Some Gene Variants for 5 S RNA Are Dispersed in the Rat Genome*

In the course of studies on genes for small nuclear RNAs, seven lambda phage clones containing sequences homologous to 5 S RNA were plaque purified from a rat genomic library. The seven clones were found to be from six different genomic loci. When the 5 S RNA hybridized to these clones was digested by T1 RNase, only clone 5S-2 protected the RNA completely. Moreover, clone 5S-2 which has five nucleotide substitutions in the internal control region was transcribed 10 times more efficiently than a bonafide Chinese hamster 5S gene. The other clones were less efficiently transcribed than a bonafide 5S gene or not transcribed at all. The number of gene variants for 5 S RNA in the rat genome was approximately 3000. In contrast to the clustering of 5S genes and gene variants found in Xenopus, Drosophila, hamster, mouse, and human cells, the 5S gene variants in the rat genome are dispersed and most contained conserved 3'-flanking sequences. These naturally occurring 5S gene variants may be useful in binding transcription factors that affect 5S genes.

In the course of studies on genes for small nuclear RNAs, seven X phage clones containing sequences homologous to 5 S RNA were plaque purified from a rat genomic library. The seven clones were found to be from six different genomic loci. When the 5 S RNA hybridized to these clones was digested by TI RNase, only clone 55-2 protected the RNA completely. Moreover, clone 55-2 which has five nucleotide substitutions in the internal control region was transcribed 10 times more efficiently than a bonafide Chinese hamster 5s gene. The other clones were less efficiently transcribed than a bonafide 55 gene or not transcribed at all. The number of gene variants for 5 S RNA in the rat genome was approximately 3000. In contrast to the clustering of 55 genes and gene variants found in Xenopus, Drosophila, hamster, mouse, and human cells, the 55 gene variants in the rat genome are dispersed and most contained conserved 3"flanking sequences. These naturally occurring 55 gene variants may be useful in binding transcription factors that affect 55 genes.
The genomic organization of 5 s genes has been studied in a number of eukaryotes such as yeast, Drosophila, Xenopus, and mouse (1-4). In most cases, the 5s genes are arranged in a tandem array having short repeat lengths with the bulk of the genes coding for the predominant form of 5 S RNA found in ribosomes. In addition to the developmen~lly regulated embryonic and somatic 5s genes (5,6), Xenopus also contains pseudogenes for 5s (7). The first pseudogenes described were for 5 S RNA in Xenopus (7). In contrast to the extensive information available on the genomic organization of 5 s genes in eukaryotes such as yeast (S), Neurospora (9), Drosophila (IO), and Xenopus (5, 6), there is little information available on the 5 s gene family in higher eukaryotes. The 5s gene family of hamster was characterized (4), and one 5 s gene variant from mouse and human genome was recently characterized (11,12).
With the intention of isolating 5 s genes from the rat geome, we isolated seven clones containing 5 S RNA homologous sequences from a rat genomic bank. These clones were characterized by transcription in vitro, fingerprinting of the 5s transcripts, sequencing of the 55 gene variants, and genomic arrangements in relation to other members of their family.
These studies indicate that there are many 5 s gene variants, and these are dispersed in the rat genome; one of these 55 gene variants was 10 times more efficiently transcribed in vitro than a bonafide Chinese hamster 5 s gene.
*These studies were supported by the Cancer Research Center Grant CA10893-P3, awarded by the National Cancer Institute, Department of Health and Human Services. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked " a d u e r t~e~n t " in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

MATERIALS AND METHODS
Chemicals-RNA ligase and polynucleotide kinase were from Pharmacia P-L Biochemicals. All restriction enzymes were obtained from New England Biolabs. The isotopes were obtained from Amersham Corp. The nucleotide triphospha~s and components for the buffers were obtained from Sigma.
RNA Probe-5 S RNA was prepared from Novikoff hepatoma cells (131, and the acrylamide gel-purified 5 S RNA was 3'-end labeled with [32P]pCp using RNA ligase (14). The labeled 5 S RNA was purified on acrylamide ge1s again, and 2 X IO6 to 5 X IO6 countslpg of 5 S RNA were routinely obtained.
