm DNA. Hybridization was performed at 37°C for 24 h. When DIG-labeled oligonucleotides were used, hybridization was performed at 25°C. After hybridization, the slides were rinsed three times for 20 min each in 50% formamide/2ϫ SSC at 37°C. Finally, slides were washed two times for 5 min at room temperature. For detection of the DIG-labeled probes, slides were incubated for 45 min at 37°C with 1:200 diluted fluorescein isothiocyanate-conjugated mouse anti-DIG (Roche Molecular Biochemicals). To stain the nucleus and kinetoplast, we incubated the cells with propidium iodide (10 g/ml) for 15 min. The cells were imaged by an Olympus Vanox AHBT3 microscope equipped with a fluorescence lamp and ϫ100 objective, and image analysis was performed using an Olympus DP50 camera.

RESULTS
Purification of the L. collosoma SRP Complex and Identification of sRNA-85-To study the composition of the trypanosomatid SRP complex, we devised a scheme to purify the particles. Whole-cell extracts were prepared in high salt concentration to assure the dissociation of the SRP from the ribosomes. Enrichment for small RNPs, including the SRP, was achieved by removing the ribosomes, and the resulting postribosomal supernatant (PRS) contained U2 snRNA, 5S RNA, 7SL RNA, and tRNAs (Fig. 1A, lane 1). A purification step on a DEAE-Sephacel was able to separate the SRP population into two types: (1) particles that did not bind to the column and were found in the flowthrough and wash fractions (lanes 2 and 3) and (2) particles that were bound to the column and were eluted at 0.4 M KCl (lanes 4 -6). The samples from the column fractionation were subjected to Northern analysis with the 7SL RNA probe (lower part of Fig. 1A). Quantitation of the Northern analysis indicates that only 19.5% of the SRP was bound to the column. The L. collosoma 7SL RNA exists in the cell in two stable conformations. We have previously demonstrated that 7SL I is preferentially bound to ribosomes, whereas 7SL II is found mainly in free RNPs (24). A great variation was observed with respect to the binding of different small particles to the DEAE-Sephacel column, except for the SRP particles, which in all our preparations were unable to bind to the column. The inability of the SRP complex to bind to the column matrix may stem from the fact that most of the RNA is inaccessible for interacting with the column, because it is either involved in intramolecular base pairing or bound to proteins. In several preparations, it was possible to enrich flowthrough fractions that were completely depleted of almost all small RNPs except for the SRP and other small RNAs in the size range of 60 -90 nt.
Such a fraction was further used to purify the SRP complex. In the last purification step, concentrated flowthrough column fractions were separated by using two sequential sucrose gradients, consequently yielding a fairly homogeneous population of ϳ12 S particles enriched for the two 7SL RNA forms and for an RNA of about 85 nt (Fig. 1B). The small RNA was termed sRNA-85.
To identify the sRNA-85 molecule, we labeled RNA from the sucrose-gradient peak fractions at the 3Ј-end with pCp and T4 RNA ligase and subjected it to enzymatic sequencing (Fig. 1C). Several gaps were observed in the partial alkaline ladder (marked with stars in Fig. 1C), indicating the presence of ribose-methylated nt in sRNA-85. Unfortunately, with the partial RNA sequencing data, we were unable to identify it, and to obtain the complete sequence of sRNA-85, the coding gene had to be cloned from an L. collosoma genomic library. Two subclones of 700-bp derived from DNA carrying sRNA-85 with Sau3A and RsaI were used for the sequencing. The complete sequence of sRNA-85 is presented in Fig. 2A. Sequence alignment showed a 90% identity between the gene sequence in Fig.  2A and the RNA sequence presented in Fig. 1C. RNase H cleavage with oligonucleotides complementary to either the 3Јor the 5Ј-end of the RNA indicated that sRNA-85 is a single transcript (data not shown). A GenBank TM search using the entire sequence presented in Fig These regions are confined to the T⌿C stem-loop. Interestingly, no significant homology was found for the 5Ј-half of the molecule. However, sRNA-85 can be folded like a canonical tRNA based on tRNAscan-SE 1.1 (Fig. 2B). The sequence of sRNA-85 was recognized as a tRNA molecule using this program, but the program suggested deleting nt 38 -44. However, this is not a valid suggestion, because these bases were detected in the direct RNA sequencing presented in Fig. 1C. A comparison of the tRNA-like molecule with the canonical tRNA structure indicates that the deviations from the tRNA structure are confined to the anti-codon stem. Moreover, in the canonical tRNA, there are 5 base-paired nt comprising the anti-codon stem, whereas in sRNA-85, there are seven base-paired nt with two bulges. The prediction of the location of the amino acid anti-codon is based on the presence of a conserved uridine and a pyrimidine upstream to it and a purine downstream to the anti-codon. In sRNA-85 the nt upstream to the anti-codon agreed with the consensus, however, a U instead of a purine is present downstream to the anti-codon.
