METHYLATION OF FELINE LEUKEMIA VIRUS VIRION AND INTRACELLULAR RNA By Arlen Read Thomason A DISSERTATION Submitted to Michigan State University in partial fulfillment for the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry ABSTRACT METHYLATION OF FELINE LEUKEMIA VIRUS VIRION AND INTRACELLULAR RNA By Arlen Read Thomason Several methods for the extraction of RNA from the Rickard strain of feline leukemia virus (R-FeLV) were compared, and the method which resulted in the highest yields of viral high molecular weight RNA was selected for use in subsequent experiments. Two major size classes of native RNA were detected in R-FeLV, one sedimenting at 50-608 and the other at 43. When native 50-608 RNA was denatured with dimethylsulfoxide (MeZSO) and centrifuged through sucrose gradients in 99% MeZSO, subunits of approximately 288 were obtained. It was shown that the subunits obtained from such gradients migrated more slowly during polyacrylamide gel electr0phoresis than 283 ribosomal RNA. The native 4S RNA obtained from sucrose gradients was resolved by electrophoresis in 7% polyacrylamide gels into four distinct classes. Two of these low molecular weight RNAs co-migrated with cellular 4S and SS RNA while the other two slower-migrating RNAs were probably forms of the 7-88 RNA reported in other oncornaviruses. The high molecular weight subunit RNA of R-FeLV was analyzed for the presence of methyl groups. Following purification of native 50-608 F-FeLV RNA on nondenaturing aqueous sucrose density gradients, Arlen Read Thomason R-FeLV 283 subunit RNA, doubly labeled with (14C) uridine and (methyl- 3H) methionine, was isolated by centrifugation through denaturing sucrose density gradients in dimethylsulfoxide. As calculated from their respective 3H/U‘C ratios, R-FeLV 28S RNA was methylated to the same degree as host cell poly A+ mRNA. When the 28S R-FeLV RNA was hydrolyzed to completion with RNAase T2 or alkali all of the methyl-3H chromatographed with mononucleotides on PellionexAWAX, a weak anion exchanger. The methyl-labeled material co-chromatographed with 6-methyladenosine if the mononucleotide fraction obtained by PellionexAWAX chromatography was hydrolyzed to nucleosides by bacterial alkaline phoSphatase or with 6-methyladenine if purine bases were released from the mononucleotides by acid hydrolysis. In another experiment in which R-FeLV 288 RNA uniformly labeled with 32F was hydrolyzed and then analyzed by Pellionex-WAX chromatography, all of the 32? label again co-chromatographed with mononucleotides. Thus R-FeLV 28$ RNA does not appear to contain a 5' structure, either methylated or nonmethylated, similar to those recently reported for cellular and some animal virus mRNAs. It was found that purification of a polynucleotide by hybridization to its mercurated complementary sequence and chromatography on sulfhydryl-agarose (SH-agarose) was unsatisfactory, due to the instability of SH-agarose prepared by the conventional cyanogen bromide technique. This technique was improved by the use of SH-agarose in.which the sulfhydryl group is attached to the agarose through a stable ether linkage. In addition, a method was develOped whereby hybridized RNA or DNA could be recovered from the SH-agarose column separately from the mercurated probe. This Arlen Read Thomason methodology was applied to the purification of R-FeLV intracellular RNA from host cell RNA. The F-422 line of feline thymus tumor cells, chronically infected with R-FeLV was labeled with 32p and the total cytOplasmic RNA was isolated. The RNA was centrifuged through sucrose gradients and R-FeLV virus-specific RNA (vRNA) was located by hybridization of portions of the gradient fractions to R-FeLV complementary DNA (cDNA). Virus-specific RNA classes with average sedimentation coefficients of approximately 368, 288, 238, and 158 were identified. Each class of RNA was recovered and hybridized with R-FeLV mercurated cDNA, and the hybrids were chromatographed on columns of sulfhydryl-Sepharose to separate them from unhybridized cellular RNA. While insufficient amounts of 36S and 288 vRNA were obtained for further analysis, the 238 and 153 vRNA classes were analyzed to determine the nature of their 5' termini. Each of these vRNA classes was found to contain cap structures in amounts sufficient to account for approximately one cap per molecule. The structure of the 238 vRNA cap was found to be m7GS'ppp5'GmpAp while that of the 158 vRNA cap was m7G5'ppp5'GmpGp. The possible relationship between R-FeLV and other defective RNA tumor viruses is discussed. ACKNOWLEDGMENTS I would like to thank all the members of the Biochemistry Department who made my stay here an enjoyable one. Special acknowledgments go to the following peOple: Dr. Fritz Rottman, whose ability to foster the development of a young scientist is second to none. His unfailing encouragement helped me get through the roughest times. Karen Friderici, whose friendship and confidence in my abilities have meant much to me. It was under her guidance that I learned many basic laboratory techniques. Dr. Leland Velicer, who provided the virus with which this work was concerned. The generous