Screening of Genomic Library-The screening of rat liver genomic library in X phage Charon 4A was carried out by the method of Benton and Davis (15). The partial ecoRI genomic library made from rat liver DNA in X phage Charon 4A was kindly provided by Dr. T. Sargent and Dr. J. Bonner (Division of Biology, California Institute of Technology, Pasadena, CA). Hybridizations with increasing stringency at 42, 52, or 58 "C were carried out in 50% formamide, 4 X SET (0.15 M NaCI; 1 mM EDTA; 0.03 M Tris-HC1, pH 8.0), 2 x Denhardt's reagent, 0.1% sodium dodecyl sulfate, and 10 pg/ml yeast transfer RNA. The nitrocellulose filters were hybridized for 16 h and washed two times each for 15 min with 3 X SSC, 1.5 X SSC, and 0.5 X SSC containing 0.1% sodium dodecyl sulfate at 42, 52, or 58 "C. The dried filters were autoradio~aphed using Kodak XAR-5 film and DuPont Lightning Plus screens at -70 "C.
Preparatwn and Analysis of 5 S RNA-The dot hybridizations were carried out by the method of Kafatos et al. (16). The T, RNase protection experiments were carried out as described by Weiner (17). Thioacetamide treatment of the rats was carried out as described by Ro-Choi et al. (18). Briefly, rats weighing about 200 g were injected intraperitoneally with 1 mg of thioacetamide (10 mg/ml in 0.9% NaC1) at 0 h and another 1-mg dose of thioacetamide at 12 h. 10 mCi of [32P]orthophosphate was injected at 24 h, and labeling was carried out in uiuo for 3 h. 5s RNA was isolated from these rat livers and analyzed by fingerprinting.
I n Vitro ~r u~c r i~~~~-W h o l e cell extracts from Novikoff bepatoma cells prepared by the method of Weil et al. (19) were used. The transcription conditions were as described by Sharp  Fingerprinting and Sequencing-RNA fingerprinting was carried out according to Brownlee et al. (20). The DNA sequencing was carried out by the method of Maxam and Gilbert (21).

RESULTS
About 50,000 phage plaques were screened with 5 S RNA as a probe. The hybridization and washings were carried out as described under "Materials and Methods" at 42 "6. Four hundred positive signals were observed after a 3-day exposure of the filters. The X Charon 4A library used for screening in this study contains 15-20 kilobase pairs of rat genomic DNA as inserts (22). Based on 6 x lo9 base pairs in the rat haploid ' The abbreviation used is: HEPES, 4-(2-hydroxyethyl)-l-pipera-_I__.
Dispersed 5s Gene Variants in the Rat Genome genome, screening 400,000 phage plaques is equivalent to screening rat haploid genome once. Therefore, it is estimated that the rat haploid genome contains about 3,000 copies of DNA that are homologous to 5 S RNA. Several of the plaques corresponding to areas showing strong autoradiographic signals from the first screening were screened at 52 "C and then were plaque-purified at 56 "C. Seven phage clones were chosen a t random and studied further. Fig. 1 shows the small RNAs that hybridized to each of the six different genomic clones of 5 S RNA. Lane 1 (STD) shows the gel pattern of whole cells 4-8 S RNA used for hybridization and lanes 2-8 (5s-1 to 6s-B) show the analysis of RNAs that hybridized to different DNAs. As expected, all clones bound 5 S RNA; in addition, clones 1, 3, and 5 hybridized to 4.5 S RNA and clone 2 hybridized to two minor 6 S-sized RNAs (Fig. 1). Since 4.5 S RNA is homologous to type I Alu DNA and 6 S RNA is homologous to type I1 Alu DNA (23-25), these clones contain Ah-type repeated DNA in the vicinity of 5 S DNA. These DNAs do not appear to contain sequences homologous to other small RNAs. 7s-u3-

FIG. 1. Analysis of small RNAs that hybridized to different
5 s cloned DNAs. 5 pg each of X phage DNA containing rat DNA inserts was immobilized on nitrocellulose circles and hybridized to a mixture of whole cell 4-8 S RNA labeled in uiuo. The filters were washed at 42 "C, and bound RNAs were analyzed on 10% polyacrylamide-7 M urea gels. The hybridization solutions and washing solutions used are described under "Materials and Methods." STD, 4-8 S RNA used for hybridization; lanes 5s-1 to 5S-6B, seven DNA preparations containing 5 S RNA homologous sequences. 6s-A and 6s-B were found to be from the same genomic locus. The 4.5 S RNA was found to be identical to mouse 4.5 S RNA described by Jelinek and Leinwand (23). 6s-A and 6s-B RNA gave similar fingerprints and may be homologous to 6 S size type I1 Alu RNA described by Haynes and Jelinek (24).