Affinity Purification of the SRP Complex Using Antisense 2Ј-O-Methyl Biotinylated Oligonucleotide to sRNA-85-The copurification of sRNA-85 with 7SL RNA through several steps of purification and its co-release together with the 7SL RNA from the ribosomes using puromycin (data not shown) suggest that these two RNAs co-exist in the same RNP particle. To rigorously examine this hypothesis, we used an affinity selection technique. Briefly, fractions enriched with 7SL RNP were subjected to affinity selection with 2Ј-O-methyl biotinylated oligonucleotide complementary to the 3Ј-end of sRNA-85. After the unbound materials were extensively washed, RNA was eluted from the column, and the samples were analyzed on a 10% polyacrylamide denaturing gel. The results, presented in Fig.  3A, indicated selecting sRNA-85 did not only result in its own selection but, rather, selected its tightly bound partner, the 7SL RNA (Fig. 3A, panel a, lane 7). Because 7SL RNA is stained less efficiently with silver compared with sRNA-85, the co-selection was quantitated. RNA from an identical experiment (shown in Fig. 3A, panel a) was subjected to Northern analysis with 7SL RNA and sRNA-85 probes. The results indicated that sRNA-85 and 7SL RNA were co-selected in a 1:1 ratio (Fig. 3A, panel b). In selecting a control for the specificity of the selection procedure, we examined the partitioning of Gradients were centrifuged at 4°C for 22 h at 35,000 rpm using a Beckman SW41 rotor, and the peak fraction was collected and dialyzed against buffer A and re-fractionated on a sucrose gradient. RNA was analyzed on a 10% denaturing gel and visualized by silver staining. The two forms of 7SL RNA and sRNA-85 are indicated with arrows. The S value markers were tRNA (4 S), catalase (11 S), and rRNA (28 S). C, a partial enzymatic sequence determination of sRNA-85. RNA obtained from the peak fraction, presented in B, was labeled at the 3Ј-end with pCp and RNA ligase and subjected to partial digestion with base-specific RNases (29). A partial alkaline hydrolysis ladder was produced by boiling the RNA in 50 mM NaOH and 1 mM EDTA for 2 min. Lanes: G, guanine-specific RNase T1 digest; A, adenine-specific RNase U2 digest; AU, adenine-uridine-specific RNase phyM digest; CU, cytosine-uridine-specific RNase digest; C, cytosine-specific RNase CL3 digest; L, partial alkaline hydrolysis ladder. The RNA sequence is indicated on the right. Positions carrying 2Ј-O-methyl groups are indicated with stars, and regions of sequence compression are indicated by triangles. Panels a and b represent two separate runs (a is the first run and b is the second run, and the sequence is presented from 3Ј-to 5Ј-end).
snoRNA-2, an abundant unrelated small nucleolar RNA (38), also enriched in this fraction. As shown in Fig. 3A (panel b), the results indicated that the affinity procedure was specific, because snoRNA-2 was not selected and only two RNAs, sRNA-85 and 7SL RNA, were affinity-purified. The specificity of co-purification was also observed when a crude post-ribosomal supernatant (PRS) containing all the cellular small RNPs was used in the experiment. The samples were analyzed as previously described, except that the gel was stained with ethidium bromide. The results suggest that even when a crude prepara- tion was used to select the particle, both RNAs, 7SL RNA and sRNA-85, were co-selected in a 1:1 ratio (Fig. 3B). Interestingly, as shown in (Fig. 3, A and B), sRNA-85 associates with the two forms of the 7SL RNA.
