Primary and secondary structure of U8 small nuclear RNA.

U8 small nuclear RNA is a new, capped, 140 nucleotides long RNA species found in Novikoff hepatoma cells. Its sequence is: m3GpppAmUmCGUCAGGA GGUUAAUCCU UACCUGUCCC UCCUUUCGGA GGGCAGAUAG AAAAUGAUGA UUGGAGCUUG CAUGAUCUGC UGAUUAUAGC AUUUCCGUGU AAUCAGGACC UGACAACAUC CUGAUUGCUU CUAUCUGAUUOH. This RNA is present in approximately 25,000 copies/cell, and it is enriched in nucleolar preparations. Like U1, U2, U4/U6, and U5 RNAs, U8 RNA was also present as a ribonucleoprotein associated with the Sm antigen. The rat U8 RNA was highly homologous (greater than 90%) to a recently characterized 5.4 S RNA from mouse cells infected with spleen focus-forming virus (Kato, N., and Harada, F. (1984) Biochim. Biophys. Acta, 782, 127-131). In addition to the U8 RNA, three other U small nuclear RNAs were found in anti-Sm antibody immunoprecipitates from labeled rat and HeLa cells. Each of these contained a m3GpppAm cap structure; their apparent chain lengths were 60, 130, and 65 nucleotides. These U small nuclear RNAs are designated U7, U9, and U10 RNAs, respectively.

nuclear and nucleolar RNAs were done as described previously (20). Estimation of the copy number of US RNA was obtained from the ratio of radioactivity in U8 RNA and other U-snRNAs like U1 and U2 RNA. Fig. 1 shows the fractionation of small RNAs present in the immunoprecipitates obtained using the anti-Sm antibodies (lanes 2 and 3). Lanes 1 and 4 show the RNAs present in the starting whole cell sonicates of the HeLa cells and Novikoff hepatoma cells, respectively. In addition to the major U-snRNAs (Ul, U2, U4, U5, and U6 RNAs) (2), several minor RNA bands were observed. Three bands which contain the U-snRNAs designated U8, U9, and U7/U10 were found consistently both in HeLa cells and in Novikoff hepatoma cells. RNA bands with similar mobility were observed when monoclonal anti-Sm antibodies (a kind gift from Dr. J. A. Steitz, Department of Molecular Biophysics, Yale University) were used or when sodium dodecyl sulfate/phenol-extracted total RNA was immunoprecipitated with anti-maG cap antibodies (a kind gift from Dr. R. Luhrmann, Max-Planck-Institute, Berlin, West Germany) (results not shown). These RNAs were analyzed further.

RESULTS
On a 50-cm long, thin 12% polyacrylamide gel (Fig. 2), the RNA band with an apparent chain length of 60 nucleotides (designated U7/U10 RNA) was separated into several distinct bands which contained capped small RNAs. All the RNA bands were analyzed for the presence of cap structure and by fingerprinting after T, RNase digestion. The faster migrating bands had similar fingerprints and were uridylic acid-rich (35%). Following the designation of Strub et al. (25), these RNAs were designated U7 RNAs and these may be homologous to U7 RNA of sea urchin. Fig. 3 shows the analysis of cap structure of U7/U10 RNAs in the anti-Sm antibody immunoprecipitates. After each RNA was digested with P1 nuclease, the products were fractionated on DEAE-cellulose paper. W7 to U10 RNAs contained a nuclease PI-resistant structure with the mobility of maGpppAm of rat U2 RNA (Fig. 3, lane I). The cap structures of U7 to U10 RNAs were also analyzed by chromatography following the method of Silberklang et al. (26), and the cap structures had the same mobility as that of msGpppAm obtained from U2 RNA (results not shown). The fingerprints of U7, U9, and U10 RNAs were distinct and did not match the fingerprints of other known U-snRNAs (results not shown). Fig. 4 shows the analysis of 4-8 S RNAs isolated from nucleoli (lane I ) and nuclei (lane 2) of Novikoff hepatoma. As found earlier (20), the nucleoli are enriched in 5 S RNA, U3 RNA, 7-1 RNA, 7-2 RNA, and 8 S RNA; the nucleoplasmic U1, U2, U4, U5, and U6 RNAs were present in lower concentrations. Since nucleolar RNA was enriched in U8 RNA (lane I ) , U8 RNA, like U3 RNA, is probably localized to the nucleolus. ["'Pjorthophosphate, and immunoprecipitations were carried out as described hy Lerner and Steitz (2). The cell supernatant used was obtained after centrifugation a t 100,000 X g for 1 h. The RNAs in the immunoprecipitates were analyzed on a 10% acrylamide, 7 M urea gel. The xylene cyanol dye marker migrated slightly slower than U7/U10 RNAs. Lane I , HeLa cell supernatant, used as starting material; lane 2, RNAs found in anti-Sm antibody immunoprecipit.ate from HeLa cells; lane 3, RNAs found in anti-Sm antibody immunoprecipitate from Novikoff hepatoma cells; lane 4. Novikoff hepatoma cell supernatant used as starting material. RNA bands, designated U8, U9. and U7/U10, were observed consistently. Several other bands were observed in some experiments, especially between LJ4 RNA and 5 S RNA (see lune 2). These may be other minor U-snRNAs or degradation products of major U-snRNAs or RNAs precipitated nonspecifically. 7SL and 7SK RNAs are wellcharacterized cytoplasmic RNAs (3, 4).