use of his laboratory and the patience and encouragement he has given me have been greatly appreciated. Dr. David Brian, who as a collaborator shared with me the ups and downs of the early part of this work. Last, but not least, I would like to thank Margit Susan Thomason. Her patience and understanding have far exceeded what anyone could have expected. ii TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES LIST OF ABBREVIATIONS LITERATURE REVIEW ONCORNAVIRUSES Definition and Classification Composition Replication Virion RNA Intracellular Viral RNA Synthesis of Viral DNA METHYLATION OF mRNA Distribution of Caps PrOperties of Caps Synthesis and Degradation of Caps Biological Function of Caps Internal Methylation of mRNA REFERENCES PART I PRELIMINARY CHARACTERIZATION OF FELINE LEUKEMIA VIRUS RNA INTRODUCTION METHODS Cells and Virus Viral RNA Extraction Centrifugation in Aqueous Sucrose Gradients Centrifugation in Sucrose Gradients in MeZSO Polyacrylamide Gel ElectrOphoresis RESULTS iii Page vi vii ~o2 >~ .t .2 5 > > 'z I; '0'- “ g n .1 5 g o i I .° 0 . .6 .- : 13 5 's 0 O 5 'o a a: U I 2 0 -25 l l b 285 2 '- I r p x I51- 8 n O. J: 2‘ '; ‘2 I — U I O .2 .O O 8 m N n .51— 1» l 1 IO 2 0 Fraction Figure 1. Isolation of R-FeLV native high molecular weight RNA by centrifugation through aqueous sucrose gradients. Figure 2. 64 Isolation of R-FeLV high molecular weight subunit RNA by centrifugation through sucrose gradients in MeZSO. R-FeLV high molecular weight RNA obtained as in Figure 1 was sedimented through a 5-20% sucrose gradient in 99% MeZSO as described in Materials and Methods. F-422 cellular 28S rRNA was run on a parallel gradient. Fractions denoted by the bars were pooled and the RNA was ethanol precipitated and saved for further analysis. (a) RNA labeled with (methyl-3H) methionine and (14C) uridine. (b RNA labeled with 32P0 . ) 4 65 "C Radioactivity (cpm x 10") out: (I I I0 _. "IO Tc; 285 x I E O. 3 3: '; 5 — n 5 '3: - D U P : O .' '- 9 5 a ‘0 .-' '-. ° 5 '-. °‘ :5 1 .I °~. :' 11.0 D" ' ‘O-D-D-fldl .0. . u u LJ'b an I .d .u.‘ 6‘ 2: IO L- 2:5 X E O. .3 Z? '; I: ‘6 5 - .2 .0 O a: Q. at In 1 IC) 2() Fraction Figure 2. Isolation of R-FeLV high molecular weight subunit RNA by centrifugation through sucrose gradients in MezSO. 66 Table 1 Extent of Methylation of R-FeLV 288 RNA Relative to F-422 Cellular Poly A+>mRNA.a Experiment 3H cpm/140 cpm FeLV 28S RNA 3H cpm/I40 cpm F-422 poly A+'mRNA l 0.91 2 1.04 3 1.00 4 1.04 Average 1.00 a R-FeLV 28S RNA labeled with (methyl-3H) methionine and (14C) uridine was purified as described in Figures 1 and 2. The 3H/U‘C ratio was determined by liquid scintillation counting and was compared to the 3H/14C ratio of F-422 cellular mRNA obtained by oligo (dT)-cellulose chromatography in the same experiment. Figure 3. 67 MeZSO/sucrose gradient analysis of RSV RNA added to R-FeLV prior to RNA extraction. RSV native high molecular weight RNA labeled with (3H) uridine was added to (14C) uridine-labeled R-FeLV and the RNA was immediately extracted as described in Materials and Methods. The native high molecular weight RNA was purified by aqueous sucrose gradient centrifugation and then centrifuged through a sucrose gradient in MeZSO as described in Figure 2. Ribosomal 28S RNA was run as a marker on a parallel gradient. 68 no.iv. 1 |()'- 96 510— x Eauv >:>:uoo:oa~_ U... n I :1. A70— x Eauv >:>:ooo_ro~_ In Fraction MeZSO/sucrose gradient analysis of RSV RNA added to R-FeLV prior to RNA extraction. Figure 3. 69 (Figure 3). Therefore it seems unlikely that the 28S RNA we observe from R-FeLV is derived from a larger 30-408 molecule, similar to the high molecular weight subunit RNA of avian oncornaviruses, by degradation during RNA purification. Furthermore it is unlikely that the R-FeLV 28S RNA is derived from a large precursor that is cleaved within the virus as a function of time, since virus labeled for a short time (60 min) with (3H) uridine also yields RNA sedimenting at 288 (unpublished results). Absence of Methylated Caps in 28S RNA Within the last year several animal virus RNAS (l, 10, ll, 12, 15, 17, 31) and eukaryotic cellular mRNAs (2, 7, 24, 29) have been reported to contain at their 5' ends "caps" of the general type m7G5'ppp5'Nmp(1_2)Np. It was therefore of interest to determine if R-FeLV 28S RNA contains a similar structure. Methyl labeled 50-608 R-FeLV RNA was first obtained by centrifugation through aqueous sucrose gradients as in Figure la to reduce or eliminate possible contamination by host cell mRNA (see below). The 288 R-FeLV RNA was then obtained by centrifuging the 50-608 through a sucrose gradient in MeZSO as shown in Figure 2a. The R-FeLV 288 RNA thus purified was hydrolyzed to completion with alkali or RNAase T2. As shown in Figure 4a no resistant structures larger than mononucleotides were found in R-FeLV 28S RNA by Pellionex-WAX chromatography. When analyzed by the same procedures poly A+ mRNA obtained from the cytoplasm of host cells in the same experiments contained oligonucleotides migrating between (Up)5 and (Up)7 markers (Figure 4b), as expected for structures such as described above. Thus R-FeLV 28S RNA does not contain a detectable quantity of methylated caps at its Figure 4. 