Mapping the location of psoralen crosslinks on RNA by mung bean nuclease sensitivity of RNADNA hybrids ( ribosomal RNA / Shine - Dalgarno sequence / RNA secondary structure / nucleotide sequence )

An indirect high resolution method has been developed for finding the location of intrastrand crosslinks in RNA. An end-labeled DNA strand that overlaps the approximate crosslink position is hybridized to the RNA and then treated with mung bean nuclease. The resulting digest is analyzed on a sequencing-type gel. The method was tested with the major psoralen crosslink seen in the 16S rRNA of inactivated Escherichia coli 30S ribosomal subunits. This crosslink was previously mapped between residues 930 ± 25 and a region close to the 3' end by electron microscopy. The new indirect method reveals that the crosslink occurs between residues 919 and 923 and residues 1530 and 1534. When these results are examined in the light of existing consensus secondary structure models for the 16S rRNA, it appears that the Shine-Dalgarno sequence is located close to the peptidyl tRNA binding site. Psoralen crosslinking is one of the few available techniques for examining the secondary and tertiary structures of large RNAs (1-4). Together with nucleotide sequence information (5) and many less direct methods (6), it has provided a fairly detailed picture of the structure of the 16S rRNA of Escherichia coli. Psoralen crosslinking freezes features of a native RNA structure and preserves them during subsequent manipulations. However, a technique is needed to map the location of the psoralen crosslinks. Electron microscopy is particularly effective for revealing crosslinks between RNA regions distant in the primary structure, but it cannot provide the positions of these crosslinks more accurately than approximately ±20 nucleotides (3). Two alternate techniques use T1 RNase digestion of crosslinked RNA: diagonal polyacrylamide gel electrophoresis and enzymatic sequencing (4, 7). These methods have provided the precise location of a number of psoralen crosslinks but they appear to be fairly laborious. In addition, there is no obvious way to use these methods to focus on crosslinks between distant regions that are likely to be particularly crucial in the evaluation of secondary and tertiary structure models. A simple method is needed to determine the location of interesting psoralen crosslinks without interference from psoralen monoadducts. MATERIALS AND METHODS Materials. Restriction enzymes, Xma III, Xma I, and HgiAI were obtained from New England Biolabs, mung bean nuclease was from P-L Biochemicals, and Klenow fragment of DNA Pol I was from Bethesda Research Laboratories. All enzymes were used as indicated by manufacturers except where specified. [3H]Aminomethyltrimethylpsoralen (9.5 X 104 cpm/,ug) was obtained from HRI Associates (Emeryville, CA), [a-32P]dCTP (3000 Ci/mmol; 1 Ci = 37 GBq) and [a-32P]dGTP (3000 Ci/mmol) were from Amersham, formamide was from MCB Manufacturing Chemists (Norwood, OH), and RG 501-X8 mixed bed resin (20-50 mesh), Bio-Rex RG grade, was from Bio-Rad. Preparation of Crosslinked RNA. 30S ribosomal subunits were isolated from E. coli MRE 600 cells by the method of Amils et al. (8). They were irradiated with [3H]aminomethyltrimethylpsoralen in inactivation buffer [0.5 mM Mg(OAc)2/100 mM NH4Cl/1 mM 2-mercaptoethanol/10 mM Tris*HCl, pH 7.6], and the crosslinked 16S rRNA was isolated as described (3). The crosslinked 16S rRNA was separated into a series of distinct species by electrophoresis on a 4.4o-4.8% linear gradient polyacrylamide gel in the presence of 98% formamide (3). Incorporation of [3H]aminomethyltrimethylpsoralen into the RNA ([3H]aminomethyltrimethylpsoralen/RNA, 8:1) was determined by 3H counting and A260 measurement. Preparation of Complementary DNA Fragments. Plasmid pFA25.1 was obtained from S. R. Fahnestock. This plasmid has the rrnB BamHI fragment (7.5 kilobases) of X rifdJ8 cloned into the BamHI site of pBR322, and it is identical to pKK3535 (9). The plasmid was transformed into a dcm6 dam3 E. coli strain 6M119, obtained from C. Smith, grown in M9 medium (10), and isolated by the method of Marko et al. (11). The DNA fragment used to analyze the 5' 16S rRNA crosslink location was from an Xma III digest of pFA25.1; that for the 3' location was from an Xma I and HgiAI double digest. The DNA fragments were purified on polyacrylamide gels of various percentages, depending on the sizes of the fragments, in buffer A (89 mM Tris borate/2.5 mM EDTA, pH 8.3) (10). RNADNA Hybridization and Mung Bean Nuclease Mapping. The DNA fragments were 3'-labeled with the Klenow fragment of DNA Pol I in the presence of [a-32P]dCTP and [a-32P]dGTP (10). Formamide was deionized once with RG 501-X8 mixed bed resin and recrystallized twice at PC. RNA (0.31 Ag) and [3'-32P]DNA (0.1 ug) were placed in hybridization solution (400 mM NaCl/1 mM EDTA/50 mM Pipes, pH 6.8/80% formamide), heated at 90°C for 5 min, then at 45°C for 12 hr (12). The hybridization mixture was chilled in ice water, then applied directly onto a 3.5% polyacrylamide (acrylamide/bisacrylamide, 29:1)/0.5% agarose composite gel in buffer A, and electrophoresed at 150 V for 6 hr. Hybrids were electroeluted by the method of Barnes (13). The purified RNA-DNA hybrid (0.3 Ag) was subjected to mung bean nuclease digestion in 200 ,ul of mung bean nuclease reaction buffer [50 mM NaCl/1 mM ZnSO4/1 mM L-serine/5% (vol/vol) glycerol/50 mM NaOAc, pH 5.0/mung bean nuclease at 500 units or 1000 units/ml], at 37°C for various times (14, 15). The mung bean nuclease-treated RNADNA hybrid was ethanol-precipitated in the presence of carrier DNA (5 ,g/100 ,ul) and analyzed on a denaturing sequencing-type gel (16).