Because we only detected sRNA-85 in association with 7SL RNA, the two RNAs must associate soon after transcription. Moreover, because sRNA-85 associates with both 7SL RNA conformations, the binding and release of SRP from the ribosome did not seem to affect the interaction of these two RNAs.
sRNA-85 Is a tRNA-like Molecule That Possesses a CCA Sequence at the 3Ј-End but Is Uncharged-As shown in Fig. 2B, sRNA-85 carries almost all the conserved tRNA nt. The predicted anti-codon of sRNA-85 is that of asparagine. The only part of the molecule that deviates strongly from the canonical structure is the anti-codon stem-loop. In the canonical structure, the anti-codon stem is composed of five base-paired nt, whereas sRNA-85 possesses a 7-bp stem that is disrupted by two bulges. In addition, the C residue in position 35 does not carry a 2Ј-O-methyl group as expected, and position 40, located 1 nt downstream from the anti-codon, is occupied by a U instead of a purine (39). Based on these observations, the actual structure of this domain may differ from that theorized according to the canonical structure (Fig. 2B). The gaps in the alkaline ladder, presented in Fig. 1C, are consistent with the presence of three 2Ј-O-methyl ribose nt: Gm18, Am40, and Um52. Gm18 and Um52 are canonically modified positions in tRNA molecules (39).
All tRNAs carry a CCA sequence at the 3Ј-end and an amino acid linked to the CCA. To examine whether sRNA-85 possesses a CCA sequence at the 3Ј-end, we used the splint labeling technique, because it enables labeling of the RNA only if the oligonucleotide used as a template exactly complements the 3Ј-end of the RNA (34). In this technique, a primer complementary to the RNA carrying runs of T nucleotides at the 3Ј-end of the oligonucleotide was used as a template to incorporate [␣-32 P]ATP at the 3Ј-end of the RNA using the enzyme T7 polymerase. Two antisense primers to sRNA-85 with CCA (20404) and without CCA (25109) were used in the splint labeling experiments. The results, presented in Fig. 4A, clearly show that labeling was achieved only with oligonucleotide carrying a CCA-complementary sequence, indicating that sRNA-85 indeed contains CCA at its 3Ј-end.
The aminoacylation status of sRNA-85 was analyzed by a method that enables separation of aminoacyl-tRNA from its uncharged form (35). RNA was prepared directly from cells under acidic conditions and subjected to Northern analysis. As a control, the RNA was subjected to mild alkaline hydrolysis to deacylate the tRNAs. A ratio of 1:1 was observed between the charged and uncharged forms of control tRNAs, namely tRNA Arg , tRNA Gln , and tRNA Lys (Fig. 4B). However, this analysis revealed only one form of sRNA-85 under acidic conditions, and mild alkaline hydrolysis did not affect its mobility (Fig. 4B, compare lanes 1 and 2 probed with sRNA-85), suggesting that most of sRNA-85 is uncharged. However, note that the assay used in this study does not directly demonstrate the capacity of sRNA-85 to be charged. This assay has been usually used to examine the acylation status of tRNAs (40); however, one major concern in using this assay is that it measures the steady-state level of acylation, and different tRNAs may be deacylated at different rates. Thus the level of charged tRNAs may not only reflect the ability of tRNAs to become charged but also the rate by which they are deacylated. In the absence of an in vitro system for aminoacylation in trypanosomatids, it is not possible to directly assess the ability of sRNA-85 to be charged by radiolabeled amino acids.
7SL RNA and sRNA-85 Interact by Base-pairing-sRNA-85 and 7SL RNA may interact directly through base pairing of complementary sequences or held together by protein interactions. Using the computer program BestFit, we identified a remarkable potential for base-complementary between the 5Јend of the 7SL RNA and the D stem-loop of sRNA-85, as presented in Fig. 3C. To test whether these two RNAs interact by base pairing, we tested the ability of these two molecules to become cross-linked in vivo. To this end, L. collosoma cells were treated with the RNA cross-linking agent AMT-psoralen and were irradiated with UV light. RNA from control and crosslinked cells was labeled with the splint labeling technique (34). As shown in Fig. 5, both RNAs were specifically labeled in both the control and after irradiation, as evident by the detection of two 7SL RNA molecules and a major sRNA-85 molecule. Additional minor sRNA-85 transcripts of 120 -160 nt were also observed, which could represent partial denatured sRNA-85 molecules and dimmers produced during labeling. In addition to labeling the cognate individual RNA molecules, a single RNA species of ϳ650 nt was detected that appeared when labeling was performed with either 7SL RNA or sRNA-85 oligonucleotide but only in cells that were irradiated with UV. To characterize the ϳ650 nt RNA and show that it is a crosslinked species that contains both 7SL RNA and sRNA-85, we scaled up the UV cross-linking experiment, fractionated the RNA in a preparative denaturing gel, and the RNA extracted from gel slices was subjected to RNase protection assay with antisense RNA probes to sRNA-85 and 7SL RNA. The data (not shown) revealed an RNA species (ϳ600 nt) that was protected by both the sRNA-85 and 7SL RNA probes, supporting the hypothesis that these two RNA molecules interact by base pairing. To map the cross-linking sites on both sRNA-85 and 7SL RNA, we subjected the RNA eluted from the gel slice carrying the cross-linked species to primer extension experiment. Due to multiple structural stops and the poor recovery of the cross-linked species, it was technically impossible to further map the position of the cross-linking on the two RNA molecules.