70 Analysis of the RNAase T2 digestion products of methyl labeled RNA. RNA labeled with (methyl-3H) methionine and (14C) uridine was hydrolyzed to completion with RNAase T2 as described in Materials and Methods. Mixtures of oligomers of U, containing (Up)3 through (Up)7 (e.g., (Up)3 is UpUpUp), were mixed with the sample and quickly applied to a Pellionex-WAX high speed liquid chromatography column equilibrated with 0.005 M sodium phOSphate, pH 7.8, 7 M urea. The column was developed with a linear gradient of 0.0 M (N114) 2804 (40 ml) to 0.2 M (NH!) 2804 (40 ml) in 0.005 M‘sodium phoSphate, pH 7.8, 7 M_urea. Markers were monitored by UV absorbance and radioactivity in the collected fractions was determined by scintillation counting of aliquots. (a) R-FeLV 28S RNA. (b) F-422 cellular poly A+’mRNA. 71 0 l0- Np Pd 5— (I 22 Nmpr '5 C) U U LI U U '- p 9999 x 13 i‘t’t‘t’ E a 0 J: .1 _- 1..., , .; b I: IN”, a .o .0 a a: I: 2' 0‘) '~ Nmpr Up Up Up Up Up a so 7 131111 I. l l l 1 IO 20 30 4o 50 so Fraction Figure 4. Analysis of the RNAase T2 digestion products of methyl labeled RNA. 72 5' ends nor does it contain internal nucleosides which are methylated at the 2' position, since such structures would be resistant to these hydrolysis procedures and would migrate with dinucleotide or larger markers. It was found that if the aqueous sucrose gradient centrifugation step for the isolation of native 50-608 R-FeLV RNA was omitted (i.e., if total viral RNA was applied directly to the MeZSO sucrose gradient) then variable amounts of 2'-4-methylated nucleosides appeared in the final Aminex high Speed liquid chromato- graphic analysis of R-FeLV 28S RNA. These 2'-4-methylated nucleosides may have been derived from contaminating cellular mRNA since chromatography of the RNA on oligo(dT)-cellulose did not result in their removal. However, when the aqueous sucrose gradient centrifugation Step was included no 2'-4-methylated nucleosides were found in R-FeLV 288 RNA. m°A Is the Only Methylated Nucleoside in 28S RNA Since after alkaline or RNAase T2 hydrolysis all of the methyl labeled nucleotides co-chromatographed with mononucleotide standards it was concluded that only base methylated nucleotides are present in R-FeLV 288 RNA. The mononucleotides derived from R-FeLV 28S RNA and purified by high speed liquid chromatography (Pellionex-WAX) were desalted by adsorption to DEAE-cellulose (carbonate form) and elution.with 1.! ammonium carbonate. After removing the ammonium carbonate by ly0philization the mononucleotides were dephoSphorylated with bacterial alkaline phosphatase and chromatographed on Aminex A-S, a cation exchanger which separates nucleosides. AS shown in Figure 5a virtually all (90%) of the methyl-3H co-chromatographed with 6- methyladenosine standard. In another experiment, purine bases were Figure 5. 73 Determination of the methylated components of R-FeLV 288 RNA. (a) Mononucleotides obtained as in Figure 4a were desalted and treated with alkaline phosphatase. The nucleosides thus generated were analyzed by Aminex high Speed liquid chroma- tography as described in Materials and Methods. Standards are uridine (U), guanosine (G), adenosine (A), and 6- methyladenosine (m°A). (b) Mononucleotides obtained as in Figure 4a were desalted and purine bases were released by acid hydrolysis as described in Materials and Methods. The sample was analyzed by Aminex high speed liquid chromatography. Standards are guanine (Gua), adenine (Ade), and 6- methyladenine (m6Ade). 74 I "'3‘ a U G A C it i t t 50*- N (J I + s Q. 8. 2‘ IE 3 (3 .2 '3 50.; I I I a: I 0 25- IO 20 30 4o Fraction Figure 5. Determination of the methylated components of R-FeLV 28S RNA. 75 removed from the mononucleotides by acid hydrolysis and analyzed by high Speed liquid chromatography on Aminex A-5. More than 96% of the methyl-3H co-chromatographed with authentic 6-methyladenine (Figure 5b). The absence of any methyl-3H radioactivity chromatographing in the position of adenine or guanine confirms that no ring-labeling of purine bases occurred. Thus m°A is the only methylated nucleoside in R-FeLV 288 RNA. Absence of nonmethylated Caps in 288 RNA Although R-FeLV 288 RNA did not contain a methylated 5' cap, we considered the possibility that the RNA of feline leukemia virus (grown in F-422 cells) might differ from other viral and cellular mRNAs by containing a nonmethylated cap of the type Ns'ppps'Np. We would not have detected such a structure in our experiments with methyl-labeled RNA. Therefore we examined R-FeLV RNA which had been uniformly labeled with 32P014. RNA purified as in Figures lb and 2b was hydrolyzed with alkali and examined by pellionex-WAX chromato- 32P-labeled material eluted in the graphy. As shown in Figure 63 no position of the (Up)4 marker as would be expected for a structure such as above. A structure with 4 phOSphates should have contained 0.05 to 0.08% of the total cpm (600,000 cpm in the experiment shown in Figure 6a), well within our limits of detection. The small peak eluting near the (Up)6 marker represents only 40 cpm and was not reproducible. Again in control eXperiments with host cell poly A+ mRNA, alkali-resistant 32P-labeled structures from 300,000 cpm of hydrolyzed RNA eluted between (Up)5 and (Up)7 markers on Pellionex-WAX (Figure 6b). Therefore R-FeLV 28S RNA lacks nonmethylated as well as methylated 5' caps. We have not, however, been able to determine what 76 Figure 6. Analysis of the alkaline digestion products of RNA labeled with 32P04. RNA labeled with 32P04 was hydrolyzed to completion with alkali and analyzed by Pellionex-WAX high Speed “liquid chromatography as in Figure 4. (a) R-FeLV 28S RNA. (b) F-422 cellular poly A+ mRNA. 77 Ufiu‘au‘ps It". 