Volume 3 no-2 February 1 976
Nucleic Acids Research
Reaction of HeLa cell methyl-labelled 28S ribosomal RNA with sodium bisulphite: a conformatlonal probe for methylated sequences.
J. P. Goddard and B. E. H. Maden
Department of Biochemistry, lniversity of Glasgow, Glasgow G12 8QQ, UK. Received 15 December 1975
ABSTRACT The reaction of 14C methyl-labelled HeLa cell 28 S ribosomal RNA with sodium bisulphite was studied. Using conditions under which 30% of the total cytidine residues were de-aminated to uridine, the reactivities of individual cytidine residues near particular methylation sites differed widely; some underwent almost quantitative reaction, some showed intermediate reactivity and others were almost inert. The possible value of this method as a conformational probe for ribosomal RNA is discussed.
INTRODUCTION Investigation of the chemical reactivity of nucleotides in RNA has proved to be an effective probe of conformation, as shown convincingly for yeast tRNAPhe(l,2). One such method involves the use of sodium bisulphite, which reacts with non-hydrogen-bonded pyrimidine residues Of the resulting adducts, 5-6 dihydrocytidine-6(3,4). sulphonate is slowly hydrolyzed to the uridine analogue. This may be converted, by adduct removal at slightly The reaction has been used to elevated pH, to uridine. identify non-hydrogen-bonded cytidine residues in tRNA (5). Such studies have been confined mainly to small molecules of known sequence, since analysis to define reactive sites in large molecules is generally much more Recently, however, the reactivity of the polycomplex. cytidylate tract in encephalomyocarditis virus RNA has been studied (6). Eukaryotic ribosomal RNA contains numerous methylated The sequences surrounding most of these have nucleotides. 0 Information Retrieval Limited 1 Falconberg Court London W1 V 5FG England
431
Nucleic Acids Research been determined in HeLa rRNA (7). Methylation occurs principally at the level of ribosomal precursor RNA (7-9), and is necessary for ribosome maturation (10). The chemical basis of enzymic "recognition" of particular nucleotides for methylation is not known, nor is the precise biological role of this process. Insight into these questions might be gained from knowledge of polynucleotide conformation(s) at the methylation sites. Here we explore the potential of sodium bisulphite modification as a conformational probe for several of the methylation sites in HeLa cell 28 S rRNA. METHODS HeLa cell rRNA, 32p labelled or 14C methyl-labelled, was prepared as described previously (7). 28 S RNA (20 - 50 4g) was modified at 250 in lml 3M bisulphite, pH 6.0, for 24 hrs, after which bisulphite was removed by two dialyses against 0.15M NaCl, 0.01M Tris-HCl, pH 7.5. Adducts were removed by dialyses against 0.1M Tris-HCl (pH 9.0 at 370) for 9 hrs. at 37e. The RNA solution was neutralized by dialysis against 0.05M Tris-HCl, pH 7.0. Salt was removed by exhaustive dialysis against distilled water. The RNA was then lyophilized. Digestion with ribonuclease T1, fingerprinting, numbering of (unreacted) spots and characterization of methyl-labelled alkaline hydrolysis products by electrophoretic mobility at pH 3.5 were carried out as previously described (7). RESULTS Reaction conditions The reaction conditions were chosen on the basis of experience with tRNA (5) and the results of kinetic The latter showed experiments with 32p labelled 28 S RNA. that after 22 hours of reaction, followed by adduct removal, some 30% of the cytidine had been deaminated to uridine (Table 1). Adenosine and guanosine were not affected by the reaction. After 70 hours some 50% of the cytidine had been deaminated, and time points up to 100 hours (not shown) indicated that a plateau had been reached, beyond which little or no further reaction was occurring under these 432
Nucleic Acids Research Table 1
Ap Gp Cp Up+Vp % conversion Unreacted (ref.17) 15.8 35.1 32.1 17.0 Unreacted 16.1 35.5 31.6 16.8 2.5 hr 16.0 35.8 29.6 18.6 6.3 22 hr 16.1 35.9 21.6 26.4 31.6 70 hr 16.3 35.0 15.7 33.0 50.3 Change in nucleotide composition of 28 S RNA after reaction with 3M bisulphite pH 6.0. RNA, unreacted or reacted for the periods indicated, was subjected to alkaline hydrolysis, and the products were separated by electrophoresis on Whatman 52 paper at pH 3.5. Radioactivity in the four major bands was determined, that in the minor alkali-stable bands being ignored. The unreacted values agree well with those published (ref. 17; values shown for unlabelled RNA; and ref. 18). Our unreacted value for Cp (31.6%) was used for computing the percentage of Cp -Up conversion in the reacted samples. The non-reactivity of Ap and Gp is evident. These results were selected for illustrative purposes from a more detailed kinetic analysis (J.P. Goddard,
unpublished).
