J'l.":q"" '75 -."\“' Link-b): 3.5%; 12%; “J" >~ .- ' -:E“rY L. I Ln" '- ' 1211"?" ‘2‘ I w t “WY-HI. : é’ififk w” .3 {A . ‘;~. I, 3%. ‘ 'IJ I I ‘W 3p A 1: v' L. w I I ,3 . ‘ 3:5: . ' .L‘ A." ”i :m-{t’ . KW“. “...-if. ' ' _ - ‘ , ..IA 3 ' u.‘\"l".'v'.‘_1{ : . II .121, ' - ‘ I , mugs: . n ‘7’? "5:321” ‘ ' 4 :_ $3.”: gk'A’. I2! 5*?” 7 5:6 II ‘5 |':I;‘ “I "., .'. "M”? Q '. «II I 'N‘i'fi‘. . ; - ”VT-S'J- 7.372: K MICE?“ '2". ”MI: '33“. Y'H'IHIWK' I::EI:I I a '(:. :35. "‘ V: :".:-f 1:. W ,".:...I' ‘ 'l'IW II? \ . . N»? $3.23,}. . , IV“ " i? - A. I,“ , ' .‘I v ' .I‘ .. R). . . {IL ' ' VIE": . ..' I .‘w I 55:4, “ Exp-é": ‘ZJXE'ZE: k :"" :I U‘ -. ‘ " u ‘ ” I.” jy‘if » -r ‘I‘ c Q..." I 63.3%}: " ,ZL 1. .4. ‘4' £01:- 5‘43"" $.52} " “513$. ‘.‘é law: I: ~- 3.2:“. '1. ~. m: .:< ..., . " ”2:4 ?&3 "$.33 \ o. I - vol “ ‘n ‘ Q , 9‘1)" «'33:. IIl‘Hih‘qgt...‘ 'g.'.}::‘|'Il l.‘ .;C"-' .YILM) .‘ '.~- 5- ". .1I ..r‘. '20 I. %1}‘c N' "H. II: I‘ (R 't ‘I I W1?!“ ”Min QI'JfinaaiIS'“ - z 3333"“ gag? . ."V; "JR. ~ @242 “x '.'..‘1 n ' '. ‘é . ' L". Gigi-I . 9’32: “37:5: 323 3%. ' ‘ ":‘n’u‘ V! <4.“ *3" ..v. 4“ I I ‘I. .. ~. .. ..It ',M g u JQI’fi.‘ :‘u 1.2;}. 32:3" " «313‘ . ... JI . ‘ ”3:3 :35 .-I “‘52: ‘r. “1"“ In - I,,II\| ', ‘ I'I,‘ . ._- .4.} Ag- r_ j v . -,;. ..'.:l‘:; W: '\‘-' . ,L . . ' ‘» “VFW .‘l I .. . . ~ .. .. .. . III-.... . stub full" .“ ' . -‘.' ".0 "5’1.“ ‘ .. . .. " IV" «:1» I}. 3 3 ‘;...‘.._~.. : .‘:.. Imam)“ ‘ ' m .' ".I ...: ‘. '..’.' ._ .. J‘ \ "¢{'.I;I“ I ‘I- '.‘II' I' “it"! '. "‘ .I, "I; LIBRARY Michigan Scab University . NOVIKOFF MESSENGER RNA METHYLATION: IMPLICATIONS OF METHYLATION IN PROCESSING BY Marian Maxine Kaehler A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 1978 .fl%:\ ABSTRACT »\&;;£>\ NOVIKOFF MESSENGER RNA METHYLATION: L£\ IMPLICATIONS OF METHYLATION IN PROCESSING By Marian Maxine Kaehler The specific function(s) of the 5'-terminal cap structure (cap 1, m7GpppN'mpN"p; and cap 2, m7GpppN'mpN"mpr) in viral and eukaryotic messenger RNA are unknown. Ribosome binding studies and cell-free translational assays have shown that cap structures are involved in the recognition and translation of mRNA. However, the data suggests that caps function as facilitators rather than requisites for these processes. Cap structures have also been observed in nuclear poly (A)-oontaining RNA, suggesting an involvement of capping and methyla- tion in processing. We have investigated this possibility via two approaches: 1) kinetic analysis of the formation of cap structures; and 2) in zizg perturbation of mRNA methylation by S-tubercidinylhomo- cysteine. Novikoff hepatoma cells were maintained in culture for these experiments. Methylation was monitored by incorporation of L[3H- methyl]methionine; synthesis was measured by [U-1u 32F}- Cjuridine or [ orthophosphate incorporation. RNA was isolated by standard procedures, including proteinase K digestion prior to extraction with phenol- chloroform. Poly (A)-containing RNA was isolated by oligo(dT)-cellulose affinity chromatography. Sucrose gradient sedimentation analysis was employed to determine the size of the mRNA. Analysis of the methylation patterns of mRNA involved specific nuclease digestions followed by various chromatographic procedures. RNase T2 effectively degraded mRNA to mononucleotides plus the RNase- resistant cap structures. These digestion products were separated by DEAE-Sephadex (7 M urea) chromatography prior to further analysis. Cap structures were analyzed either as intact caps by Partisil-SAX high speed liquid chromatography (HSLC) or by further degradation to mononucleosides, which were resolved by Aminex A-S HSLC. The kinetics of cap formation were studied by monitoring incor- poration of L-[BH-methyllmethionine into specific sites of mRNA as a function of labeling time. After a short labeling period of 20 min, approximately 80% of the label incorporated into caps was located at the penultimate nucleoside to the pyrophosphate, i.e. in the (N"m) position of cap 2 structures. The ratio of cap 2/cap 1 was observed to change with time, and the amount of internal N6- methyladenosine decreased, relative to cap, with longer labeling times. These results are consistent with a model in which methylation at three sites - generating 7-methy1guanosine, the first Z'egemethyl- nucleoside (N'm), and internal N6-methyladenosine - occur in the nucleus. The second 2'-Q¢methy1ation (N"m) appears to be a cytoplasmic event. Perturbation of methylation in Novikoff cells occurred in zigg in the presence of S-tubercidinylhomocysteine (STH). Characterization of the partially methylated mRNA indicated that all sites of methyla- tion, except the 7-position of guanine, were inhibited. STH diminished the levels of internal N6-methyladenosine and of 2'ggfmethylnucleosides in cap structures. The cytoplasmic presence of cap structures devoid of Z'figgmethylation, i.e. cap zero (m7GpppN'), suggested that ribose methylation is not required for processing and transport of mRNA. The base composition of cap 1 structures from normal and STH-exposed mRNAs were considerably different. The composition of caps zero and 1 from inhibited samples were comparable, however, indicating that inhibition by STH at the N' position was not base specific. In order to assess the functional nature of cap zero-bearing mRNA molecules, monosomal and polysomal poly (A)-containing RNA was isolated from both normal and inhibited cells. Cap zero structures were the predominant cap species from both monosomal and polysomal mRNAs. These results suggest that 2'—Q¢methylation is not requisite for ribosome binding and subsequent translation. Nuclear poly (A)-containing RNA was also analyzed for its methylation patterns. The base composition of nuclear cap 1 structures was comparable to the corresponding cytoplasmic cap 1 structures. Cap zero sturctures were present in nuclear RNA isolated only from STH-inhibited cells. No accumulation of totally unmethylated cap structures was observed. Dedicated to Dick for his love and tolerance during my graduate training. 11 ACKNOWLEDGEMENTS I would like to express my appreciation to the Biochemistry Department personnel for the constructive and congenial atmosphere maintained in the department during my graduate work. The people with whom I have interacted have been very cooperative and helpful. In particular I thank the members of my guidance committee, Drs. Pam Fraker, Debbie Delmer, Ron Patterson, John Boezi and Arnold Revzin, for their insight and advice. A special thanks is extended to my coworkers in Dr. Fritz Rottman's laboratory - Karen Friderici, Dr. John Nilson, Bruce Coffin, Drs. Sarah Stuart, Ron Desrosiers and Arlen Thomason, Nancy Sasavage and Don Bodeau. Their friendships are very valuable to me. I especially thank Karen for her continual assistance and helpful advice throughout my graduate career. Part II of this thesis was a collaborative effort with Karen, who was the principal investigator in that study. Parts III and IV of the dissertation describe the research our laboratory performed in collaboration with Dr. James Coward of Yale University. He has been most helpful and communicative during our collaboration. iii I am especially grateful to Fritz for his role in the development of my career. He has been an excellent mentor and friend. His insight and expertise has been invaluable in many helpful discussions. In addition, his sensitivity to others has resulted in very positive interactions with people within and outside the laboratory. I have been fortunate to work with Fritz and hope to continue our friendship in the future. Parts II and III are reprinted with permission from Biochemistry, lg, 523A (1976) and Biochemsitry 1Q, 5077 (1977). Copyright is by the American Chemical Society. iv TABLE OF CONTENTS LIST OF TABLES . LIST OF FIGURES LIST OF ABBREVIATIONS PART I LITERATURE SURVEY . I. Methylation of Viral and Eukaryotic Messenger RNA . mRNA Methylation: An Historical Perspective . Distribution of Cap Structures Internal Methylation in mRNA and in hnRNA . Enzymatic Mechanisms for Capping and Methylation II. Possible Functions of Capping and Methylation A. Involvement of Methylated Cap Structures in mRNA Translation . . . . . . . . . . Translation studies . . . . . ‘. Ribosome binding studies . Analog inhibition studies . . . . Regulatory Function at translational levels Role of Methylation in mRNA Processing . . Theoretical and technical considerations Page viii ix xi 11 12 18 18 18 20 24 27 29 3O Evidence implicating methylation in processing 31 Inhibition of mRNA Methylation in vivo . Increased resistance to degradation III. An Argument for Control of Genetic Expression at the Posttranscriptional Level A Speculative "Control Hierarchy" References PART II THE KINETICS OF NOVIKOFF CYTOPLASMIC MESSENGER RNA METHYLATION Abstract Introduction Materials and Methods Cell Culture and Labeling Conditions . Isolation and Characterization of Poly (A)-Containing Cytoplasmic mRNA . . . Nucleotide Pyrophosphatase Treatment of Whole mRNA Preparation of mRNA for Methyl Nucleoside Distributional Analysis . . . . . . Acid Hydrolysis . Results Discussion References . PART III IN VIVO INHIBITION OF NOVIKOFF CYTOPLASMIC MESSENGER RNA METHYLATION BY S-TUBERCIDINYLHOMOCYSTEINE . . Abstract . . . . . . . . . Introduction . . . . . . . . . . vi 3” H1 42 H7 “9 6O 61 62 63 63 6M 65 66 67 67 87 93 95 96 96 Materials and Methods . . . . . . . . . . . . 98 Cell Culture and Labeling Conditions . . . . . 98 Isolation and Characterization of Poly (A)-Containing RNA 0 O O O O C U C O O O O O I O I 99 Enzymatic and Acid Degradation of Poly (A)-Containing RNA . . . . . . . . . . . . . . . . 100 Distribution Analysis of Methylation . . . . . 101 Results . . . . . . . . . . . . . . . . . 101 Discussion . . . . . . . . . . . . . . . . 112 References . . . . . . . . . . . . . . . . 120 PART IV CYTOPLASMIC LOCATION OF UNDERMETHYLATED MESSENGER RNA MOLECULES 123 Abstract . . . . . . . . . . . . . . . . 12H Introduction . . . . . . . . . . . . . . . 12H Materials and Methods . . . . . . . . . . . . 126 Cell Culture and Labeling Conditions . . . . . 126 Isolation of Poly (A)-Containing RNA , . . . . 126 Enzymatic Digestion and Analysis of Poly (A)-Containing RNA . . . . . . . . . . . . 128 Results . . . . . . . . . . . . . . . . . 130 Discussion . . . . . . . . . . . . . . . . 1A8 References . . . . . . . . . . . . . . . . 153 vii LIST OF TABLES Table Page Part I , . . . . . . . . . . . . . . . . . . 1 I. Distribution of Cap Sturctures and Internal Methylation in Eukaryotic mRNA . - - . . - . 6 II. Distribution of Cap Structures and Internal Methylation in Viral RNAs - . - - - - - ~ - 8 Part II. . . . . . . . . . . . . . . . . . . 50 I. Amount of Internal m6A per Average mRNA: Variation with Time - - - ° ' ° ° 76 II. Distribution of 3H-Methylnucleosides in Various Positions of Cap 1 and Cap 2 Structures as a Function of Labeling Time - - - - - - - - - 86 Part III. . . . . . . . . . . . . . . . . . 95 I. Effect of STH on 1uC-Uridine and L-[methyl-BH] Methionine Incorporation into Cytoplasmic RNA - ' 103 II. Distribution of 3H-Methyl CPM in Cap Structures of Cytoplasmic Poly (A)- -Containing RNA from Normal and STE-Treated Novikoff Cells - - - 110 Part IV. . . . . . . . . . . . . . . . . . 123 I. Effect of STH, SAH, Homocysteine and Tubercidin on Methylation of Poly (A)-Containing RNA . . . 132 II. Distribution of 3H-Methyl Radioactivity within Cap Structures Determined by Partisil-SAX Chromatography 1A7 viii LIST OF FIGURES Figure Part I. 1. The S'-terminal end of eukaryotic mRNA 2. The structure of S-tubercidinylhomocysteine, the 7-deaza analog of S-adenosylhomocysteine . . . Part II. 1. HSLC resolution on Pellionex-WAX of XOR digestion products from mRNA which had previously been treated with nucleotide pyrophosphatase and alkaline phosphatase . . . . . . . . . . 2. Change in ratio of N'mpr to .N'mpN"mpN in mRNA with time . . . . . . . . . 3. DEAE-Sephadex column separation of RNase T2 and alkaline phosphatase digestion products from whole mRNA . . . . . . . . . . . A. - sucrose gradient profiles of poly (A)- cogtaining cytoplasmic RNA . - - 5. The distribution of methylnucleosides in cap 1 structures . . . . . . . . . . . 6. Methylnucleoside distribution analysis of N"m nucleoside of mRNA labeled for 2“ h by Aminex A-S HSLC o o o o o o o o o o 0 Part III. . . . . . . . . . . . . . . 1. DEAE-Sephadex column separation of RNase T2 and alkaline phosphatase digestion products from whole mRNA 0 O O O O O O O o 0 O I 0 ix Page 37 6O 7O 73 75 79 81 8A 95 106 Part IV. 1. Distribution of 3H-radioactivity in cap zero species Analysis of cap 1 structures obtained from mRNA of STE-treated cells , , , , , , Absorbance profiles of postnuclear supernatant sedimented through 10-AO% sucrose gradients DEAE-Sephadex (7 M urea) elution profiles of RNase T2 digestion products from poly (A)—containing RNAs Acetylated DBAE-cellulose chromatography of RNase T2 digestion products from STH-inhibited cytoplasmic poly (A)-containing RNA . . DEAE-Sephadex (7 M urea) elution profile of the DBAE- cellulose-bound fraction from RNase T2 digestion of poly (A)-containing RNA . . . . . . DBAE-cellulose profile of the highly charged material from STH-treated cytoplasmic mRNA . . . . . . 