sRNA-85 Mutants in the Region Comprising the Putative Interaction with 7SL RNA Do Not Form a Complex with 7SL
RNA and Therefore Cannot Bind to Ribosomes-To further study the structure-function aspects of sRNA-85 and to identify the domains necessary for its interaction with 7SL RNA and/or the ribosomes, we tagged the sRNA-85 molecule by site-directed mutagenesis. The tagged gene was cloned into the pX expression vector. Transgenic parasites were selected on elevated levels of G418. Several mutations were introduced in sRNA-85, including mutation that deleted the anti-codon region (mutant I) and insertion mutations in the acceptor stem and the putative interaction domain with 7SL RNA (mutant II and III). Only mutant I, lacking nt 40 -44 (covering the anticodon), was expressed not the two other mutations (mutant II and III), suggesting that insertion in those two positions rendered the sRNA-85 unstable (results not shown). We found that the mutant I was expressed only from a construct carrying a 200-bp upstream sequence but not if only a 55-bp upstream sequence was present in the construct, suggesting that the presence of an extragenic promoter is essential for the expression of the gene (results not shown).
We have utilized this information in designing a strategy to rigorously examine the sRNA-85 domains, which are essential for their interaction with the 7SL RNA. To this end, we generated three mutations in sRNA-85 in the putative interaction domain of sRNA-85 with 7SL RNA. Because the sRNA-85 could not tolerate insertion in this domain, three mutants with base substitutions were designed, as depicted in Fig. 6A. Mutation IV carried three base substitutions in positions 27, 28, and 30 of sRNA-85. Mutation V carried three base substitutions in positions 16, 17, and 20 of sRNA-85, and mutant VI carried the substitutions of mutants IV and V as well as an additional substitution in position 23 of sRNA-85. To examine the expression of the mutated sRNA-85 transcripts, we designed a primer extension assay that can specifically prime only on the mutated but not on the wild-type transcript. For each of the mutations, oligonucleotide was designed such that the 3Ј-end matches only the mutated sRNA-85 but not the wild-type transcript. The results, shown in Fig. 6C, indicate that the mutation-specific primer elongates the mutated transcript but not the wild-type RNA. To investigate the distribution of the mutated and wildtype sRNA-85 and their association with 7SL RNA and ribosomes, we examined the distribution of these two RNAs on RNPs fractionated on sucrose gradients (the fractionation was for 3 h). The gradient fractions were subjected to Northern analysis with 7SL-and sRNA-85-specific probes. The results, presented in Fig. 6B, demonstrate that most of the 7SL and sRNA-85 co-fractionated at the top of the gradient as free RNPs. However, a distinct fraction (10%) was found on 80 S ribosomes. The two RNAs always co-fractionated either when present as free RNPs or when the complex was associated with the ribosomes, suggesting that, in the cell, the sRNA-85 is found only complexed to 7SL RNA but not as free RNP. The fractionation of RNPs from the mutant VI is presented in Fig.  6D. The RNA from the gradient fractions was analyzed by Northern analysis with mutated sRNA-85 and 7SL RNA oligonucleotides (upper panel) or with wild-type sRNA-85 and 7SL RNA oligonucleotides (lower panel). The results, presented in Fig. 6D (3-h fractionation), indicate that the mutated sRNA-85 fractionates differently than the wild-type RNA does; the S value of the peak fraction was smaller for the mutant. In addition, the mutant sRNA-85 was not found on the 80 S peak, suggesting that if sRNA-85 does not associate with the 7SL RNA it does not bind to ribosomes. With longer fractionation, better separating the particles, the results indicate that, although the wild-type sRNA-85 was found only in fractions carrying the 7SL RNA (Fig. 6E, lower panel), the mutated sRNA-85, which cannot bind to the 7SL RNA, was present only on the top of the gradient with the bulk of free tRNAs (ϳ4 S) (Fig. 6E, upper panel), suggesting that only when associated with 7SL RNA was sRNA-85 found in the ϳ14 S particle. Note that the mutated sRNA-85 is very susceptible to degradation, as opposed to the wild-type transcript, because specific smaller size transcripts were observed upon longer fractionation of the particles (Fig. 6E, upper panel). Similar results were also obtained for mutants IV and V, suggesting that indeed these two RNAs interact by base pairing at the proposed interaction domain depicted in Figs. 3C and 6A.