conditions. Thus somewhat more than half of the potentially reactive cytidines reacted within 24 hours. Their average rate of reaction was much less than that found for poly C (5), in which half the cytidine reacted within 5 hrs. This suggested a spectrum of reaction rates for individual cytidine residues in 28 S RNA. As will be shown, this appeared to be the case for the sequences studied. 14C methyl-labelled 28 S RNA Figure 1 shows T1 ribonuclease fingerprints of 14C methyl-labelled 28 S RNA, untreated and after reaction with bisulphite. The reacted sample shows relative decreases in intensity of many spots, and also many new spots. The diminished spots represent oligonucleotides with reactive The new spots cQntain uridines derived by cytidines. reaction of bisulphite with these cytidines. Products in the T1 fingerprinting system segregate into graticules, products in the first graticule containing no uridines, products in the second graticule one uridine, (A second guanosine, due to 2'-0and so on (11). 433
Nucleic Acids Research .::
:.:!;j...jj. '.. M %;..
.:j
k ;:! :.
;i.%.
:'
V
: :.
.A.-
.41
:
-
*
C.0°
0
°
pH 3.5
O 80t 00,
o
0
03036 0
300
31
2 0t26
,7 b
0 26
23
-0
/ /29~*Q >32 15b 18
028 027 25
%40o
00
b-b220 :°23/
/
22
3 -C
21//
LI
E1.
0
1%
/
/
21
0013 1lo 012 :10010
/
/
90
70
Unreected
Key
: 7 ,@
30
(2e
v'
90 '
/
04
~ ~ ~14 013
/
06
30
160
04 622
Bisulphite
Key
Figure 1. T1 fingerprints of HeLa 14C methyl-labelled 28 S RNA, control (top left) and bisulphite reacted. In Keys, interrupted lines demarcate graticules. Bisulphite key shows new spots (black) in second graticule, and diminished, cytidine-containing spots with dotted circles. Spot 18 (low yield) is unidentified and spots 24' 25, 29 and 32 contain Above second no cytidine and are not discussed (75. graticule reacted sample yields many new products, not distinguished in key. 434
Nucleic Acids Research methylation, confers a roughly similar effect on mobility Therefore in the first graticule the as a uridine). bisulphite reaction can only cause loss of material from In the second graticule not cytidine-containing products. only are some normal cytidine-containing spots diminished, but new spots appear by cytidine -uridine conversions within sequences which derived originally from the first The two or more uridines"graticules are poorly graticule. resolved here, but may be resolved by other methods (7, 11, The effects of bisulphite on sequences in these 12). graticules become increasingly complex due to reaction of many normal products and also acquisition of new spots by Consequently, modification from both previous graticules. here we shall confine attention to effects within the first two graticules, in which a representative sample of possible effects was found. Reference products Three products containing neither cytidine nor uridine were used as reference products for quantitating the degree of reaction of other, cytidine-containing products in the These products were spots 3 bisulphite-treated samples. The molar (Am-G), 9 (A-Am-G) and 21 (Gm-G) (Table 2). yields of these products in control 28 S RNA were determined previously by a two stage analysis: - (i) Using the combined T plus pancreatic ribonuclease fingerprinting system, with P labelled RNA, it was shown that many methylated sequences occur once per molecule (15, 14). (ii) The individual yields of these three and all other Ti products in 14C methyl fingerprints were then computed from an assumed mean of one for twenty-two well resolved sequences (all those with approximately unimolar values and On the one methyl group in the first two graticules (7)). basis of these assumptions, the yields of spots 3, 9 and 21 were non-integral, possibly because, being short products and hence very stable towards over-digestion, their recovery was slightly better than that of the others. In computing the yields of other products in the bisulphite reacted RNA it was rational to attribute an equal value
435
Nucleic Acids Research Table 2 Molar yield (a) Untreated
Reference Products:3 (G)-Am-G 9 (G )-A-Am-G (G )-Gm-G 21 Sum of 3, 9, 21:-
Single 2'-0-methyl C:2 (G)-Cm-G 7 (G) -A-Cm-G (G -Cm-U-G 23 (G)-Cm-A-U-G 27
Single non-methyl C:6 (G -C-Am-G 11 (G -A-Am-C-G 15 (G)-A-A-C-Am-G 26 (G)-Um-A-C-G 28 ()- (mA-A-C,U)-G 32 ()- (C,U)-A-Am-A-G 35 () -A-A-C-A-A-Um-G
Reacted
4.51 1.17 7.06
4.87 1.18 6.69
12.74
12.74
1.08
1.01 0.52
1.16 0.85
0.63
0.99
0.59
1.17
0.28 0.39 0.05 0.03 0.09 0.60 0.13
1.08 0.82 0.96 0.95 1.01 0.91
Reactivity
% conversion(c)
6 55 26 40
76
64 94 97 91 41 86
± ++ + ++ +++
++ ++++ ++++ ++++
++ +++
Two C residues:-
4
G)-C-Cm-G
10 (G)-A-Cm-C-G 13 (G) -Cm-C-A-A-G 16 (G)-C-C-Gm-G
Others:2a (T)-mC-G 12 (G)-C-C-Cm-A-G 14 ...mA+Am-C... 33 .. mC+C?+A-A-Am-U...
