Efficiency and pattern of UV pulse laser-induced RNA–RNA cross-linking
http://www.100md.com
《核酸研究医学期刊》
Department of Molecular and Structural Biochemistry, North Carolina State University, Raleigh, NC 27695-7622, USA and 1 Department of Chemistry, North Carolina State University, Raleigh, NC 27695-8204, USA
*To whom correspondence should be addressed. Tel: +1 919 515 5703; Fax: +1 919 515 2047; Email: paul_wollenzien@ncsu.edu
ABSTRACT
Escherichia coli ribosomes were irradiated with a KrF excimer laser (248 nm, 22 ns pulse) with incident pulse energies in the range of 10–40 mJ for a 1 cm2 area, corresponding to fluences of 4.5 to 18 x 109 W m–2, to determine strand breakage yields and the frequency and pattern of RNA–RNA cross- linking in the 16S rRNA. Samples were irradiated in a cuvette with one laser pulse or in a flow cell with an average of 4.6 pulses per sample. The yield of strand breaks per photon was intensity dependent, with values of 0.7 to 1.3 x 10–3 over the incident intensity range studied. The yield for RNA–RNA cross-linking was 3 x 10–4 cross-links/photon at the intensity of 4.5 x 109 W m–2, an 4-fold higher yield per photon than obtained with a transilluminator. The cross-link yield/photon decreased at higher light intensities, probably due to intensity-dependent photoreversal. The pattern of cross-linking was similar to that observed with low intensity irradiation but with four additional long-range cross-links not previously seen in E.coli ribosomes. Cross- linking frequencies obtained with one laser pulse are more correlated to internucleotide distances than are frequencies obtained with transilluminator irradiation.
INTRODUCTION
Despite the wide use of UV cross-linking in studies of RNA and DNA structures, nanosecond pulse irradiation has not been reported as a method to obtain RNA–RNA cross-linking. There are a number of concerns that need to be addressed in the application of high intensity sources to the determination of structural constraints in RNA. Given the extent of DNA and RNA strand breaks during irradiation with UV pulse lasers, the possibility exists that strand breakage reactions would dominate all products and would preclude the analysis of RNA–RNA products by RNA sequencing methods. Moreover, extensive strand breakage during the irradiation would disrupt the native structure of the complex. A second factor not yet investigated for high intensity irradiation concerns the photoreversal response. The rates of formation and reversal of cross-link products set the equilibrium value for cross-link frequency in model dinucleotide compounds subjected to low intensity irradiation (1,2). Photoreversal has also been shown to be important in determining the yields of RNA–RNA cross-links in 16S rRNA in low intensity irradiations (3). Therefore, it is possible that cross-link yields could be lower under high intensity irradiation than they are under low intensity irradiation, if photoreversal were strongly intensity dependent. On the other hand, there are a number of advantages in the use of pulsed laser sources for UV cross-linking if these concerns could be alleviated. Since the pulse laser experiments are completed with one or a few pulses, products that arise from covalent reaction of the nucleic acid are likely to represent native conformations and interactions (4). In addition, the ability to rapidly induce cross-links would allow kinetic experiments on time scales limited mainly by the time constants of mixing devices.
Irradiation of nucleic acids with low intensity UV light results in formation of S1 and T1 excited states (5–7), and covalent reactions are thought to involve molecules in the S1 state because intersystem crossing yields for bases are low and usually cannot account for the yields of photoproducts (7,8). Nucleic acids attain activation states higher than the S1 and T1 states (the HT and HS states) at light intensities of at least 108 W m–2 (1018 photons cm–2 s–1) due to absorption of two photons (7,9,10). These light intensities are easily obtained in nanosecond or picosecond pulse lasers. In the HT and HS states, the bases undergo reactions with greater efficiency due to their higher energy, and the specificity of the reactions may also be expanded due to the different electronic structure of these states compared with excitation with low intensity irradiation (4,10).
Nucleic acid–protein cross-linking with pulse UV laser irradiation has been described for a variety of systems (4,9,11–17). These methods are attractive because high efficiency is indeed obtained with short irradiation times compared with the extended irradiation times that are needed with low intensity UV irradiation. Bases in the HT and HS states possess energy levels higher than the ionization energy (4), and irradiation with nanosecond or picosecond laser pulses leads to nucleic acid strand breakage (5,7,18–24). The structures of some photoproducts indicate formation by free radical reactions, confirming ionization (25). In DNA, particularly at guanosines in specific sequence contexts, pulse laser irradiation induces lesions in nucleotides that are cleaved upon alkaline treatment and are attributed to biphotonic reactions (24,26). The quantum yield of the strand breakage reaction has been reported for tobacco mosaic virus (TMV) RNA with low intensity, nanosecond and picosecond UV irradiation (14), and confirms a higher extent of direct strand breakage at high intensities compared with irradiation with low intensities.
The goal of the present work was to determine the extent of strand breakage and cross-linking in the 16S rRNA (1542 nt) in the ribosome with a UV pulse laser irradiation source. A KrF excimer laser with 248 nm light with pulse intensities of 4.5 to 18 x 109 W m–2 and 22 ns pulse duration has been used for these experiments. We wanted to determine first the dependence of strand breakage on the irradiation conditions and whether nanosecond pulse UV irradiation could be used under conditions where strand breakage was negligible, or at least occurred at levels where intact full-length RNA could still be isolated after irradiations. This has proven to be the case, and the frequency and locations of intramolecular RNA–RNA cross-links have been determined to see if RNA–RNA cross-links were produced that were not seen with low intensity irradiations. Differences in the relative RNA–RNA cross-linking frequencies are also found in comparing UV pulse laser and transilluminator irradiation, and these are discussed.
MATERIALS AND METHODS
Washed 70S ribosomes were prepared according to Makhno et al. (27) from Escherichia coli cells grown to mid log phase. 30S and 50S ribosomal subunits were separated by centrifugation of 70S ribosomes on sucrose gradients prepared in T20A200M3 buffer (20 mM Tris–HCl pH 7.5, 200 mM NH4Cl, 3 mM MgCl2, 4 mM ?-mercaptoethanol). Subunits were concentrated by sedimentation and dissolved in activation buffer (20 mM Tris–HCl pH 7.5, 200 mM NH4Cl, 20 mM MgCl2, 4 mM ?-mercaptoethanol), and frozen at –85°C for storage. 70S ribosomes were formed by re-association of 30S and 50S subunits in activation buffer by incubation for 30 min at 37°C.
Irradiation of ribosomes
Ribosome samples were irradiated in a cuvette with one pulse or in a flow cell with the continuous laser pulses. For irradiation in the cuvette, a fluorescence cuvette usually with 100 μl of sample was covered with a rubber septum and treated with a stream of N2 gas before irradiation. The cuvette was turned 90° from its normal arrangement so a large cross-section area of the sample would face the incident radiation. The cuvette path length was 2 mm in this arrangement and the exposed area was 0.5 cm2. Irradiation was with a single pulse (22 ns pulse length) administered to the sample using a manual shutter of a KrF excimer laser (Lambda Physik, Fort Lauderdale, FL). The laser was operated at 1 Hz to provide a reproducible pulse intensity. The flow cell had a path length of 1 mm and a cross-section area of 0.38 cm2. Samples of 1 ml were treated with N2 gas in a closed tube, and a peristaltic pump was used to pump the sample through the cell at 0.5 ml/min. The KrF excimer laser was used at 1 Hz, so each molecule received an average exposure of 4.6 pulses. The energy was monitored continuously in all experiments. The cells were Supracil quartz, obtained from Hellma Cells, Inc. (Forest Hills, NY).
Analysis of RNA strand breaks and cross-links by gel electrophoresis
Ribosomes were deproteinated by incubation with 1.0 μg/μl proteinase K and 1.0% SDS/0.02 M EDTA at 37°C for 30 min, followed by phenol extraction and ethanol precipitation. RNA was re-dissolved in water in volumes proportional to the starting amount of RNA mass; 5 μl of these samples should contain 0.5 μg of 16S RNA or, in some experiments, 1.0 μg of 16S rRNA. A 5 μl aliquot of each of these samples was mixed with equal volumes of formamide, heated to 45°C for 10 min and immediately electrophoresed on 1% agarose gels made with 1x TBE buffer (89 mM Tris base, 89 mM boric acid, 2 mM EDTA pH 8.3) and containing 100 μg/ml ethidium bromide. RNA was visualized by fluorescence. Control samples containing 0.5 or 1.0 μg of 16S rRNA were used for comparison.
For cross-link analysis, irradiated RNA first was purified by agarose gel electrophoresis to collect the 16S RNA band as described by Juzumiene et al. (28). RNA was phenol and ether extracted, and precipitated. The RNA was treated with shrimp alkaline phosphatase and, after phenol and ether extraction and ethanol precipitation, was labeled with ATP using polynucleotide kinase. The RNA was phenol and ether extracted, and unincorporated ATP was removed by two ethanol precipitations using 2.5 M NH4OAc pH 5.1 and 2.5 vol of EtOH for precipitation (28).
Usually, 0.2 μg of RNA in 70% formamide was loaded on a 1 cm wide well of a 0.35 mm thick x 40 cm long polyacrylamide gel which is 3.6% acrylamide:bisacrylamide (70:1), 8.3 M urea in BTBE buffer . The gel, thermostatted at 45°C, was run for 18 h at 6 W constant power (29). After electrophoresis, the gel was transferred to paper, dried, and the pattern and frequency of cross-linked species were determined with a Molecular Dynamics PhosphorImager (Amersham, Piscataway, NJ). Cross-link frequencies were calculated from band intensities using ImageQuant software (Amersham, Piscataway, NJ). Correlation coefficients were calculated with Microsoft Excel.
Cross-linking sites in the 16S rRNA were determined after separation of cross-linked molecules on 3.6% acrylamide: bisacrylamide (70:1), 8.3 M urea in BTBE buffer. Typically, 50–100 μg of 32P-labeled 16S rRNA was electrophoresed on a lane 34 cm wide on the 40 cm long x 0.35 mm thick polyacrylamide gel. After 18 h of electrophoresis, one glass plate was removed and the gel was covered with plastic wrap and exposed on a phosphorimager plate to locate the bands containing the cross-linked RNA species and the linear rRNA. Gel slices containing these were cut out and RNA was isolated by centrifugation over CsCl cushions (28,29). RNA from each fraction was inspected for cross-linking sites by reverse transcription reactions using a set of oligonucleotide primers according to the procedure of Wilms et al. (30). Cross-link sites are detected as reverse transcription stopping points present in specific fractions (30–32).
