Nucleic Acids Research

In vivo psoralen crosslinking of T.brucei RNAs

Crosslinking was performed similarly to the published method of Watkins et al. (31). Trypanosoma brucei strain 427 procyclic forms were grown to 7 × 10⁶ cells/ml then harvested. Concentrated cells were incubated with 0.1 mg/ml 4[prime]-aminomethyltrioxsalen (AMT; HRI Associates) then half the cells were irradiated with 365 nm ultraviolet (UV) light at 3 mW/cm² for 20 min (Ps+); the other half were not irradiated (Ps-). RNAs were isolated by the guanidinium isothiocyanate/hot phenol method (32); a single 20 mg Ps+ RNA preparation was used for all experiments herein.

To photoreverse psoralen crosslinks, 5-10 µg of Ps+ RNA in 5 µl water was irradiated with 254 nm UV light at 12 000 µW/cm² for 5 min.

Sandwich Southern analysis

This two step hybridization was based on published methods (18,19,21). Southern membranes were first hybridized with 200 µg partially hydrolyzed Ps9+ RNA, as described (21). In the second hybridization, a cocktail of antisense U3 oligonucleotides (oU3.27, oU3.29, oU3.30, oU3.31 and oU3.35) radiolabelled by T4 polynucleotide kinase and [[gamma]-³²P]ATP, was incubated with membranes at 45°C overnight in buffer containing 5× SSPE, 5× Denhardts, 1% SDS and 50 µg/ml Escherichia coli tRNA. Membranes were washed with 2× SSPE plus 0.1% SDS at 45°C.

Oligonucleotide-directed ribonucleaseH digestions

For preparative digestion of pre-rRNAs, 400 µg RNA and 200 µmol of antisense oligonucleotides in 83 µl of 0.5 µM EDTA were denatured at 90°C for 3 min, then 4°C for 5 min. Ten µl of 10× Promega RNaseH buffer and 3.5 µl of Promega RNasin at 40 U/µl were added and samples incubated at 45°C for 15 min. Aliquots of 1.5 µl 0.2 M dithiothreitol and 4 µl Promega RNaseH at 6 U/µl were added and samples placed at 37°C for 30 min. Digestion was stopped with 1 µl 0.5 M EDTA. Ammonium acetate, pH 7.0 and isopropanol were added to final concentrations of 2 M and 50%, respectively, to selectively precipitate RNAs.

RNaseH digestion of U3-5[prime]ETS complexes was performed similarly. RNAs selected from 100 µg starting material plus 6 µg carrier E.coli tRNA or rRNA (Sigma) were digested in 20 µl reactions containing 0.5 mM oligodeoxynucleotides and 0.6 U RNaseH.

Hybrid selection of RNA fragments and U3-5[prime]ETS crosslinked complexes

Pre-rRNA fragments were hybrid-selected using 5[prime]ETS-containing plasmid DNAs immobilized on MSI Nytran (Fig. 2) or Pall Biodyne Plus Nylon membranes by standard methods (32). Pre-hybridization solution (600 µl) containing 500 µg digested Ps+ or Ps- RNAs, 50% formamide, 20 mM PIPES pH 6.4, 0.2% sarkosyl, 0.4 M NaCl and 100 µg E.coli tRNA was incubated at 70°C for 10 min, then 6-8 membrane circles, each carrying 25 µg pG.site1 or pG.site2 DNAs, were added and incubated at 37°C for 12-16 h. Supernates were collected and their RNAs precipitated with 2.5 vol ethanol. Membranes were separated by type and washed 10 times with 1 ml aliquots of 10 mM Tris-HCl pH 7.5, 75 mM NaCl, 1 mM EDTA and 0.5% sarkosyl at 37°C. Hybridized RNAs were eluted into 0.8 ml of 1 mM EDTA plus 40 µg E.coli tRNA at 95°C for 3 min, then chilled on ice. RNAs were precipitated from eluates with 1/10 vol 3 M sodium acetate pH 5.2 and 2.5 vol ethanol. RNAs were next treated with Promega RQ1 DNaseI, phenol-chloroform-isoamyl alcohol (PCIA) extracted, then precipitated with 2 M ammonium acetate pH 7.0 and 50% isopropanol to remove plasmid DNA released during the selection procedure.

Biotinylated 2[prime]-O-allyl antisense U3 oligoribonucleotides were then used to isolate U3-site1 and U3-site2 complexes. Site1 or site2 RNAs selected from 400 µg starting material, 50 pmol of oU3a and/or oU3b and 200 µg carrier E.coli tRNA or rRNA in 64 µl water were denatured at 90°C for 2 min, then 4°C for 2 min. Aliquots of 16 µl 5× hybridization buffer (500 mM KCl, 100 mM Tris-HCl pH 8.0, 25 mM MgCl₂ and 5 mM DTT) plus 8 U of RNasin were added and samples incubated at 37°C for 30 min. Pre-washed Promega avidin-magneto beads (150 µl) in 20 µl 1× hybridization buffer were added to the reactions, then incubation continued for 30 min. Unbound RNAs were collected and precipitated with ethanol. Beads were washed four times with 0.5 ml hybridization buffer at room temperature. Hybridized RNAs were eluted into 200 µl 0.5 mM EDTA plus 20 µg E.coli tRNA or rRNA by treatment at 95°C for 3 min, then ice for 5 min. Double-selected RNAs in eluates were precipitated, treated with DNaseI, PCIA extracted, and ammonium acetate-isopropanol precipitated, as described above, before undergoing analytical oligonucleotide-directed RNaseH digestions.

RNA methods

For northern RNA analysis, denaturing polyacrylamide gel electrophoresis (PAGE), transfer of RNAs to nylon membranes, and hybridization with radiolabelled probes was done as described (17). To visualize the very low amounts of purified U3-5[prime]ETS RNA complexes, however, Pall Biodyne Plus membranes and hybridization solutions consisting of 0.25 M sodium phosphate pH 7.0 and 7% SDS were used. Plasmid and oligonucleotide combinations used to generate antisense probes by asymmetric polymerase chain reaction method (41) were: U3 probe, EcoRI-digested pG.RNAB and cU3.35; site1 probe, EcoRI-digested pG.5[prime]ETS3 and cr18; site2 probe, BsmI-digested pG.5[prime]ETS and cr12. Highly sensitive Kodak BioMax film and compatible intensifying screens were requisite for detection of isolated complexes by autoradiography. Primer extension analyses were done as previously described (27).

Oligonucleotides used in RNaseH digestion and primer extension experiments

Antisense 5[prime]ETS oligodeoxynucleotides (complementary positions in the 5[prime]ETS sequence are noted in parentheses) were: cr4 (nt 672-681) CGCATATAATATCAGGCGGC; cr5 (nt 893-912) agattcctcaaggcgtcact; cr6 (nt 176-195) CACACGTATTACACACACTC; cr9 (nt 1099-1118) CGTGATCCGCTGTGGGAAC; cr10 (nt 272-291) cgcgtgtgaagcaactgatg; cr12 (nt 1019-1038) GCGCATATATGTACACACGA; cr13 (nt 1053-1072) CTGCGAGTGGGTGGTGTTTC; cr15 (nt 94-113) CGCAAAAACTGTTCCCGTAT; cr17 (nt 792-798) CCGTACGCGGCGTACA; cr18 (nt 241-261) CCTACGAGCTGTGAAaGGTAA; cr19 (nt 116-135) TTCGGCAAACCTACTGAACA; cr20 (nt 149-168) CCCTTAACTGAGGAAGTGTC. Antisense U3 RNA oligodeoxynucleotides (complementary positions in U3 are noted in parentheses) were: cU3.27 (nt 1-24) CGATTCTGTTCAGAGTACGGTCTT, cU3.28 (nt 20-39) GTTGTACTCATAAAACGATTCTG, cU3.29 (nt 35-54) TTTCTCATTTAAGAGGTTGT, cU3.29.5 (nt 40-65) GTTGTTGGTTATTTCTCATTTAAGAG, cU3.30 (nt 50-69) TTTGGTTGTTGGTTATTTCT, cU3.31 (65-84) CCTTCATCATCAGGATTTGG, cU3.35 (nt 125-144) GGATCCTTCTGGAACCGGCT.