109 114 123 134 137 1ND 1N2 1MB LIST OF ABBREVIATIONS DBAE DEAE DNase EDTA GDP GMP GTP hnRNA HSLC MeZSO,DMSO m7G m7gua* mN mRNA N" xi dihydroxyborylaminoethyl diethylaminoethyl deoxyribonuclease ethylenediaminetetraacetic acid guanosine-5'-diphosphate guanosine-S'-monophosphate guanosine-S'-triphosphate heterogeneous nuclear RNA high speed liquid chromatography N6-methyladenosine N6,2'-Q¢dimethyladenosine ‘ S-methylcytosine dimethylsulfoxide 7-methy1guanosine ring-opened 7-methy1guanine base methylated nucleosides messenger RNA nucleosides nucleoside adjacent to pyrophosphates in S'-caps penultimate nucleoside to pyrophos- phates in 5'caps Nm RNase rRNA SAH SAM SDS STH Tris tRNA UV VSV xii 2'-Q:methylnucleoside ribonuclease ribosomal RNA S—adenosylhomocysteine S-adenosylmethionine sodium dodecyl sulfate S-tubercidinylhomocysteine tris(hydroxymethyl)aminomethane transfer RNA ultraviolet vesicular stomatitis virus LITERATURE SURVEY The presence of methylnucleotides in eukaryotic messenger RNA (mRNA) was first discovered in 197". Since that time several investi- gators have been actively studying mRNA methylation patterns with respect to the structural and functional properties of this posttrans- criptional modification. This review focuses primarily on those investigations which have attempted to elucidate the function of methyl groups on viral and eukaryotic messenger RNA molecules. Reviews by Shatkin (1976) and Rottman (1976;1978) provide comprehen- sive discussions concerning the discovery and initial characterizations of messenger RNA methylation. I. Methylation of Viral and Eukaryotic Messenger RNA mRNA Methylation: An Historical Perspective Perry and Kelley (197“) first published evidence that eukaryotic mRNA was methylated. Poly (A)-containing RNA was isolated from mouse L cells and found to contain 2.2 methyl groups per 1000 nucleo- tides, or approximately one sixth of the level of ribosomal RNA methylation. Upon alkaline hydrolysis of the mRNA, a significant proportion of the methylated material chromatographed as a highly charged oligonucleotide. In a simultaneous and independent study, Desrosiers, 32 El- (197”) reported that Novikoff hepatoma polysomal mRNA was also methylated. The methylated components included the four common 2'-Qfmethy1nucleosides, N6-methyladenosine (m6A) and 1 2 an unidentified nucleoside (later shown to be 7-methylguanosine, Desrosiers, 22 al., 1975). This methyl distribution was distinct from, and much simpler than, the methylation patterns of either ribosomal or transfer RNA from Novikoff cells. In addition, studies using cytoplasmic polyhedrosis virus (Muira, ‘gt‘al., 1974a; Furuichi, 197A), reovirus (Muira, 32 al., 197ub; Shatkin, 197A) and vaccinia virus (Wei and Moss, 197A) indicated that methylation was also occurring in viral systems, and that the methylating activity was present within the virions. Localization of the methylase activity in viruses suggested that the function of methyl groups was sufficiently important to be retained in such limited genomes, and provided a source from which to isolate the enzymes responsible for methylation. Analysis of the methylated components in both eukaryotic and viral mRNAs generated seemingly contradictory results. The alkali- resistant oligonucleotide fraction was also resistant to RNase T2 digestion, but the low level of methyl groups present ruled out its identification as a series of 2'-Qfmethylnucleotides. Alkaline phosphatase digestion reduced the charge of the material by two, indicating a single phosphate was released. Attempts to phosphorylate the oligonucleotide with polynucleotide kinase were unsuccessful, suggesting that the 5'-hydroxyl group was inaccessible. These data were consistent, however, with the general structure proposed by Rottman, gt El- (197”), in which 7-methylguanosine is linked via an inverted 5'-5' pyrophosphate bond to one or two 2':Q-methy1ated nucleotides (Figure 1). These atypical nucleotide structures were termed "caps", and were structurally similar to the 5'-termini of Figure 1. The 5'-terminal end of eukaryotic mRNA. The nucleoside 7-methylguanosine is attached by an unusual 5'-5'-pyro- phosphate linkage containing three phosphates to a 2'-ngethyl- nucleoside with the base indicated as N'. This terminus is called cap 1. An additional adjacent 2'59rmethylnucleoside with the base indicated as N", is present in cap 2 structures. (Taken from Rottman, 1978). u 5'-TERMINAL END CH3 0 II “I“ g > HZN ‘\N N CH /ZO I O I n O OH OH I ‘1’ o-p-o—p-o n O l O I 0—0-11- II o I to O X 5 low molecular weight nuclear RNAs (Reddy, 92 21: 197A; Ro-Choi, e_t 2;. 197a). Distribution of Cap Structures Several eukaryotic and viral messenger RNAs and virion RNAs have been shown to possess capped 5'-termini (refer to Table I). Three types of cap structures have been observed: cap zero, which contains no 2'1Q-methylation (m7GpppN'); cap 1, in which the pen- ultimate nucleoside is ribose methylated (m7GpppN'mpN); and cap 2, which contains two adjacent 2'19-methylnucleosides (m7GpppN'mpN"mpN). All of these structures are characterized by the 5'-5' pyrophosphate linkage of 7-methylguanosine to the penultimate nucleoside (N') of the ribopolymer. Tables I and II summarize the distribution of methylation in RNA from both eukaryotic and viral sources, respec- tively. Cap zero structures are present only in a few viruses and in some lower eukaryotes. All mammalian and most viral systems studied to date contain cap 1 and/or cap 2 structures. The penultimate nucleoside in all viral mRNAs, and in some lower eukaryotic mRNAs, is always a purine, suggesting that these molecules are derived from primary transcription products. In contrast, mammalian mRNA contains both purines and pyrimidines at the N' position. The presence of pyrimidines requires that some 5'-termini must be generated from sites internal to the 5'-termini of the primary transcripts. In addition to the four common 2'f9-methylnucleosides, the dimethylated structure N6,2'-9-dimethyladenosine (m6Am) is also present at N' of caps 1 and 2. This doubly methylated nucleoside has not been found at the adjacent position N", however. Acsm_v asmucmsmnmm new cumsxooq ucomn¢ so u a=z s¢ma u E.z acomn< «zma canoawnuwnnmm Amsmpv Lossocm new magma acmmn< pcomoum uconosm pcomn< av “8.2aaouhav A.zqoac~av moocmsomom HmccoucH Mlmmm Flmmm MMMMImmm_ mmmmmmmm mmc as umsamocm o n .nu.|| Aspm.v .Hm um .csosm Asso.v .Hm um .mmoz .1..hwsm_v .Hm 8m .86: 7.2m8m.v cfixumcm cam H>mm Isa: .8... am .memhmm Amwmpv .Hm no .mnofimosmon Ampapv Loamma 88m Cause Ausmpv cmflnmumx 88m mmwmw Amsmpv .Hm no .opuoaaoso Amwm—v memod 88m 2800 “mea.v .mm no .2nnom scomna omE <08 scmmna ucomosm e:.ao.ao.a< u 3:2 aa.ao.ao.a< u 8:2 .8.2 ucmmosm sa.ao.so.aa u 8:2 a=.ao.ec.a< u 8:2 so u 9:2 ucommgm e=.eo.au.a< u 8:2 a=.ao.so.a< u 3:2 8m8 Abhmpv mmoz 88m ocoom “Amsm.v mmoz new sex Amumpv amass new anowsssm Amwopv ocoum 88m occoaoo “osmpv xooaan new msunuaoom “Amem.v .mm.mm..scofisgsa Acmwmpv omncoo Iaoxcomnm 88m sufiox 5.: WWI. ...... Ashmpv .mm.ww .mouonm Ammapv .Hm um .Lossaz Asem.v .mm.mm..xfiszsoosu Ampm.v ”cossm Amsm.v xocsa Amwm—v camaawu “Ammsopv umscoo -mecomia use case» Aesopv .mm.mm .muqsmmma pcmmna unmmn¢ scmmp< mooconouom cossmawnsmz HmcnoucH Aavaas< adage 8 8Ao.«vaaquwa s . 8:28A<08 o «Vaaaoh8 .Q.z ofiavooaoops mscfiasoan.m 88 .Q.z scanausa auqqaupa zonqwha 8 nscw8noen.m <2: coHLa> o csv “assoc .-mom 88V mo: msgs> caseHm omficfioom> “cmwamaam: manonumnzaom 088mmHaoa2o ”poomcH mnmomflv maummosoz unmumaooamm Am oswmnmv maoonmn noom unma>< Lhasa @082 80889 88:03 mamogooz ooomnoa ouaaaoumm owmmoz amazon oamnoz 80288080 owmmoz whamma¢ oammoz ooomnoa oammoz osonm "pcmam msna> no waxy mm¢2m Hmufi> 8H cofiumazcuoz HangoucH 88m mossuossum one no cowusnanumfia HH wanna .HHoo boos no msoaosc ca mouoodaaomo .Haoo poo: no Emmaqou2o ca nouooAqumc .oLoo 80HLH> 2n onuH> mm.oonfioozuc2m «228 .oocaasouoc 802 u .n.2M .mIH oHnos .Amwmrv cosouom aoou 880m ooma>os 8H coaamccm Amem.v .mmrmmrxiocsaoa .a.z .a.z scoops amo A+vowfiom Ampm. .owopv .mm.mm..comoeone 2 8 Aw.8 acomnm goo A+v8 cannaba A+v<2mmn ownvcam nowmkylhao no .omom “Aowmpv .Ho no .uconsocom <08 .Q.2 8uaqqo>8 A+v<2mnm oaaoxsoa o, ocfiosa 2ocoaoz Amsopv 2xmmoxoaox mmmlwmoaooq .a.2 .o.2 8 Aosm.v .Hm so mLonomenlwAzsmpv cflxuonm pcomn< Aevoqaoaaqcha sensuoua ¢2mmc omsnfi>oom Apsopv .Ho no .mmoz “Ao>¢Fv cosufiom new hxmoxunom < 8 28V=28A<08.o.8 I <2Q oxoaoawm moano: Aoem.v somwnmnecm mac: 8 "Amhapv .Hm no .Loseom om8.<08 ASV:2Q8A<.8 I ocob< 2.88coov HH magma 10 In some cases, a distinction must be maintained between virion and viral messenger RNA termini (of Table II). The presence of cap 2 structures in viral mRNA, but not in virion RNA, of reovirus (Desrosiers, gt al., 1976) suggests that host methylases may be involved in modifying the N" nucleotides. Some virion RNAs contain no cap structures at all, but the viral mRNAs are fully capped. Such is the case for RNA viruses whose genome consists of single- stranded, minus RNA strands (vesicular stomatitis virus (VSV), Newcastle disease virus and influenza virus), and for feline leukemia virus. The latter may be due to the defective nature of this virion (Thomason, 23 al., 1976; Rottman, 1976). In addition, the picorno- viruses (poliovirus, foot-and-mouth-disease virus, and encephalomyo- carditis virus) thus far appear to not be capped either in the virion or as messenger RNA. The significance of this is unknown. Cap structures have also been identified on the 5'-termini of eukaryotic and viral-specific heterogeneous nuclear RNA (hnRNA). hnRNA of mouse L cells was shown to contain only cap 1 structures (Perry, gt al., 1975). Similarly, cap 1 structures are found in hnRNA of HeLa cells (Salditt-Georgieff, gt al., 1976) and Erlich ascites carcinoma cells (BaJszar,ygt_§l., 1976). Nuclear sequences specific for SVHO (Lavi and Shatkin, 1975) and adenovirus (Sommers, 33 al., 1976; McGuire, gt al., 1976) are also capped. The composition of cap 1 structures in the mRNA and hnRNA of both mouse L cells (Schibler, gt_gl., 1977; Perry, gt al., 1975) and adenovirus-specific RNA (Sommer, gt al., 1976) are very similar, consistent with a precursor-product relationship between these two RNA populations. No cap 2 structures have been identified in nuclear RNA. 11 Internal Methylation in mRNA and in hnRNA In addition to methylated cap structures, several mRNAs contain base-methylated nucleosides. These modified nucleosides are not resistant to either RNase T2 digestion or alkaline hydrolysis, and can therefore be resolved from cap structures by chromatography on DEAE-Sephadex (7 M urea). In nearly all mRNAs and hnRNAs containing this modification, the base-methylated nucleoside has been identified as exclusively N-6-methyladenosine (m6A). The exceptions - cultured hamster kidney cells (Dubin and Tayler, 1975), adenovirus (Sommer, _e_t_'._ 31., 1976) and Sindbus virus (Dubin and Stollar, 1975) have been reported to contain 5-methylcytosine (m5C) in addition to m6A. Most higher eukaryotic mRNA and hnRNA molecules contain base- modified methylnucleosides. These modified bases have been located internal to the 3'-termina1 poly (A) segment and to the 5'-terminal cap structure (Sommer, gt al., 1976; Desrosiers, gt al., 1975; Perry, 32 al., 1975). It is not known yet whether the methylated bases are within the coding regions of mRNA. However, the location of m6A appears to be sequence specific in HeLa cells: Wei and Moss (1977) examined the nucleotides adjacent to m6A and found only two sequences, Gpm6ApC and Apm6ApC. Of particular significance is the fact that these same two sequences were also detected in mouse L cells (Schibler, 32 al., 1977) and in avian sarcoma virus (Dimock and Stoltzfus, 1977), suggesting conservation of this sequence. The specificity of the location of m6A may be related to its function, which is unknown. Internal methylation sites in Rous sarcoma virus have been located, however, in the 3'-terminal portion of the genome, in the region containing the sarc gene (Beeman and Keith, 1977). 12 The amount of internal m6A in both mRNA and hnRNA has been correlated with molecular size (Lavi, 22 al., 1977; Sommer, gt_§l., 1976; Perry and Kelley, 1976; Friderici, gt al., 1976). Longer polynucleotides appear to contain more base-methylated residues than do shorter molecules. Some relatively small mammalian mRNAs, such as globin mRNA (Perry and Sherrer, 1975; Adams, gt al., 1978) and histone mRNA (Stein, 32 al., 1977; Moss, gt 31., 1977) lack internal base modifications. Whether this reflects a simple size correlation, or is significant for some other reason, is unknown. This question has been complicated by the recent demonstration that the apparent levels of labeled m6A in HeLa cells decreased with labeling time, an observation which may be due to demethylase activity, processing, and/or variation in turnover rates of mRNA subpopulations (Sommer, 23.2l'v 1978). The presence of m6A in viral RNA sequences appears to depend on whether these molecules are synthesized in the nucleus of host cells (Rottman, 1978). This distinction is consistent with the hypothesis that N6-methylation of adenosine is a nuclear event, and that m6A may be significant in processing and/or transport. Lower eukaryotic mRNA also lacks internal base methylation: no modified nucleosides except cap structures have been identified in the slime mold Dictostelium (Dotten, gt al., 1976) or yeast (Sripati, 32 al,, 1976; DeKloet and Andrean, 1976). Enzymatic Mechanisms for Capping and Methylation The sequential events of capping and methylation have been studied primarily in viruses. Elucidation of enzymatic mechanisms 13 for cap formation has been facilitated by two aspects of viral systems; 1) the virions themselves contain all the enzymes which are necessary to generate capped and methylated mRNA, thus providing a source for isolation and purification of the activities; and 2) unmethylated viral RNA, used as a substrate in these studies, can be synthesized in_!itgg in the presence of S-adenosylhomocysteine (SAH), an analog of the in zigg methyl donor S-adenosylmethionine (SAM). Two basic mechanisms for capping and methylation have been postulated to function in viral systems (for details, of reviews: Shatkin, 1976; Rottman, 1976, 1978). The first has been observed to function in vaccinia virus (Martin and Moss, 1975b; Moss, 23 al., 1976) and in reovirus (Furuichi, §t_§l., 1976). It involves the transfer of a GMP residue from GTP to an RNA molecule terminated by a 5'-diphosphate, and may be depicted as follows: 1) Cleavage by polynucleotide S'-triphosphatase activity to generate 5'-diphosphate termini: pppN'pN"... ++ ppN'pN"... + Pi 2) Capping by mRNA guanylyltransferase: nan u as ppN'pN"... + GTP (PPDG)-*+ GpppN'pN"... + PPi 3) Methylation by mRNA (guanine-7-)methyltransferase: a a GpppN'pN"... + SAM 42- m7GpppN'pN"... + SAH A) Ribose methylation at N' by mRNA (nucleoside-2'-)methyl- transferase: n a m7GpppN'pN"... + SAM ++ m7GpppN'mpN"... + SAH Polynucleotide 5'-triphosphatase has been purified and char- acterized (Tutas and Paoletti, 1977). Subsequently these investi- gators have demonstrated that this enzyme activity is induced in 1H HeLa cells within one hour after infection with vaccinia virus. Its induction was shown to be dependent upon gg_gggg RNA and protein synthesis, but independent of DNA synthesis, suggesting this enzyme is a prereplicative or "early" viral product (Tutas and Paoletti, 1978). Moss and his colleagues have identified activities in vaccinia virus which correspond to each of the other enzymes involved in the capping and methylation sequence depicted above (Ensinger, gt al., 1975). mRNA guanylyltransferase and mRNA (guanine-7-)methyl- transferase activities co-purify from vaccinia virions as a single enzyme of molecular weight 127,000 (Martin 25:21:: 1975a). This enzyme contains molar amounts of two subunits whose molecular weights are 95,000 and 31,1100 (Martin e_t a_l., 1975a). Subsequent char- acterization of these enzymatic activities (Martin and Moss, 1975, 1976; Moss, gt al., 1976a) permitted determination of the above reaction sequence. Boone, gt al. (1977) have also isolated the co-purifying activities from infected HeLa cells; their subsequent studies provided evidence that guanylyltransferase and the methyl- transferases are prereplicative viral gene products. Moss (1977) has also demonstrated that the viral enzymes can be used to modify heterologous mRNAs, a technique which should be useful both for identification of 5'-terminal mRNA structures and for investigation of cap function(s). Initial studies with reovirus by Furuichi and Shatkin (1976) indicated that short (<15 nucleotides) nascent reoviral RNAs, synthe- sized in gitgg, were capped and methylated. These investigators initially concluded that RNA polymerase activity in reovirus was 15 coupled to capping and methylation. More recent studies, however, have indicated that RNA polymerase functions independently of guanylyl- transferase and the methyltransferases, and vice versa (Furuichi and Shatkin, 1977). Carter (1977) has reported that reoviral methyl- transferases are also capable of modifying the 5'-termini of single- stranded oligonucleotides which possess GpppG at their 5'- ends. These oligomers are present in the virion, and their function is unknown. A second distinct mechanism for capping has been identified in the formation of capped VSV RNA (Abraham 33 al., 1975; Colonno and Banerjee, 1976; Testa and Banerjee, 1977). This mechanism differs from that described for vaccinia virus and reovirus in two important features: 1) both the a- and B-phosphates of GTP are transferred during capping; and 2) 2'-97methylation of the penultimate nucleoside N' occurs prior to methylation of the 7-position of guanine. The mechanism is outlined below: 1) Cleavage reaction to generate monophosphorylated 5'-termini: pppN'pN"... ++ pN'pN"... + PPi (or ppprN...NpN'pN" ++ ppprNp... + pN'pN"... 