Irradiation of purine nucleosides and analysis of products
Guanosine and adenosine were purchased from Sigma Chemical Company (St Louis, MO) and used without further purification. Solutions (10 mM) of both guanosine and adenosine were prepared using freshly dispensed 18 M cm water (BARNSTEAD E-PURE) and used either singly or in combination. The nucleoside solutions were placed in a Spectrosil (Starna Cells, Inc., Atascadero, CA) cuvette, with a 1 mm path length, covered with a rubber septum and purged with a stream of N2 gas before irradiation. The solutions were then irradiated with either five or 10 pulses from the KrF excimer laser at 4.5 to 18 x 109 W m–2. These reactions were analyzed by liquid chromatographic (LC) mass spectrometry (HP LC/MSD 1100, Agilent Technologies, Wilmington, DE) using the direct injection method. The polycrystalline samples were prepared by finely grinding the powders in a ceramic mortar and used singly or combined into a 50:50 mixture by weight. The powdered samples were placed between CaF2 (International Crystal Laboratories, Garfield, NJ) and irradiated in the same manner as described above. The irradiated nucleotides were then dissolved in 18 Mcm water to 10 mM and analyzed by LC mass spectrometry using the direct injection method.
RESULTS
Dependency of RNA strand breaks on photon dose and light intensity
Escherichia coli 30S subunits and 70S ribosomes were irradiated in a flow cell or in a fluorescence cuvette in sets of increasing light intensity. Ribosome concentrations between 1.25 and 5 A260 units were used so that enough cross-linked samples for subsequent analysis would be available. After irradiation, ribosome samples were treated with proteinase K and phenol extraction, and RNA was ethanol precipitated and dissolved in water. Samples were diluted with formamide and heated for 10 min at 45°C to melt the RNA and then it was immediately subjected to electrophoresis on agarose gels and was viewed by ethidium fluorescence (Fig. 1). Strand breakage was seen in many of the samples and was more extensive for samples irradiated at higher pulse intensities in the flow cell, in which the sample is exposed to multiple pulses. The sites of strand breakage appear to be fairly random because no specific low molecular weight bands appear on the agarose gels. The sites have not been determined by reverse transcription analysis. The fraction of strand breaks per 16S RNA was estimated from the gels for these samples and for other samples under additional conditions. This was done by using sample volumes that should contain 0.5 or 1.0 μg of 16S rRNA determined by the A260 measurement and comparing the 16S rRNA band intensities with those seen in samples that were not irradiated.
Figure 1. Detection of strand breaks in irradiated RNA by agarose gel electrophoresis. 30S ribosome subunits or 70S ribosomes were irradiated in a flow cell or in a cuvette with different incident intensities as indicated. Samples were deproteinated, ethanol precipitated and redissolved in water. A volume of each sample was taken that should contain 0.5 μg of 16S RNA (lanes 1–5) or 1.0 μg of 16S RNA (lanes 7–16). Lane 1 contains 0.5 μg of 16S RNA and lane 10 contains 1.0 μg of 16S rRNA (as a component of total rRNA). There are small amounts of 23S rRNA in lanes containing RNA from 30S subunits due to contamination from 50S subunits. The two lanes on the right marked with asterisks contain RNA from samples that were not de-oxygenated. The amount of full-length 16S RNA remaining in the sample was estimated and used to calculate the extent of strand breakage.
The logarithm of the fraction of full-length RNA remaining after irradiation was plotted versus dose (absorbed photons per 16S RNA), as shown in Figure 2. There are different values for the dose (photons absorbed per 16S) in this plot even for samples irradiated at the same pulse intensity because the calculation of dose also depends on RNA concentration, path length and whether 70S or 30S ribosomes were used. The slopes of the best-fit lines indicate a probability of 0.7 x 10–3 breaks/photon for samples with A260 5 irradiated with a pulse intensity of 4.5 x 109 W m–2, and a probability of 1.3 x 10–3 breaks/photon for samples with A260 5 irradiated with a pulse intensity of 18 x 109 W m–2.
Figure 2. Dependence of strand breaks on light dose and intensity. The natural logarithm of the fraction of full-length RNA (where fraction of full-length RNA was estimated from the agarose gel electrophoresis), plotted against absorbed photons per 16S rRNA is shown. The lines labeled a and b are linear fits through data obtained at 4.5 x 109 and 18 x 109 W m–2, respectively, and have slopes of –0.7 x 10–3 and –1.3 x 10–3.
Cross-linking yields for long-range RNA–RNA cross-links at increasing light intensity
The frequency of intramolecular cross-linking was determined by gel electrophoresis of 16S RNA on denaturing polyacrylamide gels. To do this, the intact 16S RNA remaining in each sample first was isolated by agarose gel electrophoresis and was then 5' end-labeled with ATP and polynucleotide kinase. It was then electrophoresed on 3.6% polyacrylamide gels (acrylamide:bisacrylamide; 70:1) with 8.3 M urea at 45°C. Molecules separate into specific bands with reduced mobility under these conditions due to the presence and location of intramolecular cross-links (29). The frequency of the cross-links was then determined by analysis of the 32P distribution in the gel. Examples of 16S RNA cross-linked in the flow cell and cuvette are shown in Figure 3.
Figure 3. Denaturing polyacrylamide gel electrophoresis of cross-linked 16S RNA to determine cross-link frequency. 16S RNA was purified by agarose gel electrophoresis and was 5' labeled with 32P. Samples were electrophoresed on 8.3 M urea, 3.6% polyacrylamide gels at 45°C. The method and incident light intensity are indicated for each lane. The gel was dried and exposed on a phosphorimager plate to determine the image and cross-link frequencies. The heavy band in the bottom of each lane is at the position of unirradiated 16S RNA.
The sum of the band frequencies was quantified to determine the overall levels of cross-linking. The gel separates molecules that contain cross-links between nucleotides that are separated by at least 50 nucleotides; however, RNA isolated from these samples was very readily reverse transcribed, so there must not be too much short-range cross-linking in the samples. In the worst case shown in Figure 3 (30S subunits in the flow cell with 18 x 109 W m–2), it was difficult to obtain enough RNA for the analysis and instead of distinct bands there is a diffuse RNA density throughout the lane. In general, at the higher irradiation intensities, strand breaks in cross-linked molecules should produce X-shaped molecules, and this was probably the reason for the reduction in the sharpness of bands. In addition, strand breaks produce an increased amount of radioactivity throughout the gel above the normal background. Therefore, radioactivity above background was included in the calculation of the total cross-link frequencies. The dependence of cross-link frequency on absorbed dose is shown in Figure 4A, and the dependence of the cross-link yield per photon as a function of the intensity at mid cell is shown in Figure 4B.
Figure 4. Dependence of cross-linking on light dose and intensity. (A) Cross-links per 16S RNA, estimated from the polyacrylamide gel electrophoresis, plotted against absorbed photons per 16S rRNA. (B) Cross-links per 16S RNA per photon plotted against light intensity at the middle of the cell. Light intensity at mid cell was calculated from the incident light and the absorbance of the sample.
Patterns of RNA–RNA cross-linking
After irradiation, RNA samples were purified for size on agarose gels and then were 32P-labeled as described in the previous section. Samples from irradiation in the cuvette or in the flow cell were pooled separately and were electrophoresed on wide lanes of 3.6% polyacrylamide, 8.3 M urea gel (Fig. 5A). All bands were cut out and RNA was isolated by sedimentation after the gel slices were layered on top of a CsCl cushion (29). Reverse transcription was done on all fractions with a set of oligonucleotide primers to identify cross-linking sites (30). Reverse transcriptase stops that occur in the RNA of a specific fraction, but not in other fractions, indicate a cross-linking site (30–32). The patterns of cross-linking and reverse transcription were the same for samples irradiated in the cuvette and in the flow cell, except for a much higher frequency of the cross-linked RNA in fraction 6 (Fig. 5A) in RNA irradiated in the cuvette. The majority of cross-linking sites determined in these experiments had been observed previously in samples cross-linked with low intensity UV irradiation (29), except for four cross-links. Reverse transcription that revealed the new cross-links is shown in Figure 5B. The new cross-links are G976–A1362, G1964–G1190, A1196 to an unidentified site and G926 to an unidentified site. The G926 cross-link partner tentatively has been assigned to U1390 because a cross-linked RNA molecule with a similar electrophoretic mobility has been identified in Thermus thermophilus 16S rRNA and it contains a reverse transcription stop corresponding to 1390 in the 16S rRNA. The A1196 cross-link partner tentatively has been assigned to C1052 because a molecule with a similar mobility has been identified in Thermus aquaticus 16S rRNA and it is a cross-link between C1052 and A1196 (33). A summary of the cross-link assignments and location in the E.coli secondary structure is shown in Figure 6.
Figure 5. Identification of new cross-links. (A) Preparative polyacrylamide gel electrophoresis separation of cross-linked 16S RNA. About 50 μg of 16S RNA was 5' 32P-labeled and electrophoresed on the full width of a gel that was 34 cm x 40 cm x 0.35 mm. Electrophoresis was for 18 h at 10 W. An image of the 32P distribution was made with a phosphorimager and fractions were cut out of the gel as indicated. The RNA was isolated from the gel slices by ultracentrifugation. (B) Identification of cross-links by reverse transcription analysis. RNA fractions were used in a series of reverse transcription reactions. Selected reactions are shown here to illustrate the characterization of four bands that were not observed in low intensity irradiation. Reverse transcription stops unique to fraction 14 (G976 and A1362), fraction 5 (G1064 and G1190), fraction 6 (A1196) and fraction 12 (G926) are indicated on the left of the figure.
Figure 6. Summary of cross-link identification and location in the E.coli 16S rRNA secondary structure. The assignments of cross-link identity are indicated next to a lane from the denaturing polyacrylamide gel electrophoresis. The numbering used to identify each fraction is the same as in Figure 5. In three fractions (15, 13 and 3), two types of cross-linked molecules were present; in these instances, cross-linked pairs were inferred by the relationship between the loop size and mobility. RNA cross-links induced by UV pulse laser irradiation, and not seen before in E.coli with low intensity irradiation, are indicated by asterisks. The numbering of the cross-links in the right panel corresponds to the numbering on the left of the gel lane. The secondary structure diagram is from Cannone et al. (46).