Antisense 2[prime]-O-allyl-oligoribonucleotides used in U3 RNA selections were kindly provided by Angus Lamond. These were: oU3a (nt 1-16) TCCUUCAGAGUACGGUCUUCCT; oU3b (nt 13-28) TCCAAAACGAUUCUGUUCACCT; and oU3c (nt 25-39) TCCGUUGUACUCAUAAAACCT. Residues in bold font indicate biontinylated residues.

Plasmid DNAs

pG.rDNA contains a 15.5 kb contiguous rDNA sequence assembled from restriction fragments of a split rDNA gene in pR4 (25) and inserted into pGEM2 (Promega). The EcoRI site at the 5[prime] end of the 15.5 kb insert resides 3.3 kb upstream from the transcription start site; the 3[prime] HindIII site lies 2.5 kb downstream from the 3[prime] end of the LSU sr4 rRNA coding sequence. pG.5[prime]ETS contains DNA sequence from -260 to 1135 bp that includes the rRNA promoter and the entire 5[prime]ETS (SSU rRNA sequence starts at 1132 bp); it has synthetic 5[prime] XbaI and 3[prime] HindIII restriction sites. The pG.5[prime]ETS insert was digested by AluI at bp 260 and fragments were subcloned to create pG.site1A and pG.site2A. pG.site2B was assembled following BsmI digestion of the pG.5[prime]ETS insert at 833 bp. pG.5[prime]ETS3 contains a PCR-generated insert of 5[prime]ETS sequences from +1 to 260 and having synthetic 5[prime] EcoRI and 3[prime] XbaI restriction sites. pG.RNAB contained U3 coding sequences with synthetic 5[prime] EcoRI and native 3[prime] BamHI sites.

U3 RNA crosslinks at two sites within the 5[prime]ETS of pre-rRNA transcripts

Interactions between the T.brucei U3 RNA and pre-rRNAs were examined by psoralen RNA crosslinking. Psoralen intercalates into nucleic acid helical regions and, upon photactivation, forms covalent crosslinks between pyrimidines (primarily uridines) on opposite strands. In this manner, base paired RNA interactions are preferentially captured, though non-paired, closely associated RNAs are sometimes crosslinked (43). Procyclic form trypanosomes were incubated with AMT psoralen and irradiated with longwave UV light. Total RNA was isolated from cells treated with both psoralen and UV light (Ps+) and control cells that were treated with psoralen but not irradiated (Ps-). Northern analysis indicated the presence of mobility-shifted U3 RNAs in Ps+ RNAs, but not Ps- RNAs, characteristic of crosslink formation with other RNAs (data not shown; but see below). U3 crosslinks with particular pre-rRNA species were not adequately discerned due to the aberrant migration of Ps+ compared with Ps- RNAs. Instead, Sandwich Southern analysis was used to detect U3 RNA-pre-rRNA crosslinks (Fig. 1). Southern membranes carrying restriction digests of plasmid DNAs containing an rDNA gene (pG.rRNA) or 5[prime]ETS sequences (pG.5[prime]ETS) were first hybridized with partially fragmented Ps+ RNAs. Unbound RNA was removed, then a second hybridization was done using radiolabelled antisense U3 oligodeoxynucleotides. Hybridizing restriction fragments, detected by autoradiography, corresponded to pre-rRNA sequences crosslinked to U3 RNA.

Figure 1. In vivo U3 RNA-pre-rRNA psoralen crosslinks localized by Sandwich Southern analysis. (A) Restriction digests of 750 ng pG.rDNA (lanes 1-9) and 200 ng pG.5[prime]ETS (lanes 10-17) were separated on 1.2% (panels I and II) or 1.5% (panel III) agarose gels and transfered to nylon membranes. Membranes were first hybridized with unlabelled Ps- (panel I) or Ps+ (panels II and III) RNAs, then subsequently with 32P-labelled antisense U3 RNA oligodeoxynucleotides. (B) Corresponding restriction fragment maps. Diagrams of the 15.7 kb rDNA insert in pG.rDNA with digests 1-9 are shown on top, and expanded diagrams of the 5[prime]ETS sequence in pG.5[prime]ETS with digests 10-17 are shown in the lower part of the figure. Restriction fragments that hybridized to U3 RNA-pre-rRNA crosslinks are boxed: a medium gray box is indicative of a single hybridizing band; a light or dark gray box is indicative of a 5[prime] or 3[prime] hybridizing band, respectively, when two were present. The stippled box in digest 16 denotes a fragment that was too small to detect. The diagram at the bottom shows the deduced U3 RNA interactions with 5[prime] site1 and 3[prime] site2 5[prime]ETS sequences. The numeric base assignments are given with respect to the transcription start site at +1 nt; the primary cleavage is at 115-116 nt and the SSU rRNA sequence begins at 1132 nt. Restriction sites are labelled: E, EcoRI; H, HindIII; B, BglI, G, BglII; L, ApaLI; F, FspI; M, BsmI; S, SspI; X, XbaI; A, AluI; N, BstNI. Synthetic sites introduced during subcloning are in parenthesis. Markers were DNA digested with EcoRI and HindIII. A hybridization control, 5.4 ng of linearized pG.RNAB, was present in a 1:25 molar ratio compared with pG.rDNA and pG.5[prime]ETS plasmids (lane C).

Figure 2. Separation of U3-site1 and U3-site2 RNA crosslinked complexes. (A) Experimental strategy: 5[prime]ETS sequences and the two sites of interaction with U3 RNA, mapped by Sandwich Southern analysis, are pictured. Ps+ RNAs were hybridized with oligodeoxynucleotides (cr10, cr4 and cr13 are shown) complementary to sequences flanking these regions, then RNaseH digestion released site1 and site2 RNA fragments. Site-specific fragment pools were obtained by hybrid selection using pG.site1A and pG.site2A or pG.site2B plasmid DNAs that were subcloned from p.5[prime]ETS. (B) Northern analysis of 5[prime]ETS site1 and site2 fragment pools. Pre-rRNA fragments generated using cr10, cr4 and cr13 were selected with pG.site1A and pG.site2A DNAs immobilized on MSI Nytran membranes. RNAs were electrophoresed through a 4% polyacrylamide/7 M urea gel, then transfered to Nytran. Lane 1 contains 12.5 µg digested, non-psoralen-treated, non-selected RNA; lanes 2-4 and lanes 5-6 contain site1- and site2-selected RNA fragments, respectively, from 50 µg of starting material; lanes 8-10 contain 12.5 µg of non-hybridized, supernate RNAs. Ps- RNAs are in lanes 2, 5 and 8; and Ps+ RNAs in lanes 3, 6 and 9. Ps+ RNAs in lanes 4, 7 and 10 were exposed to 254 nm UV light to photoreverse crosslinks, post hybrid-selection. Panels show autoradiographs of northerns probed with rabiolabelled cr6 or cr5 to detect site1 or site2 sequences, respectively; or with antisense U3 oligonucleotides.

As shown in Figure 1A, one or two hybridizing DNA fragments were detected in each restriction digest. The diagram in Figure 1B shows the location of each hybridizing fragment on restriction enzyme maps of cloned rDNA sequences. Examination of pG.rDNA digests (samples 1-9) showed that only sequences coding for pre-rRNAs corresponding to the 5[prime]ETS region were detected by labeled U3-pre-rRNA complexes. This was most clearly demonstrated in sample 3, that contains an EcoRI-BglII digest of pG.rDNA. The DNA fragment digested by EcoRI 2.5 kb upstream of, and by BglII exactly at, the 5[prime] end of the SSU rRNA sequence exclusively hybridized with U3-pre-rRNA crosslinked rRNAs. Hybridizing sequences within the 5[prime]ETS were closely examined in restriction digests of pG.5[prime]ETS (samples 10-17). Two distinct regions of the 5[prime]ETS crosslinked to U3 were delineated using the endpoints of each hybridizing restriction fragment. Site1 contained sequences bordered by FspI and AluI restriction sites (nt 84-260 of the pre-rRNA); the primary 5[prime]ETS cleavage occurred within site1 at 115-116 nt (25). Site2 contained sequences defined by BsmI and SspI sites (nt 833-1007). Significantly, the AluI-BsmI digest of pG.5[prime]ETS (sample 16) contained a 571 bp fragment lying between sites 1 and 2 that did not hybridize to U3-crosslinked RNAs, whereas the smaller 500 bp site1-containing fragment did (the small 320 bp site2-containing fragment, and generally fragments smaller than 400 bp, were not readily detected in these experiments). This demonstrated that the site1 and site2 5[prime]ETS interactions with U3 snoRNA were discrete and separable.