0.94 1.04 0.94 0.80
0.69 0.63 0.80
0.66
73 39 15 17
+++
++ + +
0.2 0 0.24 68 1.06 0.34 +++ 96 0.78(b) 0.03 ++++ 1.10 b) 0.78 + 29 92 ++++ 36 (G)-A-C3_5-Gm-A-Am-A-G 0.91 b) 0.07 Reactivities of methylated, cytidine-containing sequences in 28 S RNA. Footnotes:- (a) see text for determination of molar yields. (b) Yields corrected for double methylation. Numbers of cytidines unknown. (c) ± =0-10% conversion; + 11-35%; ++ = 36-65%; +++ = 66-89%; ++++ > 90%. The percentage conversion was determined from the degree of diminution of the spots with respect to the sum of the three reference spots, as determined by scintillation counting. (In different preparations there were 250-500 c.p.m. per methyl group per unreacted unimolar spot). The values shown are the means of quadruplicate determinations agreeing within 5%.
436
Nucleic Acids Research to the sum of the molar yields of these three (unreactive) products as for the same products in the unreacted RNA. (Two of these products occur in high yield, thereby minimizing radioactive counting errors). Assuming an equal sum of molar yields for these three products in control and reacted RNA, agreement between the individual yields was reasonably good (Table 2), providing an indication of the reproducibility of the methods. Reactivities of methylated sequences. The remaining sections of Table 2 show the reactivities of the cytidine-containing sequences. These are arranged in groups:- those containing a single 2'-0-methyl C, those with an unmethylated C near to another 2'-0-methylated nucleotide, and so on. Each product occurs approximately once per molecule in fingerprints of untreated 28 S RNA (7), and therefore each represents a unique region of the 28 S molecule. The sequences encompass a wide range of reactivities. In addition to the quantitative values the results are interpreted by numbers of pluses in the right hand column of the table (see legend). Aspects of the results are discussed below. Products of reaction. Where the original sequence contained one cytidine and no uridine, identification of the reaction product For example was straightforward; it contained one uridine. the position of the new spot, 6b, which was present in about half molar yield in modified 28 S RNA, suggested the Alkaline hydrolysis yielded a composition (A,U)Gp. dinucleotide of % (pH 3.5) corresponding to Gm-Ap or Collectively these findings indicated that Am-Gp (7). the spot is U-Am-Gp, arising from reaction of C-Am-Gp (spot 6). Three further examples of reactions involving single C--4U changes, one of them (spot 7) involving a For a sequence 2'-0-methyl C, are given in Table 3. with two or more cytidines the reaction product(s) may contain one or two "new" uridines, depending on whether one Product identification should or both cytidines reacted. distinguish between the various possible reactions, but so
437
Nucleic Acids Research Table 3 New Spot Sequence
6 b (G)-U-Am-G 7 b (G)-A-Um-G 11 b (G)-A-Am-U-G 15 b (G)-A-A-U-Am-G
Derived from
6 (G)-C-Am-G 7 (G)-A-Cm-G 11 (G)-A-Am-C-G 15 (G)-A-A-C-Am-G
Origin of selected new spots in T1 fingerprints of methyl labelled 28 S RNA after reaction with bisulphite. The products were identified from their mobilities in the two dimensional system (indicating their base compositions), the electrophoretic mobility at pH 3.5 of the respective alkali stable dinucleotides (7) and knowledge of the original (unreacted) sequences. There was approximate agreement between the molar yields of the new spots and the extent of reaction of the original products, though quantitation was slightly unreliable for spots 6b, 7b and llb due to the close proximity of other unreacted spots, 24' (A-Gm-G), 15b was well 25 (Am-U-G) and 29 (A-A-Um-G) respectively. resolved and was recovered in 0.8 molar yield, in good agreement with the almost complete disappearance of spot 15 (Table 2).