Six of the Seven Clones Are from Different Loci of the Rat Genome-The seven phage DNAs were digested with several restriction enzymes separately and fractionated on 1% agarose gels, transferred to nitrocellulose by the method of Southern and hybridized to 5 S RNA. Fig. 2 shows the results obtained with the phage DNAs digested with EcoRI. Six of the seven clones had different size EcoRI fragments that hybridize to 5 S RNA. Clones 6A and 6B appear to be from the same locus since the restriction enzyme digestion patterns were similar and fragments of the same size hybridized to 5 S RNA. Similar results were obtained when these DNAs were digested with three other restriction enzymes (results not shown). These results also showed that each of the six different clones contain only one 5s gene in the 15-20 kilobase pairs of DNA insert. In addition, when rat spleen DNA (SP.DNA) was completely digested with EcoRI and hybridized with 3'-endlabeled 5 S RNA, the DNA fragments that hybridized did not correspond to any of the six different clones that were isolated. The longer exposure of this lane showed some hybridization in regions other than the main band. One'possible explanation for these results is that the main band corresponds to the 5 s gene repeat and none of the 5 s clones isolated were from this gene repeat. 5 s Sequences in the Cloned DNAs Were Not Identical to the 5 S RNA Found in Adult Rat Liver Cells-The extent of homology between 3'-end-labeled 5 S RNA found in adult rat liver and the six different 5 s clones was analyzed (Fig. 3, panel A). 5 S RNA specifically hybridized to these clones. However, when these RNA-DNA hybrids were digested with TI RNase for 15 min, intact 5 S RNA was only recovered with one clone. Only in the case of clone 5s-2 was intact 5 S RNA found (Fig. 3, panel B). However, when the 5s-2 DNA5 S RNA hybrid was digested with TI-RNase for 1 h, instead of for 15 min, all the 5 S RNA was cleaved (results not shown).
These results show that of the six different clones isolated, "..,

FIG. 2. Southern blot analysis of 5 S DNAs digested with
restriction endonuclease EcoRI. Seven cloned X phage DNAs, designated 58-1 to 5S-6B, were digested with EcoRI enzyme and fractionated on a 1% agarose gel. SP.DNA, rat spleen DNA digested with EcoRI enzyme and probed with 5 S RNA probe. The nitrocellulose filter was hybridized at 42 "C with 3'-end-labeled 5 S RNA, and the filter was washed at 42 "C and subjected to autoradiography. The last panel shown separately is a longer exposure of SP.DNA shown to its left. only clone 5s-2 may contain 5 s sequences identical to rat liver 5 S RNA.
To test whether any of these 5 s gene variants are transcribed in uiuo, several experiments were carried out. 1) The uniformly labeled 4-8 S RNA was hybridized to the 5s-2 DNA at 56 "C, and the hybridized RNAs were digested with TI RNase (see Fig. 3 for details). If the 5s-2 gene variant is transcribed in the cells, the transcript from this gene variant would be expected to form a TI RNase-resistant hybrid with cloned 5s-2 DNA. Analysis of the protected 5 S RNA by fingerprinting showed that this 5 S RNA was the same as the bulk 5 S RNA (results not shown). 2) Earlier studies showed that when the rats were injected with thioacetamide, the RNA synthesis, including that of 5 S RNA, increases 5-10-fold (18).