Cellular Distribution of sRNA-85 and 7SL RNA-Recent studies suggest that the mammalian and yeast SRPs are assembled in the nucleolus where they form the core particle and that this core particle is transported to the cytoplasm and, only after the joining of SRP54 in the cytosol, is the assembly of the particle completed (41). It was therefore of great interest to examine the cellular distribution of both 7SL and sRNA-85 and especially to monitor the localization of the sRNA-85 mutants described in the previous section. Wild-type and mutants (IV-VI) were subjected to in situ hybridization with DIG-labeled PCR probes for detecting the wild-type sRNA-85 and 7SL RNA along with mutant-specific oligonucleotides that were end-labeled with DIG nucleotides to detect the distribution of the mutated sRNA-85. The nuclei and kinetoplasts were stained with propidium iodide. The results, presented in Fig. 7A, suggest that most of the 7SL RNA is distributed in the cytoplasm and this is also the case for sRNA-85. However, distinct staining was also observed inside the nucleus for both RNAs. Interestingly, a different pattern was observed for the sRNA-85 mutants. We chose to present the data for one of the mutants, mutant IV. The in situ hybridization pattern with the mutantspecific oligonucleotide was different from the pattern observed when the wild-type transcript distribution was examined. Very pronounced staining was observed around the nuclear envelope as well as in a distinct compartment within the nuclei of the mutant cells. In situ hybridization with small nucleolar RNA snoRNA-2 indicates that this is the nucleolus (data not shown). In addition, a yellow ring, resulting from the merging between the green fluorescence of the labeled RNA and the red propidium iodide, was also observed (Fig. 7B). These data suggest that the inability of mutant IV to associate with ribosomes changed its cellular localization and that this mutated RNA accumulated both inside and around the nucleus. DISCUSSION We have found that the trypanosomatid L. collosoma SRP, unlike all other known SRPs in nature, contains, in addition to the 7SL RNA homologue, a short RNA molecule, termed sRNA-85. We previously reported a homologous molecule that was co-purified with the T. brucei SRP complex (25), suggesting specific for the different mutants (36318 for mutant IV, 36317 for mutant V, and 36319 for mutant VI). The extension products are indicated. D, fractionation of sRNA-85 mutant VI on sucrose gradients. Low salt extract and gradient fractions were prepared from cells carrying the mutation in sRNA-85 as described in B. The gradients were fractionated for 3 h. RNA was prepared from the sucrose gradient fractions and subjected to Northern analysis with the end-labeled antisense oligonucleotides to 7SL RNA, wild-type sRNA-85, and antisense oligonucleotide-specific for mutant VI, 36319. The S values were determined as in B. E, the same as in D except that the gradients were centrifuged for 22 h. S values were determined relative to 28 S and 18 S rRNA, 4 S (Invitrogen), and catalase (11 S). The fractions are numbered from top to bottom. The identities of RNAs are indicated. that the unique composition of the SRP is conserved in the trypanosomatid family. The co-existence of these RNAs in the same complex was confirmed by (1) conventional chromatography, (2) affinity-selection using an antisense oligonucleotide to sRNA-85, and (3) in vivo cross-linking with AMT-psoralen. sRNA-85 was isolated, sequenced, and used to clone the corresponding gene. It was identified as a tRNA-like molecule that deviates from the canonical tRNA structure. sRNA-85 carries a CCA sequence at the 3Ј-end but is most probably uncharged. Mutations were introduced in sRNA-85 that disrupted its potential to interact with 7SL RNA by base pairing. These mutants could not interact with the 7SL RNA or with ribosomes and accumulated both inside and around the nucleus. We think that sRNA-85 may assist the truncated Alu domain of 7SL RNA. The discovery of sRNA-85 raises the intriguing possibility that sRNA-85 functional homologues may exist in other lower eukaryotes that possess a truncated Alu domain or eubacteria that lack this domain.