Distinct sets of U3 residues are indicated for each interaction with the 5[prime]ETS

Recognizing that U3 snoRNA has multiple roles in rRNA processing, it seemed reasonable that the two interactions detected between U3 and the 5[prime]ETS might be functionally dissimilar and involve distinct U3 RNA sequences. To examine U3 residues in each U3-5[prime]ETS RNA interaction, a strategy was developed to separate U3-site1 and U3-site2 crosslinked complexes from one another (Fig. 2A). Oligodeoxynucleotides complementary to sequences 3[prime] to site1 (cr10), and to sites flanking site2 (cr4 and cr13), were hybridized to pre-rRNAs, then corresponding 5[prime]ETS sequences were digested by RNaseH. Site1 and site2 RNA fragments released by this treatment were separated from one another and from non-crosslinked U3 RNAs by hybrid selection using plasmid DNAs containing site1 or site2 sequences. Northern analysis of site1- and site2-selected RNA fragment pools is shown in Figure 2B. cr10 directed cleavage 3[prime] to site1 sequences to produce ~271 nt fragments extending from the transcription start site (a few percent of the site1 fragment pool), and ~156 nt fragments extending from the primary cleavage site, to cr10-complementary sequences. cr4 and cr13 targeted cleavage to release ~361 nt site2-containing RNAs. Hybrid selection was efficient and very specific; no cross-contamination of site1 and site2 RNA fragment pools was apparent (Fig. 2B, compare lanes 2-4 and 5-6). Additionally, free U3 RNA was not detectable in selected RNA samples, yet was released upon exposure to shortwave UV light to reverse psoralen-RNA crosslinks (Fig. 2B, lanes 4 and 7). This showed that U3-5[prime]ETS crosslinks existed in site-selected fragment pools (see also Fig. 4, below).

Primer extension analysis of U3 RNAs in U3-site1 and U3-site2 complexes present in site-selected RNA pools denoted psoralen adducted U3 residues (Fig. 3A). AMV reverse transcriptase stops before, and to a much lesser extent at, bases carrying psoralen additions (33). A number of stops occurred in U3 RNAs present in Ps+, but not in Ps-, site-selected RNAs. Full length U3 RNAs were detected in Ps+ samples in which crosslinks were partially photoreversed, with concomitant decreases in the intensity of primer extension stops, indicating that stops were due to psoralen additions versus procedural damage to U3 RNA molecules. It was estimated, by comparison with non-selected RNAs, that <1% of U3 RNA in Ps+ samples was crosslinked to 5[prime]ETS site1 or site2 sequences.

Figure 3. Psoralen adducts in U3 RNAs crosslinked to site1 and site2 5[prime]ETS sequences. (A) Primer extension analysis of U3 RNA in U3-site1 and U3-site2 complexes to identify psoralen adducted bases. RNAs shown in Figure 2 were used, and lanes were labelled accordingly. RNAs selected from 100 µg of total Ps- or Ps+ RNA were used as template for reverse transcriptase; 1 µg of RNA was used in control and supernate lanes. Primer extensions using oligonucleotide cU3.35 as primer resolved on a 6% sequencing gel are shown; extensions using cU3.31 were also done to better resolve stops within the 5[prime] end of U3 RNA (data not shown). Marker lanes contain primer extension (N) and dideoxy sequencing lanes (T, C, G and A) of U3 RNA in untreated total RNA preparations. Primer extension stops preceding apparent psoralen adducts in crosslinked complexes are indicted. (B) Positions of psoralen adducts displayed on the T.brucei U3 RNA secondary structure model (27). Conserved U3 sequences are boxed (27 and references therein; 42). Light gray ovals indicate psoralen adducted residues in U3-site1 complexes; dark gray ovals indicate adducts in U3-site2 complexes. The size of the ovals indicates relative strengths of the detected primer extension stops. Circled bases carry adducts in non-selected, Ps+ U3 RNA; these include U8, U28 and U30 that were also detected above this background in U3-site2 complexes. This figure represents compiled data from several experiments, including primer extension analysis of U3 RNA in isolated U3-site1 and U3-site2 complexes.

Figure 4. Identification of U3-site1 and U3-site2 complexes by northern analysis. (A and B) Northerns contain 10 µg non-digested RNAs (lanes 1-3); 10 µg cr10-cr4-cr13-directed RNaseH-digested RNAs (lanes 4-6); pG.site1A- (A) or pG.site2B-selected (B) fragments from 50 µg digested RNAs (lanes 7-9); and 10 µg supernate RNA (lanes 10 and 11). Crosslink photoreversals are in lanes 3, 6 and 9. Northern blots in (A) and (B) were hybridized with U3 DNA probes. The positions of U3Xa, U3Xb and U3Xc crosslinked species are indicated by asterisks, and putative U3-site1 and U3-site2 complexes by arrows. (C and D) U3-site1 (C) and U3-site2 (D) complexes were isolated from the site-selected fragment pools shown in (A) and (B). Selections used 2[prime]-O-allyloligoribonucleotides oU3a (lanes 1 and 2) or oU3b (lanes 3). Material in each lane was derived from 200 µg RNA. (C) Hybridization with site1, and (D) with site2 DNA probes. Autoradiography of lanes 1-3 was for 3 days, and of lanes 4-6 was for ~2 h. Northerns in (A) and (C) are from 5% PAGE and those in (B) and (D) are from 4% PAGE. Marker lanes were pBr322 plasmid DNA digested with MspI.

Intriguingly, distinct, albeit somewhat overlapping, sets of U3 RNA bases were adducted with psoralen in U3-site1 and U3-site2 complexes, suggesting involvement of different U3 sequences in each crosslink. Adducts were found solely within the 5[prime] domain of U3 RNA (Fig. 3B). Those implicated in site1 interactions occurred in the hinge region, whereas those in site2 interactions occurred primarily in box A sequences. Stops corresponded to U3 psoralen adducts at U48, U55, U58 and C62 in U3-site1 complexes, and to adducts at U8, U25, U28, U30, U48 and U55 in U3-site2 complexes. These same patterns were observed when U3 RNAs in purified U3-site1 and U3-site2 complexes were analyzed (see below; data not shown). Transcription stops also occurred in U3s present in unselected Ps+ RNAs (corresponding to residues U8, U11, U13, U26, U28, U30 and U44). The majority of these were found within the flexible 5[prime] stem-loop that features elements favored by psoralen additions: a uridine-rich helix with bulged nucleotides. The results of several U3 primer extension experiments are compiled in Figure 3B.

Isolation of U3-site1 and U3-site2 RNA crosslinked complexes

Analysis of 5[prime]ETS residues crosslinked to U3 by the primer extension method, and verification of crosslink locations in both U3 and 5[prime]ETS RNAs by RNaseH analysis (see below), required isolation of U3-site1 and U3-site2 complexes. U3 RNA crosslinked species were identified by northern analysis of Ps+ RNA samples treated as in Figure 2A. The northerns shown in Figure 4 contain non-digested RNAs, RNAs digested in the presence of cr10, cr4 and cr13 oligonucleotides, and digested, hybrid-selected site1 and site2 RNA fragments, separated on 5% (Fig. 4A) or 4% (Fig. 4B) PAGE, respectively. Hybridization with U3 probes revealed slowly migrating U3-containing complexes in all Ps+ RNA samples. In both undigested (Fig. 4A and B, lanes 2) and digested, non-selected Ps+ RNAs (Fig. 4A and B, lanes 5), three prominent U3 species typically separated on 5% PAGE with approximate mobilities of 350 (U3Xa), 300 (U3Xb) and 260 nt (U3Xc) when compared with linear molecular weight markers. Only two of these were detected by 4% PAGE: U3Xa migrated at 290 nt and U3Xb migrated at 230 nt. The altered mobility of these U3 RNA complexes on acrylamide gels of different sieves was consistent with their being non-linear, crosslinked RNAs. As well, shortwave UV treatment resulted in loss or shifted mobilities of these complexes (Fig. 4A and B, lanes 3 and 6). The putative RNA species crosslinked to U3 RNA in these complexes was not examined directly, though the U3Xb component became apparent later in these studies (see below).