UV laser cross-linking of purine nucleosides
The intrinsic reactivity of purine–purine pairs under conditions similar to those used for 16S RNA studies was investigated. Guanosine and adenosine in solution (either singly or in combination) were irradiated at 10 mM concentrations for exposures of five and 10 pulses with a 248 nm laser at 4.5 x 109 W m–2. These reactions were analyzed by LC mass spectrometry. No noticeable yield of A–A, G–G or G–A cross-links was seen. Given the low solubility of purines in solution, the average distance between the purine bases may prohibit the necessary interactions. To further test for cross-links, polycrystalline guanosine and adenosine were also irradiated by 248 nm laser pulses. Analysis by LC mass spectrometry again gave no evidence of cross-links for the solid samples.
Comparison of cross-linking frequencies and the internucleotide distances
The cross-linking frequencies for each cross-link were determined from the profile of the cross-linked RNA samples on the 3.6% polyacrylamide gels (Fig. 3) for samples irradiated with 4.5 x 109 W m–2. In the flow cell, the samples received an average of 4.6 pulses; the cross-linking frequencies reported in Table 1 are calculated for one pulse assuming linearity. Cross-linking frequencies were also determined in samples irradiated with a transilluminator (results not shown). Atom coordinates from the atomic structure determined by Wimberly et al. (34) were used to calculate distances between the bonds reactive in forming covalent cross-links. These are the C5–C6 double bond for pyrimidines and the C8–N7 double bond for purines. Cross-linking frequencies and reactive bond distances are summarized in Table 1.
Table 1. Cross-linking frequencies and reactive bond distances
Differences in cross-linking frequency for each site are seen for the different irradiation methods. Frequencies are generally higher for the pulse laser experiments as expected, but the increases for different sites are not constant. Correlation coefficients were calculated to determine which frequencies were more alike and to determine whether any of the cross-linking methods produced frequencies that are correlated to the reactive bond distances. One pulse irradiation and irradiation in the flow cell produce frequencies that are highly correlated, but the frequencies obtained by either of these methods are not correlated to the frequencies produced by the transilluminator irradiation (Table 2). If nucleotide pairs that are closer are cross-linked at greater frequency, a negative correlation coefficient between cross-link frequencies and reactive bond distance is expected. A correlation like that is seen for the frequencies produced by one pulse irradiation, but there is no correlation between frequencies made in the flow cell and the reactive bond distances, and the correlation coefficient between frequencies made in the transilluminator and the reactive bond distances has a positive value (Table 2).
Table 2. Correlation coefficients between cross-link frequencies and reactive bond distances
DISCUSSION
Strand breaks occur in the 16S rRNA to an extent approximately proportional to the absorbed photon dose and are also dependent on the light intensity over the range of dose and intensities that have been studied (Fig. 2). The average probability of strand breaks per photon is 0.7 to 1.3 x 10–3 depending on the incident light intensity. This is larger than the value for strand breakage seen in TMV, 1.7 x 10–4 breaks per photon, in which 4.7 x 1010 W m–2 intensity was used (14), but smaller than the yield of photoionization for bases, nucleosides and nucleotides, in the range of 10–2, recently reported by Crespo-Hernandez and Arce (35), in which irradiation intensities were 3–15 x 1010 W m–2. RNA strand breaks in the ribosome do not seem to be mediated by oxygen since samples have been routinely treated in closed tubes with a stream of nitrogen gas before irradiations, except for two samples that were irradiated without the usual de-oxygenation. There was no increased strand breakage in those two samples.
The cross-link yield is significant even at the lowest doses and light intensities that have been used here. Levels of 0.085–0.105 cross-links per 16S rRNA were obtained with a total average absorbed photon dose in the range of 200–400 photons per 16S RNA, which corresponds to a yield per photon of 3 x 10–4. This compares favorably with the cross-link yield of 0.7 x 10–4 per photon obtained with a transilluminator. The 4-fold higher cross-link yield per photon in the high versus low intensity devices is a little larger than the difference in the yield of RNA–protein cross-links made in TMV, which was 6.5 x 10–6 with a low intensity device and 1.8 x 10–5 with a nanosecond pulse laser (14). However, the difference is significantly lower than has been measured for RNA–protein cross-linking in the ribosome in which quantum yields of 20–100 times higher in a nanosecond pulse laser were observed compared with a low intensity irradiator (10). These differences in yield in response to low versus high intensity irradiation are probably due to the fact that the frequency of nucleotide–nucleotide cross-links is strongly influenced by photoreversal reactions . However, photoreversal reactions involving the products from nucleic acid–protein photoreactions should have very different responses, if they occur at all, due to the different structures of the products (36). Thus, in the case of RNA–RNA cross-linking, photoreversal must strongly affect cross-link yields with UV laser irradiation as it does in low intensity irradiation. The decrease in the cross-link yield per photon, at higher intensities, suggests that the photoreversal response is intensity dependent.
The pattern of cross-linking within the 16S rRNA with the UV pulse laser irradiation is very similar to that seen with low intensity transilluminator UVB light. Fourteen cross-links observed previously with transilluminator irradiation are seen again in the present experiments. Four additional cross-links, not previously detected in E.coli, are seen with the UV laser irradiation. A mix of products from single- and two-photon reaction has been described for RNA–protein cross-linking (10), so it is not surprising to find such a mix here. The new cross-links are G976–A1362, G1064–G1190, G926 tentatively to U1390 and A1196 tentatively to C1052. These were not seen in a low intensity irradiation experiment that determined cross-linking pattern as a function of wavelength (3), so their occurrence here is attributable to the light intensity rather than the wavelength. Six purine nucleotides are participants in the new cross-links. There are already other examples of purine–pyrimidine and purine–purine partners in cross-links made with low intensity devices, but the base composition of the new UV laser-induced cross-links contains a higher representation of purines than was seen previously from the low intensity irradiation. If the presence of these new cross-links is attributable to two photon reactions, their frequency should be more dependent on light intensity than the other cross-links. This is not particularly obvious, perhaps because a wide enough range of light intensities has not been used.
The failure to obtain purine–purine cross-links when solutions and crystalline samples were irradiated is consistent with the rapid non-radiative decay of purines and the absence of significant photochemistry previously reported for these molecules (37–40). We conclude that purine–purine cross-links observed in irradiated 16S RNAs are attributable to a particular geometry between the bases. This is consistent with previous work done on DNA (41–44).
The distances between the bases measured from the atomic structure (34) using the positions between N7/C8 atoms of purines and the positions between the C5/C6 atoms of the pyrimidine are: G976–A1362, 9.44 ?; G1064–G1190, 10.39 ?; G926–U1390, 19.08 ?; and C1052–A1196, 6.6 ?. The first two of these distances are somewhat outside the distance range observed for the majority of cross-links made by low intensity irradiation, which is 3.7–8.0 ? (Table 1). However, there are four other cross-links made by low intensity cross-linking that are separated by >10 ?, so the separation distance is not unique to these particular cross-links. In addition, there are many hundreds of pairs of nucleotides that have atoms that could participate in a UV cross-link within 9 ?, but are not reactive (B.Huggins, S.Ghosh, K.Nanda and P.Wollenzien, unpublished results). Therefore, the appearance of the four new cross-links is still very selective. The G926–U1390 cross-link, if it is confirmed, is well outside of the expected distance for cross-linking and would require the repositioning of the nucleotide G926. This nucleotide in the atomic structure is unstacked from the helix in which it is located as an unpaired nucleotide (34), but it could be repositioned by intercalation into the interior of the helix in a stacking arrangement under some circumstances and would be close to U1390 (45).
Cross-linking frequencies obtained with the UV pulse laser for each cross-link are usually larger than those obtained with a transilluminator. The correlation coefficients indicate a strong correlation between frequencies obtained with one pulse and frequencies obtained in the flow cell, but these are not related to the frequencies obtained in the transilluminator. An important result is that in spite of the frequency differences, the same pattern of cross-links is made, except for the four additional ones made with UV pulse laser irradiation. If cross-linking frequencies were determined by de-excitation processes such as non-radiative decay or energy transfer, which change the S1 or T1 lifetimes, the cross-linking frequencies would not depend on the irradiation procedure in the way that they do. Furthermore, tertiary structure rather than photochemistry is more important in determining the pattern of cross-linking (33). Therefore, the underlying factors that may be at play are: (i) the RNA tertiary structure that somehow determines the cross-linking pattern and is a factor in determining the cross-linking frequencies; and (ii) photoreversal reactions, that affect only the cross-linking frequencies. There is an inverse correlation between the internucleotide distances and the cross-linking frequencies after one laser pulse, but this is not seen with cross-linking frequencies after multiple pulses or after the extended time of irradiation in the transilluminator. This is proabably due to less photoreversal in the experiments with the single laser pulse.
These experiments document the elevated cross-link yield per photon upon irradiation with the UV pulse laser. This may be due to enhanced photochemistry from two photon reactions, but it is more likely that it is due simply to a reduced amount of photoreversal. On the other hand, the presence of new products not seen with low intensity irradiation is evidence for two photon reactions at those sites. The yield of cross-links per photon at the lowest intensity that was studied here makes pulse laser irradiation suitable for practical experiments because strand breakage and photoreversal are limited at that intensity. Moreover, the cross-linking frequencies obtained with a single UV pulse should better reflect the details of the geometry of the cross-linking partners, so methods that use single pulses will be better able to detect conformational differences than methods that use low intensity irradiation for extended times.
ACKNOWLEDGEMENTS
We thank Professor E.I. Bodowsky for discussions of the UV pulse laser irradiation technique, Professor Angus Kingon, Department of Materials Science and Engineering, NCSU for the use of the KrFl excimer laser, and Bruce Huggins for providing nucleotide distance measurements. This work is supported by NIH grant GM43237 to P.W., and by NSF grant MCB-9874895 to S.F.
REFERENCES
Fenick,D.J., Carr,H.S. and Falvey,D.E. (1995) Synthesis and photochemical cleavage of cis-syn pyrimidine cyclobutane dimer analogs. J. Org. Chem., 60, 624–631.
Ruzsicska,B. and Lemaire,D. (1995) DNA photochemistry. In Horspool,W.M. and Song,P.S. (eds), CRC Handbook of Organic Photochemistry and Photobiology. CRC Press, Boca Raton, FL, pp. 1289–1317.
Zhirnov,O.V. and Wollenzien,P. (2003) Action spectra for UV-light induced RNA–RNA crosslinking in 16S ribosomal RNA in the ribosome. Photochem. Photobiol. Sci., 2, 688–693.