far this has not been carried out in these more complex situations. DISCUSSION Of the approximately seventy methyl groups in HeLa cell 28 S RNA (7), the reactivities of sequences encompassing twenty-three have been examined here, another (approximately) thirteen methyl groups providing reference material. No obvious general rules emerge from the results obtained It might, perhaps, have been speculated from this sample. that methylation sites, in order to be recognized by the methylase(s), would occupy exposed conformations. In fact some of the sequences are rather unreactive, notably spot 2, Ribose methylation is derived from the sequence G-Cm-G. a very early event in rRNA maturation, probably occurring It is possible that after on nascent 45 S RNA (8,7). ribose methylation has occurred further secondary or tertiary interactions intervene during 45 S RNA chain completion or rRNA maturation, thus rendering inaccessible at least some of the sites which were accessible at the time of methylation.
438
Nucleic Acids Research Nevertheless, it seems likely that information on the reactivities of particular nucleotides will become useful in several ways, in conjunction with more extensive nucleotide It is reasonable to sequence data or comparative data. infer that the most reactive sequences, spots 14, 15, 26, 28 and 36 (Table 2) contain one or more cytidines in exposed regions, within the 28 S tertiary structure. In such cases the C residue may either be unimportant for function, or it may be highly important in making contact with (for example) a ribosomal protein or another Product 26, Um-A-C-G, component of protein synthesis. a reactive contains C. Interestingly, this product very is absent in the Xenopus 28 S fingerprint, in which the related oligonucleotide, Um-A-U-G occurs instead (14). A C--AU base change is a likely explanation, suggesting that this particular (reactive) C does not play a crucial role in ribosome function (or, for that matter, in recognition of By contrast spot 36, in this sequence by the methylase). which one or more C residues are reactive, seems to be a constant vertebrate sequence (unpublished fingerprints of It is possible that in this case the Khan and Maden). In fact the doubly cytidines play an important role. methylated component, Gm-A-Am-A-Gp, has been reported in the 28 S ribosomal RNAs of such distantly related organisms as Drosophila (15) and yeast (16) and therefore seems to be a highly conserved eukaryotic sequence. There appears to be considerable scope for usefully extending the present type of analysis, for example by more detailed kinetic analysis of the reactivities of individual sequences, by analysis of reactivities of further unique sequences within 28 S and 18 S RNA, and of the same sequences within the various ribosomal precursor RNAs.
ACKNOWLEDGEMENT This work was supported by a grant from the Medical We thank T. Carr for skilled technical Research Council. assistance.
439
Nucleic Acids Research REFERENCES 1. Robertus,J.D., Ladner,J.E., Finch,J.T., Rhodes,D., Brown,R.S., Clark,B.F.C., and Klug,A. (1974). Nucleic Acid Research 1 927-932. 2. Rhodes,D. (1975). J. Mol. Biol. 94, 449-460. 3. Shapiro,R., Servis,R.E. and Welcher,M. (1970). J. Amer. Chem. Soc. 92, 422-424. 4. Hayatso, H., Wataya,Y., Kai,K. and Iida, S. (1970). Biochemistry 9, 2858-2864. 5. Goddard, J.P. and Schulman,L.H. (1972). J. Biol. Chem. 247, 3864-3867. 6. Goodchild,J., Fellner,P. and Porter,A.G. (1975). Nucleic Acids Research 2, 887-895. 7. Maden,B.E.H. and Salim,M. (1974). J. Mol. Biol. 88, 133-164. 8. Greenberg,M. and Penman,S. (1966). J. Mol. Biol. 21, 527-535. 9. Zimmerman,E.F. (1968). Biochemistry 7, 3156-3164. 10. Vaughan,M.H., Soeiro,R., Warner,J.R. and Darnell,J.E. (1967). Proc. Nat. Acad. Sci., U.S.A. 58, 1527-1534. 11. Brownlee,G.G. and Sanger,F. (1967). J. Mol.Biol. 23, 337-353. Brownlee,G.G. and Sanger, F. (1969). Eur. J. Biochem. 12. 11, 395-399. 13. Maden,B.E.H., Lees,C.D. and Salim,M. (1972). FEBS Letters 28, 293-296. Khan,M.S.N. and Mladen,B.E.H. (1976). J. Mol. Biol., 14. in press. 15. Maden,B.E.H. and Tartof,K. (1974). J. Mol. Biol. 90, 51-64. 16. Klootwijk,J. and Planta,R.J. (1974). Mol.Biol. Reports, 1, 187-191. 17. Amaldi,F. and Attardi,G. (1968). J. Mol. Biol. 21, 29-41. 18. Willems,M., Wagner,E., Laing,R. and Penman,S. (1968). J. Mol. Biol. 32, 211-220.
