Alternative 3'-end processing of U5 snRNA by RNase III

Abstract

The cellular components required to form the 3' ends of small nuclear RNAs are unknown. U5 snRNA from Saccharomyces cerevisiae is found in two forms that differ in length at their 3' ends (U5L and U5S). When added to a yeast cell free extract, synthetic pre-U5 RNA bearing downstream genomic sequences is processed efficiently and accurately to generate both mature forms of U5. The two forms of U5 are produced in vitro by alternative 3'-end processing. A temperature-sensitive mutation in the RNT1 gene encoding RNase III blocks accumulation of U5L in vivo. In vitro, alternative cleavage of the U5 precursor by RNase III determines the choice between the two multistep pathways that lead to U5L and U5S, one of which (U5L) is strictly dependent on RNase III. These results identify RNase III as a trans-acting factor involved in 3'-end formation of snRNA and show how RNase III might regulate alternative RNA processing pathways.

Figure 1

In vitro processing of a model pre-U5 transcript. (A) Sequence and a secondary structure model of the precursor used in this study. The secondary structure of mature U5 is drawn from Frank et al. (1994). The most stable potential secondary structure of the downstream genomic sequence calculated using Mfold (Zuker 1994) is shown here but has not been proved experimentally nor phylogenetically. Sequences not derived from the U5 gene are indicated in lowercase. They include two additional guanosines at the 5′ end of the molecule to facilitate transcription by T7 RNA polymerase and part of a BamHI restriction site at the 3′ end. (B) Time course of processing. Precursor-U5 (P) was incubated in a whole-cell extract for the times indicated. Shown is an autoradiograph of a 6% gel. σ and λ are generated rapidly and have the characteristics of intermediates, whereas U5L and U5S have characteristics of products (see text). The molecular weight marker (M) is a pBR322 plasmid digested with MspI and labeled with [γ-³²P]ATP.

Figure 1

In vitro processing of a model pre-U5 transcript. (A) Sequence and a secondary structure model of the precursor used in this study. The secondary structure of mature U5 is drawn from Frank et al. (1994). The most stable potential secondary structure of the downstream genomic sequence calculated using Mfold (Zuker 1994) is shown here but has not been proved experimentally nor phylogenetically. Sequences not derived from the U5 gene are indicated in lowercase. They include two additional guanosines at the 5′ end of the molecule to facilitate transcription by T7 RNA polymerase and part of a BamHI restriction site at the 3′ end. (B) Time course of processing. Precursor-U5 (P) was incubated in a whole-cell extract for the times indicated. Shown is an autoradiograph of a 6% gel. σ and λ are generated rapidly and have the characteristics of intermediates, whereas U5L and U5S have characteristics of products (see text). The molecular weight marker (M) is a pBR322 plasmid digested with MspI and labeled with [γ-³²P]ATP.

Figure 2

Mapping of 3′ ends of transcripts produced by in vitro processing. (A) Structure of the 3′ downstream genomic sequence. Shown are the diagnostic RNase T1 fragments, as well as the 3′ ends of putative intermediates, and products, deduced from the experiments shown in B and C and in Figs. 3B and 5C. Arrows from U5L and U5S indicate 3′ ends of the products of the in vitro reaction, whereas U5Long and U5Short indicate the 3′ ends of the long and short forms of U5 mapped in vivo. No diagnostic T1 fragment is found in the remaining part of the precursor; therefore, this part of the molecule is not shown. (B) RNase T1 mapping. Shown is an autoradiograph of a 10% acrylamide gel. Each of the species (precursor, intermediates and products) was subjected to total digestion by RNase T1 (see Materials and Methods). Legends and marker as in Fig. 1B. Marker is shown as an indication of molecular weight but is composed of double-stranded DNA, whereas the mapped species are single-stranded RNA. Therefore, direct comparison of the sizes of low molecular weight fragments is not possible. (C) Mapping cleavage sites to the U5 flanking sequence. A 5′-labeled precursor was incubated for 2 min (lanes 1,2) or 45 min (lane 3) to give rise to intermediates and products and was loaded on a 5% acrylamide sequencing gel, in parallel with a sequence generated from the same precursor (see Materials and Methods), substituted with purine phosphorothioates (R) or uridine (U), and cleaved with iodine. Lanes 1 and 2 are loadings of different amounts of the same reaction.