The homology of sRNA-85 to tRNA genes is only partial, and therefore it is difficult to determine from which tRNA the sRNA-85 was derived. Possibly, the extra nucleotides present in the anti-codon stem-loop of sRNA-85 represent remnants of an intron that escaped splicing, and sRNA-85 may have originated from a tRNA-splicing defective mutant. Most known amino-acyl tRNA synthetases simultaneously recognize the amino acid and the anti-codon region (42). Because the anticodon region of sRNA-85 deviates from the canonical structure, it may not be recognized by a cognate synthetase and, therefore, this tRNA-like molecule most probably remains uncharged.
Although sRNA-85 is most probably uncharged it carries a CCA sequence at the 3Ј-end. This is not surprising, because it is well-known that only the "top half " of a tRNA molecule containing the acceptor stem and the T⌿C stem-loop, which exist in sRNA-85, are essential for recognition by the CCAadding enzyme (43). Note that CCA is also important for binding uncharged tRNAs to the E site and for the translocation of tRNA from the P to the E site (44). Like a bona fide tRNA, the data, presented in Fig. 1C, is consistent with the presence of tRNA-specific modification at positions Gm18 and Um52 on sRNA-85 (39).
sRNA-85 differs from all tRNA-like structures identified so far, including the 3Ј-end of plant viruses that functions in viral replication (45), 10S␣ or tmRNA, a bacterial-charged RNA that serves to rescue truncated mRNAs (46), and Alu RNAs, whose function is currently unknown (47). The only RNA molecule that is closely related to sRNA-85 is sRNA-76, which was identified as a tRNA-like molecule that co-purifies with the T. brucei 7SL RNP (25). Interestingly, although sRNA-76 and sRNA-85 do not have similar sequences, both are tRNA-like molecules that deviate from the canonical structure in the anti-codon stem-loop. sRNA-76 resembles tRNA Asp and tRNA Gly , whereas sRNA-85 resembles tRNA His and tRNA Asn . Interestingly, the position downstream of the anti-codon in both sRNA-85 and sRNA-76 is U instead of purine.
Using the computer program bestfit, we identified remarkable potential for base pairing between the 5Ј-end of the 7SL RNA and the D stem-loop of sRNA-85, as presented in Fig. 3C. To examine whether the two RNAs are held together by base pairing, we used the in vivo cross-linking approach in the presence of AMT-psoralen. Indeed, a cross-linking species that contained these two molecules was identified by the splint labeling technique and by probing the cross-linked species with specific probes for sRNA-85 and 7SL RNA (data not shown). Unfortunately, it was impossible to map the cross-linked sites on both molecules for two major reasons: first, the cross-linking species was not abundant and it was difficult to efficiently purify it from the gel; second, both molecules have very complex secondary structures. We have previously mapped the many structural stops on both 7SL I and 7SL II in primer extension experiments (24). These many secondary structural stops precluded the efficient extension on the cross-linked species and made it impossible to map the cross-linked stops by the sensitive primer extension assay.
The importance of 15-30 nt of sRNA-85 for its function was substantiated by examining the association of sRNA-85 mutants with 7SL RNA and the ribosomes. All three mutants behaved very similarly; the mutated sRNA-85 (mutant VI), which was unable to interact by base pairing with the 7SL RNA, was found on the top of the sucrose gradient along with the bulk of the cellular tRNAs (ϳ4 S) (Fig. 6, D and E), suggesting that the presence of sRNA-85 in a ϳ14 S particle depends on its association with 7SL RNA. In fact, all the sRNA-85 mutants that could not interact with 7SL RNA could not bind to ribosomes as well, suggesting that sRNA-85 can interact with ribosomes only via its association with 7SL RNA. Indeed, the sRNA-85 mutants also accumulated around the nuclei and were not dispersed throughout the cell like the wild-type sRNA-85 and 7SL RNA (Fig. 7).
It is currently unknown where the assembly of sRNA-85 and 7SL RNA takes place. The accumulation of sRNA-85 mutants in the nucleolus and around the nuclear envelope (Fig. 7B) suggests that the first interaction between these two molecules may take place in the nucleolus. Competition for export of sRNA-85 might exist in cells overexpressing the mutated sRNA-85. This may cause sRNA-85 to accumulate in the nuclear envelope where these molecules are modified before being exported from the nucleus.