440
Nucleic Acids Research
Reaction of HeLa cell methyl-labelled 28S ribosomal RNA with sodium bisulphite: a conformatlonal probe for methylated sequences.
J. P. Goddard and B. E. H. Maden
Department of Biochemistry, lniversity of Glasgow, Glasgow G12 8QQ, UK. Received 15 December 1975
ABSTRACT The reaction of 14C methyl-labelled HeLa cell 28 S ribosomal RNA with sodium bisulphite was studied. Using conditions under which 30% of the total cytidine residues were de-aminated to uridine, the reactivities of individual cytidine residues near particular methylation sites differed widely; some underwent almost quantitative reaction, some showed intermediate reactivity and others were almost inert. The possible value of this method as a conformational probe for ribosomal RNA is discussed.
INTRODUCTION Investigation of the chemical reactivity of nucleotides in RNA has proved to be an effective probe of conformation, as shown convincingly for yeast tRNAPhe(l,2). One such method involves the use of sodium bisulphite, which reacts with non-hydrogen-bonded pyrimidine residues Of the resulting adducts, 5-6 dihydrocytidine-6(3,4). sulphonate is slowly hydrolyzed to the uridine analogue. This may be converted, by adduct removal at slightly The reaction has been used to elevated pH, to uridine. identify non-hydrogen-bonded cytidine residues in tRNA (5). Such studies have been confined mainly to small molecules of known sequence, since analysis to define reactive sites in large molecules is generally much more Recently, however, the reactivity of the polycomplex. cytidylate tract in encephalomyocarditis virus RNA has been studied (6). Eukaryotic ribosomal RNA contains numerous methylated The sequences surrounding most of these have nucleotides. 0 Information Retrieval Limited 1 Falconberg Court London W1 V 5FG England
431
Nucleic Acids Research been determined in HeLa rRNA (7). Methylation occurs principally at the level of ribosomal precursor RNA (7-9), and is necessary for ribosome maturation (10). The chemical basis of enzymic "recognition" of particular nucleotides for methylation is not known, nor is the precise biological role of this process. Insight into these questions might be gained from knowledge of polynucleotide conformation(s) at the methylation sites. Here we explore the potential of sodium bisulphite modification as a conformational probe for several of the methylation sites in HeLa cell 28 S rRNA. METHODS HeLa cell rRNA, 32p labelled or 14C methyl-labelled, was prepared as described previously (7). 28 S RNA (20 - 50 4g) was modified at 250 in lml 3M bisulphite, pH 6.0, for 24 hrs, after which bisulphite was removed by two dialyses against 0.15M NaCl, 0.01M Tris-HCl, pH 7.5. Adducts were removed by dialyses against 0.1M Tris-HCl (pH 9.0 at 370) for 9 hrs. at 37e. The RNA solution was neutralized by dialysis against 0.05M Tris-HCl, pH 7.0. Salt was removed by exhaustive dialysis against distilled water. The RNA was then lyophilized. Digestion with ribonuclease T1, fingerprinting, numbering of (unreacted) spots and characterization of methyl-labelled alkaline hydrolysis products by electrophoretic mobility at pH 3.5 were carried out as previously described (7). RESULTS Reaction conditions The reaction conditions were chosen on the basis of experience with tRNA (5) and the results of kinetic The latter showed experiments with 32p labelled 28 S RNA. that after 22 hours of reaction, followed by adduct removal, some 30% of the cytidine had been deaminated to uridine (Table 1). Adenosine and guanosine were not affected by the reaction. After 70 hours some 50% of the cytidine had been deaminated, and time points up to 100 hours (not shown) indicated that a plateau had been reached, beyond which little or no further reaction was occurring under these 432
Nucleic Acids Research Table 1
Ap Gp Cp Up+Vp % conversion Unreacted (ref.17) 15.8 35.1 32.1 17.0 Unreacted 16.1 35.5 31.6 16.8 2.5 hr 16.0 35.8 29.6 18.6 6.3 22 hr 16.1 35.9 21.6 26.4 31.6 70 hr 16.3 35.0 15.7 33.0 50.3 Change in nucleotide composition of 28 S RNA after reaction with 3M bisulphite pH 6.0. RNA, unreacted or reacted for the periods indicated, was subjected to alkaline hydrolysis, and the products were separated by electrophoresis on Whatman 52 paper at pH 3.5. Radioactivity in the four major bands was determined, that in the minor alkali-stable bands being ignored. The unreacted values agree well with those published (ref. 17; values shown for unlabelled RNA; and ref. 18). Our unreacted value for Cp (31.6%) was used for computing the percentage of Cp -Up conversion in the reacted samples. The non-reactivity of Ap and Gp is evident. These results were selected for illustrative purposes from a more detailed kinetic analysis (J.P. Goddard,
unpublished).