32? Radioactivity (cpm x IO") It" I U", U", 0:, Fraction Figure 6. Analysis of the alkaline digestion products of RNA labeled with 32P04. 78 is at the 5' end of R-FeLV 288 RNA since no structures correSponding to pr, ppr, or pppr were detected in 32PO4 labeled RNA and our attempts to label the 5' end of R-FeLV RNA using polynucleotide kinase have been unsuccessful. DISCUSSION The results presented here Show that R-FeLV 28S subunit RNA does contain methylated nucleosides. The possibility that these methylated components derive from contaminating 50-608 associated 4S RNA seems very unlikely since the R-FeLV 28S RNA was purified by centrifugation through a denaturing density gradient, and Since the methylation pattern of R-FeLV 288 RNA differs completely from that of 4S RNA (6). Furthermore it is unlikely that the R-FeLV 288 RNA was contaminated with cellular 288 rRNA since methylation of the latter molecule occurs principally on the 2' position of ribose (6), whereas the only methylated nucleoside in R-FeLV 28S RNA is a derivative of adenosine methylated at the N-6 position. This was demonstrated both by analysis of the nucleosides produced by phOSphomonoesterase digestion of the RNAase T2-generated mononucleotides and by analysis of the bases released after acid hydrolysis of the methyl-labeled mono- nucleotides. Since R-FeLV 28S RNA is methylated to the same degree, on a per uridine basis, as cellular poly A+ mRNA (Table l) the number of mgA residues per R-FeLV 288 RNA molecule can be calculated if the total number of nucleotides in the latter RNA is known. If the molecular weight of R-FeLV 288 RNA is the same as that of 28S rRNA, 6 i.e., 1.8 x 10 , then there are about 5,000 total nucleotides and 79 about 10 m6A residues per R-FeLV RNA molecule. However, if the molecular weight of R-FeLV 28S RNA is 2.8 to 3.2 x 106 (see below) then there are 8,700 to 10,000 total nucleotides and about 17 to 20 m6A residues per R-FeLV 28S RNA molecule. These calculations use the value of 0.2% methylation given by Perry and Kelly (23) for cellular mRNA. The error due to differences in the base compositions of R-FeLV 288 RNA and cellular poly A+'mRNA is assumed to be small and is neglected. Our results also indicate that R-FeLV 288 RNA contains no detectable "cap" (that is, a structure with an inverted 5'-5' pyro- phOSphate linkage (26)), either methylated or nonmethylated, at its 5' end. However, the amounts of radioactive R-FeLV 28S RNA used in these studies place limits on the conclusions that can be reached. Thus it can be concluded that less than one in ten R-FeLV 28S RNA contains a cap. Cap structures have been found recently in cellular mRNAs (2, 7, 24, 29) and in animal virus mRNAs (l, 10, ll, 17, 31) including those from two strains of an RNA tumor virus (12, 15). The latter reports both dealt with an avian sarcoma virus whereas this report concerns a mammalian leukemia virus. Whether or not this fact is important to the differences reported remains to be determined. There are several possible explanations for the uneXpected absence of a cap Structure at the 5' end of R-FeLV 288 RNA. One possibility is that a 5' cap structure is present in an earlier transcript of R-FeLV high molecular weight subunit RNA, but that after packaging in the virus the 5' cap is removed either naturally or as an artifact by endonucleolytic hydrolysis of the RNA molecule. Since the R-FeLV subunit RNA isolated in our laboratories cosediments with 28S ribosomal RNA (both an aqueous and MeZSO gradients), whereas the 80 usually accepted value for RNA tumor virus Subunit RNA is about 358, it might be argued that degradation may have occurred during RNA isolation. While this possibility cannot be rigorously excluded, the control experiment, in which RSV 30-408 RNA added to R-FeLV before RNA extraction was not reduced in Size, argues against it. In addition such degradation would have to result from a very Specific nick since we always observe R-FeLV RNA subunit molecules that are 288 and since very little material sediments between the 288 and 4S peaks (Figure 2). Furthermore, it does not appear that R-FeLV 28S RNA results from break- down of RNA within the virion as a function of time, since R-FeLV subunit RNA labeled for as short a time as l h still sediments at 288. That R-FeLV and other mammalian RNA tumor virus subunit RNAS can have a sedimentation coefficient of about 288 is supported by the results obtained by East t al. (8) for FeLV (Rickard), FeSV (Rickard), RD-ll4, and Crandell virus, and by Manning 42,42, (19) for MuSV (Moloney). Whether