did not correspond to that of any known U-snRNAs. In some experiments, oligonucleotides characteristic of U1 RNA were also found in the fingerprints of U8 RNA; these may result from co-migration of U1* RNA with U8 RNA (13). The TI oligonucleotides of U8 RNA were analyzed after T2 RNase, RNase A, Pl nuclease, or U2 RNase digestions. Some oligonucleotide sequences such as AAAAUG (spot 12 isolated and fractionated on a 50-cm long. 12% polyacrylamide, 7 M urea gel. The xylene cyanol marker was run for a total of 45 cm, and RNAs were visualized by autoradiography. Two to three major RNA hands and several minor RNA bands were observed. All these RNAs were analyzed for cap structure and by fingerprinting. The RNA bands labeled U7 RNAs gave similar fingerprints, and RNA hands labeled U10 RNAs gave similar fingerprints hut different from 117 RNAs. gonucleotide T-20 contained the cap structure and T-3 contained the 3'-end of UR RNA (Fig. SA). Fig. 6 shows sequencing gels obtained with 3'-end-laheled U8 RNA. The 3"end-labeled RNA was reacted under hasespecific conditions according to Peattie (22) and analyzed on sequencing gels. Fig. 6A shows the sequence of UR RNA from nucleotides 99 to 136; Fig. 6R shows the sequence from  hepatoma U8 RNA which is slightly different from that structure, all the S1 nuclease-sensitive sites observed are in reported by Kat0 and Harada (27) for mouse U8 RNA. In this single-stranded regions (marked by arrows). A putative Sm antigen-binding site found to be uridylic acid-rich and singlenezt to the nucleotides correspond to the nucleotide numbers in the stranded in Ul , U2, U4, and U5 RNAs (28-31) is also single-U8 RNA sequence. The sequence of the oligonucleotide T-18 is that &Ianded in this proposed secondary structure for u8 RNA of a minor sequence variant, different from the T-18 shown in the (nucleotides 85-98) and may be the Sm-binding site (Fig. 8).  RNAs and a rat U7 RNA. The TI RNase fingerprint of UlOB RNA was obtained from the U10 RNA species with most radioactivity (see Fig. 2 A ) . The fingerprints of other U10 RNAs in this size range were identical to the fingerprint of UlOR RNA. The fingerprints showed 11 oligonucleotides, 1 to 11, in molar yield and 2 oligonucleotides, 12 and 13, in less than molar yield. These oligonucleotides were digested with U2 RNase or with RNase A, and the results are as follows: 1, AUCCGp and CACUGp; 8, CUUGp; 9, UAUUGp; 10, C4AZU2Gp; and 11, CaA2UGp. In addition to the three U10 RNA species with apparent chain lengths of 60 nucleotides, an RNA species with apparent chain length of 120 nucleotides was also found in anti-Sm antibody immunoprecipitates. This RNA was designated UlOD, since it contained all the oligonucleotides, 1 to 13, found in UlOB RNA and at least 4 additional oligonucleotides, numbered 14 to 17 (Fig. 9). Both of these RNAs contained the cap structure, indicating that the 5'-end may be the same and t.hat the UlOD species may be longer at the 3'-end when compared to UlOB RNA. The analysis of RNase A or U2 RNA digests of 1 to 13 oligonucleotides from UlOR and UlOD RNA yielded the same products. The UlOD RNA contained about 1 5 2 0 % of the total radioactivity in U10 RNAs.