conditions. Thus somewhat more than half of the potentially reactive cytidines reacted within 24 hours. Their average rate of reaction was much less than that found for poly C (5), in which half the cytidine reacted within 5 hrs. This suggested a spectrum of reaction rates for individual cytidine residues in 28 S RNA. As will be shown, this appeared to be the case for the sequences studied. 14C methyl-labelled 28 S RNA Figure 1 shows T1 ribonuclease fingerprints of 14C methyl-labelled 28 S RNA, untreated and after reaction with bisulphite. The reacted sample shows relative decreases in intensity of many spots, and also many new spots. The diminished spots represent oligonucleotides with reactive The new spots cQntain uridines derived by cytidines. reaction of bisulphite with these cytidines. Products in the T1 fingerprinting system segregate into graticules, products in the first graticule containing no uridines, products in the second graticule one uridine, (A second guanosine, due to 2'-0and so on (11). 433
Nucleic Acids Research .::
:.:!;j...jj. '.. M %;..
.:j
k ;:! :.
;i.%.
:'
V
: :.
.A.-
.41
:
-
*
C.0°
0
°
pH 3.5
O 80t 00,
o
0
03036 0
300
31
2 0t26
,7 b
0 26
23
-0
/ /29~*Q >32 15b 18
028 027 25
%40o
00
b-b220 :°23/
/
22
3 -C
21//
LI
E1.
0
1%
/
/
21
0013 1lo 012 :10010
/
/
90
70
Unreected
Key
: 7 ,@
30
(2e
v'
90 '
/
04
~ ~ ~14 013
/
06
30
160
04 622
Bisulphite
Key
Figure 1. T1 fingerprints of HeLa 14C methyl-labelled 28 S RNA, control (top left) and bisulphite reacted. In Keys, interrupted lines demarcate graticules. Bisulphite key shows new spots (black) in second graticule, and diminished, cytidine-containing spots with dotted circles. Spot 18 (low yield) is unidentified and spots 24' 25, 29 and 32 contain Above second no cytidine and are not discussed (75. graticule reacted sample yields many new products, not distinguished in key. 434
Nucleic Acids Research methylation, confers a roughly similar effect on mobility Therefore in the first graticule the as a uridine). bisulphite reaction can only cause loss of material from In the second graticule not cytidine-containing products. only are some normal cytidine-containing spots diminished, but new spots appear by cytidine -uridine conversions within sequences which derived originally from the first The two or more uridines"graticules are poorly graticule. resolved here, but may be resolved by other methods (7, 11, The effects of bisulphite on sequences in these 12). graticules become increasingly complex due to reaction of many normal products and also acquisition of new spots by Consequently, modification from both previous graticules. here we shall confine attention to effects within the first two graticules, in which a representative sample of possible effects was found. Reference products Three products containing neither cytidine nor uridine were used as reference products for quantitating the degree of reaction of other, cytidine-containing products in the These products were spots 3 bisulphite-treated samples. The molar (Am-G), 9 (A-Am-G) and 21 (Gm-G) (Table 2). yields of these products in control 28 S RNA were determined previously by a two stage analysis: - (i) Using the combined T plus pancreatic ribonuclease fingerprinting system, with P labelled RNA, it was shown that many methylated sequences occur once per molecule (15, 14). (ii) The individual yields of these three and all other Ti products in 14C methyl fingerprints were then computed from an assumed mean of one for twenty-two well resolved sequences (all those with approximately unimolar values and On the one methyl group in the first two graticules (7)). basis of these assumptions, the yields of spots 3, 9 and 21 were non-integral, possibly because, being short products and hence very stable towards over-digestion, their recovery was slightly better than that of the others. In computing the yields of other products in the bisulphite reacted RNA it was rational to attribute an equal value
435
Nucleic Acids Research Table 2 Molar yield (a) Untreated
Reference Products:3 (G)-Am-G 9 (G )-A-Am-G (G )-Gm-G 21 Sum of 3, 9, 21:-
Single 2'-0-methyl C:2 (G)-Cm-G 7 (G) -A-Cm-G (G -Cm-U-G 23 (G)-Cm-A-U-G 27
Single non-methyl C:6 (G -C-Am-G 11 (G -A-Am-C-G 15 (G)-A-A-C-Am-G 26 (G)-Um-A-C-G 28 ()- (mA-A-C,U)-G 32 ()- (C,U)-A-Am-A-G 35 () -A-A-C-A-A-Um-G
Reacted
4.51 1.17 7.06
4.87 1.18 6.69
12.74
12.74
1.08
1.01 0.52
1.16 0.85
0.63
0.99
0.59
1.17
0.28 0.39 0.05 0.03 0.09 0.60 0.13
1.08 0.82 0.96 0.95 1.01 0.91
Reactivity
% conversion(c)
6 55 26 40
76
64 94 97 91 41 86
± ++ + ++ +++
++ ++++ ++++ ++++
++ +++
Two C residues:-
4
G)-C-Cm-G
10 (G)-A-Cm-C-G 13 (G) -Cm-C-A-A-G 16 (G)-C-C-Gm-G
Others:2a (T)-mC-G 12 (G)-C-C-Cm-A-G 14 ...mA+Am-C... 33 .. mC+C?+A-A-Am-U...