2) Capping by transfer of guanosine diphosphate to the RNA chain: pN'pN"pN... + 333G +*' GaspN'pN"pN... 3) Ribose methylation at N' by mRNA (nucleoside-2'-)methyltransferase: GBBpN'pN"pN... + SAM a“: G33pN'mpN"pN... + SAH A) Methylation by mRNA (guanine-7-)methyltransferase: GpppN'mpN"pN. . . + SAM ++ In7 GpppN'mpN"pN... + SAH The enzymes involved in this reaction scheme have not been purified. The capping and methylase activities are known to be 16 associated with the virion ribonucleoprotein core (Rhodes 93 al., 197A) and appear to be transcription-dependent (Abraham and Banerjee, 1976). Hefti and Bishop (1976) have suggested that the guanylyl— transferase from VSV possesses sequence specificity for the donor substrate mRNA. The two methyltransferase activities involved in the reaction scheme indicated above have recently been identified in purified virions of VSV (Testa and Banerjee, 1977). The concentration of SAM, the methyl donor, appeared to determine the number and location of the methyl groups transferred to cap structures. Limiting SAM concentrations resulted in only 2'e9-methylation of the penultimate nucleoside, whereas saturating SAM concentrations permitted base methylation at the 7-position of guanine (Testa and Banerjee, 1977). These methyltransferase activities have been identified in virions purified from hamster, mouse and human host cells (Testa and Banerjee, 1977). Since vesticular stomatitis virus replicates in the nucleus of its host cells, and since its mRNA appears to be generated from large nuclear precursors (Colonno, 92 al., 1976), this second mechanism was presumed most likely to be occurring in eukaryotes. Formation of cap structures onto a monophosphorylated terminus allowed for capping at sites located internally from the 5'-terminus of an hnRNA molecule. Such a mechanism is necessary owing to the presence of pyrimidines at the N' position of cap structures, since all trans- cription is believed to initiate with purines (Chambon, 197A; Schmincke, 33 al., 1976; Schibler and Perry, 1976). Investigation of the 5'- termini of hnRNA in mouse L cells, however, indicated that approximately 17 201 of the hnRNA population was terminated by diphosphates (Schibler and Perry, 1976; Schibler, gt al., 1977). Furthermore, the base distribution of the diphosphorylated termini was very similar to the composition of N' nucleosides in cap structures from mRNA. These findings suggested that the capping mechanism described for vaccinia viral RNAs might be functional in eukaryotes as well. Eukaryotic methylase activities have been observed in nuclear homogenates of HeLa cells (Groner and Hurowitz, 1975), mouse L cells (Winicov and Perry, 1976) and mouse myeloma (MOPC-21) cells (S. Stuart, unpublished observations). Groner, gt El: (1978) recently identified RNA polymerase II transcripts as the substrates for the capping and methylating activities present in HeLa nuclear homo- genates. These investigators further demonstrated that theEB-phosphate of the pyrophosphate bridge originates from the RNA chain (Groner, at al., 1978). These results are consistent with the capping mechanism which occurs in vaccinia virus and reovirus. An RNA (guanine-7-)methyltransferase had been partially purified from the cytoplasm of HeLa cells (Ensinger and Moss, 1976). The presence of this activity in the postnuclear supernatant was incon- sistent with the presumed nuclear location of capping and 7-methyla- tion of guanine, but the authors acknowledged the possibility of nuclear leakage during cell lysis. The enzyme was capable of modifying vaccinia viral mRNA, synthetic 5'-diphosphate-terminated ribopolymers, and GpppG, indicating its substrate specificity is comparable to the mRNA (guanine-7-)methyltransferase isolated from vaccinia virions. 18 II. Possible Function(s) of Capping and Methylation A. Involvement of Methylated Cap Structures in mRNA Translation Translation Studies. Capping and methylation of both viral and eukaryotic mRNA suggested that these posttranscriptional modifications might function at the level of translation. The influence of methylation on trans- lational efficiency of mRNA was first demonstrated by Both, 33 El- (1975). Reoviral and vesicular stomatitis viral mRNAs were synthe- sized in zitgg in the absence and presence of SAM, in order to generate unmethylated and methylated viral mRNA. The unmethylated mRNA was translated with lesser efficiency in wheat germ extracts than was methylated mRNA. Addition of S-adenosylhomocysteine (SAH) to the cell-free extract further reduced the translational efficiency of unmethylated mRNAs, suggesting that SAM-mediated methylation of this mRNA might have occurred in the wheat germ extract. Positive identification of the endogenous methyltransferase activity was made by analysis of unmethylated mRNA after incubation in the wheat germ system containing [BR-methyl]-labeled SAM (Both, §t_al., 1975). These results showed that the ability of mRNA molecules to direct protein synthesis was dependent upon the methylated state of the mRNA. Muthukrishnan and colleagues (Muthukrishnan, 32 al., 1975a) subsequently showed that the endogenous methylation produced cap zero structures, m7GpppN', at the 5'-terminus of unmethylated mRNA, suggesting that translation of viral mRNA was specifically dependent upon the presence of 7-methylguanosine in capped mRNA. Comparable translational studies using capped and fully-methylated 7 mRNAs were performed after removal of the 5'-terminal m G by periodate 19 oxidation and B—elimination (Muthukrishnan, at al., 1975a). Decreased translation efficiency in cell-free protein synthesizing extracts was observed following B-elimination of rabbit reticulocyte mRNA (Muthukrishnan, 32 al., 1975a; Rose and Lodish, 1976), brine shrimp mRNA (Muthukrishnan, at al., 1975b), bovine parathyroid mRNA (Kemper, 1976), brome mosaic virus RNA-A (Shih, §t_al., 1976) and reoviral mRNA (Muthukrishnan, 23 al., 1975a; Samuel and Lewin, 1976). In contrast, no difference in translational ability was observed after B-oxidizing satellite tobacco necrosis viral RNA (Kemper, 1976; Roman, 32 al., 1976), an uncapped RNA. 7G has The harsh conditions required for B-elimination of m been criticized with respect to the nonspecific alterations of mRNA structure which may occur during the procedure (Rose and Lodish, 1976). Alternative methods of m7G removal involving enzymes have been reported to be more specific and less damaging to the mRNA chain (Zan-Kowalczewska, 33 al., 1977; Abraham and Pihl, 1977). Enzymatic removal of pm7G from the 5'-terminus of reovirus, rabbit globin and brine shrimp mRNAs, using purified potato nucleotide pyrophosphatase, resulted in a greater than 80% decrease in each mRNA's template activity (Zan-Kowatczewska, gt al., 1977). In con- trast, Abraham and Pihl (1977) have reported that "decapping" of rabbit globin and mouse immunoglobulin light chain mRNAs by poly- nucleotide kinase does not affect translation of these messengers. The results of translational efficiency studies also appear to be dependent upon the degree of homogeneity in the assay systems. The use of wheat germ extracts with animal virus and cellular mRNAs represents a heterologous system, which may not adequately reflect 20 the regulatory functions and signals of the in 1319 environment. Rose and Lodish (1976) reported that the translation of B-eliminated vesicular stomatitis viral mRNA was one-tenth as efficient as un- treated mRNA in wheat germ systems, but was reduced to only one- fourth the control level in reticulocyte lysates. Similar results have been published by Samuel, gt a; (1977) for the translational efficiencies of methylated vs. unmethylated reovirus mRNA. In this study, unmethylated mRNA directed translation at less than 10% the rate of methylated mRNA in wheat germ extracts, but the two mRNA preparations were equally efficient in the homologous mouse ascites system (Samuel, 33 al., 1977). Toneguzzo and Ghosh (1976) obtained comparable results using vesicular stomatitis viral mRNA. Held, 32 a; (1977) also cautions against interpreting the results of translation studies as a structure-function analysis. These investigators demonstrated that uncapped reovirus and globin mRNAs can be translated up to 70% as efficiently as methylated-capped mRNAs in the presence of optimal concentrations of reticulocyte initiation factors. In contrast, optimal concentrations of ascites initiation factors or suboptimal levels of reticulocyte initiation factors resulted in a translational efficiency for unmethylated mRNA at 5 to 10% of the methylated-capped mRNA efficiency (Held, e_t_ g. , 1977). Ribosome Binding Studies. More precise evaluation of the effect of methylation on mRNA translational efficiency has been obtained by ribosome binding assays. Both, 25 21- (1975b) showed that wheat germ ribosomes selectively bound reovirus mRNA molecules which contained 7-methylguanine in 21 the 5'-terminal cap. This selectively occurred during or prior to the formation of AOS—mRNA-containing complex, suggesting that methylation functions at initiation (Both, 32 al., 1975b). A majority of cap structures was sensitive to partial RNase digestion of the 8OS-mRNA complexes, however, indicating that some caps were not physically protected by the ribosomes at the 808 level. Terminal oligonucleotides, containing 7-10 nucleotides and cap structures, did not rebind to ribosomes, suggesting that the presence of caps is insufficient for recognition by ribosomal subunits (Both, at al., 1975b). The relative importance of the 5'-terminus and the base composi- tion of the polynucleotide chain was evaluated by binding studies of synthetic ribopolymers (Both, 32 al., 1976; Muthukrishnan, at .al., 1976a,b). The ribopolymers were synthesized with polynucleotide phosphorylase, and the reactions were primed with m7GppmepC, its ring-opened derivative m7G*ppmepC, m7GpppGpC, GpppGpC or ppGpC. Preferential binding was observed for m7G-containing polymers and for those sequences rich in (A,U) (Both, £3 31., 1976; Muthukrishnan, .gt‘al., 1976a). Eleven to twenty per cent of the (A,U)-rich ribo- polymers were complexed to ribosomes without regard to the 5'-terminus, indicating that 7-methylguanosine is not an absolute requisite for ribosomal recognition. The presence of 2'eQ-methylation at the penultimate residue (N' position) enhanced ribosome binding of capped (A,U)-polymers five fold in reticulocyte lysates. Similar enhancement of binding in wheat germ extracts was observed only if 608 subunits were depleted by high speed centrifugation (Muthukrishnan, 23 al., 1976b). 22 The importance of cap structure in ribosomal binding affinity also seems to depend on the nature of the systems used. Lodish and Rose (1977) studied binding and translational efficiencies of vesicular stomatitis virus mRNAs in the heterologous plant cell- free extract and in the more homologous reticulocyte lysate. The presence of 7-methylguanosine appeared to be far more important for mRNA function in the wheat germ system (Lodish and Rose, 1977). Similar observations were made for reovirus mRNA (Muthukrishnan, gt al., 1976b) and vaccinia virus mRNA (Muthukrishnan, gt al., 1978). Because these ribosomal studies implicated methylation at the initiation level, several investigators sought to identify a ribosomal protein or initiation factor which bound to cap structures. Fillipo- wdcz, gt al. (1976) reported that a ribosome-associated protein in brine shrimp extracts was capable of binding m7GpppGpC, but could not identify this protein with any known initiation factor. Shafritz and colleagues, however, demonstrated that IF-M3, an initiation factor in reticulocyte lysates, specifically recognized cap structures and that its binding to capped mRNA was inhibited by analogues of the cap (Shafritz, 22 al., 1976). This factor is also necessary for the translation of uncapped picornovirus mRNAs, though, and therefore presumably possesses additional interactions with mRNA. Studies by Kaempfer, gt 21° (1978), however, argue that initiation factor EF-2 is involved in recognition of cap structures. A model involving primary recognition of an internal mRNA sequence, and Met f 9 is proposed which provides a molecular basis for differential trans- secondary recognition of cap structures and methionyl-tRNA lation of mRNA species (Kaempfer, at al., 1978). 23 Further information about the involvement of 5'-terminal cap structures in ribosome attachment has been obtained by analysis of viral RNA fragments which are protected by ribosomes. Dasgupta, gt 21- (1976) demonstrated that the AUG codon is located just 10 nucleotides from the 5'-terminus of brome mosaic virus RNA-M. Eight of the 10 nucleotides are adenosine and uridine (the 5'-terminal 7GpppG). Kozak and Shatkin (1976) nucleosides comprise the cap m have similarly characterized the ribosome-protected fragments of reovirus mRNAs. The sequence of fragments from six of these mRNAs has been determined (Kozak and Shatkin, 1977a,b; Kozak 1977). Comparison of the sequence and re—binding ability of these ribosome-protected fragments permitted identification of the following common features: a) in all cases the MOS-initiation complex pro- tected a significantly larger segment of the mRNA (including the cap) than did the 803 complex; b) each 8OS-complex protected a subset of the ROS-protected sequence and contained an AUG codon; c) ribo- somes re-bound to the mRNA fragments at the same initiation site irrespective of the methylated state of the cap; d) only those partial digestion products which retained the AUG codon could form initiation complexes, although the efficiency of binding was reduced if either the 5'-terminal region (including the cap structure) or the 3'-terminal region to AUG was removed (Kozak and Shatkin, 1978). These results strongly suggest that a variety of parameters are involved in ribo- somal recognition of mRNA, and that the presence of cap structures serves to facilitate binding rather than to determine it. 2H Analqgilnhibition Studies. An alternative approach to investigation into the role of cap structures in ribosome binding and translation was introduced by Hickey, 92 al. (1976a). These investigators demonstrated that 7- methylguanosine—S'-monophosphate (pm7G) inhibited cell-free trans- lation of rabbit globin mRNA, tobacco mosaic virus RNA, HeLa cell poly (A)-containing RNA and reovirus RNA. Translation of poly (A) and uncapped satellite tobacco necrosis virus RNA was unaffected by pm7G, indicating that this inhibition was specific for translation of capped RNAs (Hickey, gt al., 1976a). These findings have been con- firmed and expanded by several independent investigations (Canaani, 32. al., 1976; Levin and Samuel, 1976; Roman, 23 al., 1976; Suzuki, 1976; Groner, 32 al., 1976). The inhibitory effect of pm7G seems to be dependent upon both the methyl group in position 7 and the 5'-phosphate, since 7-methyl- guanosine-2',3'-monophosphate, 7-methylguanosine and guanosine-5'- monophosphate did not affect translation (Hickey, 33 al., 1976a,b; Canaani, 23 al., 1976). 