Candidate U3-site1 and U3-site2 RNA complexes were identified in site-selected 5[prime]ETS fragment pools. Site1 RNAs had a single U3-containing complex (Fig. 4A, lane 8) that migrated at ~300 linear nt on 5% PAGE, whereas two U3 RNA-containing complexes were seen in site2 RNAs that migrated at ~800 and ~600 nt on 4% PAGE (Fig. 4B, lane 8). Crosslinks in each complex were readily photoreversed to release free U3 RNA (Fig. 4A and B, lanes 9). The low abundance U3-site1 and U3-site2 RNA complexes could not be visualized with site1 or site2 probes, respectively, as background hybridization concealed them. A second selection step was done to unequivocally identify and to purify the complexes.

U3-site1 and U3-site2 RNA crosslinks were concentrated from site1 and site2 RNA pools by hybrid selection with biotinylated 2[prime]-O-allyloligoribonucleotides (oU3a and oU3b) having sequence complementarity to the 5[prime] end of U3 RNA. Purified U3-site1 and U3-site2 complexes, detectable by hybridization to both 5[prime]ETS (Fig. 4C and D, respectively) and U3 (data not shown) probes, were equivalent to those detected in site-selected RNAs shown in Figure 4A and B. Northern analyses showed <1% of 5[prime]ETS fragments crosslinked to U3 RNA. Furthermore, it was calculated that only fmol amounts of complex were obtained per mg of total RNA, hence, purification for detailed biochemical analysis was prohibited. Instead, isolated complexes were subjected to targeted RNaseH digestion and primer extension analysis to resolve crosslinked U3 RNA and 5[prime]ETS bases. Figure 5 diagrams the site1 and site2 regions as well as U3 RNA, and indicates positions of antisense oligonucleotides used in subsequent experiments.

Figure 5. (A) Map of the 5[prime]ETS site1 region. (B) Map of the 5[prime]ETS site2 region. In each diagram, site1 or site2 regions defined by Sandwich Southern analysis are bracketed. Sequences complementary to oligonucleotides used in RNaseH and primer extension experiments are noted. The RNaseH digestion products directed by cr10, cr4 and cr13 are indicated by a solid line. Most site1 fragments actually have 5[prime] ends at the primary cleavage site (Fig. 2). The approximate position of the 65 nt site1 RNA fragment is also shown. In both (A) and (B), lines with arrowheads indicate the location of antisense site1 and site2 probe sequences. (C) The positions of antisense U3 oligonucleotides, used in RNaseH digestion of U3-site1 and U3-site2 complexes, are shown on the secondary structure model of T.brucei U3 RNA. Antisense U3 DNA probes extended the length of the molecule (not illustrated).

U3 RNA hinge bases contact sequences 3[prime] proximal to the processed 5[prime]ETS cleavage

U3-site1 RNA complexes isolated following cr10-targeted digestion of pre-rRNA, described above, served as templates for reverse transcriptase to identify crosslinked bases in site1 RNAs. Positions of the site1 oligonucleotide primers used are shown in Figure 5A. Unexpectedly, initial transcription experiments using cr6 and cr18 as primers failed to produce extension products (data not shown). The possibility that U3 was crosslinked to an upstream, processed 5[prime]ETS fragment was then examined. Indeed, a U3-site1 complex, that co-migrated with U3Xb, could be selected from non-RNaseH-digested Ps+ RNAs by pG.site1 DNA (Fig. 6A, lane 4) and subsequently by oU3a and oU3b (Fig. 6A and B, lanes 6). As only U3-site1 complexes of this same mobility were detected in cr10-targeted Ps+ RNAs by northern analysis (Fig. 4B, lane 5), it was evident that U3 was preferentially crosslinked to a digested 5[prime]ETS fragment. The U3-site1 complex hybridized to probes specific for sequences 3[prime], but not 5[prime], to nt 115, ruling out the possibility that U3 was crosslinked to the 5[prime]ETS fragment released by primary site cleavage (data not shown). U3-site1 complexes were examined by S1 nuclease protection analysis using a uniformly labeled 5[prime]ETS probe (from 85 to 291 nt), and fragments averaging 65 nt in size were protected, as well as a small amount of long 5[prime]ETS (data not shown). These data implied that U3 crosslinked with pre-RNAs cleaved at the primary site, then 5[prime]ETS sequences were mostly digested downstream at ~180 nt. Free 65 nt fragment was detected in both Ps- and Ps+ RNA preparations (see below), suggesting that this digestion might occur during rRNA processing, however, as a stable 3[prime] processing intermediate was not detectable by reverse transcription or northern analysis (data not shown), it need pertain to rapid 5[prime]ETS processing.

Figure 6. Isolation of the U3-site1 RNA crosslinked complex from non-RNaseH-digested RNA preparations. Panels show autoradiographs of the same northern membrane, first hybridized with antisense U3 probe (A), then cleaned of this probe and secondarily hybridized with antisense site1 probe (B). 10 µg of non-selected Ps+ RNA (lanes 1), pG.site1A-selected (lanes 2 and 3) and corresponding supernate (lanes 9) RNAs from 50 µg RNA, and pG.site1A-antisense U3-selected (lanes 4-6) and corresponding supernate (lanes 7 and 8) RNAs from 100 µg RNA, were electrophoresed through 5% denaturing polyacrylamide gels.

Isolated U3-site1 complexes were examined by targeted RNaseH digestion to localize crosslinked bases. Northern blots of digested U3-site1 complexes were first hybridized with site1 probes (Fig. 7A), then secondarily probed for U3 RNA (Fig. 7C). Free 65 nt site1 fragment was released by photoreversal of U3-site1 crosslinks (Fig. 7A, lane 2), and was also detectable in the isolated complex preparations. Neither the U3-site1 complex nor free 65 nt fragment was degraded by RNaseH in the presence of cr15 or cr6, consistent with assignment of the site1 RNA fragment to 115-180 nt. cr19 and cr20 each directed digestion of site1 RNA, resulting in increased mobility of the U3-site1 complex and localization of crosslinked bases between 136 and 148 nt. Control digestions showed appropriate targeting of 5[prime]ETS site1 sequences (Fig. 7A and B) and no aberrant cleavage of U3 RNAs (Fig. 7C).

Figure 7. Localization of crosslinks in U3-site1 complexes by oligonucleotide-directed RNaseH digestion. (A and C) Identical northerns of 5% PAGE transfered to Biodyne Plus nylon membranes. These were first hybridized with a site1 DNA probe (A), and secondarily with a U3 probe (C). (B) Probed for site1 sequences. Complexes selected from 100 µg Ps+ RNAs were treated with shortwave UV light (lanes 2) or digested with RNAseH in the presence of the indicated antisense site1 (lanes 3-6) or U3 RNA (lanes 7-12) oligodeoxynucleotides (lanes 3-12). Indicated are positions of non-digested U3-site1 complex (A and C), free ~65 nt site1 fragment (A and B), control site1 fragments derived from cr10-targeted RNAs (B), and free U3 RNA (C). Carrier E.coli rRNAs versus tRNAs were used in isolation so that similar-sized 65 nt RNA was not masked by non-specific probe hybridization to tRNAs. Use of Biodyne Plus membranes to immobilize pG.site1A for hybrid selections yielded background site1 and U3 RNAs in Ps- (lanes C) and Ps+ samples, not seen when MSI Nytran was used (Fig. 2).