Budowsky,E.I. and Abdurashidova,G.G. (1989) Polynucleotide–protein cross-links induced by ultraviolet light and their use for structural investigation of nucleoproteins. Prog. Nucleic Acid Res. Mol. Biol., 37, 1–65.
Daniels,M. (1976) Excited states of the nucleic acids: bases, mononucleosides and mononucleotides. In Wang,S.Y. (ed.), Photochemistry and Photobiology of Nucleic Acids. Academic Press, New York, pp. 23–109.
Fisher,G. and Johns,H. (1976) Pyrimidine photodimers. In Wang,S.Y. (ed.), Photochemistry and Photobiology of Nucleic Acids. Academic Press, New York, NY, pp. 225–294.
Cadet,J. and Vigny,P. (1990) The photochemistry of nucleic acids. In Morrison,H. (ed.), Bioorganic Photochemistry: Photochemistry and the Nucleic Acids. John Wiley and Sons, New York, NY, Vol. 1, pp. 1–272.
Lemaire,D.G.E. and Ruszicska,B.P. (1993) Quantum yields and secondary photoreactions of the photoproducts of dTpdT, dTpdC and dTpdU. Photochem. Photobiol., 57, 755–769.
Kryukov,P., Letokhov,V., Nikogosyan,D., Borodavkin,A., Budowsky,E.I. and Simukova,N. (1979) Multiquantum photoreactions of nucleic acid components in aqueous solution by powerful ultraviolet picosecond radiation. Chem. Phys. Lett., 61, 375–379.
Budowsky,E.I., Azentyeva,M., Abdurashidova,G., Simukova,N. and Rubin,L. (1986) Induction of polynucleotide–protein cross-linkages by ultraviolet irradiation. Pecularities of the high-intensity laser pulse irradiation. Eur. J. Biochem., 159, 95–101.
Harrison,C., Turner,D. and Hinkle,D. (1982) Laser crosslinking of E.coli RNA polymerase and T7 DNA. Nucleic Acids Res., 10, 2399–2414.
Hockensmith,J., Kubasek,W., Vorachek,W. and von Hippel,P. (1986) Laser cross-linking of nucleic acids to proteins. J. Biol. Chem., 8, 3512–3518.
Angelov,D., Stefanovsky,V., Dimitrov,S., Russanova,V., Keskiniva,E. and Pashev,I. (1988) Protein–rDNA crosslinking in reconstituted nucleohistone complexes, nuclei and whole cells by picosecond UV laser irradiation. Nucleic Acids Res., 16, 4525–4538.
Dobrov,E., Arbieva,Z., Timofeeva,E., Esenaliev,R., Oraevsky,A. and Nikogosyan,D. (1989) UV laser induced RNA–protein crosslinks and RNA chain breaks in tobacco mosaic virus RNA in situ. Photochem. Photobiol., 49, 595–598.
Ho,D., Sauve,D. and Roberge,M. (1994) Detection and isolation of DNA-binding proteins using single-pulse ultraviolet laser crosslinking. Anal. Biochem., 218, 248–254.
Connor,D., Falick,A., Yuoung,M. and Shetlar,M. (1998) Probing the binding region of the single-stranded DNA-binding domain of rat DNA polymerase beta using nanosecond-pulse laser-induced crosslinking and mass spectrometry. Photochem. Photobiol., 68, 299–308.
Moss,T., Dimtrov,S. and Houde,D. (1997) UV-laser crosslinking of proteins to DNA. Methods, 11, 225–234.
Nikogosyan,D. and Gurzadyan,G. (1984) Two quantum photoprocesses in DNA and RNA biopolymers under powerful picosecond laser UV irradiation. Laser Chem., 4, 297–303.
Johnson-Thompson,M., Halpern,J., Jackson,W. and Georges,J. (1984) Vacuum UV laser induced scission of simian virus 40 DNA. Photochem. Photobiol., 39, 17–24.
Schulte-Frohlinde,D., Opitz,J., Gorner,H. and Bothe,E. (1985) Model studies for the direct effect of high energy irradiation on DNA. Mechanism of strand break formation induced by laser photoionization of poly(U) in aqueous solutions. Int. J. Radiat. Biol., 48, 397–408.
Menshonkova,T., Simukova,N., Budowsky,E.I. and Rubin,L. (1980) The effect of high intensity ultraviolet irradiation on nucleic acids and their components. FEBS Lett., 112, 299–301.
Opitz,J. and Schulte-Frohlinde,D. (1987) Laser-induced photoionization and single-strand break formation for polynucleotides and single-stranded DNA in aqueous solution: model studies for the direct effect of high energy irradiation on DNA. J. Photochem., 39, 145–163.
Blau,W., Croke,D., Kelly,J., McConnell,O.C. and van der Putten,W. (1987) Base-specific photocleavage of DNA induced by nanosecond UV pulse laser ragiation or methylene blue sensitization. J. Chem. Soc. Chem. Commun., 10, 751–752.
Croke,D., Blau,W., OhUigin,C., Kelly,J. and McConnell,D. (1988) Photolysis of phosphodiester bonds in plasmid DNA by high intensity UV laser irradiation. Photochem. Photobiol., 47, 527–536.
Rubin,L., Menshonkava,T., Simukova,N. and Budowsky,E.I. (1981) The effects of high-intensity UV-radiation on nucleic acids and their components—I. Thymine. Photochem. Photobiol., 34, 339–344.
Kovalsky,O., Panyutin,I. and Budowsky,E. (1990) Sequence-specificity of the alkali-sensitive lesions induced in DNA by high-intensity ultraviolet laser radiation. Photochem. Photobiol., 52, 509–517.
Makhno,V.I., Peshin,N.N., Semenkov,Y.P. and Kirillov,S.V. (1988) Modified method of producing ‘tight’ 70S ribosomes from Escherichia coli, highly active in individual stages of the elongation cycle. Mol. Biol., 22, 28–537.
Juzumiene,D., Shapkina,T., Kirillov,S. and Wollenzien,P. (2001) Short-range RNA–RNA crosslinking methods to determine rRNA structure and interactions. Methods, 25, 333–343.
Wilms,C. and Wollenzien,P. (1994) Purification of RNA from poly acrylamide gels by ultracentrifugation. Anal. Biochem., 221, 204–205.
Wilms,C., Noah,J.W., Zhong,D. and Wollenzien,P. (1997) Exact determination of UV-induced crosslinks in 16S ribosomal RNA in 30S ribosomal subunits. RNA, 3, 602–612.
Ericson,G. and Wollenzien,P. (1988) Use of reverse transcription to determine the exact locations of psoralen photochemical cross-links in RNA. Anal. Biochem., 174, 215–223.
Ericson,B., Chevli,K. and Wollenzien,P. (1989) Structure of synthetic unmethylated 16S ribosomal RNA as purified RNA and in reconstituted 30S ribosomal subunits. Biochemistry, 28, 6446–6454.
Noah,J.W. Shapkina,T. and Wollenzien,P. (2000) UV-induced crosslinks in the 16S rRNAs of Escherichia coli, Bacillus subtilis and Thermus aquaticus and their implications for ribosome structure and photochemistry. Nucleic Acids Res., 28, 3785–3792.
Wimberly,B.T., Brodersen,D.E., Clemons,W.M., Morgan-Warren,R.J., Carter,A.P., Vonrhein,C., Hartsch,T. and Ramakrishnan,V. (2000) Structure of the 30S ribosomal subunit. Nature, 407, 327–339.
Crespo-Hernandez,C.E. and Arce,R. (2002) Photoionization of DNA and RNA bases, nucleosides and nucleotides through a combination of one- and two-photon pathways upon 266 nm nanosecond excitation. Photochem. Photobiol., 76, 259–267.
Saito,I. and Sugiyama,H. (1990) Photoreactions of nucleic acids and their constituents with amino acids and related compounds. In Morrison,H. (ed.), Bioorganic Photochemistry. John Wiley and Sons, New York, NY, Vol. 1, pp. 317–340.
Bose,S.N., Davies,R.J.H., Anderson,D.W., Vanniekerk,J.C., Nassimbeni,L.R. and Macfarlane,R.D. (1978) Photo-chemical (8-]8) coupling of purine nucleosides to guanosine. Nature, 271, 783–784.
Gorner,H. (1994) Photochemistry of DNA and related biomolecules: quantum yields and consequences of photoionization. J. Photochem. Photobiol., 26, 117–139.
Ivanchenko,V.A., Budowsky,E.I., Simukova,N.A., Vul’fson,N.S., Tishchenko,A.I. and Askerov,D.B. (1977) The photochemistry of purine components of nucleic acids. 2 Photolysis of deoxyguanosine. Nucleic Acids Res., 4, 955–968.
Ivanchenko,V.A., Titschenko,A.I., Budowsky,E.I., Simukova,N.A. and Wulfson,N.S. (1975) The photochemistry of purine components of nucleic acids. 1 The efficiency of photolysis of adenine and guanine derivatives in aqueous solution. Nucleic Acids Res., 2, 1365–1373.
Bose,S.N., Davies,R.J.H., Sethi,S.K. and McCloskey,J.A. (1983) Formation of an adenine–thymine photoadduct in the deoxydinucleoside monophosphate d(TpA) and in DNA. Science, 220, 723–725.
Bose,S.N. and Davies,R.J.H. (1984) The photoreactivity of T–A sequences in oligodeoxyribonucleotides and DNA. Nucleic Acids Res., 12, 7903–7914.
Melvin,T., Plumb,M.A., Botchway,S.W., O’Neil,P. and Parker,A.W. (1995) 193 nm light induces single strand breakage of DNA predominantly at guanine. Photochem. Photobiol., 61, 584–591.
Davies,R.J.H. (1995) Ultraviolet-radiation damage in DNA. Biochem. Soc. Trans., 23, 407–418.
Ericson,G., Minchew,P. and Wollenzien,P. (1995) Structural changes in base-paired region 28 in 16S rRNA close to the decoding region of the 30S ribosomal subunit are correlated to changes in tRNA binding. J. Mol. Biol., 250, 407–419.