The presence of sRNA-85 as part of the trypanosomatid SRP complex raises the question why such a molecule is needed for the function of the trypanosomatid SRP. We propose that sRNA-85 may replace the truncated Alu domain of the trypanosomatid 7SL RNA. It has been speculated that the Alu domain might function in translation arrest by the RNA mimicking tRNA, consequently blocking the access of aminoacyl-tRNA to the A site of the ribosome (47,48). Elongation arrest might indeed arise from steric interference of active competition of the Alu domain with the elongation factor eukaryotic elongation factor-2 or tRNA. Although the Alu domain shows no structural resemblance to tRNA or elongation factors but instead resembles the structure of the hammerhead ribozyme, its size and shape do not preclude it from entering the ribosomal elongation factor binding site. However, it is still not clear what features are required for elongation arrest, because neither the RNA component nor the protein component of the Alu domain across species is conserved (16,49).
From studies on mammalian cells, it was suggested that SRP interacts with the ribosomes before the translocation of the peptidyl-tRNA from the A to the P site (14). Translocation is mediated by eEF-2 or its prokaryotic homologue elongation factor-G (50). SRP may therefore interfere with the elongation factor's mode of action. EF-G lacks the structure equivalent to the pseudoknot in the tRNA of the ternary complex. Therefore, at the time when SRP interacts with the ribosome, the tRNA pseudoknot binding site is likely to be vacant (51). In this way, sRNA-85 may functionally replace the truncated Alu-like domain and bind to this vacant ribosomal site.
Another possibility is that sRNA-85 interacts with the E site of the ribosome. The finding that sRNA-85 is uncharged is intriguing, especially because uncharged tRNAs exhibit high affinity to the ribosomal E site (52). If sRNA-85 interacts with the ribosomal E site, it may back up the ribosome to prevent translocation from the A to the P site. This hypothesis may explain the finding that E. coli mutants with a low level of SRP RNA can survive only if they carry an extragenic suppresser mutation in the tRNA synthetase, which leads to the accumulation of a high level of uncharged tRNAs (53). The rationale for this second-site suppression is that uncharged tRNAs may bind to the ribosome E site, prevent translocation, and prolong the time during which SRP can bind the ribosome (14,53). Analogously, sRNA-85 may bind to the E site and back up the ribosome. Understanding the role of sRNA-85 in the context of the SRP function requires mapping the site where it interacts with the ribosome.
The presence of truncated SRP RNA in many eubacteria species led to the hypothesis that these RNAs emerged from deleting an ancestral molecule (3). Our finding of a two-RNA component SRP in trypanosomatids suggests that this may have been the composition of the ancient eukaryotic SRP. It is therefore possible that trypanosomatids adapted and modified an existing tRNA molecule to replace the Alu-truncated domain or that sRNA-85 was required to assist SRP function. Perhaps, when the tRNA-like domain of SRP RNA is carried on a separate molecule, the pause in translational arrest is tighter, longer, or better regulated.
The Alu domain was shown to function in the transport of the 7SL RNA with its core proteins from the nucleolus to the cytoplasm. The export of 7SL RNA with the core proteins was shown in yeast to depend on the export receptor Xpo1p (11). However, the export of tRNA molecules is mediated by another receptor, the Los1p (54). tRNA molecules that are transcribed by RNA polymerase III interact with the La protein that localizes tRNAs to the nucleolus where it undergoes the tRNAspecific modifications. After CCA addition at the 3Ј-end and splicing, the mature tRNA is exported (55). Trypanosomatids utilize tRNA molecules for different purposes, for instance, the 7SL RNA and the U snRNAs are transcribed from an extragenic promoter, which is composed of a tRNA molecule that is divergently transcribed with respect to the small RNA transcript (56). Because trypanosomatids diverged very early from the eukaryotic lineage, they lack sophisticated promoters and therefore utilize tRNA promoters to assist in recruiting RNA polymerase III to transcribe those small RNA genes. tRNA genes also mark the termination of the long polymerase II polycistronic transcript (57). By analogy, the function of sRNA-85 could be to assist in transporting the SRP particle from the nucleolus to the cytoplasm, utilizing the efficient tRNA transport system.
The absence of an Alu-like domain in many eubacteria and its severe truncation in lower eukaryotes are puzzling, assuming that the translational arrest is an essential function of the SRP. This study raises the question whether sRNA-85 functional homologues may exist in other unicellular organisms such as eubacteria, yeast, and other protozoa.