Psoralen adducts in 5[prime]ETS RNA in U3-site1 complexes were detected by reverse transcription primed by cr20 (Fig. 8), as cr20 paired 3[prime] to crosslinked bases. A strong stop was detected only in Ps+ U3-site1-selected RNAs at G141 and a weaker stop at G143, corresponding to adducts at U140 and U142. Extension products were also detected at U116 and G117. The transcription assay could not distinguish between the possibilities that adducts occurred on U115 and U116 in crosslinked RNAs, or that the stops, which coincided with primary cleavage sites, indicated the 5[prime] ends of contaminating free 65 nt RNAs (Fig. 7A). Nonetheless, RNaseH digestion of U3-site1 complexes positioned crosslinks between 136 and 148 nt in site1 RNAs, establishing that U140 and U142, and not U115 and U116, were crosslinked with U3 RNA. Products extending to the transcription start site were present only in non-selected Ps- and Ps+ samples.

Figure 8. (A) Primer extension analysis of 5[prime]ETS sequences within the U3-site1 complex to identify adducted bases. cr20 was used to prime reverse transcription reactions of: 0.5 µg untreated RNAs (lane 1); 4.5 µg carrier E.coli rRNAs (lane 2); U3-site1 complexes selected from 100 µg Ps+ (lane 3) or Ps- (lane 4) RNAs; and 0.5 µg of supernate Ps+ and Ps- RNAs (lanes 5 and 6), and non-selected Ps+ and Ps- (lanes 7 and 8). Positions of all stops are indicated, including transcription start (A1), the primary cleavage sites, and psoralen adduct positions. Marker lanes are dideoxy sequencing lanes of pG.5[prime]ETS. (B) Model of the U3 RNA interaction with site1 5[prime]ETS sequences. An extended base pairing scheme is diagrammed, showing positions of psoralen adducts and possible crosslinked bases.

The location of crosslinked U3 bases in U3-site1 complexes was confirmed by directed RNaseH digestion (Fig. 7A and C, lanes 7-12). As detailed earlier, hinge region U3 bases were crosslink candidates (Fig. 3). Each antisense U3 oligonucleotide tested (Fig. 5C) directed digestion of free U3, but not site1 RNA. cU3.27, cU3.28 and cU3.31 targeted U3 RNA with concomitant decreases in mobility of the U3-site1 complex, confirming that U3 bases 40-64 contained crosslinks to site1 RNA. In the presence of cU3.29.5 and cU3.30, and perhaps cU3.29, that span this U3 region, the U3-site1 complex was degraded into two faster migrating species. Comparison of the mobility of these complexes (Fig. 7A and C, lanes 9-11) to the cU3.31-targeted complex containing the 5[prime] 64 nt of U3 (Fig. 7A and C, lanes 12), and the intensity of their hybridization to U3 probes (Fig. 7C), indicated that slower migrating complexes contained large 3[prime] fragments of U3 RNA and faster migrating complexes contained small 5[prime] fragments of U3. Evidently, hybridization of these antisense 20mers to U3-site1 complexes was not prevented, yet complete digestion of U3 RNA was inhibited, consistent with the presence of crosslinks in complementary U3 sequences. RNaseH requires only 4 bp as a target for digestion, so paired U3 sequences flanking crosslinked bases were likely substrates. The digestion patterns suggest that multiple crosslinks occured in U3-site1 complexes that are close in proximity, as undigested complexes co-migrated. Together, these results implied that U54, C58 and C62, and possibly U48, are in crosslinks with site1 RNA.

These data showed that U3 RNA hinge region bases crosslinked with 5[prime]ETS sequences lying 3[prime] to the processed primary cleavage site. One model for U3 RNA interaction with 5[prime]ETS site1 sequences is shown in Figure 8B. The C62 residue of U3 RNA and the U140 and U142 residues of site1 sequences are diametrically opposed, the predominant arrangement noted for psoralen crosslinks (43), in a scheme involving 9 of 11 bases paired. Notably, the scheme could be extended to include 12 of 14 bases paired between U3 nt 41-54 and 5[prime]ETS nt 154-167. A previously noted complementarity between U3 RNA nt 36-51 and 5[prime]ETS nt 114-129 (27) could accomodate a crosslink between U3 U48 and 5[prime]ETS U116 (not shown), but this was not supported by RNaseH experiments.

U3 RNA interacts via box A residues with site2 5[prime]ETS sequences

U3-site2 RNA crosslinked complexes were likewise investigated to identify bases involved in crosslink formation. Antisense site2 oligonucleotides used in targeted digestions of U3-site2 complexes are shown in Figure 5B; northerns comprising these experiments are shown in Figure 9. Initially, two U3-site2 complexes were isolated from pools of the 361 nt-sized site2 RNA produced by cr4-cr13-directed digestion of Ps+ RNAs (Fig. 4D). Two smaller complexes were similarly selected from site2 RNAs produced by cr17-cr13-targeted digestion (Fig. 9A and B). Photoreversal of the crosslinks released only the expected 244 nt-sized site2 fragment (Fig. 9A, lane 2), implying that the two detected complexes represented two different U3 crosslinks to the 244 nt fragment, rather than a single U3-site2 contact on two differently sized site2 RNAs. It was somewhat surprising that subsequent cr5-directed digestion of the isolated complexes resulted in a single faster migrating species (Fig. 9A and B, lanes 3). Possibilities included: (i) two different U3-site2 crosslinks containing the same sized site2 RNA were co-migrating; (ii) each half of the digested 244 nt site2 fragment contained a U3-site2 crosslink and these were comigrating; (iii) cr5-targeted digestion had destroyed one U3-site2 interaction and a single U3-site2 complex remained. The first possibility was favored by subsequent analyses. U3-site2 complexes were directly isolated from cr5-cr13-targeted Ps+ RNAs (Fig. 9C and D, lanes 5), verifying that sequences between 913 and 1052 nt crosslinked with U3. Photoreversal released only the expected ~140 nt fragment. Yet when these isolated U3-site2 complexes were digested in the presence of cr12 (Fig. 9C and D, lanes 7), two species of increased mobility were produced. Two equivalent complexes were also isolated directly from cr5-cr12-targeted Ps+ RNAs (data not shown). These results indicated that at least one, perhaps two, U3-site2 crosslinks occurred between 913 and 1018 nt in 5[prime]ETS sequences.

Figure 9. Localization of crosslinks in U3-site2 complexes by oligonucleotide-directed RNaseH digestion. (A and B) and (C and D) Identical northerns of 5% denaturing PAGE gels transfered to Biodyne Plus nylon membranes. These were first hybridized with antisense site2 probes (A and C), and subsequently with antisense U3 probes (B and D). Complexes isolated from cr17-cr13- (A and B) or cr5-cr13-targeted (C and D) Ps+ RNAs were treated to photoreverse crosslinks (lanes 2 and 6) or digested by RNaseH in the presence of the indicated antisense site2 (lanes 3 and 7) or U3 RNA (lanes 8-10) oligodeoxynucleotides. Indicated are positions of U3-site2 complexes (all panels), and free site2 (A and C) and U3 RNAs (B and D) (see also Fig. 7 legend).

Psoralen adducts within site2 5[prime]ETS RNAs crosslinked to U3 were next located by reverse transcription. U3-site2 complexes concentrated from cr4-cr13-targeted Ps+ RNAs served as templates, and cr12, shown to pair 3[prime] to crosslinked residues, was the primer (Fig. 10A). A strong extension stop was seen at G946 and, together with the aforementioned RNaseH data, indicated that U945 was crosslinked with U3 RNA. If a second U3-site2 crosslink occurred within the 913-1018 nt region as suggested by complex digestion patterns, it was not detected, yet background extension products attributed to E.coli tRNA (shown) or rRNA carrier (not shown) could have obscured its presence. Attempts to develop other experimental conditions and to better examine the 913-1018 nt region were not successful.