Cannone,J., Subramanian,S., Schnare,M., Collett,J., S’Souza,L., Du,Y., Feng,B., Lin,N., Madabusi,L., Muller,K., Pande,N., Shang,Z. and Gutell,R. (2002) The comparative RNA Web (CRW) site: an online database of comparative sequence and structure information for ribosomal, intron and other RNAs. BioMed. Central Bioinform., 3, 2.(Tatjana Shapkina, Simon Lappi1, Stefan F)
*To whom correspondence should be addressed. Tel: +1 919 515 5703; Fax: +1 919 515 2047; Email: paul_wollenzien@ncsu.edu
ABSTRACT
Escherichia coli ribosomes were irradiated with a KrF excimer laser (248 nm, 22 ns pulse) with incident pulse energies in the range of 10–40 mJ for a 1 cm2 area, corresponding to fluences of 4.5 to 18 x 109 W m–2, to determine strand breakage yields and the frequency and pattern of RNA–RNA cross- linking in the 16S rRNA. Samples were irradiated in a cuvette with one laser pulse or in a flow cell with an average of 4.6 pulses per sample. The yield of strand breaks per photon was intensity dependent, with values of 0.7 to 1.3 x 10–3 over the incident intensity range studied. The yield for RNA–RNA cross-linking was 3 x 10–4 cross-links/photon at the intensity of 4.5 x 109 W m–2, an 4-fold higher yield per photon than obtained with a transilluminator. The cross-link yield/photon decreased at higher light intensities, probably due to intensity-dependent photoreversal. The pattern of cross-linking was similar to that observed with low intensity irradiation but with four additional long-range cross-links not previously seen in E.coli ribosomes. Cross- linking frequencies obtained with one laser pulse are more correlated to internucleotide distances than are frequencies obtained with transilluminator irradiation.
INTRODUCTION
Despite the wide use of UV cross-linking in studies of RNA and DNA structures, nanosecond pulse irradiation has not been reported as a method to obtain RNA–RNA cross-linking. There are a number of concerns that need to be addressed in the application of high intensity sources to the determination of structural constraints in RNA. Given the extent of DNA and RNA strand breaks during irradiation with UV pulse lasers, the possibility exists that strand breakage reactions would dominate all products and would preclude the analysis of RNA–RNA products by RNA sequencing methods. Moreover, extensive strand breakage during the irradiation would disrupt the native structure of the complex. A second factor not yet investigated for high intensity irradiation concerns the photoreversal response. The rates of formation and reversal of cross-link products set the equilibrium value for cross-link frequency in model dinucleotide compounds subjected to low intensity irradiation (1,2). Photoreversal has also been shown to be important in determining the yields of RNA–RNA cross-links in 16S rRNA in low intensity irradiations (3). Therefore, it is possible that cross-link yields could be lower under high intensity irradiation than they are under low intensity irradiation, if photoreversal were strongly intensity dependent. On the other hand, there are a number of advantages in the use of pulsed laser sources for UV cross-linking if these concerns could be alleviated. Since the pulse laser experiments are completed with one or a few pulses, products that arise from covalent reaction of the nucleic acid are likely to represent native conformations and interactions (4). In addition, the ability to rapidly induce cross-links would allow kinetic experiments on time scales limited mainly by the time constants of mixing devices.
Irradiation of nucleic acids with low intensity UV light results in formation of S1 and T1 excited states (5–7), and covalent reactions are thought to involve molecules in the S1 state because intersystem crossing yields for bases are low and usually cannot account for the yields of photoproducts (7,8). Nucleic acids attain activation states higher than the S1 and T1 states (the HT and HS states) at light intensities of at least 108 W m–2 (1018 photons cm–2 s–1) due to absorption of two photons (7,9,10). These light intensities are easily obtained in nanosecond or picosecond pulse lasers. In the HT and HS states, the bases undergo reactions with greater efficiency due to their higher energy, and the specificity of the reactions may also be expanded due to the different electronic structure of these states compared with excitation with low intensity irradiation (4,10).
Nucleic acid–protein cross-linking with pulse UV laser irradiation has been described for a variety of systems (4,9,11–17). These methods are attractive because high efficiency is indeed obtained with short irradiation times compared with the extended irradiation times that are needed with low intensity UV irradiation. Bases in the HT and HS states possess energy levels higher than the ionization energy (4), and irradiation with nanosecond or picosecond laser pulses leads to nucleic acid strand breakage (5,7,18–24). The structures of some photoproducts indicate formation by free radical reactions, confirming ionization (25). In DNA, particularly at guanosines in specific sequence contexts, pulse laser irradiation induces lesions in nucleotides that are cleaved upon alkaline treatment and are attributed to biphotonic reactions (24,26). The quantum yield of the strand breakage reaction has been reported for tobacco mosaic virus (TMV) RNA with low intensity, nanosecond and picosecond UV irradiation (14), and confirms a higher extent of direct strand breakage at high intensities compared with irradiation with low intensities.
The goal of the present work was to determine the extent of strand breakage and cross-linking in the 16S rRNA (1542 nt) in the ribosome with a UV pulse laser irradiation source. A KrF excimer laser with 248 nm light with pulse intensities of 4.5 to 18 x 109 W m–2 and 22 ns pulse duration has been used for these experiments. We wanted to determine first the dependence of strand breakage on the irradiation conditions and whether nanosecond pulse UV irradiation could be used under conditions where strand breakage was negligible, or at least occurred at levels where intact full-length RNA could still be isolated after irradiations. This has proven to be the case, and the frequency and locations of intramolecular RNA–RNA cross-links have been determined to see if RNA–RNA cross-links were produced that were not seen with low intensity irradiations. Differences in the relative RNA–RNA cross-linking frequencies are also found in comparing UV pulse laser and transilluminator irradiation, and these are discussed.
MATERIALS AND METHODS
Washed 70S ribosomes were prepared according to Makhno et al. (27) from Escherichia coli cells grown to mid log phase. 30S and 50S ribosomal subunits were separated by centrifugation of 70S ribosomes on sucrose gradients prepared in T20A200M3 buffer (20 mM Tris–HCl pH 7.5, 200 mM NH4Cl, 3 mM MgCl2, 4 mM ?-mercaptoethanol). Subunits were concentrated by sedimentation and dissolved in activation buffer (20 mM Tris–HCl pH 7.5, 200 mM NH4Cl, 20 mM MgCl2, 4 mM ?-mercaptoethanol), and frozen at –85°C for storage. 70S ribosomes were formed by re-association of 30S and 50S subunits in activation buffer by incubation for 30 min at 37°C.
Irradiation of ribosomes
Ribosome samples were irradiated in a cuvette with one pulse or in a flow cell with the continuous laser pulses. For irradiation in the cuvette, a fluorescence cuvette usually with 100 μl of sample was covered with a rubber septum and treated with a stream of N2 gas before irradiation. The cuvette was turned 90° from its normal arrangement so a large cross-section area of the sample would face the incident radiation. The cuvette path length was 2 mm in this arrangement and the exposed area was 0.5 cm2. Irradiation was with a single pulse (22 ns pulse length) administered to the sample using a manual shutter of a KrF excimer laser (Lambda Physik, Fort Lauderdale, FL). The laser was operated at 1 Hz to provide a reproducible pulse intensity. The flow cell had a path length of 1 mm and a cross-section area of 0.38 cm2. Samples of 1 ml were treated with N2 gas in a closed tube, and a peristaltic pump was used to pump the sample through the cell at 0.5 ml/min. The KrF excimer laser was used at 1 Hz, so each molecule received an average exposure of 4.6 pulses. The energy was monitored continuously in all experiments. The cells were Supracil quartz, obtained from Hellma Cells, Inc. (Forest Hills, NY).
Analysis of RNA strand breaks and cross-links by gel electrophoresis
Ribosomes were deproteinated by incubation with 1.0 μg/μl proteinase K and 1.0% SDS/0.02 M EDTA at 37°C for 30 min, followed by phenol extraction and ethanol precipitation. RNA was re-dissolved in water in volumes proportional to the starting amount of RNA mass; 5 μl of these samples should contain 0.5 μg of 16S RNA or, in some experiments, 1.0 μg of 16S rRNA. A 5 μl aliquot of each of these samples was mixed with equal volumes of formamide, heated to 45°C for 10 min and immediately electrophoresed on 1% agarose gels made with 1x TBE buffer (89 mM Tris base, 89 mM boric acid, 2 mM EDTA pH 8.3) and containing 100 μg/ml ethidium bromide. RNA was visualized by fluorescence. Control samples containing 0.5 or 1.0 μg of 16S rRNA were used for comparison.
For cross-link analysis, irradiated RNA first was purified by agarose gel electrophoresis to collect the 16S RNA band as described by Juzumiene et al. (28). RNA was phenol and ether extracted, and precipitated. The RNA was treated with shrimp alkaline phosphatase and, after phenol and ether extraction and ethanol precipitation, was labeled with ATP using polynucleotide kinase. The RNA was phenol and ether extracted, and unincorporated ATP was removed by two ethanol precipitations using 2.5 M NH4OAc pH 5.1 and 2.5 vol of EtOH for precipitation (28).
Usually, 0.2 μg of RNA in 70% formamide was loaded on a 1 cm wide well of a 0.35 mm thick x 40 cm long polyacrylamide gel which is 3.6% acrylamide:bisacrylamide (70:1), 8.3 M urea in BTBE buffer . The gel, thermostatted at 45°C, was run for 18 h at 6 W constant power (29). After electrophoresis, the gel was transferred to paper, dried, and the pattern and frequency of cross-linked species were determined with a Molecular Dynamics PhosphorImager (Amersham, Piscataway, NJ). Cross-link frequencies were calculated from band intensities using ImageQuant software (Amersham, Piscataway, NJ). Correlation coefficients were calculated with Microsoft Excel.
Cross-linking sites in the 16S rRNA were determined after separation of cross-linked molecules on 3.6% acrylamide: bisacrylamide (70:1), 8.3 M urea in BTBE buffer. Typically, 50–100 μg of 32P-labeled 16S rRNA was electrophoresed on a lane 34 cm wide on the 40 cm long x 0.35 mm thick polyacrylamide gel. After 18 h of electrophoresis, one glass plate was removed and the gel was covered with plastic wrap and exposed on a phosphorimager plate to locate the bands containing the cross-linked RNA species and the linear rRNA. Gel slices containing these were cut out and RNA was isolated by centrifugation over CsCl cushions (28,29). RNA from each fraction was inspected for cross-linking sites by reverse transcription reactions using a set of oligonucleotide primers according to the procedure of Wilms et al. (30). Cross-link sites are detected as reverse transcription stopping points present in specific fractions (30–32).