Figure 2

Mapping of 3′ ends of transcripts produced by in vitro processing. (A) Structure of the 3′ downstream genomic sequence. Shown are the diagnostic RNase T1 fragments, as well as the 3′ ends of putative intermediates, and products, deduced from the experiments shown in B and C and in Figs. 3B and 5C. Arrows from U5L and U5S indicate 3′ ends of the products of the in vitro reaction, whereas U5Long and U5Short indicate the 3′ ends of the long and short forms of U5 mapped in vivo. No diagnostic T1 fragment is found in the remaining part of the precursor; therefore, this part of the molecule is not shown. (B) RNase T1 mapping. Shown is an autoradiograph of a 10% acrylamide gel. Each of the species (precursor, intermediates and products) was subjected to total digestion by RNase T1 (see Materials and Methods). Legends and marker as in Fig. 1B. Marker is shown as an indication of molecular weight but is composed of double-stranded DNA, whereas the mapped species are single-stranded RNA. Therefore, direct comparison of the sizes of low molecular weight fragments is not possible. (C) Mapping cleavage sites to the U5 flanking sequence. A 5′-labeled precursor was incubated for 2 min (lanes 1,2) or 45 min (lane 3) to give rise to intermediates and products and was loaded on a 5% acrylamide sequencing gel, in parallel with a sequence generated from the same precursor (see Materials and Methods), substituted with purine phosphorothioates (R) or uridine (U), and cleaved with iodine. Lanes 1 and 2 are loadings of different amounts of the same reaction.

Figure 2

Mapping of 3′ ends of transcripts produced by in vitro processing. (A) Structure of the 3′ downstream genomic sequence. Shown are the diagnostic RNase T1 fragments, as well as the 3′ ends of putative intermediates, and products, deduced from the experiments shown in B and C and in Figs. 3B and 5C. Arrows from U5L and U5S indicate 3′ ends of the products of the in vitro reaction, whereas U5Long and U5Short indicate the 3′ ends of the long and short forms of U5 mapped in vivo. No diagnostic T1 fragment is found in the remaining part of the precursor; therefore, this part of the molecule is not shown. (B) RNase T1 mapping. Shown is an autoradiograph of a 10% acrylamide gel. Each of the species (precursor, intermediates and products) was subjected to total digestion by RNase T1 (see Materials and Methods). Legends and marker as in Fig. 1B. Marker is shown as an indication of molecular weight but is composed of double-stranded DNA, whereas the mapped species are single-stranded RNA. Therefore, direct comparison of the sizes of low molecular weight fragments is not possible. (C) Mapping cleavage sites to the U5 flanking sequence. A 5′-labeled precursor was incubated for 2 min (lanes 1,2) or 45 min (lane 3) to give rise to intermediates and products and was loaded on a 5% acrylamide sequencing gel, in parallel with a sequence generated from the same precursor (see Materials and Methods), substituted with purine phosphorothioates (R) or uridine (U), and cleaved with iodine. Lanes 1 and 2 are loadings of different amounts of the same reaction.

Figure 3

U5L and U5S arise from distinct intermediates generated by endonucleolytic cleavage. (A) The long and the short forms of U5 arise from different intermediates. Gel-purified σ, λ, and U5L species were incubated in extracts for the amount of time indicated and fractionated on a 6% acrylamide gel. Legends as in Figure 1B. (B) Time course of 3′ processing of internally labeled or 3′-labeled precursors. Legends as in Fig. 1B. In the mock incubation points (Mk), samples were incubated for 15 min at 30°C with extract, buffer, and 25 mm EDTA. The sums of σ′ (60 nucleotides) plus σ (270 nucleotides), or λ′ (90 nucleotides) plus λ (240 nucleotides) are about the same as the total length of pre-U5 (332 nucleotides), supporting the interpretation that the short unstable RNAs are the downstream products of endonucleolytic cleavages that generate the intermediates.

Figure 3

U5L and U5S arise from distinct intermediates generated by endonucleolytic cleavage. (A) The long and the short forms of U5 arise from different intermediates. Gel-purified σ, λ, and U5L species were incubated in extracts for the amount of time indicated and fractionated on a 6% acrylamide gel. Legends as in Figure 1B. (B) Time course of 3′ processing of internally labeled or 3′-labeled precursors. Legends as in Fig. 1B. In the mock incubation points (Mk), samples were incubated for 15 min at 30°C with extract, buffer, and 25 mm EDTA. The sums of σ′ (60 nucleotides) plus σ (270 nucleotides), or λ′ (90 nucleotides) plus λ (240 nucleotides) are about the same as the total length of pre-U5 (332 nucleotides), supporting the interpretation that the short unstable RNAs are the downstream products of endonucleolytic cleavages that generate the intermediates.