0.94 1.04 0.94 0.80
0.69 0.63 0.80
0.66
73 39 15 17
+++
++ + +
0.2 0 0.24 68 1.06 0.34 +++ 96 0.78(b) 0.03 ++++ 1.10 b) 0.78 + 29 92 ++++ 36 (G)-A-C3_5-Gm-A-Am-A-G 0.91 b) 0.07 Reactivities of methylated, cytidine-containing sequences in 28 S RNA. Footnotes:- (a) see text for determination of molar yields. (b) Yields corrected for double methylation. Numbers of cytidines unknown. (c) ± =0-10% conversion; + 11-35%; ++ = 36-65%; +++ = 66-89%; ++++ > 90%. The percentage conversion was determined from the degree of diminution of the spots with respect to the sum of the three reference spots, as determined by scintillation counting. (In different preparations there were 250-500 c.p.m. per methyl group per unreacted unimolar spot). The values shown are the means of quadruplicate determinations agreeing within 5%.
436
Nucleic Acids Research to the sum of the molar yields of these three (unreactive) products as for the same products in the unreacted RNA. (Two of these products occur in high yield, thereby minimizing radioactive counting errors). Assuming an equal sum of molar yields for these three products in control and reacted RNA, agreement between the individual yields was reasonably good (Table 2), providing an indication of the reproducibility of the methods. Reactivities of methylated sequences. The remaining sections of Table 2 show the reactivities of the cytidine-containing sequences. These are arranged in groups:- those containing a single 2'-0-methyl C, those with an unmethylated C near to another 2'-0-methylated nucleotide, and so on. Each product occurs approximately once per molecule in fingerprints of untreated 28 S RNA (7), and therefore each represents a unique region of the 28 S molecule. The sequences encompass a wide range of reactivities. In addition to the quantitative values the results are interpreted by numbers of pluses in the right hand column of the table (see legend). Aspects of the results are discussed below. Products of reaction. Where the original sequence contained one cytidine and no uridine, identification of the reaction product For example was straightforward; it contained one uridine. the position of the new spot, 6b, which was present in about half molar yield in modified 28 S RNA, suggested the Alkaline hydrolysis yielded a composition (A,U)Gp. dinucleotide of % (pH 3.5) corresponding to Gm-Ap or Collectively these findings indicated that Am-Gp (7). the spot is U-Am-Gp, arising from reaction of C-Am-Gp (spot 6). Three further examples of reactions involving single C--4U changes, one of them (spot 7) involving a For a sequence 2'-0-methyl C, are given in Table 3. with two or more cytidines the reaction product(s) may contain one or two "new" uridines, depending on whether one Product identification should or both cytidines reacted. distinguish between the various possible reactions, but so
437
Nucleic Acids Research Table 3 New Spot Sequence
6 b (G)-U-Am-G 7 b (G)-A-Um-G 11 b (G)-A-Am-U-G 15 b (G)-A-A-U-Am-G
Derived from
6 (G)-C-Am-G 7 (G)-A-Cm-G 11 (G)-A-Am-C-G 15 (G)-A-A-C-Am-G
Origin of selected new spots in T1 fingerprints of methyl labelled 28 S RNA after reaction with bisulphite. The products were identified from their mobilities in the two dimensional system (indicating their base compositions), the electrophoretic mobility at pH 3.5 of the respective alkali stable dinucleotides (7) and knowledge of the original (unreacted) sequences. There was approximate agreement between the molar yields of the new spots and the extent of reaction of the original products, though quantitation was slightly unreliable for spots 6b, 7b and llb due to the close proximity of other unreacted spots, 24' (A-Gm-G), 15b was well 25 (Am-U-G) and 29 (A-A-Um-G) respectively. resolved and was recovered in 0.8 molar yield, in good agreement with the almost complete disappearance of spot 15 (Table 2).