Irradiation of purine nucleosides and analysis of products
Guanosine and adenosine were purchased from Sigma Chemical Company (St Louis, MO) and used without further purification. Solutions (10 mM) of both guanosine and adenosine were prepared using freshly dispensed 18 M cm water (BARNSTEAD E-PURE) and used either singly or in combination. The nucleoside solutions were placed in a Spectrosil (Starna Cells, Inc., Atascadero, CA) cuvette, with a 1 mm path length, covered with a rubber septum and purged with a stream of N2 gas before irradiation. The solutions were then irradiated with either five or 10 pulses from the KrF excimer laser at 4.5 to 18 x 109 W m–2. These reactions were analyzed by liquid chromatographic (LC) mass spectrometry (HP LC/MSD 1100, Agilent Technologies, Wilmington, DE) using the direct injection method. The polycrystalline samples were prepared by finely grinding the powders in a ceramic mortar and used singly or combined into a 50:50 mixture by weight. The powdered samples were placed between CaF2 (International Crystal Laboratories, Garfield, NJ) and irradiated in the same manner as described above. The irradiated nucleotides were then dissolved in 18 Mcm water to 10 mM and analyzed by LC mass spectrometry using the direct injection method.
RESULTS
Dependency of RNA strand breaks on photon dose and light intensity
Escherichia coli 30S subunits and 70S ribosomes were irradiated in a flow cell or in a fluorescence cuvette in sets of increasing light intensity. Ribosome concentrations between 1.25 and 5 A260 units were used so that enough cross-linked samples for subsequent analysis would be available. After irradiation, ribosome samples were treated with proteinase K and phenol extraction, and RNA was ethanol precipitated and dissolved in water. Samples were diluted with formamide and heated for 10 min at 45°C to melt the RNA and then it was immediately subjected to electrophoresis on agarose gels and was viewed by ethidium fluorescence (Fig. 1). Strand breakage was seen in many of the samples and was more extensive for samples irradiated at higher pulse intensities in the flow cell, in which the sample is exposed to multiple pulses. The sites of strand breakage appear to be fairly random because no specific low molecular weight bands appear on the agarose gels. The sites have not been determined by reverse transcription analysis. The fraction of strand breaks per 16S RNA was estimated from the gels for these samples and for other samples under additional conditions. This was done by using sample volumes that should contain 0.5 or 1.0 μg of 16S rRNA determined by the A260 measurement and comparing the 16S rRNA band intensities with those seen in samples that were not irradiated.
Figure 1. Detection of strand breaks in irradiated RNA by agarose gel electrophoresis. 30S ribosome subunits or 70S ribosomes were irradiated in a flow cell or in a cuvette with different incident intensities as indicated. Samples were deproteinated, ethanol precipitated and redissolved in water. A volume of each sample was taken that should contain 0.5 μg of 16S RNA (lanes 1–5) or 1.0 μg of 16S RNA (lanes 7–16). Lane 1 contains 0.5 μg of 16S RNA and lane 10 contains 1.0 μg of 16S rRNA (as a component of total rRNA). There are small amounts of 23S rRNA in lanes containing RNA from 30S subunits due to contamination from 50S subunits. The two lanes on the right marked with asterisks contain RNA from samples that were not de-oxygenated. The amount of full-length 16S RNA remaining in the sample was estimated and used to calculate the extent of strand breakage.
The logarithm of the fraction of full-length RNA remaining after irradiation was plotted versus dose (absorbed photons per 16S RNA), as shown in Figure 2. There are different values for the dose (photons absorbed per 16S) in this plot even for samples irradiated at the same pulse intensity because the calculation of dose also depends on RNA concentration, path length and whether 70S or 30S ribosomes were used. The slopes of the best-fit lines indicate a probability of 0.7 x 10–3 breaks/photon for samples with A260 5 irradiated with a pulse intensity of 4.5 x 109 W m–2, and a probability of 1.3 x 10–3 breaks/photon for samples with A260 5 irradiated with a pulse intensity of 18 x 109 W m–2.
Figure 2. Dependence of strand breaks on light dose and intensity. The natural logarithm of the fraction of full-length RNA (where fraction of full-length RNA was estimated from the agarose gel electrophoresis), plotted against absorbed photons per 16S rRNA is shown. The lines labeled a and b are linear fits through data obtained at 4.5 x 109 and 18 x 109 W m–2, respectively, and have slopes of –0.7 x 10–3 and –1.3 x 10–3.
Cross-linking yields for long-range RNA–RNA cross-links at increasing light intensity
The frequency of intramolecular cross-linking was determined by gel electrophoresis of 16S RNA on denaturing polyacrylamide gels. To do this, the intact 16S RNA remaining in each sample first was isolated by agarose gel electrophoresis and was then 5' end-labeled with ATP and polynucleotide kinase. It was then electrophoresed on 3.6% polyacrylamide gels (acrylamide:bisacrylamide; 70:1) with 8.3 M urea at 45°C. Molecules separate into specific bands with reduced mobility under these conditions due to the presence and location of intramolecular cross-links (29). The frequency of the cross-links was then determined by analysis of the 32P distribution in the gel. Examples of 16S RNA cross-linked in the flow cell and cuvette are shown in Figure 3.
Figure 3. Denaturing polyacrylamide gel electrophoresis of cross-linked 16S RNA to determine cross-link frequency. 16S RNA was purified by agarose gel electrophoresis and was 5' labeled with 32P. Samples were electrophoresed on 8.3 M urea, 3.6% polyacrylamide gels at 45°C. The method and incident light intensity are indicated for each lane. The gel was dried and exposed on a phosphorimager plate to determine the image and cross-link frequencies. The heavy band in the bottom of each lane is at the position of unirradiated 16S RNA.
The sum of the band frequencies was quantified to determine the overall levels of cross-linking. The gel separates molecules that contain cross-links between nucleotides that are separated by at least 50 nucleotides; however, RNA isolated from these samples was very readily reverse transcribed, so there must not be too much short-range cross-linking in the samples. In the worst case shown in Figure 3 (30S subunits in the flow cell with 18 x 109 W m–2), it was difficult to obtain enough RNA for the analysis and instead of distinct bands there is a diffuse RNA density throughout the lane. In general, at the higher irradiation intensities, strand breaks in cross-linked molecules should produce X-shaped molecules, and this was probably the reason for the reduction in the sharpness of bands. In addition, strand breaks produce an increased amount of radioactivity throughout the gel above the normal background. Therefore, radioactivity above background was included in the calculation of the total cross-link frequencies. The dependence of cross-link frequency on absorbed dose is shown in Figure 4A, and the dependence of the cross-link yield per photon as a function of the intensity at mid cell is shown in Figure 4B.
Figure 4. Dependence of cross-linking on light dose and intensity. (A) Cross-links per 16S RNA, estimated from the polyacrylamide gel electrophoresis, plotted against absorbed photons per 16S rRNA. (B) Cross-links per 16S RNA per photon plotted against light intensity at the middle of the cell. Light intensity at mid cell was calculated from the incident light and the absorbance of the sample.
Patterns of RNA–RNA cross-linking
After irradiation, RNA samples were purified for size on agarose gels and then were 32P-labeled as described in the previous section. Samples from irradiation in the cuvette or in the flow cell were pooled separately and were electrophoresed on wide lanes of 3.6% polyacrylamide, 8.3 M urea gel (Fig. 5A). All bands were cut out and RNA was isolated by sedimentation after the gel slices were layered on top of a CsCl cushion (29). Reverse transcription was done on all fractions with a set of oligonucleotide primers to identify cross-linking sites (30). Reverse transcriptase stops that occur in the RNA of a specific fraction, but not in other fractions, indicate a cross-linking site (30–32). The patterns of cross-linking and reverse transcription were the same for samples irradiated in the cuvette and in the flow cell, except for a much higher frequency of the cross-linked RNA in fraction 6 (Fig. 5A) in RNA irradiated in the cuvette. The majority of cross-linking sites determined in these experiments had been observed previously in samples cross-linked with low intensity UV irradiation (29), except for four cross-links. Reverse transcription that revealed the new cross-links is shown in Figure 5B. The new cross-links are G976–A1362, G1964–G1190, A1196 to an unidentified site and G926 to an unidentified site. The G926 cross-link partner tentatively has been assigned to U1390 because a cross-linked RNA molecule with a similar electrophoretic mobility has been identified in Thermus thermophilus 16S rRNA and it contains a reverse transcription stop corresponding to 1390 in the 16S rRNA. The A1196 cross-link partner tentatively has been assigned to C1052 because a molecule with a similar mobility has been identified in Thermus aquaticus 16S rRNA and it is a cross-link between C1052 and A1196 (33). A summary of the cross-link assignments and location in the E.coli secondary structure is shown in Figure 6.
Figure 5. Identification of new cross-links. (A) Preparative polyacrylamide gel electrophoresis separation of cross-linked 16S RNA. About 50 μg of 16S RNA was 5' 32P-labeled and electrophoresed on the full width of a gel that was 34 cm x 40 cm x 0.35 mm. Electrophoresis was for 18 h at 10 W. An image of the 32P distribution was made with a phosphorimager and fractions were cut out of the gel as indicated. The RNA was isolated from the gel slices by ultracentrifugation. (B) Identification of cross-links by reverse transcription analysis. RNA fractions were used in a series of reverse transcription reactions. Selected reactions are shown here to illustrate the characterization of four bands that were not observed in low intensity irradiation. Reverse transcription stops unique to fraction 14 (G976 and A1362), fraction 5 (G1064 and G1190), fraction 6 (A1196) and fraction 12 (G926) are indicated on the left of the figure.
Figure 6. Summary of cross-link identification and location in the E.coli 16S rRNA secondary structure. The assignments of cross-link identity are indicated next to a lane from the denaturing polyacrylamide gel electrophoresis. The numbering used to identify each fraction is the same as in Figure 5. In three fractions (15, 13 and 3), two types of cross-linked molecules were present; in these instances, cross-linked pairs were inferred by the relationship between the loop size and mobility. RNA cross-links induced by UV pulse laser irradiation, and not seen before in E.coli with low intensity irradiation, are indicated by asterisks. The numbering of the cross-links in the right panel corresponds to the numbering on the left of the gel lane. The secondary structure diagram is from Cannone et al. (46).