To test whether any of these 5 s gene variants are activated and transcribed under these drug-induced conditions, the 5 S RNA, synthesized in livers of rats treated with thioacetamide, was analyzed by fingerprinting. In this case also, the 5 S RNA was the same as in normal rat liver, and no minor oligonucleotides corresponding to 5s-1 or 5s-2 gene variants were transcribed in vitro, the phage or plasmid DNAs containing clones 5s-1 and 5s-2 were used for transcription in vitro (Fig.  4, lanes 3-6) using Novikoff hepatoma whole cell Weil extracts (19). Clones 5s-3, 5s-4, 5s-5, and 5s-6A or -6B did not support 5 S RNA synthesis (compare PBR322, lane 12, or no DNA added, lane 1). Clone 5s-2 gave a single 5 S size RNA and clone 5s-1 produced two RNA bands of 5 S size. These results show that of the six different 5 s gene variants isolated, four (clones 3-6) did not support 5 S RNA synthesis and two (clones 5s-1 and 5s-2) were transcribed with different efficiencies. The amount of 5 S RNA synthesized was quantitated by excising the 5 S RNA bands and determining the amount of radioactivity in each band. Based on these determinations, in three different experiments, the 5s-2 DNA was 10 times more efficiently transcribed when compared to Chinese hamster 5 s gene. The 5s-1 gene was transcribed less efficiently (30%) when compared to a bonafide 5 s gene. 1 and 2-To ascertain whether the 5 S size transcripts were derived from 5 S-related DNA or from some other region of the insert, the in vitro transcripts were analyzed by fingerprinting. Fig. 5 shows the fingerprints of 5 s transcripts obtained from clone 5s-1 (panel B ) and clone 5s-2 (panel C) compared to transcripts obtained from Chinese hamster 5 s gene (panel A). The nucleotide sequences of Chinese hamster 5 S RNA and rat 5 S RNA are identical, and, therefore, the fingerprints are same. The fingerprints of 5 S RNA from clone 5s-1 and 5s-2 contained many oligonucleotides common to 5 S RNA but also showed many different oligonucleotides (Fig.  5). The T, RNase fingerprint of the 5 S RNA obtained from 5s-1 DNA was found to be consistent with the 5s-1 DNA sequence. For example, the oligonucleotide 20 (AACG,) is missing in 5s-1 RNA, since there are two nucleotide substi--5s  Fig. 4 were isolated and fingerprinted after TI RNase treatment as described by Brownlee et al. (20). [LI-~'P]GTP was used as the RNA precursor for these experiments. A , 5 S RNA from Chinese hamster 5 s gene (4); B, 5s-1, 5 S RNA from clone 1. The two bands observed in the case of clone 1 in the region of 5s (see Fig. 4) gave similar fingerprints. C, 5S-2,5 S RNA from clone 2. The oligonucleotides were numbered according to Forget and Weissman (34). The oligonucleotides 53', 54', 55', and 56' in 5s-1 RNA and oligonucleotides 37' and 54' in 5s-2 RNA had altered mobility compared to 5 S RNA (panel A ) . Other oligonucleotides with altered mobility are not indicated.   85 (27, 28), is underlined.

. -f f t t t f t t f f f f f f t t t t t t t t t t t t t f t t t t f t t t t t f ~f f t~~f
tutions in this region of 5s-1 DNA. In addition, several like 37 and 54, had altered mobility as expected. The 5 S RNA oligonucleotides like 53, 54, 55, and 56 had altered mobilities, transcripts from 5s-1 and 5s-2 DNAs contained pppGp (olias would be predicted from the nucleotide sequence of 5s-1 gonucleotide 1'), the first nucleotide in 5 S RNA. These data DNA. Similarly, the oligonucleotides 8 (AAG,) and 17 show that the 5 S-sized transcripts were initiated and com-(ACCG,) are missing in the fingerprint of 5s-2 RNA, as pleted properly from 5s-1 and 5s-2 gene variants. expected because of nucleotide substitutions in 5s-2 DNA at 5 S DNA Sequences in Clones 1 and 2 Are Highly Homolopositions 49 and 93 (see Fig. 6B). Many other oligonucleotides, gous to but Not Identical to 5 S RNA-The 5s DNA-contain-1 2 3 4 5 1 1 6 A 66 FIG. 7. Hybridization of different 5 s clones with 5 S RNA and a probe prepared from 3"flanking sequences from clone 5s-1. Phage DNAs prepared from clones 1 to 6B were digested with EcoRI and fractionated on a 1% agarose gel and transferred to nitrocellulose papers. One filter was hybridized to 3"end-labeled 5 S RNA from Novikoff hepatoma cells, and another filter was hybridized with a 5'-end-labeled 75-nucleotide-long HinfI-Am11 DNA fragment flanking the 3'-side of 5s-1 DNA sequences (see Fig. 6A).
ing portions from the phage clones 1,2, and 6 were subcloned into PBR322 and the sequences corresponding to 5 S RNA and flanking sequences were determined. Fig. 6A shows the restriction enzyme sites and the DNA sequencing strategy used. The nucleotide sequences of the 5 S DNA and flanking are shown in Fig. 6B. The rat liver 5 S RNA sequence (26) is shown numbered 1 to 123 nucleotides, and the nucleotides in the cloned DNA identical to 5 S RNA are indicated by dashes; only the nucleotide mismatches are shown. Clone 5s-1 differed in eight positions from 5 S RNA, and seven of these were C to T transitions. Clone 5s-2 differed in nine positions and contained one nucleotide deletion at position 100 or 101. The overall homology was 95% for clone 5s-1 and 93% for clone 5s-2. The 5"flanking sequences were different between these two clones, but the 3"flanking sequences were 90% homologous. The 5s-6 DNA contained a truncated 5 s gene; this DNA contained sequences corresponding to nucleotide 1 through 41 of 5 S RNA (Fig. 6B). There was one nucleotide substitution at position 6, and the overall homology was 97%.