far this has not been carried out in these more complex situations. DISCUSSION Of the approximately seventy methyl groups in HeLa cell 28 S RNA (7), the reactivities of sequences encompassing twenty-three have been examined here, another (approximately) thirteen methyl groups providing reference material. No obvious general rules emerge from the results obtained It might, perhaps, have been speculated from this sample. that methylation sites, in order to be recognized by the methylase(s), would occupy exposed conformations. In fact some of the sequences are rather unreactive, notably spot 2, Ribose methylation is derived from the sequence G-Cm-G. a very early event in rRNA maturation, probably occurring It is possible that after on nascent 45 S RNA (8,7). ribose methylation has occurred further secondary or tertiary interactions intervene during 45 S RNA chain completion or rRNA maturation, thus rendering inaccessible at least some of the sites which were accessible at the time of methylation.
438
Nucleic Acids Research Nevertheless, it seems likely that information on the reactivities of particular nucleotides will become useful in several ways, in conjunction with more extensive nucleotide It is reasonable to sequence data or comparative data. infer that the most reactive sequences, spots 14, 15, 26, 28 and 36 (Table 2) contain one or more cytidines in exposed regions, within the 28 S tertiary structure. In such cases the C residue may either be unimportant for function, or it may be highly important in making contact with (for example) a ribosomal protein or another Product 26, Um-A-C-G, component of protein synthesis. a reactive contains C. Interestingly, this product very is absent in the Xenopus 28 S fingerprint, in which the related oligonucleotide, Um-A-U-G occurs instead (14). A C--AU base change is a likely explanation, suggesting that this particular (reactive) C does not play a crucial role in ribosome function (or, for that matter, in recognition of By contrast spot 36, in this sequence by the methylase). which one or more C residues are reactive, seems to be a constant vertebrate sequence (unpublished fingerprints of It is possible that in this case the Khan and Maden). In fact the doubly cytidines play an important role. methylated component, Gm-A-Am-A-Gp, has been reported in the 28 S ribosomal RNAs of such distantly related organisms as Drosophila (15) and yeast (16) and therefore seems to be a highly conserved eukaryotic sequence. There appears to be considerable scope for usefully extending the present type of analysis, for example by more detailed kinetic analysis of the reactivities of individual sequences, by analysis of reactivities of further unique sequences within 28 S and 18 S RNA, and of the same sequences within the various ribosomal precursor RNAs.
ACKNOWLEDGEMENT This work was supported by a grant from the Medical We thank T. Carr for skilled technical Research Council. assistance.
439
Nucleic Acids Research REFERENCES 1. Robertus,J.D., Ladner,J.E., Finch,J.T., Rhodes,D., Brown,R.S., Clark,B.F.C., and Klug,A. (1974). Nucleic Acid Research 1 927-932. 2. Rhodes,D. (1975). J. Mol. Biol. 94, 449-460. 3. Shapiro,R., Servis,R.E. and Welcher,M. (1970). J. Amer. Chem. Soc. 92, 422-424. 4. Hayatso, H., Wataya,Y., Kai,K. and Iida, S. (1970). Biochemistry 9, 2858-2864. 5. Goddard, J.P. and Schulman,L.H. (1972). J. Biol. Chem. 247, 3864-3867. 6. Goodchild,J., Fellner,P. and Porter,A.G. (1975). Nucleic Acids Research 2, 887-895. 7. Maden,B.E.H. and Salim,M. (1974). J. Mol. Biol. 88, 133-164. 8. Greenberg,M. and Penman,S. (1966). J. Mol. Biol. 21, 527-535. 9. Zimmerman,E.F. (1968). Biochemistry 7, 3156-3164. 10. Vaughan,M.H., Soeiro,R., Warner,J.R. and Darnell,J.E. (1967). Proc. Nat. Acad. Sci., U.S.A. 58, 1527-1534. 11. Brownlee,G.G. and Sanger,F. (1967). J. Mol.Biol. 23, 337-353. Brownlee,G.G. and Sanger, F. (1969). Eur. J. Biochem. 12. 11, 395-399. 13. Maden,B.E.H., Lees,C.D. and Salim,M. (1972). FEBS Letters 28, 293-296. Khan,M.S.N. and Mladen,B.E.H. (1976). J. Mol. Biol., 14. in press. 15. Maden,B.E.H. and Tartof,K. (1974). J. Mol. Biol. 90, 51-64. 16. Klootwijk,J. and Planta,R.J. (1974). Mol.Biol. Reports, 1, 187-191. 17. Amaldi,F. and Attardi,G. (1968). J. Mol. Biol. 21, 29-41. 18. Willems,M., Wagner,E., Laing,R. and Penman,S. (1968). J. Mol. Biol. 32, 211-220.
440