the 288 value represents a difference in chain length or conformation remains to be determined, but it is interesting to note that although R-FeLV subunit RNA cosediments with 28S ribosomal RNA it nevertheless migrates more slowly than the same marker during polyacrylamide gel electrOphoreSis (with an estimated molecular weight of about 3.2 x 106) (5). In this regard RD-ll4 subunit RNA, which also cosediments with 28S rRNA (8), has a molecular weight of 2.8 x 10° as determined by electron microscopy (16). Therefore R-FeLV 288 RNA could have a molecular weight in the 2.8 to 3.2 x 106 range. Another possible explanation for the absence of a 5' cap in R-FeLV 28S RNA is that the R-FeLV high molecular weight RNA destined for packaging into virions is not methylated by the normal post- 81 transcriptional modification enzymes or that a methylated 5' end is removed from the RNA before packaging. This possibility cannot be eliminated, and in fact the finding of no caps in packaged viral RNA does not preclude their presence in Similar polysomal viral RNA Species. Studies to determine if polysomal R-FeLV RNA does indeed contain a 5' cap structure are being actively pursued. Recent pre- liminary results in our laboratories indicate, however, that our stocks of feline leukemia virus (Rickard), produced by F-422 feline thymus tumor cells, have only a low level of infectivity. Therefore the present results on R-FeLV RNA methylation may pertain mainly to defective particles. Experiments are in progress to determine if this is indeed the case. If so, it will be interesting to examine the relationship between the absence of a cap structure in R-FeLV 28S RNA and the defective nature of the virus. REF ERENCE S 10. 11. 12. l3. 14. 15. 82 REFERENCES Abraham, G., D. Rhodes and A. K. Banerjee. (1975) Cell 4, 51-58. Adams, J. M. and 8. Cory. (1975) Nature 244, 28-33. Bishop, J. M., W. E. Levinson, N. Quintrell, D. Sullivan, L. Fanshier, and J. Jackson. (1970) Virology 42, 182-195. Both, G. W., A. K. Banerjee and A. J. Shatkin. (1975) Proc. Natl. Acad. Sci. U.S.A. 22, 1189-1193. Brian, D. A., A. R. Thomason, F. M. Rottman, and L. F. Velicer. (1975) Virol. 24, 535-545. Desrosiers, R., K. Friderici and F. Rottman. (1974) Proc. Natl. Acad. Sci. U. S. A. 22, 3971-3975. Desrosiers, R. C., K. H. Friderici, and F. M. Rottman. (1975) Biochem. 24, 4367-4374. East, J. L., J. Knesek, P. Allen, and L. Dmochowski. (1973) J. Virol. 22, 1085-1091. Erikson, R. L. (1969) Virology 22, 124-131. Furuichi, Y., and K.-I. Miura. (1975) Nature 244, 374-375. Furuichi, Y., M. Morgan, 8. Muthukrishman, and A. J. Shatkin. (1975) Proc. Natl. Acad. Sci. U.S.A.‘22, 362-366. Furuichi, Y., A. J. Shatkin, E. Stavnezer, and J. M. Bishop. (1975) Nature 242, 618-620. GilleSpie, D., 8. Marshall, and R. C. Gallo. (1972) Nat. N. Biol. 244, 227-231. Green, M. (1970) Ann. Rev. Biochem. 22, 701-756. Keith, J. and H. Fraenkel-Conrat. (1975) Proc. Natl. Acad. Sci. U.S.A. 12, 3347-3350. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 83 Kung, H.-J., J. M. Bailey, N. Davidson, M. 0. Nicolson, and R. M. McAllister. (1975) J. Virol. 24, 397-411. Lavi, S. and A. J. Shatkin. (1975) Proc. Natl. Acad. Sci. U.S.A. ‘22, 2012-2016. Long, P.A., P. Kaveh-Yamini, and L. F. Velicer. (1975) J. Virol. 24, 1182-1191. Manning, J. S., F. Schaffer, and M. Soergel. (1972) Virology 44, 804-807. McCain, B., N. Biswal, and M. BenyeSh-Melnick. (1973) J. Gen. Virol. 24, 69-74. Miura, K.-I., K. Watanabe, and M. Sugiura. (1974) J. Mol. Biol. _8_6_, 31-48. Moyer, S., G. Abraham, R. Adler, and A. Banerjee. (1975) Cell 4, 59-67. Perry, R. P. and D. E. Kelley. (1974) Cell 2, 27-42. Perry, R., D. Kelley, K. Friderici, and F. Rottman. (1975) Cell 4, 387-394. Pike, L. M. and F. Rottman. (1974) Anal. Biochem.‘42, 367-378. Rottman, F., A. J. Shatkin, and R. P. Perry. (1974) Cell 4, 197-199. Shatkin, A. (1974) Proc. Natl. Acad. Sci. U.S.A. 22, 3204-3207. Van Der Helm, K. and P. H. Duesberg. (1975) Proc. Natl. Acad. Sci. U.S.A. 22, 614-618. Wei, C.-M., A. Gershowitz, and B. Moss. (1975) Cell 4, 379-386. Wei, C.-M. and B. Moss. (1974) Proc. Natl. Acad. Sci. U.S.A. 1, 3014-3018. Wei, C.-M. and B. Moss. (1975) Proc. Natl. Acad. Sci. U.S.A. _2, 318-322. PART III A METHOD FOR PURIFICATION OF MERCURATED NUCLEIC ACID HYBRIDS ON ETHER-LINKED SULFHYDRYL-AGAROSE INTRODUCTION The importance of purification of nucleic acid sequences from mixtures in which the sequence of interest represents only a fraction of the sample has stimulated the development of a number of techniques for this purpose. Isolation of specific sequences on the basis of size or base composition has been practical in only a few unusual cases. Purification schemes which rely on the base sequence itself for specificity are potentially much more powerful techniques. One of the oldest of such methods involves hybridization of a nucleic acid probe to a mixture of sequences and subsequent separation of single-stranded and double-stranded nucleic acids on hydroxyapatite (l). A limitation of this method is that double-stranded regions in extraneous sequences may co-purify with the hybrid. Purification has also been achieved by