Figure 10. (A) Primer extension analysis of 5[prime]ETS sequences within the U3-site2 complex to identify adducted bases. cr12 was used to prime reverse transcription reactions of: 0.5 µg untreated RNAs (lane 1); U3-site2 complexes selected from 100 µg cr17-cr13-targeted Ps+ (lane 2) or Ps- (lane 3) RNAs; 0.5 µg of supernate Ps+ and Ps- RNAs (lanes 5 and 6), and 4.5 µg carrier E.coli tRNAs (lane 6). Stop positions indicated are the 5[prime] end of the RNA fragment (~U688) and preceding the U945 adduct. Marker lanes are dideoxy sequencing lanes of pG.5[prime]ETS. (B) Model of the U3 RNA interaction with site2 5[prime]ETS sequences. A base pairing scheme is diagrammed, showing positions of psoralen adducts and possible crosslinked bases.

Positions of U3 bases crosslinked to the cr5-cr13-targeted 5[prime]ETS fragment were localized by RNaseH digestion of U3-site2 complexes (Fig. 9C and D, lanes 8-10). Several U3 5[prime] end bases had been implicated in crosslinks with site2 RNAs (Fig. 3). Antisense U3 oligonucleotides (Fig. 5C) directed appropriate digestion of U3 and did not lead to digestion of site2 RNAs. cU3.27 targeted cleavage of the U3-site2 complex indicating that crosslinks lay 3[prime] to nt 24. Digestion in the presence of hinge region cU3.29.5 resulted in a fast migrating complex that hybridized weakly with antisense U3 probes, indicating that the 5[prime] 39 nt of U3 remained crosslinked to site2 RNA. Supporting this conclusion, cU3.31-targeted digestion resulted in a comparatively slower complex containing the 5[prime] 64 nt of U3 (data not shown). Digestion of the U3-site2 complex targeted by cU3.28 that spans the box A region resulted in a distinctive pattern: two species that only weakly hybridized with site2 and U3 probes migrated between cU3.27- and cU3.29.5-targeted complexes, attesting that much of the complex was completely degraded and what remained held 3[prime] fragments of U3. A small, putative complex that hybridized to site2 probes might contain 5[prime] U3 fragments. That the complexes in each digestion did not resolve tightly by PAGE might be due to the presence of multiple U3-site2 crosslinks, as suggested above. This RNaseH data indicated that U3 bases 25-39 contained all crosslinks with the cr5-cr13-generated site2 RNA. The psoralen adducted U3 RNA bases: U25, U28 and U30 (Fig. 3) found within box A sequences were indicated in these crosslinks. The digestion patterns of U3 RNA in U3-site1 and U3-site2 complexes were clearly dissimilar, establishing that discrete regions of U3 RNA contacted different 5[prime]ETS sequences.

One base pairing model that accommodates this U3-site2 crosslink data is presented in Figure 10B. Site2 residue U945 is diametrically opposed to U3 residues U28 and U30. Psoralen additions to U25 could be accounted for in this model due to its presence in a helical region, however, a crosslink with site2 RNA is not suggested, and other models are possible. Interestingly, sequences with strong base pairing potential to U3 were noted upstream of the actual site2 crosslink: 5[prime]ETS nt 920-937 and U3 nt 25-44 could pair in 13 of 16 positions (data not shown). A similar phenomenon was noted for sequences flanking the site1 residues crosslinked to U3 RNA (see above). Potentially, these multiple 5[prime]ETS sequence complementarities with U3 RNA have a role in actual U3-5[prime]ETS interactions.

Some psoralen additions to U3 RNA in U3-site2 complexes containing cr17-cr13-generated site2 fragments (Fig. 3) were not accounted for in crosslinks to the cr5-cr13 RNA fragment. Notably, the 5[prime] most adduct noted in U3 RNA at U8 must indicate a crosslink to 5[prime]ETS sequences, rather than a monoaddition or U3 internal crosslink. A faint transcription stop at G862 was often noted in the cr17-cr13-fragment in U3-site2 complexes that possibly indicates a candidate contact between U861 and U3 base U8. However, as no U3 interaction with the cr17-cr5 fragment that contained U861 was resolved in RNaseH studies (see discussion of Fig. 9, above), further experimentation is required to count or discount this possibility.

DISCUSSION

In this work, in vivo interactions between T.brucei U3 snoRNA and pre-rRNAs were investigated by psoralen crosslinking. Molecular biological approaches were used to identify, dissect, and analyze two distinctive crosslinks between U3 RNA and 5[prime]ETS sequences. U3-pre-rRNA interactions first mapped by Sandwich Southern analysis were corroborated by subsequent resolution of individual crosslinks. U3-site1 or U3-site2 complexes were isolated from digested 5[prime]ETS RNA fragment pools, and bases involved in each crosslink were localized in U3 and 5[prime]ETS RNAs by targeted digestion of complexes and reverse transcription of each RNA component. These methods unequivocally determined that different U3 RNA sequences contacted separate regions of the 5[prime]ETS. U3 hinge bases crosslinked to processed 5[prime]ETS RNAs just downstream from the major cleavage site (site1), whereas box A residues contacted downstream 5[prime]ETS sequences (site2). These are unique crosslinks in comparison to those noted in other organisms (Fig. 11), and expand our understanding of the multifunctional potential of U3 RNA in rRNA processing.

Figure 11. Comparison of trypanosomal U3-5[prime]ETS crosslinks to those mapped in yeast (21) and mouse (20). 5[prime]ETS sizes and positions of major cleavage site are indicted. Regions of U3 in each interaction are noted by U3A[prime] for box A[prime], U3H for hinge, and U3A for box A regions. U3? indicates that the U3 residues in this putative yeast interaction have not been specified.

The T.brucei U3-site1 5[prime]ETS interaction is novel in that it is the first evidence of a close association between U3 RNA and a major 5[prime]ETS processing site. This result suggests a direct action for U3 in readying this site for cleavage. The mechanism by which U3, or any snoRNA, stimulates pre-rRNA cleavage is unknown. In yeast, an RNaseIII activity has been described that cleaves A0 in vitro and is required for this U3-dependent 5[prime]ETS cleavage in vivo, attesting that the role of U3 in processing is non-catalytic (12). U3 snoRNP instead might recruit processing enzymes to cleavage sites, or might effect pre-rRNA structure and substrate availability through base paired interactions. The latter notion fits with chaparone-like roles proposed for U3 in 18S rRNA pseudoknot assembly (6), and for U8 snoRNA in directing 5.8S and 28S rRNA interactions during pre-28S rRNA processing (34). In this regard, it is intriguing that isolated trypanosome U3-site1 crosslinks contained a small 5[prime]ETS fragment digested not only at the primary site, but at a site ~65 nt downstream as well. It is conceivable that U3-site1 crosslink formation led to downstream RNA degredation via a U3-induced conformational change in 5[prime]ETS structure.

Early pre-rRNA processing sites in T.brucei and metazoans are similarly located in the 5[prime] portion of their 5[prime]ETS sequences, suggesting these might be related cleavage events. Mouse and Xenopus cultured cell extracts support U3-dependent cleavage of this sequence in short exogenous 5[prime]ETS substrates, on which ribonucleoprotein complexes containing U3 are detected (7,8,35). Minimal-sized substrates contain an indispensable, 11 nt conserved sequence, and well-conserved sequences up to 200 nt downstream stimulate processing (8,10,35). The base similarities between metazoan 5[prime]ETS processing sites do not extend to protist 5[prime]ETS sequences. Yet in comparison to metazoans, only sequences proximal to the major cleavage site of T.brucei 5[prime]ETS are essential for its processing. In trypanosomes carrying chimeric gene constructs wherein mRNAs are transcribed via the rRNA promoter, certain pre-rRNA-mRNA transcripts having 5[prime]ETS sequences extending to 260 nt are efficiently processed at 115-116 nt (36). Thus, sequences containing the U3-site1, but not the downstream U3-site2, interaction are critical for primary 5[prime]ETS cleavage. U3 must contact intact 5[prime]ETS to have a role in its cleavage, yet interactions with low abundance unprocessed pre-rRNAs might be undetectable. Instead, trypanosome U3 crosslinks to cleaved pre-rRNAs were identified, possibly in accordance with the observation in metazoan systems that U3-containing complexes remain associated with processing competent substrates following primary site cleavage (7,20,35).