Sequences Flanking the 3 -End of 5 s Clones 1 and 2 Were Also Present in 5 s Clones 3, 4, and 5"To determine whether the conserved sequences flanking the 3'-end of 5 s clones 1 and 2 were also found in nontranscribable 5 s clones, a probe was prepared from the 3"flanking region, the HinfI to Am11 site (see Fig. 6, A and B ) , and all the cloned DNAs were analyzed. The results obtained with EcoRI-digested phage DNAs are shown in Fig. 7. When the digested DNA was probed with 5 S RNA, all the clones hybridized (panel A ) .
When the 3"flanking sequence (Fig. 6B, nucleotide +3 to +70) was used as a probe only clones 1,2,3,4, and 5 hybridized but not clones 6A or 6B. These results show that the 3'flanking sequence is common to many, but not all, 5 s gene variants in the rat genome. Interestingly, clones 2, 4, and 5 also contained DNA sequences homologous to the 3'-flanking region of clone 1 in regions unrelated to 5 S DNA.

DISCUSSION
Seven clones containing sequences homologous to 5 S RNA were purified and analyzed. Of these, six were found to be from different loci of the rat genome. These DNAs hybridized to 5 S RNA when a mixture of whole cell 4-8 S RNA was used. The sequence analysis of two transcribable 5 S DNAs showed that clone 5s-1 contained eight substitutions, but there were none in the sequence corresponding to the internal control region nucleotides 47 to 85 (27,28). Clone 2, however, had five substitutions in this region. Interestingly, the 5s-2 clone was transcribed 10 times more efficiently than in authentic Chinese hamster 5 s clone (4). Therefore, characterization of these naturally occurring 5 s gene variants may be useful in understanding the interactions between 5s genes and the transcription factors.
In many species, the bonafide 5 S RNA genes were shown to be repeated tandemly (1-4). In some species, the pseudogenes for 5 S RNA are found clustered as part of the tandem repeat (7-10). The differing restriction enzyme digestion patterns obtained for the six 5 S DNA clones and the observation that the 15-20-kilobase pair inserts contained only one 5 s homologous sequence shows that these 5 s gene variants are dispersed in the rat genome and are probably not part of a tandem repeat.
Active and inactive forms of 5 S RNA genes were recently reported by Emerson and Roeder (11,12) in mouse and human genomes. The analysis of the six clones in an in uitro transcription assay showed that two were transcribed and four were not. These present results are similar to those obtained by Emerson and Roeder (11,12) and show that the rat genome contains both active and inactive 5 s gene variants. However, the 5 s gene variants reported by Emerson and Roeder (11, 12) were from the 5 s gene cluster, and the gene variants characterized in the present study appear to be from loci dispersed in the rat genome and not from the 5 s gene cluster.
The aim of this study was to isolate a true 5 S RNA gene from the rat liver genomic library. After screening the rat genomic equivalent once with authentic 5 S RNA, the isolated clones contained only 5 s gene variants but no real 5s gene.
The reason for this result is not clear. One possibility is that the partial EcoRI digestion of the rat genomic DNA did not produce suitable 5 s gene-containing DNA fragments for cloning into the X phage. When the rat genomic DNA was digested with EcoRI enzyme, fractionated on 1% agarose gel, and hybridized with labeled 5 S RNA, the most intense signal observed corresponded to DNA fragments of more than 20 kilobase pairs in length (Fig. 2, lune SP.DNA). It is possible that the 5 S RNA genes are clustered in these long EcoRI fragments and were under-represented in the genomic bank screened. Recent studies by Emerson and Roeder (11,12) and earlier studies by Jacq et al. (7) and Sharp et al. (10) showed that 5s gene variants are interspersed with real 5 s genes in mouse, human, Drosophila, and Xenopus genomes. The present study shows that some 5 s gene variants are dispersed in the rat genome, separated from real 5 s genes. However, these results do not rule out the presence of 5s gene variants that may be clustered with the real 5 s genes.
The origin and function of these 5 s gene variants are not clear. Since the 3'-flanking sequence was conserved between two 5 s clones and was probably conserved in three other 5 s clones, these gene variants are unlikely to have arisen by RNA-mediated DNA synthesis and integration into the genome (29,30). Direct repeats, a, characteristic of many RNAmediated pseudogenes, were Got found flanking these 5 s sequences. Therefore, these 5 s gene variants may have had a common progenitor molecule and drifted through evolution.