hybridization of a nucleic acid mixture to a probe covalently bound to a solid Support (2-4). While these techniques are very useful, they have some disadvantages. Coupling of the probe to the support is often an inefficient process, making it difficult to achieve high concentrations of the probe. In addition, hybridization to a probe bound to a solid support does not proceed as rapidly or go as far toward completion as hybridization in solution (5, 6). I Recently Dale and Ward (7) have described a technique in which mercurated polynucleotides can be separated from their unmercurated counterparts by chromatography on columns of sulfhydryl-agarose (SH-agarose). Using their technique a mercurated nucleic acid probe can be annealed in solution to a complementary component in a nucleic 84 85 acid mixture and the resulting hybrid can be bound to SH-agarose. After removal of unhybridized nucleic acid by extensive washing, the hybrid is recovered by elution with a reducing agent such as 2- mercaptoethanol. In practice, we have found that the instability of the bond linking the sulfhydryl group to the agarose, when the SH- agarose is synthesized by the conventional cyanogen bromide technique (8), can be a serious disadvantage. This is eSpecially true when the nucleic acid sequence to be purified constitutes only a small fraction of the sample.‘ In this case, the extensive washing of the SH-agarose required to remove unhybridized sequences can result in the loss of the hybrid due to breakage of sulfhydryl-agarose bonds. We describe here the use of this method with SH-agarose in which the sulfhydryl group is bound to the agarose 424 a stable ether linkage. In addition, we report conditions under which hybridized nucleic acid can be recovered from the column separately from the mercurated probe, allowing the probe to be re-used in subsequent hybridizations. MATERIALS AND METHODS SH-agarose with the sulfhydryl group bound to the agarose support through an ether linkage was synthesized essentially as described by Axen.4£_42, (9). Cross-linked Sepharose 2B (Pharmacia Fine Chemicals, Uppsala, Sweden) was washed extensively with distilled water and freed of interstitial water by vacuum filtration. The agarose (6 g) was suSpended in 4.8 ml lM.NaOH and 0.6 m1 of epichlorohydrin (Aldrich Chemical Co., Inc., Milwaukee, Wis.) was added slowly with stirring at room temperature. The temperature was 86 raised to 60°C and stirring was continued for 2 h after which the gel was washed extensively with distilled water and then with 200 m1 of 0.5 M phOSphate buffer (8.2g NaHzPOh-HZO + 8.44gNa2HPO4-7H20 dissolved in 200 ml distilled water; pH - 6.3). After suspending the agarose in 12 m1 of the same buffer, 6 m1 of 2'M sodium thiosulfate was added and the mixture was shaken at room temperature for 6 h. The gel was again washed by vacuum filtration with distilled water and suSpended in 0.1 M sodium bicarbonate to a final volume of 12 ml. The S-alkyl- thiosulfate gel was reduced by addition of 6 ml of a solution of dithiothreitol (200 mg/ml) in lmM EDTA for 30 min at room temperature. Finally, the product was washed with 60 ml of 0.1 M sodium bicarbonate, 1M NaCl, 1 mM EDTA, and then with 200 ml 1 mM EDTA. Synthesis of SH-agarose by the cyanogen bromide technique was as described (8). The sulfhydryl content of the SH-agarose was determined as described by Ellman (10), in which thiol groups are titrated colorimetrically with dithionitrobenzene. SH-agarose synthesized as described here contained 2.6 umol SH/ml gel. Complementary 3H-DNA (3H-cDNA) was synthesized with AMV DNA poly- merase using bovine 188 ribosomal RNA (rRNA) for a template as described (A. Thomason 44,42., submitted for publication). Bovine pituitary ribosomal RNA was a gift from John Nilson. Novikoff cell ribosomal RNA, labeled with (3H) adenosine, was a gift from Ronald Desrosiers. RNA was mercurated as described by Dale and Ward (7) in a reaction containing 0.05 M sodium acetate, pH 6.0, 0.02 Mgmercuric acetate, and 60 mg/ml RNA incubated at 65°C for 20 min. The reaction was st0pped by addition of 0.4 vol cold quench buffer (0.01 M Tris-HCl, 87 pH 7.5, 1 M NaCl, 0.1 M EDTA) and unreacted mercuric acetate was removed by chromatography on a column of Sephadex G-50. Hybridization conditions and chromatography on SH-agarose are described in the figure legends. RESULTS When SH-agarose synthesized by cyanogen bromide activation (8) was used for affinity chromatography of mercurated polynucleotides at room temperature, it was found that the capacity of the column decreased progressively over a period of several months. Figure 1 shows that this decrease in capacity was accompanied by a Steady loss of sulfhydryl groups from the agarose. After 113 days the sulfhydryl content of the gel (0.55 umol SH/ml packed agarose) was less than one- fourth of the original level (2.58 umol/m1). SH—agarose stored at -4°C underwent a slower loss of sulfhydryl groups (data not shown). Under comparable conditions, SH-agarose with the sulfhydryl group attached through an ether linkage did not undergo any decrease in sulfhydryl content (Figure 1). The capacity of this gel, which contained approximately 2.60 umol SH/ml, was high, When 3H-labeled cDNA in which all of the thymidine residues were replaced by mercurated uridine was mixed with 200 ug of mercurated calf thymus DNA and passed over a 3- m1 column of ether-linked SH-agarose, 92% of the radioactivity was retained by the column (data not Shown). Thus the capacity of the column was not exceeded by this amount of mercurated nucleic acid. In order to develop a method in which a mercurated polynucleotide can be removed from the SH-agarose separately from its hybridized Figure 1. 