7-Methylinosine-5'-monophosphate was also a poor inhibitor, suggesting that the amino group at position 2 of guanine may be important (Hickey, at al., 1977). However, addition 7 7G resulted in of 5'-phosphate groups to generate ppm G and pppm an enhancement of inhibition relative to the monophosphorylated nucleotide (Hickey, 23 al., 1977). Cap structures and their analogs have also been shown to inhibit protein synthesis, and are much more effective than the mononucleotide derivatives (Hickey, gt_§l., 1977; Suzuki, 22 al., 1977; Canaani, 25 ‘gt‘al., 1976). The presence of 7-methylguanosine appears to be most important, since GpppN and Gppme were not inhibitory, and m7Gppme and m7 7 GpppN were comparable to each other and to pppm7G and ppm G in inhibitory effect (Hickey, g£_al., 1977). Results from our laboratory have indicated that methylated tetraphosphate cap analogs are also potent inhibitors of protein synthesis (N. Sasavage, K. Friderici and F. Rottman, unpublished observations). The doubly- methylated analog, m7Gppppm7G, was more inhibitory than the singly- methylated structure m7GpppG; no inhibition was observed in the presence of the unmethylated tetraphosphate cap (GppppG). Inhibition by pm7G has been determined to occur at the conversion from the ROS-initiation complex to the 8OS-initiation complex, a transition which is mRNA-dependent (Roman, at al., 1976). The mono- phosphorylated derivative has been shown to equally inhibit the translation of all species present in HeLa cell poly (A)-containing RNA (Weber 33 al., 1976), and to be noninhibitory for the translation of unmethylated reoviral mRNAs (Levin and Samuel, 1977). The use of these cap analogs in both translation and ribosome binding studies has indicated that the m7G analogs compete with capped mRNA molecules, as predicted. However, the results of these studies must also be subject to the considerations described above, i.e. the use of heterologous vs. homologous systems. In addition, the observed effect of pm7G on mRNA translation in both wheat germ and reticulocyte systems has also been shown to be dependent upon the potassium ion concentration used (Kemper and Stolarsky, 1977; Weber, 32 al., 1977, 1978). The effective inhibition by pm7G increased with increased potassium concentrations up to the optimum K+ level 26 for protein synthesis. The counterion of the potassium salt (Kemper and Stolarsky, 1977) and the incubation temperature (Weber, at al., 1978) used in these assays also influenced the apparent inhibition of protein synthesis by 7-methylguanosine-5'-monophosphate. NMR studies by Hickey, at El- (1977) have generated an hypothesis which correlates inhibitory strength with the structure of a given analog. 7-Methylguanosine possesses a flexible conformation which becomes increasingly rigid by the introduction of one or two 5'- phosphate groups. This is due to an electrostatic interaction between the positively charged N-7 position of 7-methylguanine and the negatively charged phosphates, which results in a stable "W-shaped" conformation of the m7G backbone. Use of this structural criterion has enabled accurate predictions of the inhibitory effect of analogs on mRNA translation. This hypothesis is also consistent with the similar inhibtory effects of ppm7G, pppm7G, m7GpppN' and m7GpppN'm: the conformational stability is maximized in the diphosphate mono- nucleotide (Hickey, gt al., 1977). Adams, 23 a; (1978) also concluded that the positively charged imidazole moiety of m7G and negatively charged phosphate groups were the important structural features of cap analogs. These investi- gators synthesized a variety of substituted 7-methylguanosine-5'- diphosphate compounds and determined the effect of these derivatives on reovirus mRNA binding to wheat germ ribosomes. Alkylation at the 7-position of guanine with methyl, ethyl and benzyl groups generated active cap analogs, but loss of inhibitory activity was observed if the positive charge on the imidazole ring was eliminated. The 2-amino group of guanine also influenced the inhibitory effect of 27 the analog, whereas the 2',3'-cis-diol moiety was not critical for ribosomal binding. It should be noted that two enzymes have been detected which are capable of hydrolyzing cap structures. HeLa cell extracts contain a pyrophosphatase activity which cleaves m7GpppN' to pm7 G and ppN', and possesses substrate specificity for m7G-terminated oligonucleotides up to 10 residues in length (Nuss, gt al., 1975). A distinct enzymatic activity has been identified in tobacco cells which is capable of hydrolyzing various phosphodiester and pyrophosphate bonds without degrading polynucleotides (Shinshi, gt al., 1976). These enzymes may function to protect cells from the inhibitory effects of residual caps generated by degradation of mRNAs. Removal of the 5'-terminal pm7G from tobacco mosaic virus by the tobacco enzyme has been shown to virtually destroy viral infectivity (Ohno, 35 al., 1976). Regulatory Function at Translational Levels. Differential translation of methylated and unmethylated mRNAs prompted investigations of the possible role of methylation in development. Nontranslated, stable "maternal mRNA" is known to be present in unfertilized oocytes of sea urchins and in brine shrimp cysts and embryos. Capping and/or methylation represented potential signals for activation of this mRNA population. However, embryonic brine shrimp mRNA (Muthukrishnan, §t_al., 1975b) and sea urchin oocyte and embryonic mRNAs (Hickey, gt al., 1976b; Faust, at al., 1976; Sconzo, gt_al., 1977) were shown to already possess methylated cap structures. The stored maternal mRNA in tobacco hornworm oocytes, though, is capped but not methylated (Kastern and Berry, 1976); whether the cap becomes methylated after fertilization is unknown. 28 The appearance of 7-methylguanosine-5'-phosphate in RNA of mouse one-cell embryos has been reported (Young, 1977). A regulatory function of methylation has been suggested by the impairment of reovirus mRNA methylation in interferon-treated Ehrlich ascites tumor cells (Sen, gt al., 1975,1977). The relevance of these in gitgg studies remains to be established. The above experimental approaches used to assess the significance of cap structures at the level of translation have been facilitated by several characteristics of viral systems. These features include 1) the relatively simple, and generally well-characterized, viral genome, which permits monitoring of distinct RNA species and protein products; 2) the ability to synthesize viral mRNAs in 31239, which not only provides access to virtually pure RNAs, but also permits generation of unmethylated mRNAs to serve as substrate for enzyme assays and to enable comparative studies on the function(s) of methylation; and 3) the presence of the capping and methyltransferase activities within the virion, affording substantial purification (relative to a typical cellular enzyme) prior to the onset of isola- tion procedures. As a result, most experiments thus far performed have used viral mRNAs and enzyme preparations. A notable exception is globin mRNA, which is unusually simple to isolate relative to most eukaryotic mRNAs. A massive amount of data has been generated as a result of these efforts to identify a role for cap structures in mRNA transla- tion. An overall assessment of these results indicates that cap structures do seem to function in translation at the level of initia- tion, but its role is quantitative rather than qualitative. The 29 conformation afforded by electrostatic interaction of 7-methylguanosine with the adjacent phosphate groups appears to stabilize the ribosomal binding of caps, which would account for preferential translation of capped mRNAs. Perhaps the most serious criticism of these experimental approaches, taken as a whole, is the heterologous nature of the in gitgg assays. Fidelity of regulatory mechanisms is essential for structure-function studies, and an in 31359, heterologous system may not meet this criterion. A corollary to this criticism is that the factors important for recognition of viral-specific mRNA may be different than those involved in eukaryotic mRNA translational control. Future experimental designs must be developed to enable assessment of these more subtle aspects of translational control. B. Role of Methylation in mRNA Processing The concept of mRNA processing - that large, heterogeneous nuclear RNA molecules (hnRNA) are precursors of cytoplasmic mRNAs - was first proposed by Sherrer and Marcaud (1968). Processing encompasses posttranscriptional polyadenylation of the 3'-terminus, capping and methylation at the 5'-terminus, internal base methylation, and cleavage (or degradation) of the precursor in order to reduce its size (cf reviews: Darnell, 23 al., 1973; Greenberg, 1975; Lewin, 1975; Rottman, 1978). Only a few per cent of the transcribed se- quences present in the nucleus reach the cytoplasm. Thus processing appears to control genetic expression at the posttranscriptional level by regulating cytoplasmic entry of mRNA sequences. Considerable efforts have been directed at elucidating the nature of processing events, but the control mechanisms involved remain ambiguous. 30 Theoretical and Technical Considerations. Identification of a relationship between methylation and other posttranscriptional events is difficult for both theoretical and technical reasons. The apparent contradiction that both 3'-poly (A) and 5'-cap structures of hnRNA seemed to be conserved during processing, has most likely been resolved by the recent discovery of intragenic sequences in unique genes (Jeffreys and Flavell, 1977; Tilghman, 22 al., 1977; Breathnach, g£_al., 1977; Brack and Tonegawa, 1977). The demonstration that the intragenic sequences in B-globin genes are transcribed into a B-globin precursor RNA molecule, but are not present in B-globin mRNA (Tilghman, 33 al., 1978), requires that cleavage occur from within the hnRNA and thus permits conserva- tion of both termini. Although these data have clarified the overall mechanism of processing, the precise sequential and regulatory inter- relationships between specific posttranscriptional events remains vague. Characterization of hnRNA has proven to be a difficult and often controversial task. As indicated by its name, hnRNA consists of a heterogenous population of RNA molecules both with respect to size and sequence. The sequence complexity of hnRNA is extremely high, and certain sequences appear to be restricted to the nucleus. Less than ten per cent of the hnRNA synthesized is believed to reach the cytoplasm, with the remainder being rapidly degraded and/or performing unknown functions. The halflife of most hnRNA molecules is believed to be extremely short - 3 to 23 minutes - and thus pro- cessing intermediates are also presumably short-lived. These char- acteristics make kinetic studies extremely complex, particularly if the design is toward elucidation of a precursor-product relationship. 31 Isolation of hnRNA is a technical problem in itself. Massive amounts of DNA and rRNA precursors must be separated from hnRNA. Eukayrotic nuclei contain a variety of ribonucleases and the large size of hnRNA renders it particularly susceptible to nicks and degra- dation during rigorous purification procedures. Isolation of specific hnRNA sequences requires additional sensitivity for its detection. Also, considerable controversy still exists as to the actual size distribution of hnRNA. In addition to being sensitive to degradation, this RNA class appears to form aggregates on gels and sucrose gradients. The recent and rapid development of cloning procedures, coupled with the sensitivity of complementary DNA hybridization techniques, should facilitate studies of hnRNA and mRNA processing. Evidence Implicating Methylation in Processing. The possibility that methylated cap structures are involved in processing was suggested by Rottman, 33 31- (197A) in the first communication which identified the 5'-terminal structure. Involvement of methylation in processing has precedent in rRNA processing of ABS precursor molecules. Although the 18S and 288 cytoplasmic rRNA species represent less than half the nucleotide sequences of the HSS precursor, all methylated sites of the precursor are retained in mature rRNA (cfl,review: Perry, 1976). The identification of cap structures in hnRNA molecules isolated from mouse L cells (Perry, gt_al., 1975b), HeLa cells (Salditt- Georgieff, gt al., 1976), adenovirus-infected cells (Sommer, 33 al., 1976) and SVAO-infected cells (Lavi and Shatkin, 1975) has been described above. In all cases, only cap 1 structures were observed. The presence of cap 1 in hnRNA, coupled with the similarity 32 in methylnucleotide composition of caps 1 from mouse L cell hnRNA and mRNA (Perry, at al., 1975b; Schibler and Perry, 1976; Schibler, 22 al., 1977) raised the possibility that caps might be conserved in processing. Perry and Kelley (1976) performed pulse-chase exper- iments with mouse L cells to determine the kinetics of cap formation. The data indicated that minimum cap turnover occurred in the nucleus and that virtually all the labeled hnRNA caps were chased into cyto- plasmic mRNA after 3 h. The 2'eQ-methylation at the N" nucleoside appeared to occur in the cytoplasm (Perry and Kelley, 1976; Perry, _e_t_ 3;. , 1976). Similar results were observed by Friderici, gt al. (1976; repro- duced as part two of this thesis) for Novikoff mRNA methylation. These investigators analyzed the methylation patterns of mRNA as a function of labeling time. After a 20 min pulse, 80% of the radio- activity in mRNA was in N"m positions of cap 2. With longer labeling periods, a greater proportion of the mRNA radioactivity was observed in m7G, N'm, and m6 A, i.e. in those sites which appear to be methylated in the nucleus. The kinetics of cap formation in both Novikoff and mouse L cells suggests that capping and methylation of terminal guanosine, N' nucleosides, and internal bases occur in the nucleus. The methylated hnRNA molecules are processed and transported into the cytoplasm prior to additional methylation at N" to generate cap 2 structures. The sequential methylation pattern of globin mRNA in mice, as determined by Cheng and Kazazian (1978), is con- sistent with this model. 