UV laser cross-linking of purine nucleosides
The intrinsic reactivity of purine–purine pairs under conditions similar to those used for 16S RNA studies was investigated. Guanosine and adenosine in solution (either singly or in combination) were irradiated at 10 mM concentrations for exposures of five and 10 pulses with a 248 nm laser at 4.5 x 109 W m–2. These reactions were analyzed by LC mass spectrometry. No noticeable yield of A–A, G–G or G–A cross-links was seen. Given the low solubility of purines in solution, the average distance between the purine bases may prohibit the necessary interactions. To further test for cross-links, polycrystalline guanosine and adenosine were also irradiated by 248 nm laser pulses. Analysis by LC mass spectrometry again gave no evidence of cross-links for the solid samples.
Comparison of cross-linking frequencies and the internucleotide distances
The cross-linking frequencies for each cross-link were determined from the profile of the cross-linked RNA samples on the 3.6% polyacrylamide gels (Fig. 3) for samples irradiated with 4.5 x 109 W m–2. In the flow cell, the samples received an average of 4.6 pulses; the cross-linking frequencies reported in Table 1 are calculated for one pulse assuming linearity. Cross-linking frequencies were also determined in samples irradiated with a transilluminator (results not shown). Atom coordinates from the atomic structure determined by Wimberly et al. (34) were used to calculate distances between the bonds reactive in forming covalent cross-links. These are the C5–C6 double bond for pyrimidines and the C8–N7 double bond for purines. Cross-linking frequencies and reactive bond distances are summarized in Table 1.
Table 1. Cross-linking frequencies and reactive bond distances
Differences in cross-linking frequency for each site are seen for the different irradiation methods. Frequencies are generally higher for the pulse laser experiments as expected, but the increases for different sites are not constant. Correlation coefficients were calculated to determine which frequencies were more alike and to determine whether any of the cross-linking methods produced frequencies that are correlated to the reactive bond distances. One pulse irradiation and irradiation in the flow cell produce frequencies that are highly correlated, but the frequencies obtained by either of these methods are not correlated to the frequencies produced by the transilluminator irradiation (Table 2). If nucleotide pairs that are closer are cross-linked at greater frequency, a negative correlation coefficient between cross-link frequencies and reactive bond distance is expected. A correlation like that is seen for the frequencies produced by one pulse irradiation, but there is no correlation between frequencies made in the flow cell and the reactive bond distances, and the correlation coefficient between frequencies made in the transilluminator and the reactive bond distances has a positive value (Table 2).
Table 2. Correlation coefficients between cross-link frequencies and reactive bond distances
DISCUSSION
Strand breaks occur in the 16S rRNA to an extent approximately proportional to the absorbed photon dose and are also dependent on the light intensity over the range of dose and intensities that have been studied (Fig. 2). The average probability of strand breaks per photon is 0.7 to 1.3 x 10–3 depending on the incident light intensity. This is larger than the value for strand breakage seen in TMV, 1.7 x 10–4 breaks per photon, in which 4.7 x 1010 W m–2 intensity was used (14), but smaller than the yield of photoionization for bases, nucleosides and nucleotides, in the range of 10–2, recently reported by Crespo-Hernandez and Arce (35), in which irradiation intensities were 3–15 x 1010 W m–2. RNA strand breaks in the ribosome do not seem to be mediated by oxygen since samples have been routinely treated in closed tubes with a stream of nitrogen gas before irradiations, except for two samples that were irradiated without the usual de-oxygenation. There was no increased strand breakage in those two samples.
The cross-link yield is significant even at the lowest doses and light intensities that have been used here. Levels of 0.085–0.105 cross-links per 16S rRNA were obtained with a total average absorbed photon dose in the range of 200–400 photons per 16S RNA, which corresponds to a yield per photon of 3 x 10–4. This compares favorably with the cross-link yield of 0.7 x 10–4 per photon obtained with a transilluminator. The 4-fold higher cross-link yield per photon in the high versus low intensity devices is a little larger than the difference in the yield of RNA–protein cross-links made in TMV, which was 6.5 x 10–6 with a low intensity device and 1.8 x 10–5 with a nanosecond pulse laser (14). However, the difference is significantly lower than has been measured for RNA–protein cross-linking in the ribosome in which quantum yields of 20–100 times higher in a nanosecond pulse laser were observed compared with a low intensity irradiator (10). These differences in yield in response to low versus high intensity irradiation are probably due to the fact that the frequency of nucleotide–nucleotide cross-links is strongly influenced by photoreversal reactions . However, photoreversal reactions involving the products from nucleic acid–protein photoreactions should have very different responses, if they occur at all, due to the different structures of the products (36). Thus, in the case of RNA–RNA cross-linking, photoreversal must strongly affect cross-link yields with UV laser irradiation as it does in low intensity irradiation. The decrease in the cross-link yield per photon, at higher intensities, suggests that the photoreversal response is intensity dependent.
The pattern of cross-linking within the 16S rRNA with the UV pulse laser irradiation is very similar to that seen with low intensity transilluminator UVB light. Fourteen cross-links observed previously with transilluminator irradiation are seen again in the present experiments. Four additional cross-links, not previously detected in E.coli, are seen with the UV laser irradiation. A mix of products from single- and two-photon reaction has been described for RNA–protein cross-linking (10), so it is not surprising to find such a mix here. The new cross-links are G976–A1362, G1064–G1190, G926 tentatively to U1390 and A1196 tentatively to C1052. These were not seen in a low intensity irradiation experiment that determined cross-linking pattern as a function of wavelength (3), so their occurrence here is attributable to the light intensity rather than the wavelength. Six purine nucleotides are participants in the new cross-links. There are already other examples of purine–pyrimidine and purine–purine partners in cross-links made with low intensity devices, but the base composition of the new UV laser-induced cross-links contains a higher representation of purines than was seen previously from the low intensity irradiation. If the presence of these new cross-links is attributable to two photon reactions, their frequency should be more dependent on light intensity than the other cross-links. This is not particularly obvious, perhaps because a wide enough range of light intensities has not been used.
The failure to obtain purine–purine cross-links when solutions and crystalline samples were irradiated is consistent with the rapid non-radiative decay of purines and the absence of significant photochemistry previously reported for these molecules (37–40). We conclude that purine–purine cross-links observed in irradiated 16S RNAs are attributable to a particular geometry between the bases. This is consistent with previous work done on DNA (41–44).
The distances between the bases measured from the atomic structure (34) using the positions between N7/C8 atoms of purines and the positions between the C5/C6 atoms of the pyrimidine are: G976–A1362, 9.44 ?; G1064–G1190, 10.39 ?; G926–U1390, 19.08 ?; and C1052–A1196, 6.6 ?. The first two of these distances are somewhat outside the distance range observed for the majority of cross-links made by low intensity irradiation, which is 3.7–8.0 ? (Table 1). However, there are four other cross-links made by low intensity cross-linking that are separated by >10 ?, so the separation distance is not unique to these particular cross-links. In addition, there are many hundreds of pairs of nucleotides that have atoms that could participate in a UV cross-link within 9 ?, but are not reactive (B.Huggins, S.Ghosh, K.Nanda and P.Wollenzien, unpublished results). Therefore, the appearance of the four new cross-links is still very selective. The G926–U1390 cross-link, if it is confirmed, is well outside of the expected distance for cross-linking and would require the repositioning of the nucleotide G926. This nucleotide in the atomic structure is unstacked from the helix in which it is located as an unpaired nucleotide (34), but it could be repositioned by intercalation into the interior of the helix in a stacking arrangement under some circumstances and would be close to U1390 (45).
Cross-linking frequencies obtained with the UV pulse laser for each cross-link are usually larger than those obtained with a transilluminator. The correlation coefficients indicate a strong correlation between frequencies obtained with one pulse and frequencies obtained in the flow cell, but these are not related to the frequencies obtained in the transilluminator. An important result is that in spite of the frequency differences, the same pattern of cross-links is made, except for the four additional ones made with UV pulse laser irradiation. If cross-linking frequencies were determined by de-excitation processes such as non-radiative decay or energy transfer, which change the S1 or T1 lifetimes, the cross-linking frequencies would not depend on the irradiation procedure in the way that they do. Furthermore, tertiary structure rather than photochemistry is more important in determining the pattern of cross-linking (33). Therefore, the underlying factors that may be at play are: (i) the RNA tertiary structure that somehow determines the cross-linking pattern and is a factor in determining the cross-linking frequencies; and (ii) photoreversal reactions, that affect only the cross-linking frequencies. There is an inverse correlation between the internucleotide distances and the cross-linking frequencies after one laser pulse, but this is not seen with cross-linking frequencies after multiple pulses or after the extended time of irradiation in the transilluminator. This is proabably due to less photoreversal in the experiments with the single laser pulse.
These experiments document the elevated cross-link yield per photon upon irradiation with the UV pulse laser. This may be due to enhanced photochemistry from two photon reactions, but it is more likely that it is due simply to a reduced amount of photoreversal. On the other hand, the presence of new products not seen with low intensity irradiation is evidence for two photon reactions at those sites. The yield of cross-links per photon at the lowest intensity that was studied here makes pulse laser irradiation suitable for practical experiments because strand breakage and photoreversal are limited at that intensity. Moreover, the cross-linking frequencies obtained with a single UV pulse should better reflect the details of the geometry of the cross-linking partners, so methods that use single pulses will be better able to detect conformational differences than methods that use low intensity irradiation for extended times.
ACKNOWLEDGEMENTS
We thank Professor E.I. Bodowsky for discussions of the UV pulse laser irradiation technique, Professor Angus Kingon, Department of Materials Science and Engineering, NCSU for the use of the KrFl excimer laser, and Bruce Huggins for providing nucleotide distance measurements. This work is supported by NIH grant GM43237 to P.W., and by NSF grant MCB-9874895 to S.F.
REFERENCES
Fenick,D.J., Carr,H.S. and Falvey,D.E. (1995) Synthesis and photochemical cleavage of cis-syn pyrimidine cyclobutane dimer analogs. J. Org. Chem., 60, 624–631.
Ruzsicska,B. and Lemaire,D. (1995) DNA photochemistry. In Horspool,W.M. and Song,P.S. (eds), CRC Handbook of Organic Photochemistry and Photobiology. CRC Press, Boca Raton, FL, pp. 1289–1317.
Zhirnov,O.V. and Wollenzien,P. (2003) Action spectra for UV-light induced RNA–RNA crosslinking in 16S ribosomal RNA in the ribosome. Photochem. Photobiol. Sci., 2, 688–693.
