Cell, Vol. 29, 265-274, May 1982, Copyright 0 1982 by MIT Transcription of a Gene for Human Ul Small Nuclear RNA James T. Murphy and Richard R. Burgess McArdle Laboratory for Cancer Research James E. Dahlberg and Elsebet Lund Department of Physiological Chemistry University of Wisconsin-Madison Madison, Wisconsin 53706 Summary Transcription of Ul small nuclear RNA from a 592 bp fragment of human DNA was analyzed in vivo and in vitro. When injected into Xenopus laevis oocyte nuclei, the cloned DNA is transcribed by RNA polymerase II to make human Ul snRNA. Thus the sequences of this fragment are sufficient for expression of the Ul snRNA gene. Moreover, injec- tion of templates carrying deletions of flanking se- quences demonstrates that the DNA sequences re- quired for in vivo transcription are located at least 100 nucleotides upstream from the point corre- sponding to the 5’ end of mature Ul snRNA. In vitro transcription in a HeLa cell extract leads to synthe- sis not of mature Ul snRNA, but of a larger molecule starting 183 nucleotides upstream from the site corresponding to the 5’ end of mature Ul snRNA. Transcription from this upstream promoter also is catalyzed by RNA polymerase II, and is comparable in efficiency with the very strong major late pro- moter of adenovirus 2. We propose that Ul snRNA is synthesized in vivo as a precursor that is proc- essed by an enzyme or enzymes missing from our extracts. Introduction The small nuclear RNA Ul is present in the nuclei of most, if not all, eucaryotic cells (Reddy and Busch, 1981; Zieve, 1981). In mammalian cells this 164 nucleotide RNA (Branlant et al., 1980) is very abun- dant, with estimates ranging as high as one million copies per HeLa cell (Weinberg and Penman, 1968; Zieve and Penman, 1976). Recent measurements of the number of genes for Ul snRNA vary from 40 to 125 per cell (Denison et al., 1981; Manser and Geste- land, 1981, 1982; E. Lund and J. E. Dahlberg, man- uscript in preparation). Even though there are multiple genes for this RNA, the promoters for the individual genes must be quite efficient to allow for the synthesis of that many RNA molecules per generation. The genes for Ul snRNAs are unusual in that most of them appear to have highly conserved 5’-flanking sequences extending for as much as a few thousand nucleotides upstream from the actual coding region (Manser and Gesteland, 1981, 1982). The biological significance of this observation is unknown. Accumulation of Ul snRNA depends either directly or indirectly on active RNA polymerase II (Frederiksen et al., 1978; Gram-Jensen et al., 1979; Eliceiri, 1980; Roop et al., 1981). Although RNA polymerase II tran- scripts normally contain N7-methylguanosine-Y-tri- phosphate caps (mG caps) at their 5’ ends, 111 snRNA (as well as several other snRNAs) contains an N2,N2,N7-trimethylguanosine-5’-triphosphate cap (m3G cap) (Cory and Adams, 1975; Recldy et al., 1974; Branlant et al., 1980; Reddy and Busch, 1981). In those cases where initiation of messenger RNA synthesis has been studied in detail, the mG cap was found attached to the initiating nucleotide (Ziff and Evans, 1978; Contreras and Fiers, 1981; Gidoni et al., 1981; Hagenbtichle and Schibler, 198’1). It is not clear whether this same rule holds for molecules, such as Ul snRNA, that bear the m,G cap. We have undertaken a detailed study of the struc- ture and transcription of one of the Ul snRNA genes. Injection of a cloned human Ul snRNA gene into Xenopus laevis oocyte nuclei led to accumulation of human Ul snRNA. Unexpectedly, we found that the flanking DNA sequences required for this in vivo tran- scription were located more than 100 nlucleotides upstream from the coding region. In addition, we found that in vitro transcription of the clone yielded products containing 183 extra nucleotides on the 5’ side of the Ul snRNA sequences, The fact that mature Ul snRNA was made in vivo but the larger molecule was made in vitro indicates that specific processing or transcription factors are needed for production of mature Ul snRNA. Results Figure 1 shows the nucleotide sequence of the 592 bp DNA fragment insert in the plasmid pHUl-1 D, which we have used in these transcription studies. This fragment contains the entire Ul snRNA gene plus 393 and 35 nucleotides of the 5’- and 3’-flanking regions, respectively. This Dde l-generated fragment was cloned into the Barn HI site of pBR322 after addition of linkers. It is referred to as the 592 bp fragment, although the addition of linkers makes it slightly longer than that. Also included in Figure 1 are the restriction endonuclease cleavage sites that are referred to in the experiments described below. Transcription in Vivo In an effort to determine whether the human DNA in pHUl-1 D contains a functional gene, we purified the 592 bp fragment shown in Figure 1, and ligated its ends to form circular DNA molecules. We injected these molecules into X. laevis oocyte nuclei in the presence of al-32P-GTP, and after 20 hr of incubation, extracted the RNA from these and control oocytes, which had received only the radioactively labeled GTP. Analysis of the RNA by one- and two-dimen- sional polyacrylamide gel electrophoresis yielded the results shown in Figure 2. Clearly, injection of the