33 The presence of internal base methylation in both hnRNA and mRNA (see above) is also suggestive of an involvement of methylation in processing. In this regard it is of interest to note that only those viruses which replicate in the nucleus contain internal base methyl groups. The fact that m6A is present in the same specific sequence (G,A)pm6ApC in mouse L cell mRNA and hnRNA (Schibler, at $1., 1977) implies that m6A-containing hnRNA sequences are conserved in mRNA. As mentioned above, the m6A content in both mRNA and hnRNA has been correlated with molecular size (Perry and Kelley, 1976; Friderici, 23 al., 1976; Sommer, 32 al., 1977). In contrast, Lavi and Shatkin (1977) found that the m6 A residues per nucleotide number in HeLa cell mRNA is three to four times greater than that of HeLa hnRNA, suggesting that all internal base methylations in hnRNA are conserved during processing. Sommers, gt al.,(1978) however, have shown that the apparent levels of m6A in RNA molecules may reflect labeling periods, and, as pointed out by Schibler, at al. (1977), the current data on m6A labeling kinetics may be interpreted in a variety of ways. Further studies are thus necessary to determine the fate of internal base methylnucleosides during processing. Two other studies have implicated methylation in processing. McGuire, gt 21- (1976) analyzed adenovirus-specific RNA and found two types of 5'-termini. Both cytoplasmic and nuclear viral-associated RNA, believed to not function as viral mRNA, were terminated by guanosine tetraphosphate (pppGp). Viral nuclear precursor RNA, however, contained both pppGp and cap structures, whereas polysome- associated viral mRNA was terminated exclusively by cap structures. These findings suggest that capping and methylation of virus-specific 3A transcripts only occurs on those molecules which will perform mRNA functions (McGuire _e_t 31., 1976). Rose, e_t. £41977) synthesized in zitgg VSV mRNA and analyzed the products by gel electrophoresis. Analysis of the polyadenylated (and methylated) RNA produced in the presence of SAM showed discrete VSV mRNA species, while unmethylated mRNA, synthesized in the presence of SAH, electrophoresed as large, heterogeneous RNA species. Further analyses demonstrated that in the presence of SAH, very large heterogeneous poly (A) was present on the VSV mRNA species. These results implicate that a relationship does exist between the posttranscriptional events of methylation and polyadenylation, and that pertubation of methylation has disrupted regulation of polyadenylation. Further studies concerning the relationship between methylation, polyadenylation and cleavage of hnRNA will be of considerable interest with regard to posttranscrip- tional control of genetic expression. Inhibition of mRNA Methylation In Vivo Since the discovery of S-adenosylmethionine (Cantoni, 1952), a variety of SAM-dependent biological transmethylation reactions have been identified (of. review: Borchardt, 1977). Most of these SAM-dependent methyltransferases are inhibited by the demethylated product S—adenosylhomocysteine (SAH), including catechol Q-methyl- transferase, phenolethanolamine Nemethyltransferase, histamine H? methyltransferase, glycine Ngmethyltransferase, homocysteine §f methyltransferase, indoleethylamine Nrmethyltransferase and tRNA methyltransferases (Borchardt, 1977). Due to the biological importance of these SAM-dependent reactions, several laboratories have, in recent years, systematically synthesized and evaluated the inhibitory 35 properties of numerous SAH analogs in order to study SAM and SAH binding characteristics (Coward at 31., 197A; Borchardt, 1977; Pugh 32 al., 1977). These studies have indicated that inhibitory activity requires the ribose and amino acid moieties remain virtually intact. Modifications of adenine, however, can result in potent inhibitory analogs if the planarity and aromaticity of the purine is retained (Coward, 197M; Borchardt, 1977). The use of SAH analogs to inhibit mRNA methylation represented an alternative approach to the study of the function(s) of methyl groups. The generation of undermethylated eukaryotic mRNA was not feasible via in zitgg RNA synthetic reactions, as has been described above for viral methylation studies. In addition, SAH cannot be used to inhibit methylation of cells in culture, since it is virtually impermeable to cellular membranes and is also a naturally-occurring metabolite subject to in 1339 degradation. Our laboratory and others have thus initiated methylation studies which employ SAH analogs to inhibit eukaryotic mRNA methylation in cell cultures, in order to assess the functional nature of this posttranscriptional modifica- tion. S-tubercidinylhomocysteine (STH), the 7-deaza analog of SAH, is structurally depicted in Figure 2. This analog exhibited equivalent or greater potency of inhibition than SAH when used to inhibit catechol gymethyltransferase, indoleethylamine Nemethyltransferase, and tRNA methyltransferases 32.11232 (Coward 22 al., 1974). Of greater signif- icance was the demonstration of the in 3139 inhibition by STH on tRNA methylation in phytohemagglutinin—stimulated lymphocytes (Chang and Coward, 197A) and on catecholamine methylation in neuroblastoma 36 Figure 2. The structure of S-tubercidenylhomocysteine, the 7-deaza analog of S-adenosylhomocysteine. 37 Figure 2 38 cells (Michelot, §t_al., 1977). The capability of STH to be inhibi- tory in zigg indicated both permeability and metabolic stability of the compound. Our laboratory has employed STH as a means of perturbing Novikoff mRNA methylation. Preliminary studies suggested that the presence of STH in culture media was nontoxic to Novikoff cells, and that by 2” h after incubation with the drug, the cells had overcome the inhibitory effects of STH (M. Kaehler, J. Coward and F. Rottman, unpublished observations). At 250 ”M, STH inhibited messenger RNA synthesis to 321 the level of control cells. Analysis of 3H-methyl mRNA indicated that methylation at N',N" and internal base sites were affected by STH. Inhibition of 2'egrmethylation at N' positions generated cap zero structures (m7GpppN') in both cytoplasmic and nuclear poly (A)-containing RNA, indicating that ribose methylation of the penultimate nucleoside is not necessary for processing and transport of Novikoff mRNA (Kaehler gt al., 1977 (reproduced as part III of this thesis)). Subsequent location of cap zero-bearing mRNA molecules on both monosomes and polysomes of STH-inhibited Novikoff cells suggests that this 2'1Qemethylation is not requisite for ribosomal binding or for translation (Kaehler, M., Coward, J. and Rottman, F., manuscript submitted and reproduced as Part IV of this dissertation). Non-poly (A)-containing RNA from normal and STH-treated cells was also analyzed and shown to contain virtually identical methylation patterns (M. Kaehler, J. Coward and F. Rottman, unpublished results). Analysis of tRNA methylation is in progress (P.Crooks, personal communication). Pugh at al. (1977) have recently found STH to be the most potent inhibitor of all SAH analogs tested 39 22.22222 against Newcastle disease virus mRNA (guanine-7-)methyl- transferase. Studies in our laboratory indicate that if 7-methyl- guanosine formation is blocked in 3112 by STH, the resultant GpppN' structures are not stable (Kaehler, M., Coward, J. and Rottman, F., manuscript submitted). Another SAH analog, S-isobutyladenosine (SIBA) has been reported by Jacquemont and Huppert (1977) to inhibit viral mRNA methylation l2.!£!2- 1mM SIBA reversibly blocked Herpex simplex type 1 virus multiplication, viral protein synthesis, and viral mRNA methylation; the data is interpreted to establish a correlation of capping and methylation inhibition to decreased protein synthesis (Jacquemont and Huppert, 1977). It should be noted, however, that this inter- pretation does not explain the reversible inhibition of production observed for encephalomyocarditis virus, which does not contain capped mRNA (Jacquemont and Huppert, 1977). Cantoni (1977) has suggested that SIBA is an inhibitor of SAH hydrolase, rather than a competitive inhibitor for SAM binding sites. Robert-Gero gt al. (1975) and Michelot gt al. (1976) have demonstrated that SIBA, in contrast to both SAH and STH, is a very weak inhibitor of tRNA trans- methylases ig_gitgg in both normal and transformed cells. Other compounds than SAH analogs have been shown to perturb mRNA methylation. Cycloleucine is an in 31359 competitive inhibitor of methionine for SAM-synthetase (Lombardini, gt_gl.,1970), and its mode of action in vizg presumably is to reduce intracellular concentrations of SAM (Caboche, 1977). Cycloleucine has been shown to quantitatively affect rRNA maturation processes (Caboche and U0 Bachellerie, 1977). More recently, Dimock and Stoltzfus (K. Dimock and C. Stoltzfus, personal communication) have demonstrated that the presence of 40 mM cycloleucine in low methionine medium results in greater than 90% inhibition of internal and N'methylations in avian sarcoma virus B-77. No inhibition of 7-methylation of guanine was detected (K. Dimock and C. Stoltzfus, personal communication). It should be noted that although similar inhibition patterns are observed for STH and cycloleucine, the two compounds are believed to function differently at the macromolecular level. Glazer and Peale (1978) have recently reported that cordycepin (3'-deoxyadenosine) and xylosyladenine (9-512-xylofuranosyladenosine) inhibit methylation of nuclear RNA to a greater extent than RNA synthesis in L1210 cells. Cordycepin has been shown to be an effec- tive inhibitor of nuclear RNA synthesis (Kann and Kohn, 1972; Darnell, at al., 1971), and xylosyladenine appears to effect similar results but with greater potency. Inhibition was also observed for both base-methylated mononucleotides and 2'-Q¢methylated dinucleotides, the latter presumably of rRNA origin. An oligonucleotide containing 7-methylguanosine and 2'-9emethylnucleotides was also present at decreased levels in the presence of either drug. These results are of considerable interest in light of the anticancer usage of cordycepin and xylosyladenine, but must be considered as preliminary since the entire nuclear RNA population was analyzed together (Glazer and Peale, 1978). Kredich and Martin (1977) have recently studied adenosine- mediated toxicity in cultured mouse T lymphoma cells, and have con- cluded that SAH accumulation is the direct effect of adenosine H1 cytotoxicity. High levels of adenosine inhibit the enzyme SAH- hydrolase, and thus increase intracellular levels of SAH. In 1132 DNA methylation was inhibited in correlation with increased SAH concentrations, suggesting the primary toxic effect is due to inhibi- tion of SAM-mediated methyltransferases (Kredich and Martin, 1977). RNA methylation in human myeloma cells also appears to be inhibited by high adenosine concentrations (J.Bynum, personal communication). It would be of interest to determine what, if any, qualitative changes adenosine exposure might effect on mRNA methylation. Finally, N6-methyladenosine has been found to be a substrate for S-adenosylhomocysteine hydrolase both ig_gi£§g and in gixg, resulting in the synthesis of S-N6-methyladenosylhomocysteine (Sm6AH) (J. Hoffman, personal communication). This analog has been shown to be a potent inhibitor of both tRNA methyltransferases (Trewyn and Kerr, 1976) and mRNA (guanine-7-) methyltransferase (Pugh gt al. 1977). Thus the potential usefulness of SAH analogs continues to expand. Increased Resistance to Degradation A straightforward yet nontrivial function of mRNA capping and methylation might be to provide additional resistance of the RNA Inclecules to ribonuclease attack. Jervis and DeBusk (1975) have :3hown that undermethylated tRNAs are more susceptible to nuclease aattack than are normal tRNA samples. Methylation has been demonstrated 1:0 enhance the stability of both rRNA (Liau, 23 al., 1976) and synthetic ribopolymers (Stuart and Rottman, 1973). "2 III. An Argument for Control of Genetic Expression at the Posttrans- criptional Level The control of genetic expression is a major problem in cellular biology. The processes of development, tissue specialization, regeneration, and normal and abnormal cell growth must reflect those DNA sequences which are permitted to be expressed via eukaryotic mRNA species. Control mechanisms have been established at the trans- criptional level, and reannealing experiments have indicated that most mRNA sequences are transcribed from unique DNA genes (cf. reviews: Lewin, 1975a,b). However, the sequence complexity of hnRNA, the primary transcript of DNA, is several fold higher than the complexity of cytoplasmic mRNA, suggesting a second level of control. This brief (and certainly incomplete) review will delineate evidence for posttranscriptional control of genetic expression. One of the most dramatic demonstrations implicating posttrans- criptional control of gene expression involved the characterization of two clones of the Friend cell line. The Friend erythroleukaemic cell line was isolated in 1966 (Friend, gt al., 1966), and the addition of dimethylsulfoxide (DMSO) to cultures of virus-transformed Friend cells has been shown to induce erythroid differentiation and hemoglobin accumulation (Friend, gt al., 1974). Paul and colleagues (Gilmore, 23 al., 197A; Harrison, 32 al., 197“) analyzed the basal and induced levels of globin cytoplasmic and nuclear globin-specific RNA sequences in two clones: M2 and 707. In clone M2, the globin mRNA on polysomes increased 50 to 100 times during induction, whereas the hnRNA sequences complimentary to globin mRNA increased only 5- to 6-fold (Gilmore, M3 32 al., 197"). In addition, no detectable globin mRNA was synthesized from chromatin templates of uninduced M2 cells, but significant levels of globin mRNA were synthesized from induced cell chromatin. These data pointed to primary control of gene expression at the transcriptional level, with some secondary control mechanisms func- tioning posttranscriptionally. In contrast, although clone 707 cells were inducible by DMSO exposure, and induced globin mRNA levels in the cytoplasm increased to levels comparable to that of induced M2 clones, the basal level of cytoplasmic globin mRNA in uninduced 707 cells was much higher than that observed in M2 cells. No differ- ences were observed in the induced and uninduced 707 cells for globin- specific nuclear RNA sequences or the capacity of the chromatin template to synthesize globin sequences. These results suggested that transcriptional control in clone 707 has been relaxed, and posttranscriptional mechanisms were responsible for the observed induction (Harrison, 33 al., 1978). In subsequent studies, annealing experiments were performed to determine the complexity of Friend cell hnRNA and mRNA. At least five times more unique DNA gene sequences were represented in nuclear poly (A)-containing RNA than in the polysomal poly (A)-containing RNAs. Furthermore, some gene transcripts were enriched in the poly- somal RNA relative to its concentration in hnRNA (Birnie, gt al., 1979; Getz, gt al., 1975). These results suggested that posttrans- criptional mechanisms alter the relative concentration of some gene transcripts between nucleus and cytoplasm. These results have been verified by Kleiman, §£_§l. (1977) using more sensitive hybridization AH techniques. Similar conclusions, i.e. that there is a considerably greater number of unique DNA sequences represented in hnRNA than in mRNA, have been made for sea urchin embryos (Smith, gt al., 1979), Xenopus liver cells (Ryffel, 1976), HeLa cells (Herman, 33 al., 1976), and mouse brain (Bantle and Hahn, 1976). Evidence is also accumulating which indicates that cell phenotype may be determined more