Budowsky,E.I. and Abdurashidova,G.G. (1989) Polynucleotide–protein cross-links induced by ultraviolet light and their use for structural investigation of nucleoproteins. Prog. Nucleic Acid Res. Mol. Biol., 37, 1–65.
Daniels,M. (1976) Excited states of the nucleic acids: bases, mononucleosides and mononucleotides. In Wang,S.Y. (ed.), Photochemistry and Photobiology of Nucleic Acids. Academic Press, New York, pp. 23–109.
Fisher,G. and Johns,H. (1976) Pyrimidine photodimers. In Wang,S.Y. (ed.), Photochemistry and Photobiology of Nucleic Acids. Academic Press, New York, NY, pp. 225–294.
Cadet,J. and Vigny,P. (1990) The photochemistry of nucleic acids. In Morrison,H. (ed.), Bioorganic Photochemistry: Photochemistry and the Nucleic Acids. John Wiley and Sons, New York, NY, Vol. 1, pp. 1–272.
Lemaire,D.G.E. and Ruszicska,B.P. (1993) Quantum yields and secondary photoreactions of the photoproducts of dTpdT, dTpdC and dTpdU. Photochem. Photobiol., 57, 755–769.
Kryukov,P., Letokhov,V., Nikogosyan,D., Borodavkin,A., Budowsky,E.I. and Simukova,N. (1979) Multiquantum photoreactions of nucleic acid components in aqueous solution by powerful ultraviolet picosecond radiation. Chem. Phys. Lett., 61, 375–379.
Budowsky,E.I., Azentyeva,M., Abdurashidova,G., Simukova,N. and Rubin,L. (1986) Induction of polynucleotide–protein cross-linkages by ultraviolet irradiation. Pecularities of the high-intensity laser pulse irradiation. Eur. J. Biochem., 159, 95–101.
Harrison,C., Turner,D. and Hinkle,D. (1982) Laser crosslinking of E.coli RNA polymerase and T7 DNA. Nucleic Acids Res., 10, 2399–2414.
Hockensmith,J., Kubasek,W., Vorachek,W. and von Hippel,P. (1986) Laser cross-linking of nucleic acids to proteins. J. Biol. Chem., 8, 3512–3518.
Angelov,D., Stefanovsky,V., Dimitrov,S., Russanova,V., Keskiniva,E. and Pashev,I. (1988) Protein–rDNA crosslinking in reconstituted nucleohistone complexes, nuclei and whole cells by picosecond UV laser irradiation. Nucleic Acids Res., 16, 4525–4538.
Dobrov,E., Arbieva,Z., Timofeeva,E., Esenaliev,R., Oraevsky,A. and Nikogosyan,D. (1989) UV laser induced RNA–protein crosslinks and RNA chain breaks in tobacco mosaic virus RNA in situ. Photochem. Photobiol., 49, 595–598.
Ho,D., Sauve,D. and Roberge,M. (1994) Detection and isolation of DNA-binding proteins using single-pulse ultraviolet laser crosslinking. Anal. Biochem., 218, 248–254.
Connor,D., Falick,A., Yuoung,M. and Shetlar,M. (1998) Probing the binding region of the single-stranded DNA-binding domain of rat DNA polymerase beta using nanosecond-pulse laser-induced crosslinking and mass spectrometry. Photochem. Photobiol., 68, 299–308.
Moss,T., Dimtrov,S. and Houde,D. (1997) UV-laser crosslinking of proteins to DNA. Methods, 11, 225–234.
Nikogosyan,D. and Gurzadyan,G. (1984) Two quantum photoprocesses in DNA and RNA biopolymers under powerful picosecond laser UV irradiation. Laser Chem., 4, 297–303.
Johnson-Thompson,M., Halpern,J., Jackson,W. and Georges,J. (1984) Vacuum UV laser induced scission of simian virus 40 DNA. Photochem. Photobiol., 39, 17–24.
Schulte-Frohlinde,D., Opitz,J., Gorner,H. and Bothe,E. (1985) Model studies for the direct effect of high energy irradiation on DNA. Mechanism of strand break formation induced by laser photoionization of poly(U) in aqueous solutions. Int. J. Radiat. Biol., 48, 397–408.
Menshonkova,T., Simukova,N., Budowsky,E.I. and Rubin,L. (1980) The effect of high intensity ultraviolet irradiation on nucleic acids and their components. FEBS Lett., 112, 299–301.
Opitz,J. and Schulte-Frohlinde,D. (1987) Laser-induced photoionization and single-strand break formation for polynucleotides and single-stranded DNA in aqueous solution: model studies for the direct effect of high energy irradiation on DNA. J. Photochem., 39, 145–163.
Blau,W., Croke,D., Kelly,J., McConnell,O.C. and van der Putten,W. (1987) Base-specific photocleavage of DNA induced by nanosecond UV pulse laser ragiation or methylene blue sensitization. J. Chem. Soc. Chem. Commun., 10, 751–752.
Croke,D., Blau,W., OhUigin,C., Kelly,J. and McConnell,D. (1988) Photolysis of phosphodiester bonds in plasmid DNA by high intensity UV laser irradiation. Photochem. Photobiol., 47, 527–536.
Rubin,L., Menshonkava,T., Simukova,N. and Budowsky,E.I. (1981) The effects of high-intensity UV-radiation on nucleic acids and their components—I. Thymine. Photochem. Photobiol., 34, 339–344.
Kovalsky,O., Panyutin,I. and Budowsky,E. (1990) Sequence-specificity of the alkali-sensitive lesions induced in DNA by high-intensity ultraviolet laser radiation. Photochem. Photobiol., 52, 509–517.
Makhno,V.I., Peshin,N.N., Semenkov,Y.P. and Kirillov,S.V. (1988) Modified method of producing ‘tight’ 70S ribosomes from Escherichia coli, highly active in individual stages of the elongation cycle. Mol. Biol., 22, 28–537.
Juzumiene,D., Shapkina,T., Kirillov,S. and Wollenzien,P. (2001) Short-range RNA–RNA crosslinking methods to determine rRNA structure and interactions. Methods, 25, 333–343.
Wilms,C. and Wollenzien,P. (1994) Purification of RNA from poly acrylamide gels by ultracentrifugation. Anal. Biochem., 221, 204–205.
Wilms,C., Noah,J.W., Zhong,D. and Wollenzien,P. (1997) Exact determination of UV-induced crosslinks in 16S ribosomal RNA in 30S ribosomal subunits. RNA, 3, 602–612.
Ericson,G. and Wollenzien,P. (1988) Use of reverse transcription to determine the exact locations of psoralen photochemical cross-links in RNA. Anal. Biochem., 174, 215–223.
Ericson,B., Chevli,K. and Wollenzien,P. (1989) Structure of synthetic unmethylated 16S ribosomal RNA as purified RNA and in reconstituted 30S ribosomal subunits. Biochemistry, 28, 6446–6454.
Noah,J.W. Shapkina,T. and Wollenzien,P. (2000) UV-induced crosslinks in the 16S rRNAs of Escherichia coli, Bacillus subtilis and Thermus aquaticus and their implications for ribosome structure and photochemistry. Nucleic Acids Res., 28, 3785–3792.
Wimberly,B.T., Brodersen,D.E., Clemons,W.M., Morgan-Warren,R.J., Carter,A.P., Vonrhein,C., Hartsch,T. and Ramakrishnan,V. (2000) Structure of the 30S ribosomal subunit. Nature, 407, 327–339.
Crespo-Hernandez,C.E. and Arce,R. (2002) Photoionization of DNA and RNA bases, nucleosides and nucleotides through a combination of one- and two-photon pathways upon 266 nm nanosecond excitation. Photochem. Photobiol., 76, 259–267.
Saito,I. and Sugiyama,H. (1990) Photoreactions of nucleic acids and their constituents with amino acids and related compounds. In Morrison,H. (ed.), Bioorganic Photochemistry. John Wiley and Sons, New York, NY, Vol. 1, pp. 317–340.
Bose,S.N., Davies,R.J.H., Anderson,D.W., Vanniekerk,J.C., Nassimbeni,L.R. and Macfarlane,R.D. (1978) Photo-chemical (8-]8) coupling of purine nucleosides to guanosine. Nature, 271, 783–784.
Gorner,H. (1994) Photochemistry of DNA and related biomolecules: quantum yields and consequences of photoionization. J. Photochem. Photobiol., 26, 117–139.
Ivanchenko,V.A., Budowsky,E.I., Simukova,N.A., Vul’fson,N.S., Tishchenko,A.I. and Askerov,D.B. (1977) The photochemistry of purine components of nucleic acids. 2 Photolysis of deoxyguanosine. Nucleic Acids Res., 4, 955–968.
Ivanchenko,V.A., Titschenko,A.I., Budowsky,E.I., Simukova,N.A. and Wulfson,N.S. (1975) The photochemistry of purine components of nucleic acids. 1 The efficiency of photolysis of adenine and guanine derivatives in aqueous solution. Nucleic Acids Res., 2, 1365–1373.
Bose,S.N., Davies,R.J.H., Sethi,S.K. and McCloskey,J.A. (1983) Formation of an adenine–thymine photoadduct in the deoxydinucleoside monophosphate d(TpA) and in DNA. Science, 220, 723–725.
Bose,S.N. and Davies,R.J.H. (1984) The photoreactivity of T–A sequences in oligodeoxyribonucleotides and DNA. Nucleic Acids Res., 12, 7903–7914.
Melvin,T., Plumb,M.A., Botchway,S.W., O’Neil,P. and Parker,A.W. (1995) 193 nm light induces single strand breakage of DNA predominantly at guanine. Photochem. Photobiol., 61, 584–591.
Davies,R.J.H. (1995) Ultraviolet-radiation damage in DNA. Biochem. Soc. Trans., 23, 407–418.
Ericson,G., Minchew,P. and Wollenzien,P. (1995) Structural changes in base-paired region 28 in 16S rRNA close to the decoding region of the 30S ribosomal subunit are correlated to changes in tRNA binding. J. Mol. Biol., 250, 407–419.
Cannone,J., Subramanian,S., Schnare,M., Collett,J., S’Souza,L., Du,Y., Feng,B., Lin,N., Madabusi,L., Muller,K., Pande,N., Shang,Z. and Gutell,R. (2002) The comparative RNA Web (CRW) site: an online database of comparative sequence and structure information for ribosomal, intron and other RNAs. BioMed. Central Bioinform., 3, 2.(Tatjana Shapkina, Simon Lappi1, Stefan F)