The trypanosome U3-site1 RNA crosslink bears resemblance to the functionally significant U3-5[prime]ETS interaction examined in S.cerevisiae, in that U3 single-stranded hinge sequences are implicated in each (22,23; Fig. 11). However, trypanosome U3 contacts bases 3[prime] proximal to the primary 5[prime]ETS cleavage, whereas the yeast U3 pairs 130 nt 5[prime] to the A0 cleavage; 5[prime]ETS RNA secondary structure may orient the yeast U3 binding site spatially closer to the downstream A0 and A1 cleavages it affects. Primer extension stops were noted in 5[prime]ETS 5[prime] proximal to the proposed U3-5[prime]ETS duplex in two species of yeast (22,37), suggestive of a conserved, U3-associated cleavage in yeasts that extends to trypanosomes. However, recent evidence shows that A0 cleavage releases intact upstream 5[prime]ETS sequences in yeast lacking exosome or Dop1p helicase activities, therefore endonucleolytic cleavage near the site of U3 interaction apparently does not have a major role in yeast rRNA processing (38). Perhaps the commonality between the yeast and trypanosome U3-5[prime]ETS interactions is that U3 snoRNP bound to 5[prime]ETS via hinge bases is in a configuration that corresponds with alteration of 5[prime]ETS structure and/or processing factor recruitment to a spatially nearby cleavage site.

The 5[prime] end sequences of all U3 RNAs share similarity and contain the highly conserved box A residues (nt 17-31), expected to have important functional roles in pre-rRNA processing. The T.brucei U3-site2 interaction is the first clearly delineated crosslink involving U3 RNA box A bases and 5[prime]ETS sequences, and it occurs relatively near the 3[prime] end of the 5[prime]ETS. In yeast, several adducts occurring within and 5[prime] of box A, in addition to hinge bases implicted in the functional U3-5[prime]ETS interaction, were noted in U3 RNAs crosslinked to the 35S pre-rRNA (21). A non-paired interaction between unaccounted for U3 crosslinks and a base lying between A0 and A1 was postulated, but the exact contacts were not resolved. Alternatively, yeast U3 hinge region crosslinks could represent recently indicated interactions with 5[prime] end 18S rRNA sequences. In one report, phylogenetic comparisons noted three blocks of conserved sequence in the 5[prime] end of U3 having base complementarity to pseudoknot elements in the 5[prime] end of 18S rRNAs (6). Yeast mutants expressing a U3 with four base substitutions in the box A-containing element were cold sensitive for growth and were deficient in processing at A1 and A2, but not at the 5[prime]ETS A0 site. These results demonstrated a role for box A bases in 18S formation, and functional separation of U3 roles in 5[prime]ETS cleavage and 18S maturation (6). Of late, a base paired interaction between U3 box A and 5[prime] 18S rRNA pseudoknot sequences has been shown to be required for A1 and A2 cleavage and 18S accumulation (24). As these U3-18S rRNA pairings occur across species, U3 functions in 18S rRNA maturation are surely conserved. It is intriguing, then, that trypanosome box A sequences clearly crosslink 5[prime]ETS sequences. Together, these data suggest that U3 snoRNP interactions with pre-rRNAs via box A bases serve multiple functions.

U3 RNA bases 5[prime] to box A have been detected in U3-5[prime]ETS or U3-pre-rRNA crosslinks in rat, mouse, yeast and trypanosome systems. 5[prime] U3 bases including box A base U23 contacted a non-specified 5[prime]ETS sequence within 350 nt of the primary processing site in rat cells (19), whereas in mouse cultured cell extracts, U3 bases C4, U5 and U8 crosslinked an exogenous, processing competent substrate at a position 362 nt downstream of the primary cleavage (20). This latter interaction was not required for cleavage, yet might have in vivo roles in downstream processing events. Perhaps a related interaction occurs in T.brucei, as the conserved U8 base appeared crosslinked, at an undiscerned location, with 5[prime]ETS site2 sequences that are likewise not required for primary cleavage. Altogether, it is apparent that the 5[prime]-most sequences of U3 RNA interact with pre-rRNAs, though the function of these contacts is presently unknown.

U3 is the only snoRNA essential for 5[prime]ETS processing, yet it may function together with other snoRNAs. In yeast, U3 is postulated to work in a multi-snoRNP complex with U14, snR30 and snR10; the latter snoRNAs influence A0 cleavage and conjointly with U3 are critical for A1 and A2 cleavages and 18S rRNA maturation (3,4). In mouse cell extracts, U3-dependent processing of 5[prime]ETS substrates was stimulated by U14, E3 and U17 snoRNAs, implying that these act together in a complex (14). Possibly, interactions between U3 and other snoRNAs occur. In this study, three T.brucei U3-crosslinked species were seen in Ps+ RNAs. The small U3Xa might represent internally crosslinked U3, and U3Xb contained the site1 65 nt fragment. U3Xc could contain an RNA of ~80 nt that was not related to 5[prime]ETS sequences by hybrid selection and northern analyses. The possibility that U3Xc included a snoRNA in rRNA processing is being investigated. One candidate might be the conserved U14 snoRNA. U14 shares two regions of base complementarity with SSU rRNA, one of which is essential for function and can be crosslinked to 18S rRNA sequences in yeast (39,40). A search for a T.brucei U14 equivalent using oligonucleotide probes complementary to conserved regions of U14, or similar to appropriate trypanosome SSU rRNA sequences, proved futile (unpublished data). Thus, if a trypanosome U14 exists, it does not possess conserved pairing motifs to SSU rRNA seen in more recently evolved organisms.

A model for U3 roles in trypanosome rRNA processing that incorporates these crosslinking results and data from other systems is proposed: U3 snoRNP first pairs via hinge bases with sequences 3[prime] adjacent to the 5[prime]ETS cleavage site and alters pre-rRNA conformation for riboendonuclease recognition and/or promotes formation of a processing complex in which cleavage occurs. Subsequent U3 interactions via box A bases with downstream 5[prime]ETS sites (and possibly SSU rRNA sequences) might affect events leading to 18S rRNA maturation. As distinct U3-5[prime]ETS contacts involve different U3 sequences, one U3 snoRNP could conceivably link sites of interaction and possibly juxtapose an upstream processing complex to a 3[prime] location, or different U3 snoRNPs might interact at each site. 5[prime] end, box A-containing sequences occur in a possible stem-loop structure, and it is feasible that alternative conformations of U3 RNA are involved in dissimilar interactions with pre-rRNAs. To test aspects of this model, molecular genetic approaches are being used to determine the functional significance of U3-5[prime]ETS interactions in T.brucei.

ACKNOWLEDGEMENTS

I am grateful to Nina Agabian for providing a stimulating laboratory environment during the early phases of this work. I thank Jan Dungan for in depth discussions of RNA methodologies, David Tollervey for communication of unpublished results, and Angus Lamond for the generous gift of 2[prime]-O-allyl-oligoribonucleotides. This work was supported by NIH grant AI34093.