88 Stability of SH-agarose prepared with an ether linkage and by cyanogen bromide activation. Each of the SH-agarose preparations was in use in a chromatography column at room temperature during the time Span in which measurements were made. In order to check the sulfhydryl content, a portion of the gel was removed from the column and incubated with 0.15 M 2-mercaptoethanol for 30 min to reduce all of the sulfhydryl groups. The gel was washed extensively with distilled water and interstitial water was removed by vacuum filtration. The sulfhydryl content of 0.1 g portions of the gel was determined by procedure of Ellman (10). SH-agarose prepared by the cyanogen bromide technique, Open circles; ether-linked SH-agarose, closed circle. 89 " 477 I I I a: 0) o r 2 . a: ‘~ {I ‘5. 19 ‘~..\ ‘1 L, ‘~\~ E Q 2 \ O E? ‘~\. :3 ‘4 " ‘~ F- ‘\‘~ 2 \\ “J .— *_ LC) ‘qu ‘_ I! ‘~ ‘x CD ‘~ C) ¢P~ I: (I) l l l 1 36 so 90 IZO T I M E (DA Y 8) Figure 1. Stability of SH-agarose prepared with an ether linkage and by cyanogen bromide activation. 90 complementary sequence, experiments were undertaken to determine if the complementary sequence can be eluted with formamide and heat without removing the mercurated polynucleotide. Figure 2 Shows an experiment in which unlabeled, mercurated bovine ribosomal RNA (rRNA) was hybridized to unmercurated 3H-labeled cDNA complementary to 18S rRNA ‘ and chromatographed on ether-linked SH-agarose. Approximately 67% of the 3H-cDNA was bound to the column. The column was then washed successively with 99% formamide, 0.25 mM EDTA, 0.2% SDS at room temperature (24°C), 40°C, 50°C, 65°C, and 75°C; finally, the mercurated RNA (along with any 3 H-cDNA remaining hybridized) was eluted with 2-mercaptoethanol. It was necessary to elute the column with starting buffer after each formamide and heat treatment in order to remove all of the released 3H-cDNA; the reason for this is unknown, but it may be that the NaCl in the starting buffer diSplaces nucleic acid which becomes attached to positively charged groups on the agarose. The Tm (temperature at which 50% of the hybrid is disso- ciated) of the 3H-cDNA18S rRNA:Hg-rRNA complex was found to be about 33°C under these conditions; greater than 95% of the bound 3H-cDNA was released at 65°C, and this temperature was used for dissociation of hybrids in subsequent experiments. AS shown in Figure 3, only 0.15% of the 3H-cDNA18$ rRNA was bound to the column and released by formamide at 65°C when chromatographed in the absence of complementary Hg-RNA. Chromatography of a variety of unmercurated nucleic acids, both DNA and RNA, indicated that non- Specific binding to the column under these conditions is in the range of 0.1 to 0.3% (data not shown). Figure 4 demonstrates that little mercurated RNA is released from 91 .NooH an: ooom .m: an voxums oEHu sea on aeaaoo osu awaouau say was an mH.oV HoamnuoouamouoEnu .uom some SH sauna paooom ozu hp poxums aowuwmoa asu um aESHoo onu swsounu can haoumwpoeaa was Mommas magnumum .aws m How asaaoo map wawvaaouuam uoxomn noun? a swsousu nouns paumon wafimeam kn panama mmB ouaumuoasou oau can poaaOum was 36am unnaao may .aeaaoo one ouao saw was mam N~.o . aeafioo oao .m he woumoapaa mosau asu u< .vouooHHoo ones He mN.H mo maowuomum .:\HE m.o mo ouau 36am a up omoumwmtmm poxawfiuuoauo mo aSSHoo Eu o.m x o.o n on ooaaooo can Amen x~.o .Hooz_m m.o .Am.a mov Lom-naua_m.ao.ov summon masseuse as m.o oboe oouoaeo no: coauooos use .oomc so a N sou mom N~.o .Hosz m m.o .Aw.o woe mmomm_m no.0 .uoasnsuoa gem as ~.o ca azmu-wm season so choose sauna n sues censuses»; mp3 .oumaasou on a Eoum ammuos%aoa z< nuwa ponamosuahm .«znoumm .oausss ooom .m ouawam aw nonauomop mm omoumwmumm aoxafiatuoguo ao aosmaquuanuso was .mpocuoz can mamwuouaz SH ponauomop mm poumusouoe pan mammoaopm Ammv suaa poaonaa oz .oomo can opaaashom %n amoumwmumm voxawaouosuo Eouw powwoaou u>_._.<4mm ”-2 3O 4O 20 FRACTION a zoo in» Sucrose gradient sedimentation of F-422 32P-labeled Figure 2. cytoplasmic RNA and location of R-FeLV virus-Specific RNA. Figure 3. 114 Purification of F-422 virus-Specific 238 RNA hybridized to Hg-cDNA by chromatography on SH-Sepharose. Fractions from the sucrose gradient in Figure 2a containing 238 RNA (peak III) were ethanol precipitated, and the RNA was hybridized for 4 h in a 0.2 ml reaction volume to (3H)Hg-cDNA and chromatographed on SH-Sepharose as described in Materials and Methods. At the times marked by the arrows, buffer containing 0.15 M 2-mercaptoethanol was applied to the column. Aliquots (2.5 ul) of each fraction (1.25 ml) were assayed for 32F and 3H by scintillation counting. (a) SH-Sepharose chromatography of the original hybrid- ization reaction. (b) Re-chromatography on SH-Sepharose of the hybrid eluted by 2-mercaptoethanol in (a), after dissociation by formamide and heat as described in Materials and Methods. 