by the relative abundance of mRNA species on polysomes than by the absence or presence of specific RNA sequences. A high level of homology between the poly (A)-containing polysomal RNA of mouse brain, embryo, and liver tissue was demonstrated by Young, g£_al. (1976). Getz, 32 al. (1976) has studied the relationship between cell proliferation and the amount and diversity of mRNA sequences in mouse embryo cells. Within limits of detection, all species of poly (A)-containing mRNA present in growing cells are also present in resting cells. Humphries, gt El- (1976) detected mouse globin RNA sequences present in hnRNA from nonerythroid tissues (including adult brain and liver, and lymophoma, untransformed, and transformed fiborblast cell cultures) as well as in erythroid tissues (reticulocytes and fetal liver). The proportion of the globin RNA sequences in hnRNA containing poly (A) was similar for all species, but in erythroid cells the cytoplasm contained a much greater percentage of the total globin sequences (Humphries, 23 al., 1976). The association of poly (A) with nuclear globin sequences indicated that poly (A) does not play a major role in selecting sequences for transport into the cytoplasm. This is consistent with the demonstration by Perry, gt El. (197“) that nuclear poly (A) is not quantitatively converted to cytoplasmic poly (A), even ”5 though these investigators have demonstrated the potential exists for most nuclear hnRNA molecules to function as mRNA precursors (Hames and Perry, 1977). A second perspective which may be of importance than with regard to the presence of posttranscriptional control mechanisms is the rate of selection, processing, and transport of mRNA molecules into the cytoplasm. Studies which compare resting and induced fibroblasts (3T6 cells) are useful for this hypothesis. As summarized by Green (197“), a characteristic of growing fibroblasts and other cell types is a high messenger RNA to ribosome ratio. The increased mRNA levels in the cytoplasm has not been correlated with stabilization of the RNA, since the halflife of mRNA molecules is not significantly altered by induction (Abelson, 23 al., 197A). Nor can the higher mRNA levels be attributed to increased transcription rates (Mauch and Green, 1973), since the rate of transcription of hnRNA per unit of DNA does not change during induction or if DNA synthesis is blocked. However, the increased mRNA levels in the cytoplasm of serum-stimulated fibroblasts has been correlated with an increased rate of processing (Johnson, 33 al., 1979, 1975). These data taken together are con- sistently supportive of posttranscriptional control of gene expression, both with respect to qualitative and quantitative perspectives. It should be emphasized, however, that transcriptional control of genetic expression is certainly not excluded or minimized by the additional control mechanisms which may function posttranscriptionally. In this regard, Williams and Penman (1975) reported that approximately 3% of the mRNA in resting 3T6 cells will not crosshydridize with cDNA of mRNA from growing cells. Similar results were obtained #6 by crosshybridization of growing cell mRNA to cDNA from restng cell mRNA. Egyhazi (1976) has quantitatively studied the intranuclear metabolism and transport of hnRNA transcribed in the Balbiani rings of Chironomus tentans salivary gland cells. Of the total amount of the 753 RNA synthesized at the ring loci, only 1H—171 can be recovered in the nuclear sap whereas 4-72 is present in the cytOplasm. The remainder is presumably degraded, and experiments using inhibitors of transcription indicate very little of the 758 RNA can be chased into the cytoplasm. Since no size reduction of the 753 RNA occurs prior to export of this RNA molecule, the nucleotide sequences of 753 hnRNA and 758 mRNA are presumably the same. These results demonstrate for the first time that a specific, protein-coding sequence in hnRNA is degraded within the nucleus, and implicate posttranscriptional regulatory mechanisms. This system provides a unique opportunity for studying what role, if any, methylation and/or polyadenylation plays in determining the metabolic fate of the three pools of 758 RNA. Regulation of virus gene expression at transcriptional and posttranscriptional levels has been demonstrated as well. The appearance and quantity of 47 viral RNA species and 35 viral proteins present in frog virus 3 was monitored by Willis, 33 al. (1977) as a function of time post-infection. Although proteins could be classified as early, intermediate or late polypeptides, viral mRNAs could not be classified according to time of maximum synthesis. Once the RNA synthesis began, most RNA species continued to be made at the same or elevated rates, implying that posttranscriptional I17 mechanisms controlled the shut-off of early and intermediate protein synthesis. The realization that more than 90% of those RNA sequences trans- cribed in the nucleus are not permitted to enter the cytoplasm for ultimate gene expression has been historically puzzling. The in- efficiency of such a method seemed incredible in view of the efficiency of the enzymatic metabolism and catabolism occurring elsewhere within cells. Relatively recent investigations such as those described above, however, make such a "wasteful" synthetic process plausible. A rationale for the necessity of posttranscriptional as well as transcriptional regulation of gene expression is presented below. A Speculative "Control Hierarchy". The eukaryotic genome contains a high complexity of gene sequences which must be conserved in each and every cell of the organism. At any given time in the life of the cell, a very small percentage of these genes are required for normal cell function. Both gross and subtle alterations in the cellular environment, however, require different combinations of gene products for adequate cellular response. The overall "off-on" control of genes occurs at the transcriptional level. The sensitivity of these control mechanisms is determined largely by the developmental or specialized state of the cell, and by gross environmental changes. Its response occurs throughout the genome. Transcriptional control mechanisms may be envisioned to grossly monitor "sets" of genes which, due to the very bulk of the genome, include several gene sequences which are either undesirable or unnecessary in a given cell state. However, all genes of the set #8 must be transcribed in order to express those sequences which are needed. Thus posttranscriptional control mechanisms have evolved to provide finer, more sensitive regulation of those sequences which enter the cytoplasm. The events of processing - methylation at internal sites, capping and methylation of 5'-termini, polyadenylation of 3'-termini, cleavage of intragenic sequences, and other unidentified modifications of nuclear RNA - provide a molecule-by-molecule screening process of transcribed gene sequences. Such screening control may be sequential in nature, or the selection responsibility at a given time may rest with one specific processing event, which in turn is determined by the transcriptional regulation in effect at the time. The appearance of wasteful and inefficient RNA transcription by eukaryotic cells may actually reflect an extremely complicated and sophisticated set of control mechanisms which permit both reten- tion of the entire genomic information in the nucleus and extreme sensitivity for response to the cells needs. The adaptability and extent of control afforded by posttranscriptional regulation of gene expression makes such systems very attractive. REFERENCES Abelson, H., Johnson, L., Penman, S. and Green, H. (197”) Cell 1, 161. Abraham, G. and Banerjee, A. (1976) Virology 11, 230. Abraham,G., Rhodes, D. and Banerjee, A. (1975a) Cell 5, 51. Abraham, G., Rhodes, D. and Banerjee, A. (1975b) Nature (London) 255, 37. Abraham, K. and Pihl, A. (1977) Eur. J. Biochem. 11, 589. Adams, B., Morgan, M., Muthukrishnan, S., Hecht, S. and Shatkin, A. (1978) J. Biol. Chem. 253, 2589. Bajszar, G., Samarina, O. and Georgiev, G. (1976) Cell 2, 323. Bantle, J. and Hahn, W. (1976) Cell 8, 139. Bartkoski, M. and Roizman, B. (1976) J. Virology g9, 583. Bartkoski, M. and Roizman, B. (1978) Virology 85, 1N6. Beeman, K. and Keith, J. (1977) J. Mol. Biol. 113, 165. Birnie, G., MacPhail, E., Young, B., Getz, M. and Paul, J. (197A) Cell Differentiation 3, 22. Boone, R., Ensinger, M. and Moss, B. (1977) J. Virology 21, A75. Boone, R. and Moss, E. (1977) Virology 19, 67. Bondurant, M., Hashimoto, S. and Green, M. (1976) J. Virology 12, 998. Borchardt, R. (1977) The Biochemistry of Adenosylmethionine (eds. F. Salvatore, E. Borek, V. Zappia, H. Williams-Ashman, F. Schlenk) Columbia University Press:New York, p. 151. Both, G., Banerjee, A. and Shatkin, A. (1975a) Proc. Natl. Acad. Sci. USA 2, 1189. H9 50 Both, G., Furuichi, Y., Muthukrishnan, S. and Shatkin, A. (1975b) Cell 8, 185. Both, G. Furuichi, Y., Muthukrishnan, S. and Shatkin, A. (1976) J. Mol. Biol. 10”, 637. Brack, C. and Tonegawa, S. (1977) Proc. Natl. Acad. Sci. USA 15, 5652. Breathnach, R., Mandel, J. and Chambon, P. (1977) Nature (London) 270, 3114. Caboche, M. (1977) J. Cell Physiol. 22, A07. Caboche, M. and Bachellerie, J.-P. (1977) Eur. J. Biochem. 12, 19. Canaani, D.,Revel, M. and Groner, Y. (1976) FEBS Letters 88, 326. Cantoni, G. (1952) J. Amer. Chem. Soc. 12, 29A2. Carter, C. (1977) Virology 82, 2A9. Chambon, P. (197“) The Enzymes 12 (P. Boyer, ed.) New York:Academic Press, p. 261. Chang, C.-D. and Coward, J. (1975) Mol. Pharmacol. 11, 701. Cheng, T.-C. and Kazazian, H. (1977) J. Biol. Chem. 252, 1758. Cheng, T.-C. and Kazazian, H. (1978) J. Biol. Chem. 253, 2A6. Colonno, R., Abraham, G. and Banerjee, A. (1976) Prog. Nucl. Acids Res. Mol. Biol. 12, 83. Colonno, R. and Banerjee, A. (1976) Cell 8, 197. Colonno, R. and Stone, H. (1975) Proc. Natl. Acad. Sci. USA 12, 2611. Colonno, R. and Stone, H. (1976) Nature (London) 281, 611. Cory, S. and Adams, J. (1975) J. Mol. Biol. 22, 519. Coward, J., Bussolotti, D. and Chang, C.-D. (197A) J. Med. Chem. _11, 1286. Darnell, J., Jelinek, W. and Molloy, G. (1973) Science 181, 1215. Darnell, J., Philipson, L. Wall, R. and Adesnik, M. (1971) Science 17“. 507. Dasgupta, R., Harada, F. and Kaesberg, P. (1976) J. Virology 18, 260. Dasgupta, R., Shih, D. Saris, C. and Kaesberg, P. (1975) Nature (London) 256, 62”. 51 DeKloet, S. and Andrean, B. (1976) Biochim. Biophys. Acta A25, A01. Desrosiers, R., Friderici, K. and Rottman, F. (197A) Proc. Natl. Acad. Sci. USA _7_1. 3971. Desrosiers, R., Friderici, K. and Rottman, F. (1975) Biochemistry 13, A367. Desrosiers, R., Sen, G. and Lengyel, P. (1976) Biochem. Biophys. Res. Commun. 12, 32. Dimock, K. and Stoltzfus, C. (1977) Biochemistry 18, A71. Dottin, R., Weiner, A. and Lodish, H. (1976) Cell 8, 233. Dubin, D. and Stollar, V. (1975) Biochem. Biophys. Res. Commun. 99: 1373- Dubin, D. and Taylor, R. (1975) Nucl. Acids Res. 2, 1653. Egyhazi, E. (1976) Cell 1, 507. Ensinger, M., Martin, S., Paoletti, E. and Moss, B. (1975) Proc. Natl. Acad. Sci. USA 12, 2525. Ensinger, M. and Moss, B. (1976) J. Biol. Chem. 221, 5283. Faust, M., Millward, S., Duchastel, A. and Fromson, D. (1976) Cell 2; 597. Fellner, P., Frisby, D., Goodchild, A., Porter, A. and Carey, N. (1975) Third International Congress of Virology Abstracts, 161. Fernandez-Munoz, R. and Darnell, J. (1976) J. Virology 18, 719. Fillipowicz, W., Furuichi, Y., Sierra, J., Muthukrishnan, S., Shatkin A. and Ochoa, S. (1976) Proc. Natl. Acad. Sci. USA 12, 1559. Friderici, K., Kaehler, M. and Rottman, F. (1976) Biochemistry 12, 523A. Friend, C., Patuleia, M. and DeHarven, E. (1966) National Cancer Inst. Monograph 22, 505. Friend, 0., Scher, W., Holland, J. and Sato, I. (1971) Proc. Natl. Acad. Sci. USA 88, 378. Frisby, D., Eaton, M. and Fellner, P. (1976) Nucl. Acids Res. 2, 2771. Furuichi, Y. (197A) Nucl. Acids Res. 1, 809. 52 Furuichi, Y. (1978) Proc. Natl. Acad. Sci. USA 12, 1086. Furuichi, Y. and Muira, K. (1975) Nature (London) 222, 37A. Furuichi, Y., Muthukrishnan, S., Tomasz, J. and Shatkin, A. (1976) Prog. Nucl. Acids Res. Mol. Biol. 12, 3. Furuichi, Y. and Shatkin, A. (1977) Virology 11, 566. Furuichi, Y., Shatkin, A., Stavnezer, E. and Bishop, J. (1975) Nature (London) 257, 618. Getz, M., Birnie, G., Young, B. MacPhail, E. and Paul, J. (1975) Cell 1, 121. Getz, M., Elder, P., Benz, B., Stephens, R. and Moses, H. (1976) Cell 1, 255. Gilmour, R., Harrison, P., Windass, J., Affara, N. and Paul, J. (197A) Cell Differentiation 2, 9. Glazer, R. and Peale, A. (1978) Biochem. Biophys. Res. Commun. 81, 521. Green, H. (197A) Cell Surfaces and Malignancy, (ed. P. Mora) DHEW Publ. No. (NIH) 75-796:Washington, D.C., p. 73. Greenberg, J., (1975) J. Cell Biol. 82, 269. Groner, Y., Gilboa, E. and Aviv, H. (1978) Biochemistry 11, 977. Groner, Y., Grosfeld, H. and Littauer, U. (1976) Eur. J. Biochem. u, 281. Groner, Y. and Hurowitz, J. (1975) Proc. Natl. Acad. Sci. USA 12, 2930. Haegaman, G. and Fiers, W. (1978) J. Virology 22, 82A. Hames, B. and Perry, R. (1977) J. Mol. Biol. 122, A37. Harrison, P., Gilmour, R., Affara, N., Conkie, D. and Paul, J. (197A) Cell Differentiation 2, 23. Hefti, E. and Bishop, D. (1976) Biochem. Biophys. Res. Commun. 88, 393. Hefti, E. and Bishop, D. (1975) J. Virology 12, 90. Hefti, E., Bishop, D., Dubin, D. and Stollar, V. (1976) J. Virology 11, 1A9. Held, W., West, K. and Gallagher, J. (1977) J. Biol. Chem. 222, 8A89. 53 Herman, R., Williams, J. and Penman, S. (1976) Cell 1, A29. Hewlett, M., Rose, J. and Baltimore, D. (1976) Proc. Natl. Acad. Sci. USA 13, 327. Hickey, E., Weber, L. and Baglioni, C. (1976a) Proc. Natl. Acad. Sci. USA 13, 19. Hickey, E., Weber, L. and Baglioni, C. (1976b) Nature (London) 261, 71. Hickey, E., Weber, L. and Baglioni, C. (1977) J. Mol. Biol. 109, 173. Humphries, S., Windass, J. and Williamson, R. (1976) Cell 1, 267. Jacquemont, B. and Huppert, J. (1977) J. Virology 22, 160. Jeffreys, A. and Flavell, R. (1977) Cell 12, 1097. Jervis, H. and DeBusk, A. (1975) Nature (London) 228, 160. Johnson, L., Abelson, H., Green, H. and Penman, S. (197A) Cell 1, 95. Johnson, L., Williams, J., Abelson, H., Green, H. and Penman, S. (1975) Cell 2, 69. Kaempfer, R., Rosen, H. and Israeli, R. (1978) Proc. Natl. Acad. Sci. USA 12, 650. Kann, H. and Kohn, K. (1972) Mol. Pharmacol. 8, 551. Kastern, W. and Berry, S. (1976) Biochem. Biophys. Res. Commun. ‘11. 37. Keith, J. and Fraenkel-Conrat, H. (1975a) FEBS Letters 21, 31. Keith, J. and Fraenkel-Conrat, H. (1975b) Proc. Natl. Acad. Sci. USA 13, 33117. Kemper, B. (1976) Nature (London) 282, 321. Kemper, B. and Stolarsky, L. (1977) Biochemistry 18, 5676. Kleiman, L., Birnie, G., Young, B. and Paul, J. (1977) Biochemistry 18, 1218. Klootwijk, J., Klein, I., Zabel, P. and van Kammen, A. (1977) Cell .11, 73- Kozak, M. (1977) Nature (London) 282, 390. SA Kozak, M. and Shatkin, A. (1976) J. Biol. Chem. 221, A259. Kozak, M. and Shatkin, A. (1977a) J. Biol. Chem. 222, 6895. Kozak, M. and Shatkin, A. (1977b) J. Mol. Biol. 112, 75. Kozak, M. and Shatkin, A. (1978) Cell 12, 201. Kredich, N. and Martin, D. (1977) Cell 12, 931. Krug, R., Morgan, M. and Shatkin, A. (1976) J. Virology 22, A5. Lavi, U., Fernandez-Munoz, R. and Darnell, J. (1977) Nucl. Acids Res. 2, 63. Lavi, S. and Shatkin, A. (1975) Proc. Natl. Acad. Sci. USA 12, 2012. Leppert, M. and Kolakofsky, D. (1978) J. Virology 22, A27. Levin, K. and Samuel, C. (1977) Virology 11, 2A5. Lewin, B. (1975a) Cell 2, 11. Lewin, B. (1975b) Cell 2, 77. Liau, M., Hunt, M. and Hurlbert, R. (1976) Biochemistry 12, 3158. Lockhard, R. and Rathandary, U. (1976) Cell 2, 7A7. Lodish, H. (1976) Ann. Rev. Biochem. 