Figure 4

Intermediates and products of the processing reaction associate with Sm proteins. After 2 or 45 min of incubation, RNAs were immunoprecipitated with various antibodies (see text and Materials and Methods). T (total) represents of the amount of RNAs used in each reaction. The other lanes contain RNAs isolated from immunoprecipitates. Shown is an autoradiograph of a 6% acrylamide gel. Legends as in Fig. 1B.

Figure 5

RNase III is required for correct U5 3′-end processing. (A) The rnt1 mutant strain does not produce the long form of U5 in vivo. Total RNAs were extracted from wild-type or rnt1 strains, grown at 26°C or shifted to 37°C for 4 hr. The RNAs were loaded onto a denaturing 6% acrylamide gel, transferred to a nylon membrane, and hybridized to DNA fragments complementary to snRNAs. (B) Extracts from rnt1 cells are deficient in the production of intermediates and U5L. Internally-labeled transcript was incubated in extracts made from wild-type (WT) or mutant rnt1 strains for the times indicated, and products were loaded in a denaturing 6% acrylamide gel. Legends as in Fig. 1B. (C) 3′-end trimming of the short form of U5 in rnt1 extracts. A 5′-labeled transcript was incubated for 5, 15, or 45 min in rnt1 extract and loaded on a 5% sequencing gel, in parallel with a phosphorothioate generated sequence with purines (R) and uridine (U) as in Fig. 3C. Arrowheads on the sequence indicate the 3′ ends after 5 and 45 min of incubation. Note that the trimming must be at the 3′ end, because the 5′ end of the molecule is labeled.

Figure 5

RNase III is required for correct U5 3′-end processing. (A) The rnt1 mutant strain does not produce the long form of U5 in vivo. Total RNAs were extracted from wild-type or rnt1 strains, grown at 26°C or shifted to 37°C for 4 hr. The RNAs were loaded onto a denaturing 6% acrylamide gel, transferred to a nylon membrane, and hybridized to DNA fragments complementary to snRNAs. (B) Extracts from rnt1 cells are deficient in the production of intermediates and U5L. Internally-labeled transcript was incubated in extracts made from wild-type (WT) or mutant rnt1 strains for the times indicated, and products were loaded in a denaturing 6% acrylamide gel. Legends as in Fig. 1B. (C) 3′-end trimming of the short form of U5 in rnt1 extracts. A 5′-labeled transcript was incubated for 5, 15, or 45 min in rnt1 extract and loaded on a 5% sequencing gel, in parallel with a phosphorothioate generated sequence with purines (R) and uridine (U) as in Fig. 3C. Arrowheads on the sequence indicate the 3′ ends after 5 and 45 min of incubation. Note that the trimming must be at the 3′ end, because the 5′ end of the molecule is labeled.

Figure 5

RNase III is required for correct U5 3′-end processing. (A) The rnt1 mutant strain does not produce the long form of U5 in vivo. Total RNAs were extracted from wild-type or rnt1 strains, grown at 26°C or shifted to 37°C for 4 hr. The RNAs were loaded onto a denaturing 6% acrylamide gel, transferred to a nylon membrane, and hybridized to DNA fragments complementary to snRNAs. (B) Extracts from rnt1 cells are deficient in the production of intermediates and U5L. Internally-labeled transcript was incubated in extracts made from wild-type (WT) or mutant rnt1 strains for the times indicated, and products were loaded in a denaturing 6% acrylamide gel. Legends as in Fig. 1B. (C) 3′-end trimming of the short form of U5 in rnt1 extracts. A 5′-labeled transcript was incubated for 5, 15, or 45 min in rnt1 extract and loaded on a 5% sequencing gel, in parallel with a phosphorothioate generated sequence with purines (R) and uridine (U) as in Fig. 3C. Arrowheads on the sequence indicate the 3′ ends after 5 and 45 min of incubation. Note that the trimming must be at the 3′ end, because the 5′ end of the molecule is labeled.

Figure 6

RNase III cleaves pre-U5 to produce intermediates in the absence of other factors. 5′-Labeled or internally-labeled pre-U5 transcripts were incubated for 5 and 45 min with GST–RNT1 or for 45 min with GST alone (Mk). Markers are the phosphorothioate-generated sequence with purines (R) and pyrimidines (Y) as in Fig. 3C.

Figure 7

RNA processing pathways for generating the 3′ ends of yeast U5 snRNA. The 5′ part of U5 snRNA is represented by the shaded ovals.