The U7 RNA from the Novikoff hepatoma cells was fingerprinted and shown in Fig. 9. The RNA species still contained some U10 RNA as a contaminant, and the oligonucleotides consistently observed were designated 1 to 11 and shown in Fig. 9. Since all the U-snRNAs previously characterized in higher eukaryotes contained modified nucleotides (1, 3), the U7 RNA was analyzed for modified nucleotides. The human U7 RNA contained 2"O-methylated cytidine and pseudouridine in 0.6-0.8 molar yield, in addition to the cap structure (Fig. 9). Other modified nucleotides were observed in less than 0.5 molar yield. ; the first dimension was electrophoresis on cellulose acetate strips, and the second dimension was homochromatography using C -15 homomixture. The analysis of modified nucleotides was by the method of Silherklang et a1 (26). The molar yield of p$, pCm. and m&pppAm was between 0.6 and 0.8. The molar yield of pm6A was 0.3

DISCUSSION
In this study, four minor U-snRNAs were analyzed. The number of U-snRNAs identified in higher eukaryotes is therefore extended from 6 to 10. The US RNA identified in this study probably corresponds to the U7 RNA shown in the 12 S ribonucleoprotein particle involved in the accurate processing of sea urchin histone mRNAs (18,25,32,33). Rirnstiel and colleagues (2.5) showed this U7 RNA was immunoprecipitable with anti-Sm antibodies. Our results confirm this observation and the indication that U7 RNA containn a trimethylguanosine cap structure (18); therefore, U7 RNA is a member of the U-snRNA series by the criteria that it has a cap structure and is present in Sm-containing ribonucleoprotein particle.
U8 RNA was first identified in nucleolar preparations and was designated 5.4 S RNA because of its electrophoretic migration between 5 S RNA and 5.8 S RNA (20). Although the function of this RNA is not known, the characteri711tion of U8 RNA (5.4 S RNA) in mouse cells (27) and in Novikoff hepatoma cells in this investigation establishes this RNA to be a distinct member of the U-snRNA series. These results also show that U8 RNA is present in a ribonucleoprotein particle in association with Sm antigen and may be localized to nucleoli because it, like U3 RNA, was enriched in nucleolar preparations. However, its association with high molecular weight, nucleolar ribosomal RNAs was weak because no detectable U8 RNA was associated with high molecular weipht nucleolar RNAs extracted a t room temperature (20).
Although U3 RNA was conclusively shown to be present in nucleoli (1,3,4) and U8 RNA is found in nucleoli, no detectable nucleolar localization was observed with m3G cap-specific antibodies (34) or anti-Sm antibodies (35,36). The reason(s) for these apparently inconsistent results are not clear. One possibility is that the U8 RNA is a minor RNA and may not be contributing significant nucleolar immunofluorescence compared to nucleoplasm using anti-Sm antibodies. The other possibility is that nucleolar snRNAs may be tightly bound to proteins and are not available for binding to antibodies. Yet another, although unlikely, possibility is that both U3 and U8 RNA are only co-purifying with nucleoli and are really not inside the nucleolus. More studies will be needed to distinguish between these possibilities. The subcellular localization of U7 and U10 RNAs is not known since the U7 and U10 RNAs were run off this gel. Further experiments are needed to clarify this question. The amounts of U9 RNA were not sufficient to be visualized by methylene blue staining in both total nuclear and nucleolar 4-8 S RNA preparations.
All U-snRNAs associated with Sm antigen were shown to contain a uridylic acid-rich, single-stranded region that appears to be the Sm antigen-binding site (29-31). The secondary structure proposed for U8 RNA in this study is consistent with that hypothesis. The nucleotides 85 to 97 of U8 RNA are uridylic acid-rich, single-stranded, and flanked by two stem regions. These features are found in other Sm antigenbound snRNAs including U1, U2, U4/U6, and U5 RNAs (28-31). Therefore, this region of US RNA may be the Sm antigenbinding site.
The fingerprints of U9 and U10 RNAs were not similar to those of the other known U-snRNAs although they contained cap structures. Adams et al. (7) found a distinct trimethylguanosine-containing small RNA unrelated to other known U-snRNAs in silk worms. The RNA described by Adams et al. (7) was slightly larger than 5 S RNA and smaller than U4 RNA; it may be homologous to the U9 RNA. The U9 and U10 RNAs are new, distinct U-snRNA species in rat and human cells (Fig. 9).
The U10 RNA consisted of at least three RNAs, about 60 nucleotides long, and one RNA species, about 120 nucleotides long. The 5'-ends of UlOB RNA and UlOD RNA are similar, and it appears that UlOD RNA is longer on the 3'-end. It is not known whether UlOD is a precursor for other U10 RNAs or whether these are two functionally distinct but structurally related U-snRNAs.
The designation of U-snRNAs U7 to U10 was done using the following criteria. U-snRNAs U1 to U3 were first designated in the ascending order by Hodnett and Busch (37), and these were extended by Lerner and Steitz (2) to include U4 to U6 RNAs in descending order. A 60-nucleotide long Sm antigen-associated RNA was added by Strub et al. (25) to this series as U7 RNA. The designation of U7 RNA is supported by its cap structure. The other three distinct, capped, Sm antigen-associated RNAs are designated U8 to U10 RNAs, again in descending order. Although this nomenclature is imperfect, these symbols reflect molecular species with a variety of similar features.