22, 39. Lodish, H. and Rose, J. (1977) J. Biol. Chem. 222, 1181. Lombardini, J. Coulter, A. and Talalay, P. (1970) Mol. Pharmacol. Martin, S. and Moss, B. (1975) J. Biol. Chem. 222, 9330. Martin, S. and Moss, B. (1976) J. Biol. Chem. 221, 7313. Martin, S., Paoletti, E. and Moss, B. (1975) J. Biol. Chem. 250, 9322. Mauck, J. and Green, H. (1973) Proc. Natl. Acad. Sci. USA 12, 2819. McGuire, P., Piatak, M. and Hodge, L. (1976) J. Mol. Biol. 121, 379. Michelot, R., Legraverend, M., Farrugia, G. and Lederer, E. (1976) Biochimie 28, 201. Michelot, R. Lesko, N., Stout, R. and Coward, J. (1977) Mol. Pharmacol. 12, 368. 55 Muira, K., Furuichi, Y., Shimotohno, K., Urushibara, T., Watanabe, K. and Suguira, M. (1975) INSERM g, 153. Muira, K., Watanabe, K. and Suguira, M. (197Aa) J. Mol. Biol. 88, 31. Muira, K., Watanabe, K., Suguira, M. and Shatkin, A. (197Ab) Proc. Natl. Acad. Sci. USA 11, 3979. Moss, 8. (1977) Biochem. Biophys. Res. Commun. 12, 37A. Moss, B., Gershowitz, A., Strenger, J., Holland, L. and Wagner, E. (1977a) J. Virology 22, 23A. Moss, E., Gershowitz, A., Weber, L. and Baglioni, C. (1977b) Cell 18, 113. Moss, B. Gershowtiz, A., Wei, C.-M. and Boone, R. (1976a) Virology 12, 3A1. Moss, B., Keith, J., Gershowitz, A., Ritchey, M. and Palese, P. (1978) J. Virology 22, 312. Moss, B. and Koczot, F. (1976) J. Virology 11, 385. Moss, B., Martin, S., Ensinger, M., Boone, R. and Wei, C.-M. (1976b) Proc. Nucl. Acids Res. Mol. Biol. 12, 63. Moyer, 8., Abraham, G., Alder, R. and Banerjee, A. (1975) Cell 2, 59. Moyer, S. and Banerjee, A. (1976) Virology 12, 339. Muthukrishnan, 8., Both, G., Furuichi, Y. and Shatkin, A. (1975a) Nature (London) 255, 33. Muthukrishnan, S., Filipowicz, W., Sierra, J., Both, G., Shatkin, A. and Ochoa, S. (1975b) J. Biol. Chem. 250, 9336. Muthukrishnan, S., Furuichi, Y., Both, G. and Shatkin, A. (1976a) Prog. Nucl. Acid Res. Mol. Biol. 11, A73. Muthukrishnan, S., Morgan, M., Banerjee, A. and Shatkin, A. (1976b) Biochemistry 12, 5761. Muthukrishnan, 8., Moss, B., Cooper, J. and Maxwell, E. (1978) J. Biol. Chem. 223, 1710. Nomoto, A., Lee, Y. and Wimmer, E. (1976) Proc. Natl. Acad. Sci. USA 12, 375. Nuss, D., Furuichi, Y., Koch, G. and Shatkin, A. (1975) Cell 8, 21. 56 Ohno, T., Okada, Y., Shimotohno, K., Miura, K., Shinshi, H., Miwa, M. and Sugimura, T. (1976) FEBS Letters 81, 209. Ouellette, A., Frederick, D. and Malt, R. (1975) Biochemistry 12, A361. Perry, R. (1976) Annual Rev. Biochem. 22, 621. Perry, R., Bard, E., Hames, B., Kelley, D. and Schibler, U. (1976) Prog. Nucl. Acids Res. Mol. Biol. 11, 275. Perry, R. and Kelley, D. (197A) Cell 1, 37. Perry, R. and Kelley, D. (1976) Cell 8, A33. Perry, R., Kelley, D., Friderici, K. and Rottman, F. (1975a) Cell 2, 387. Perry, R., Kelley, D., Friderici, K. and Rottman, F. (1975b) Cell 6, 13. Perry, R., Kelley, D. and LaTorre, J. (197A) J. Mol. Biol. 82, 315. Perry, R. and Sherrer, K. (1975) FEBS Letters 21, 73. Pinck, L. (1975) FEBS Letters 22, 2A. Pugh, C., Borchardt, R. and Stone, H. (1977) Biochemistry 18, 3928. Reddy, R., Ro-Choi, T., Henning, D. and Busch, H. (197A) J. Biol. Chem. 2A9, 6A86. Rhodes, D., Moyer, S. and Banerjee, A. (197A) Cell 2, 327. Rhodes, D., Reddy, D., MacLeod, R., Black, L. and Banerjee, A. (1977) Virology 18, 55A. Robert-Gero, M., Lawrence, F., Farrugia, G., Berneman, A., Blanchard, P., Vigier, P. and Lederer, E. (1975) Biochem. Biophys. Res. Commun. 82, 12A2. Ro-Choi, T., Choi, Y., Henning, D., McCloskey, J. and Busch, H. (1975) J. Biol. Chem. 250. 3921. Roman, R., Brooker, J., Seal, S. and Marcus, A. (1976) Nature (London) 260, 359. Rose, J. (1975) J. Biol. Chem. 222, 8098. Rose, J. and Lodish, H. (1976) Nature (London) 282, 32. Rose, J., Lodish, H. and Brock, M. (1977) J. Virology 21, 683. 57 Rottman, F. (1976) Trends Biol. Sci. 1, 217. Rottman, F. (1978) International Review of Biochemistry, Biochemistry of Nucleic Acids II, 11 (ed. B.F.C. Clark) University Park Press:Baltimore, in press. Rottman, F., Desrosiers, R. and Friderici, K. (1976) Prog. Nucl. Acid Res. Mol. Biol. 12, 21. Rottman, F., Shatkin, A. and Perry, R. (197A) Cell 2, 197. Ryffel, G. (1976) Eur. J. Biochem. 82, A17. Salditt-Georgieff, M., Jelinek, W. and Darnell, J. Cell 1, 227. Samuel, C., Farris, D. and Levin, K. (1977) Virology 81, A76. Sangar, D., Rowlands, D., Harris, T. and Brown, F. (1977) Nature (London) 268, 6A8. Scherrer, K. and Marcaud, L. (1968) J. Cell Physiol. 12, 181. Schibler, U., Kelley, D. and Perry, R. (1977) J. Mol. Biol. 115, 695. Schibler, U. and Perry, R. (1976) Cell 2, 121. Schmincke, C., Herrmann, K. and Hausen, P. (1976) Proc. Natl. Acad. Sci. USA 12, 199A. Sconzo, G., Roccheri, M., diLiberto, M. and Giudice, G. (1977) Cell Differentiation 2, 323. Sen, G., Lebleu, B., Brown, G., Rebello, M., Furuichi, Y., Morgan, M., Shatkin, A. and Lengyel, P. (1975) Biochem. Biophys. Res. Commun. 82, A27. Shafritz, D., Weinstein, J., Safer, B., Merrick, W., Weber, L., Hickey, E. and Baglioni, C. (1976) Nature (London) 281, 291. Shatkin, A. (197A) Proc. Natl. Acad. Sci. USA 11, 320A. Shatkin, A. (1976) Cell 2, 6A5. Shih, D., Dasgupta, R. and Kaesberg, P. (1976) J. Virology 12, 637. Shinshi, H., Miwa, M., Sugimura, T., Shimotohno, K. and Miura, K. (1976) FEBS Letters 82, 25A. Smith, M., Hough, B., Chamberlin, M. and Davidson, E. (197A) J. Mol. Biol. 82, 103. 58 Sommer, S., Lavi, U. and Darnell, J. (1978) J. Mol. Biol., in press. Sommer, S., Salditt-Georgieff, M., Bachenheimer, S., Darnell, J., Furuichi, Y., Morgan, M. and Shatkin, A. (1976) Nucl. Acids Res. 2, 7A9. Sripati, C., Groner, Y. and Warner, J. (1976) J. Biol. Chem. 251, 2898. Stein, J., Stein, G. and McGuire, P. (1977) Biochemistry 18, 2207. Stoltzfus, C. and Dimock, K. (1976) J. Virol. 18, 586. Stuart, S. and Rottman, F. (1973) Biochem. Biophys. Res. Commun. 22, 1001. Suzuki, H. (1976) FEBS Letters 12, 309. Suzuki, H. (1977) FEBS Letters 12, 11. Symons, R. (1975) Mol. Biol. Reports 2, 277. Testa, D. and Banerjee, A. (1977) J. Virology 22, 786. Thomason, A., Brian, D., Velicer, L. and Rottman, F. (1976) J. Virology 29, 123. Thomason, A., Friderici, K., Velicer, L. and Rottman, F. (1978) J. Virol., in press. Tilghman, 8., Curtis, P., Tiemeier, D., Leder, P. and Weissmann, C. (1978) Proc. Natl. Acad. Sci. USA 12, 1309. Tilghman, S., Tiemeier, D., Polsky, F., Edgell, M., Seidman, J., Leder, A., Enquist, L., Norman, B. and Leder, P. (1977) Proc. Natl. Acad. Sci. USA 12, AA06. Toneguzzo, F. and Ghosh, H. (1976) J. Virology 11, A77. Trewyn, R. and Kerr, S. (1976) Onco-Developmental Gene Expression (eds. W. Fishman and S. Sell) Academic Press, Inc.; New York p. 101. Tutas, D. and Paoletti, E. (1977) J. Biol. Chem. 252, 3092. Tutas, D. and Paoletti, E. (1978) J. Virology 22, 37. Weber, L., Fernan, E., Hickey, E., Williams, M. and Baglioni, C. (1976b) J. Biol. Chem. 221, 5657. Weber, L., Hickey, E. and Baglioni, C. (1978) J. Biol. Chem. 222, 178. 59 Weber, L., Hickey, E., Nuss, D. and Baglioni, C. (19763) Proc. Natl. Acad. Sci. USA 12, 325A. Wei, C.-M., Gershowtiz, A. and Moss, B. (1976) Biochemistry 12, 397. Wei, C.-M. and Moss, B. (197A) Proc. Natl. Acad. Sci. USA 11, 301A. Wei, C.-M. and Moss, B. (1975) Proc. Natl. Acad. Sci. USA 12, 318. Wei, C.—M. and Moss, B. (1977) Biochemistry 18, 1672. Williams, J. and Penman, S. (1975) Cell 8, 197. Willis, D., Goorha, R., Miles, M. and Granoff, A. (1977) J. Virology 2_u, 326. Wimmer, E., Chang, A., Clark, J. and Reichmann, M. (1968) J. Mol. Biol. 28, 59. Winicov, I. and Perry, R. (1976) Biochemistry 12, 5039 Yang, N., Manning, F. and Gage, L. (1976) Cell 1, 339. Young, R. (1977) Biochem. Biophys. Res. Commun. 18, 32. Young, B., Birnie, G. and Paul, J. (1976) Biochemistry 12, 2823. Zan—Kowalczewska, M., Bretner, M., Sierakowska, H., Szczesna, E., ggéépowicz, W. and Shatkin, A. (1977) Nucl. Acids. Res. 2, Zimmern, D. (1975) Nucl. Acids Res. 2, 1189. PART II THE KINETICS OF NOVIKOFF CYTOPLASMIC MESSENGER RNA METHYLATION 60 61 ABSTRACT Methylation patterns of Novikoff cytoplasmic mRNA were deter- mined as a function of labeling time with L-(methyl-3H1methionine. The 5'-terminal m7G could be released from whole mRNA by treatment with nucleotide pyrophosphatase. Subsequent alkaline phosphatase treatment of this mRNA followed by KOH digestion yielded N'mpr and N'mpN"mpr from cap 1 (m7GpppN'mpN) and cap 2 (m7GpppN'mpN"mpN) respectively. Our results indicate that the relative amounts of labeled cap structures do change with time and that the amount of internal N6- methyladenosine decreases, relative to 5'-cap structures, as the cytoplasmic mRNAs age and the average size decreases. The formation of cap 2 structures by the addition of the second 2'382methyl group at position N"m appears to be a cytoplasmic event. Thus, after very short labeling times greater than 80% of the labeled methyl groups in cap 2 are found in this position. These results, along with earlier data obtained on L-cell heterogeneous nuclear RNA methylation, are consistent with a model in which the nucleus is the cellular site of three mRNA methylation events producing 5'- terminal m7G, the first 2'eQemethylnucleoside (N'm) found in cap 1 structures and internal N6-methyladenosine. Subsequently these nuclear methylations are followed by the cytoplasmic methylation at N"m. Analysis of the methylnucleoside composition of cap 1 structures, along with comparable "core" structures (m7GpppN'm) generated from cap 2 by removal of N"m, indicates that at any single labeling time the methylnucleoside composition of a given cap 1 and the cap 2 62 "core" structure is remarkably similar. On the other hand, comparisons of the methylnucleoside composition of the cap structures at different labeling times indicate an increase in Cm in the first 2'582methyl- nucleoside (N'm) with time. INTRODUCTION There is an increasing body of evidence to suggest that the majority of eukaryotic mRNA molecules from both cells (Rottman 88 81., 197A; Adams and Corey, 1975; Desrosiers £2.2l-: 1975; Furuichi 2£.§l-I 1975a; Perry 23.22;: 1975a; Wei 22.21:: 1975) and viruses (Furuichi and Miura, 1975; Furuichi £2.2i" 1975b; Furuichi 88_81., 19750; Keith and Fraenkel-Conrat, 1975; Meyer EE.§l-: 1975; Wei and Moss, 1975) contain a unique methylated cap structure, m7GpppN'mp(N"mp)Np on their 5'-termini. Furthermore, recent reports indicate that similar structures are also present on hnRNA (Perry .2£.él-I 1975b; Salditt-Georgieff 88 81., 1976), a class of molecules that contains mRNA precursors. In addition to a 5'-terminal cap structure, both cytoplasmic mRNA and hnRNA also contain N6-methyl- adenylic acid located internally between the cap and 3'-poly(A) segment. Although this mononucleotide is apparently absent in many viral RNAs (Furuichi and Miura, 1975; Mbyer 88_81., 1975; Wei and Moss, 1975), globin mRNA (Perry and Scherrer, 1975) and yeast mRNA (Sripati 22.22-2 1976; Dekloet and Andrean, 1976), it can account for nearly 50% of the labeled methylnucleotides in cellular mRNA (Desrosiers EE.El-: 197A). In a recent characterization of Novikoff cytoplasmic mRNA we noted the presence of two types of cap structures: m7GpppN'mpr (cap 1) and m7GpppN'mpN"mpr (cap 2) (Desrosiers £2.21-r 1975). 63 Similar results have been reported for mouse myeloma cells (Adams and Cory, 1975), L-cells (Perry 88_81., 1975a) and HeLa cells (Wei 22.2l-I 1975; Furuichi 88_81., 1975a). In an earlier publication (Desrosiers §£.§l-I 1975) it was suggested that the two cap structures might in fact represent separate classes of mRNA with different degrees of cytoplasmic stability. Previous studies on cap structures in cytoplasmic mRNA have utilized cells labeled for a single time period. Therefore, it was reasoned that the relative abundance of cap 1 and cap 2 structures might depend upon the length of labeling time, possibly reflecting different kinetics of labeling or turnover. The presence of such time-dependent changes might, however, also reflect other factors such as additional cytoplasmic methylation of partially methylated precursors. To study the possibility that differential cytoplasmic mRNA stability affects the distribution of cap structures, and at the same time to permit continual cytoplasmic methylation to occur, we employed continuous labeling of Novikoff cells with L-[methyl- 3H]methionine for periods up to 2A h. The methylation patterns of mRNA were then analyzed as a function of time. MATERIALS AND METHODS Cell Culture and Labeling;Conditions Novikoff hepatoma cells (N1S1 strain) were grown in Swim's S-77 medium (GIBCO) containing 10% calf serum essentially as described (Desrosiers 88_81., 197A). For labeling with L-[methyl—3H1methionine (Amersham/Searle, 5 Ci/mole), cells in midlogarithmic growth phase were harvested aseptically and resuspended in fresh warm medium 6A at a concentration of approximately 7.5 x 105 cells/ml. Labeling for the time course study was performed in the presence of 20 mM sodium formate and A0 HM each adenosine and guanosine to supress nonmethyl purine ring labeling; normal methionine levels were present for all labeling times except for 20 min, in which medium without methionine was used. Labeling conditions for the distributional analysis experiments were altered to increase the specific activity of the methylated nucleotides in mRNA. Adenosine, guanosine and formate were omitted and the 5 and 2A h samples were labeled in 50 uM methionine, one half the normal concentration. Under these conditions there was no discernable change either in cell doubling time or in cell appear- ance when examined by phase microscopy. In most experiments 15-20 uCi/ml of L-[methyl-BHJmethionine was used. At each time point an aliquot of the cells was harvested aseptically and the radioactive medium returned to the growing culture. Final cell concentrations at the time of harvest never exceeded 1.3 x 106/ml. Isolation and Characterization of Poly (A)-Containing Cytoplasmic 5% Total cytoplasmic RNA was isolated as previously described (Desrosiers 88_81., 197A). Poly (A)-containing mRNA was isolated by oligo (dT)-ce11ulose chromatography, including a heat step prior to a second passage over the column (Desrosiers 88H81., 1975). This step was necessary to eliminate traces of rRNA that otherwise interfere with methylation analysis. 