115 I 1 ME , O l20~ -‘ I50 80- I00 40 - 50 I —JL 0 I l t'. ‘.“ 6 - <3 - 2 a. :x 4'_’ t) 2 I a. '0 (J 2 - ‘ I00 .3' f). '0 I, - .1 50 I _.’ _ C ‘ IOO - 50 1 20 FRACTION Figure 3. Purification of F-422 virus-Specific 23S RNA hybridized to Hg-cDNA by chromatography on SH-Sepharose. Figure 4. 116 SH-Sepharose chromatography of Hg-cDNA hybridized with 32P-labeled R-FeLV RNA and FLF-3 cytOplasmic RNA. (a) (3H)Hg-cDNA was hybridized with 32P-labeled R-FeLV 288 subunit RNA as described in the text and chromatographed on SH-Sepharose as described in Materials and Methods. (b) (3H)Hg-cDNA was incubated with 32P-labeled uninfected FLF-3 cytoplasmic RNA and chromatographed on SH-Sepharose as in (a). At the times denoted by F, one column volume of 99% formamide, 0.2% SDS was run onto the column; the temperature was raised to 70°C for 4 min; and the column was immediately washed with starting buffer. Starting buffer containing 2-mercaptoethanol was applied at the times marked by ME. “P cm x 10" o 8" F ME .4 ' i 6 II gr II '3 II :1 4*. Ilia . :1 ' 11‘ 5 2' . :"'|'? II 2 '3 80 4O 20 b _ I r F ME i i i F 3 '33 It I‘. r :I -2 I t I I ' I I I ~ , I “I ' t 1 1 I i. . IO 20 30 FRACTION Figure 4. SH-Sepharose chromatography of Hg-cDNA hybridized with 32P-labeled R-FeLV RNA and FLF-3 cyt0plasmic RNA. 311 cm x 118 some of the Hg-cDNA bound to SH-Sepharose is eluted by the formamide and heat step shown in Figure 4. This problem is much reduced when the hybrid is denatured before application to the column, as in Figure 2. In the control experiment shown in Figure 4a, 0.3 ug of Hg-cDNA was hybridized to 3.0 ug of 32p-labeled R-FeLV 233 subunit RNA; 0.36 ug, or 12% of the input RNA was bound to SH-Sepharose along with the Hg-cDNA. This amount of RNA represents complete saturation of the Hg-cDNA. When 7.0 ug of 32 P-labeled uninfected FLF-3 total cytOplasmic RNA was hybridized to 0.3 ug of Hg-cDNA, only 4 ng or 0.06% of the RNA was retained by SH-Sepharose and eluted by formamide and heat. These results demonstrate the specificity of this purification technique, which can be improved Still further by use of the modifications described in Figure 3. Analysis of F-422 "158" and 23S vRNA for 5' Termini The 32P-labeled "158" and 238 vRNA species were each mixed with reovirus RNA (labeled 2r_1 m with (3H)S-adenosylmethionine) and digested with ribonucleases A, T1, and T2. The products were characterized by chromatography on.DEAE-Sephadex, 7 M_urea, as shown in Figure 5 for 238 vRNA. Essentially all of the 3H migrated near the (pU’m)5 marker, which was eXpected Since all of the methyl groups in reovirus RNA are in cap I structures (12). Most of the 32P-labeled material eluted at the position characteristic of 3' mononucleotides. Much Smaller amounts of material eluted with dinucleotides and near (pUm)4 and (pU’m)5 markers. The material eluting with dinucleotides is presumed to be dinucleotides derived from contaminating ribosomal RNA (7). It appears that vRNA has some complementarity with rRNA, since little or no FLF-3 cell (containing no vRNA) rRNA co-purifies 119 .muoxume ma uoaaaoaa ouo3 AmAEDav swaousu EDQV mopauooaoaa nowwao unavawpamum .waauaaoo aoaumaawuawoa %n >ua>wuomoavmu pom pohmmmm mums muoauwam HE Ha.o pan wouooaaoo ones He H.H mo maoauoaum .£\He m haoumEonuaam mo moan scam a up aeaaoo osu awaounu saw was nous 2.5 .w.o ma .Hum.maua 2 No.0 aw Homz z ¢.o ouuz H.o mo Ampam\as omv uaoapauw < .xouanaomum a auw3 wouaawv mp3 ouauxfla coauomau osH .mpoauoz van mamwuoumz papa: ponfiuomop on my use .HH .< ommoou woaonmaum cows noxwe was m ouowwm aw vanauomov mm commensa mmw 03H m .azm oamaooan-mawa> mmm «Norm mo muoapoua aofiumomwu NH van .HH .< ommmzm mo coaunuaaom asaaoo xopmnaom1m mmu 62o //-—0 .HH .< ommwzm mo coaumumaom aEsHoo xovnnaomumoou mm oSu uom .Nmm .aoA_mvaaaA.mvonE cam “NmN .«om moum "Rm .Aunv mumsamocaocosnm maamoamsw "XNH A mmm scum mousuosuum amo woeamoua mo muosvoua cofiumowflv Hm ommmfiosc ozu mo mam%amc< .n ouswfim 126 Eoum mousuosuum mmo bosomoua mo muusooua . m2 :oHumova Hm ammoaosc onu wo mammamc< .m ouswfim 20:04": . 00 00 0? ON - t.-- . . \ - 1-- . ... . r x. — j c. a g c H . __ OOMT . O 4 .0 .. w .. . . ..... now ” 00?. a + + a .. on an oa + C a a 9 9 + s. _ 66833»... Emacs—one l 00. + + Bananas—c Engages... . . pianos... Von _ b — — dzcz '——° WdO 127 .n ouswam a“ mm mama osu mama muoxumz .mowfimoofiosa mg was Nag cam 50A.mvanaa.mvous a“ mp3 mzm msuw>oou osu scum mm map mo unwouoa ocwcumuswwm .NHm .EUA.mvmaaA.mvuhE vow "gem .qom moum .NoH .wm “gm . :mmH: Scum monsoonuum moo vossmoua mo muosvoua nowummwwv Hm mmmmaosc mnu mo mwmhama< .w ouswwm 128 o :mmH: Eouw mousuosuum amo vasomoua mo muosvoua coHumowwv Hm ommmaosc mnu mo mam%amc< .w ouswfim °""° Wd 3 H9 00. CON 00m zo_._.ooom .n ouswfim cw mo oEmm osu who: mwumwcoum .Nma .EGA.mvmmmA.vanE a; £3 .afiuomovau onH .asu wasp a“ vo>Homou maouoamsoo uoa ouos mouosmmosmocoa u.m ovwmooflosz .n ouawam a“ ponwuomov mo x