REFERENCES

1. Eichler,D.C. and Craig,N. (1994) Prog. Nucl. Acid Res. Mol. Biol. 49, 197-239.

2. Maxwell,E.S. and Fournier,M.J. (1995) Annu. Rev. Biochem. 35, 897-934.

3. Venema,J. and Tollervey,D. (1995) Yeast 11, 1629-1650. MEDLINE Abstract

4. Tollervey,D. and Kiss,T. (1997) Curr. Opin. Cell Biol. 9, 337-342. MEDLINE Abstract

5. Hughes,J.M.X. and Ares,M. (1991) EMBO J. 10, 4231-4239. MEDLINE Abstract

6. Hughes,J.M.X. (1996) J. Mol. Biol. 259, 645-654. MEDLINE Abstract

7. Kass,S., Tyc,K., Steitz,J.A. and Sollner-Webb,B. (1990) Cell 60, 897-908. MEDLINE Abstract

8. Mougey,E.B., Pape,L.K. and Sollner-Webb,B. (1993) Mol. Cell. Biol. 13, 5990-5998. MEDLINE Abstract

9. Savino,R. and Gerbi,S.A. (1990) EMBO J. 7, 2299-2308.

10. Craig,N., Kass,S. and Sollner-Webb,B. (1991) Mol. Cell. Biol. 11, 458-467. MEDLINE Abstract

11. Shumard,C.M. and Eichler,D.C. (1988) J. Biol. Chem. 263, 19346-19352. MEDLINE Abstract

12. Elela,S.A., Igel,H. and Ares,M.J. (1996) Cell 85, 115-124. MEDLINE Abstract

13. Lafontaine,D. and Tollervey,D. (1995) Biochem. Cell Biol. 73, 803-812. MEDLINE Abstract

14. Enright,C.A., Maxwell,E.S. and Sollner-Webb,B. (1996) RNA 2, 1094-1099. MEDLINE Abstract

15. Epstein,P., Reddy,R. and Busch,H. (1984) Biochemistry 23, 5421-5425. MEDLINE Abstract

16. Tollervey,D. (1987) EMBO J. 6, 4169-4175. MEDLINE Abstract

17. Hartshorne,T. and Agabian,N. (1994) Mol. Cell. Biol. 13, 144-154.

18. Maser,R.L. and Calvet,J.P. (1989) Proc. Natl. Acad. Sci. USA 86, 6523-6527. MEDLINE Abstract

19. Stroke,I. and Weiner,A.M. (1989) J. Mol. Biol. 210, 497-512. MEDLINE Abstract

20. Tyc,K. and Steitz,J.A. (1992) Nucleic Acids Res. 20, 5375-5382. MEDLINE Abstract

21. Beltrame,M. and Tollervey,D. (1992) EMBO J. 11, 1531-1542. MEDLINE Abstract

22. Beltrame,M. and Tollervey,D. (1995) EMBO J. 14, 4350-4356. MEDLINE Abstract

23. Beltrame,M., Henry,Y. and Tollervey,D. (1994) Nucleic Acids Res. 22, 5139-5147. MEDLINE Abstract

24. Sharma,K. and Tollervey,D. (1998) personal communication.

25. White,T.W., Rudenko,G. and Borst,P. (1986) Nucleic Acids Res. 14, 9471-9489.

26. Campbell,D.A., Kubo,K., Clark,C.G. and Boothroyd,J.C. (1987) J. Mol. Biol. 196, 113-124. MEDLINE Abstract

27. Hartshorne,T. and Agabian,N. (1994) Nucleic Acids Res. 22, 3354-3364. MEDLINE Abstract

28. Mereau,A., Fournier,R., Gregoire,A., Mougin,A., Fabrizio,P., Lührmann,R. and Branlant,C. (1997) J. Mol. Biol. 273, 552-571. MEDLINE Abstract

29. Parker,K.A. and Steitz,J.A. (1987) Mol. Cell. Biol. 7, 2899-2913. MEDLINE Abstract

30. Jeppesen,C., Stebbins-Boaz,B. and Gerbi S.A. (1988) Nucleic Acids Res. 16, 2127-2148. MEDLINE Abstract

31. Watkins,K.P., Dungan,J.M. and Agabian,N. (1994) Cell 76, 171-182.

32. Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual. 2nd edition. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

33. Ericson,G. and Wollenzien,P. (1988) Anal. Biochem. 174, 215-223. MEDLINE Abstract

34. Peculis,B.A. (1997) Mol. Cell. Biol. 17, 3702-3713. MEDLINE Abstract

35. Kass,S. and Sollner-Webb,B. (1990) Mol. Cell. Biol. 10, 4920-4931. MEDLINE Abstract

36. Chapman,A. and Agabian,N. (1995) personal communication.

37. Brule,F., Venema,J., Segault,V. and Tollervey,D. (1996) RNA 2, 183-197. MEDLINE Abstract

38. de la Cruz,J., Kressler,D., Tollervey,D. and Linder,P. (1998) EMBO J. 17, 1128-1140. MEDLINE Abstract

39. Morrissey,J.P. and Tollervey,D. (1997) Chromosoma 105, 515-522.

40. Jarmolowski,A., Zagorski,J., Li,H.V. and Fournier,M.J. (1990) EMBO J. 9, 4503-4509. MEDLINE Abstract

41. Dungan,J.D., Watkins,K.P. and Agabian,N. (1996) EMBO J. 15, 4016-4029.

42. Myslinski,E., Segault,V. and Branlant,C. (1990) Science 247, 1213-1216. MEDLINE Abstract

43. Cimino,G.D., Gamper,H.B., Isaacs,S.T. and Hearst,J.E. (1985) Annu. Rev. Biochem. 54, 1151-1193. MEDLINE Abstract

---

*Tel: +1 518 262 6654; Fax: +1 518 262 5689; Email: toinette_hartshorne{at}ccgateway.amc.edu

---

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 19 May 1998
Copyright©Oxford University Press, 1998.
CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				S. Barth, B. Shalem, A. Hury, I. D. Tkacz, X.-h. Liang, S. Uliel, I. Myslyuk, T. Doniger, M. Salmon-Divon, R. Unger, et al. Elucidating the Role of C/D snoRNA in rRNA Processing and Modification in Trypanosoma brucei Eukaryot. Cell, January 1, 2008; 7(1): 86 - 101. [Abstract] [Full Text] [PDF]

				O. Fernandez-Miragall and E. Martinez-Salas In vivo footprint of a picornavirus internal ribosome entry site reveals differences in accessibility to specific RNA structural elements J. Gen. Virol., November 1, 2007; 88(11): 3053 - 3062. [Abstract] [Full Text] [PDF]

				S. Barth, A. Hury, X.-h. Liang, and S. Michaeli Elucidating the Role of H/ACA-like RNAs in trans-Splicing and rRNA Processing via RNA Interference Silencing of the Trypanosoma brucei CBF5 Pseudouridine Synthase J. Biol. Chem., October 14, 2005; 280(41): 34558 - 34568. [Abstract] [Full Text] [PDF]

				A. V. BOROVJAGIN and S. A. GERBI Xenopus U3 snoRNA docks on pre-rRNA through a novel base-pairing interaction RNA, June 1, 2004; 10(6): 942 - 953. [Abstract] [Full Text] [PDF]

				A. Colley, J. D. Beggs, D. Tollervey, and D. L. J. Lafontaine Dhr1p, a Putative DEAH-Box RNA Helicase, Is Associated with the Box C+D snoRNP U3 Mol. Cell. Biol., October 1, 2000; 20(19): 7238 - 7246. [Abstract] [Full Text]

				M. N. Schnare, J. C. Collings, D. F. Spencer, and M. W. Gray The 28S-18S rDNA intergenic spacer from Crithidia fasciculata: repeated sequences, length heterogeneity, putative processing sites and potential interactions between U3 small nucleolar RNA and the ribosomal RNA precursor Nucleic Acids Res., September 15, 2000; 28(18): 3452 - 3461. [Abstract] [Full Text] [PDF]

				M. Antal, A. Mougin, M. Kis, E. Boros, G. Steger, G. Jakab, F. Solymosy, and C. Branlant Molecular characterization at the RNA and gene levels of U3 snoRNA from a unicellular green alga, Chlamydomonas reinhardtii Nucleic Acids Res., August 1, 2000; 28(15): 2959 - 2968. [Abstract] [Full Text] [PDF]

				K. Sharma and D. Tollervey Base Pairing between U3 Small Nucleolar RNA and the 5' End of 18S rRNA Is Required for Pre-rRNA Processing Mol. Cell. Biol., September 1, 1999; 19(9): 6012 - 6019. [Abstract] [Full Text] [PDF]

				D. A. Dunbar, S. Wormsley, T. M. Lowe, and S. J. Baserga Fibrillarin-associated Box C/D Small Nucleolar RNAs in Trypanosoma brucei. SEQUENCE CONSERVATION AND IMPLICATIONS FOR 2'-O-RIBOSE METHYLATION OF rRNA J. Biol. Chem., May 5, 2000; 275(19): 14767 - 14776. [Abstract] [Full Text] [PDF]

This Article Abstract Alert me when this article is cited Alert me if a correction is posted Services