65 Sedimentation analysis of poly (A)-containing mRNA was performed using A.8 ml gradients of 5-20% sucrose in 99% MeZSO, 10 mM LiCl, 1 mM EDTA. The mRNA was made 91% M6280, 10 mM LiCl, 1 mM EDTA in a total volume of 100 ul and heated at 60°C for 2 min prior to layering onto the gradient. Centrifugation was for 1A.5 h at 25°C and A5,000 rpm in a Beckman SW 50.1 rotor. Nucleotide Pyrophosphatase Treatment of Whole mRNA Poly (A)-containing cytoplasmic mRNA, essentially free of tRNA and rRNA contamination, was digested with nucleotide pyrophosphatase from Crotalus atrox (Sigma). A 200 pl reaction contained 0.25 units enzyme, 9 A units RNA, 20 umoles Tris HC1, pH 7.8 and 0.2 umoles 260 of magnesium acetate. After incubation at 37°C for 35 min the reaction was stopped by heating in a boiling water bath for 5 min. The high level of carrier RNA was added to suppress nonspecific diesterase activity which contaminates this enzyme. The RNA was separated from the released pm7G by chromatography on Biogel P2 (100-200 mesh; 1.5 x 22 cm column) and treated with 0.25 units bacterial alkaline phosphatase (PL Biochemicals, electrophoretically pure) in 0.05 M Tris HC1, pH 7.8, 0.001 M magnesium acetate for A5 min at 37°C to remove the newly exposed 5'-terminal phosphates. The dephos- phorylated RNA was then digested for 18 h at 37°C with 0.A N KOH to obtain N'mpr from cap 1 and N'mpN"mpr from cap 2 structures. These oligonucleotides were resolved from each other and m6Ap by chromatography on Pellionex WAX in the presence of 7 M urea (Desrosiers 91 318., 1975). 66 Prgparation of mRNA for Methyl Nucleoside Distributional Analysis Internal methylnucleosides and intact 5'-terminal caps were produced by enzymatic digestion of poly (A)-containing cytoplasmic mRNA with RNase T2 (Sigma) at 2 units/A26 unit of RNA in 0.9 M 0 NaCl, 0.15 M sodium acetate pH A.5, 0.01 M EDTA, for 2 h at 37°C. The reaction mixture was then adjusted to pH 8 with 1 M NaOH and made 0.017 M in magnesium acetate. Alkaline phosphatase that had been dialyzed against 0.05 M NH HCO was added (0.25 units/A26 u 3 0 unit of RNA) and the reaction continued for 30 min at 37°C. The products of this reaction were resolved on DEAE-Sephadex (7 M urea). Intact cap structures to be used for subsequent analysis were desalted on Biogel P-2 (100—200 mesh). Nucleosides were adsorbed to charcoal and eluted with 20% pyridine. Cap 2 structures (m7GpppN'mpN"mpN) were purified on Pellionex- WAX to remove remaining traces of urea which were found to inhibit subsequent digestion with nuclease P1. Cap 2 oligonucleotides con- tained in a volume of 500 pl were applied to Pellionex-WAX (1/7" x 30 cm) and urea was removed by eluting with 10 ml 0.1 M ammonium acetate. The buffer was changed to 6 M ammonium acetate and the cap structure eluted in 2 ml total volume. Ammonium acetate was removed by lyophilization. Cap 2 structures were digested with 160 ug/ml nuclease P1 (Yamasa Shoyl Co., Ltd.) in 0.01 M sodium acetate, pH 6.1. After A5 min at 37°C the sample was made 0.05 M Tris HC1, pH 7.8, 0.001 M in magnesium acetate and 0.3 units alkaline phosphatase/100 ul were added. Incubation was continued for 30 min at 37°C. The reaction mixture was diluted to 500 ul with water and reapplied to the Pellionex-WAX column described above. N"m 67 was eluted with 0.1 M ammonium acetate (2 ml total volume) and "core" oligonucleotide, m7GpppN'm, eluted with 6 M ammonium acetate. Both cap structures and "core" oligonucleotides generated from cap 2 were completely digested to nucleosides by incubation for A5 min at 37°C with 0.25 units nucleotide pyrophosphatase and 0.A units alkaline phosphatase in 100 pl reactions containing 0.1 M Tris HC1, pH 7.8, 0.1 mM magnesium acetate. Acid Hydrolysis Acid hydrolysis of whole mRNA, 5'-terminal oligonucleotides and mononucleotides can be used to cleave the N-glycosidic bond of purine-containing nucleotides, thereby releasing free purine bases. Generally 1.5 A2 unit of RNA was dissolved in 0.5 ml concen- 60 trated formic acid, the tube sealed and the hydrolysis carried out at 100°C for 2 h, similar to the procedure of Munns ££.§l-v(197u)- The released bases were resolved by high speed liquid chromatography [1uC1Adenosine (HSLC) on Aminex A-5 (Desrosiers £2.2l" 1975). was added as an internal standard to the [BHJ-labeled RNA before hydrolysis to permit determination of 1uC/3H ratios after digestion and thus provide a measure of the methanol lost from 2'ggfmethyl groups in the presence of strong acid. RESULTS Cap 1 and cap 2 structures differ from each other by containing one and two 2'52rmethylnucleosides respectively. These structures can be resolved in the presence of 7 M urea on DEAE-Sephadex columns or on Pellionex-WAX. Frequently the resolution obtained on intact whole cap structures on Pellionex-WAX is not satisfactory, even 68 in the presence of 7 M urea which is added to suppress base composi- tion effects. Removal of terminal m7G with nucleotide pyrophosphatase followed by treatment with bacterial alkaline phosphatase leaves the mRNA with a 5'-terminal end of N'mpr... or N'mpN"mpr ..... , corresponding to cap 1 and cap 2 structures, respectively. Subsequent alkaline hydrolysis of the remaining portion of mRNA produces N'mpr and N'mpN"mpr, which are easily and quickly separated by HSLC on Pellionex-WAX. Novikoff mRNA was labeled for varying times with L-[methyl-3H]- methionine and rigorously purified to eliminate rRNA as described in MATERIALS AND METHODS. Whole mRNA treated with nucleotide pyro- phosphatase yielded a mononucleotide which was hydrolyzed with formic acid and the hydrolysate chromatographed on Aminex A-5. Greater than 85% of the released material eluted as 7-methylguanine (data not shown). The mRNA remaining after nucleotide pyrophosphatase treatment was further hydrolyzed with alkaline phosphatase and KOH. The internal base-methylated mononucleotide m°Ap and the 5'-terminal oligonucleotides N'mpr and N'mpN"mpr were separated on Pellionex- WAX (Figure 1). One of the main objectives of these studies was to examine the relative distribution of cap 1 and cap 2 structures as a function of time of continuous labeling with L—[methyl-3Hlmethionine. A pronounced change in labeling of these cap structures was observed with different labeling times (Figure 1). After a short exposure of only 20 minutes, most of the label in cytoplasmic mRNA was contained in cap 2 while further labeling yielded an increase in cap 1. Similar determinations were made at later time points and the ratios of 69 Figure 1. HSLC resolution on Pellionex-WAX of KOH diges- tion products from mRNA which had previously been treated with nucleo- tide pyrophosphatase and alkaline phosphatase. A 1/8" x A0 cm column was developed at room temperature with a 100 ml gradient of 0-0.2 M (NHA)ZS°A in 7 M urea, 0.005 sodium phosphate, pH 7.7 at a flow rate of 25 ml/h. The position of the oligo(Up) standards added as UV markers and carrier rRNA digestion products are shown. Poly (A)-containing RNA was isolated at the times indicated. 3H counISI min. 70 Nmpr Up3 Up4 Up Up 5 6 200' Nmpr Up3 Up, Ups Up6 H1 1 ll q .11. db db H 24 m. 200T Nmpr Up3 Up. Ups Ups loo-H1111 J l l 1 250 500 750 1000 070p No. Figure 1 71 radioactivity between cap 1 and cap 2 were observed to change (Figure 2). To further identify and study the methylnucleoside distribution in internal and cap positions in poly (A)-containing mRNA, three periods of labeling were chosen: 20 min, 5 h and 2A h. Although Pellionex-WAX efficiently resolves N'mpr and N'mpN"mpr derived from cap 1 and cap 2, respectively, DEAE-Sephadex (7 M urea) is better suited for obtaining intact cap structures. Whole mRNA was digested with RNase T2 and alkaline phosphatase to produce nucleosides from internal base methylations plus the 5'-terminal caps 1 and 2. These products were resolved on DEAE-Sephadex developed in the presence of 7 M urea (Figure 3). It is again apparent that the distribution of 3 fl-radioactivity between nucleoside, cap 1 and cap 2 varied significantly with time of labeling. The nucleosides recovered from urea buffer by charcoal adsorption were analyzed on Aminex A-5 under conditions that resolve base-modified nucleosides (Desrosiers £2.2l-2 197A). N6-Methyladenosine comprised greater than 951 of the base methylated nucleosides at all times of labeling, but small amounts of radioactivity eluting with 5-methyl- cytidine were observed at later times (data not shown). To correlate the number of internal methylations per messenger RNA molecule with labeling time, the amount of internal m°A was compared to the amount of m7G in whole poly (A)-containing mRNA which had been labeled for the times indicated (Table I). The amount of m7 G in mRNA should reflect the absolute number of modified 5'- termini without being complicated by the number of ribose methylations in each message (i.e., cap 1 vs. cap 2). 72 Figure 2. Change in ratio of N'mpr to N'mpN"mNp in mRNA with time. The data was obtained from mRNA that had been labeled with L-[methyl-3H1methionine for the times indicated and treated as described in the legend to Figure 1. Dinucleotide contamination from rRNA was determined by KOH hydrolysis of intact mRNA and chroma- tography on Pellionex-WAX as in Figure 1. Appropriate corrections were made. Ratio of radioactivity in N'mpN"mpr, eluting with the (Up)3 marker, vs N'mpr was calculated. NmmepN / NmpNP 73 3.0 ' 2.0- 1.0L .17 .41 IO 20 Time of Labeling (ha) Figure 2 7A Figure 3. DEAE-Sephadex column separation of RNase T2 and alkaline phosphatase digestion products from whole mRNA. Poly (A)-containing mRNA which had been incubated with L-[methyl-BH]- methionine for various time periods was digested with RNase T2 and alkaline phosphatase as described in MATERIALS AND METHODS. Following digestion the reaction mixture was diluted with 9 volumes of 7 M urea, 0.02 M Tris HC1, pH 7.A and applied to a 0.9 x 25 cm DEAE- Sephadex column. The mononucleosides were eluted with 0.1 M NaCl, 7 M urea, 0.02 M Tris HC1, pH 7.A. To resolve oligonucleotides a 200-ml gradient of 0.1 M to 0.A M NaCl in 0.02 M Tris HC1, pH 7.A, 7 M urea was used at a flow rate of 12 ml/h; 2 ml fractions were collected. Standard oligonucleotides (pUm)3 and (pUm)u were included as markers to indicate approximate charge. IOO 50 400 200 3H CPM 100 400 200 IOO I (A) I 20 minutes (B) 5 hours I 100 (C) 24 hou I—I—H—n—I 1 IS 20 40 Figure 3 i 60 FRACTION NUMBER 76 Table I Amount of Internal m6A Per Average mRNA: Variation with Time 1 of Label in mRNA as: m7Ga m°Ab Ratio m6A/m7G 20 min 5.5 28 5.1 5 h 17 58 3.A 13 h 19 53 2.8 2A h 27 35 1.3 aDeterminations were made by acid hydrolysis of whole mRNA and analyzed on Aminex A5 HSLC. bDetermination of m°A was the amount of mononucleotide from the DEAE-Sephadex (7 M urea) column of T2 and alkaline phosphatase digest (of. Figure 3) corrected for ring labeling when necessary. 77 As indicated in Table I the ratio of internal m6A to terminal m7G decreases with time, indicating that the average number of internal methylations in poly (A)-containing mRNA is reduced in longer labeling periods. Since other reports have shown that larger hnRNA (Salditt-Georgieff 22.21-9 1976) and mRNA (Perry and Kelley, 1976) have a higher average number of internal m°A residues, it was of interest to examine the size of [3H-methyl]-labeled mRNA as a function of time. Poly (A)- containing mRNA from three labeling periods was analyzed by MeZSO- sucrose gradient centrifugation (Figure A). Messenger RNA labeled for shorter times was substantially larger in size than mRNA labeled for longer periods, i.e., with increased labeling time the per cent of the molecules sedimenting at less than 188 changes from 32% to 5A%. Since there is as time dependent change in m°A content, size of message, and cap 1 to cap 2 ratio, the methylnucleoside distribution within the caps was also studied. Cap 1 and cap 2 structures produced by RNase T2 and alkaline phosphatase were resolved on DEAE-Sephadex (7 M urea) (Figure 3) and desalted on Biogel as described in MATERIALS AND METHODS. Cap 1 structures were digested with a mixture of nucleotide ,pyrophosphatase and alkaline phosphatase to produce 2'ggymethyl- nucleosides and m7G. The separation of methylnucleosides found 111 cap 1 structures is presented in Figure 5. Only results from mRNA obtained at 5 and 2A h are included since the amount of radio- actiwdty in cap 1 at 20 min is too small to analyze (Figure 3). In“: important aspects of this data should be mentioned: first, the major change in methylnucleoside composition as a function of time 78 Figure A. MeZSO-sucrose gradient profiles of poly (A)- containing cytoplasmic RNA. Preparation of mRNA and gradient centri— fugation was as described in MATERIALS AND METHODS. The percentage of the total radioactivity present in fractions smaller than 188, between 188 and 28S, and greater than 288 was calculated for cacti RNA sample. Messenger RNA was isolated after labeling with L- [methyl‘ 3H]methionine for A) 20 min B) 5 h and C) 2A h, as described in MATERIALS AND METHODS . 79 20 s 'III' 4lv|||| Ia. m s .. 8|..le III m & Zlvllll. 10 a m % a % n s 2 u I .I r I o '- m w — h . O h 4 2 5 2 . mm mm mm . _ b _ h _ v 0 O O 0 O O 4 4 2 4 2 zao In FRACTION NUMBER Figure A 80 Figure 5. The distribution of methylnucleosides in cap 1 structures. Cap 1 structures produced by RNase T2 and alkaline phosphatase treatment were eluted from a DEAE-Sephadex (7 M urea) column in a volume of 10-20 ml and desalted on a 1.9 x A2 cm Biogel P2 column by elution with 0.02 M NHuHCO3. Material in the void volume was made 20% with ethanol and evaporated. Cap 1 structures were then digested with nucleotide pyrophosphatase and alkaline phosphatase as described in MATERIALS AND METHODS. The reaction mixture was dried with N2 and dissolved in 125 pl column buffer. HSLC on Aminex A-5 (1/8" x 90 cm) was in 0.A M ammonium formate, pH A.25, A01 ethylene glycol at A0°C. Flow rate was 7 ml/h (A750 psi) for remainder of the run. Fraction size was 10 drop ( 0.A ml) until Cm was eluted; fraction size was then doubled. A) Cap 1 from mRNA labeled for 5 h. Inset is the acid hydrolysis of the same 5 h cap 1 structure analyzed on Aminex A-5 in 0.A M ammonium formate at pH 5.3. B) Cap 1 from mRNA labeled for 2A h. 2'ggymethyl- 7 nucleosides and m guanosine were added as markers and detected at 260 nm. 81 2'-O-me I I 'oJ m7Guo A Gun IAdc 100- I ' I 200- 50 m7G Um Gm Am 1 l l 1 ° IOO- Cm I :E 25 o : I I 1 ":31: mIG B 100— F' Um Gm Am Cm l l I I (I 50‘ 11 ‘vah I I I I O 20 4O 60 80 FRACTION NUMBER Figure 5 82 is the increase in Cm content and second, it appears that most of the Am is present as a doubly-methylated derivative, N6, 2'ggrdimethyl- adenosine (m6Am). Verification of the m°A content in this dimethylated nucleoside was obtained by subjecting an aliquot of isolated cap 1 to acid hydrolysis and subsequently analyzing the free bases pro- duced, as described earlier. After 5 h labeling, 50% of the label 6 I‘— in m Am was converted to m°Ade (of. Figure 5, inset). Also the percentage of label in m7 Gua equals the percentage of label in 2'- 1 O-methyl products. Similar results were obtained with acid hydrolysis of cap 1 derived from 2A h mRNA, i.e., 87% of the Am is found in the form of m°Am. The methylnucleoside distribution in the N"m position of cap 2 was determined by digesting cap 2 structures with nuclease P1 which produces m7GpppN'm. The released methylnucleoside N"m can be separated from the remainder of the cap structure by HSLC on Pellionex-WAX in ammonium acetate and subsequently analyzed on Aminex A-5. Of the label in cap 2 after 20 min, 80% was released as N"m. The N"m position of 2A h labeled cap 2 appears to be particularly rich in Um, and contains a significant amount of Am (Figure 6). Acid hydrolysis of this N"m nucleoside produced no m°Ade (data not shown). The overall distribution of methylnucleosides at each specific site of Novikoff mRNA methylation after 20 min, 5 h and 2A h of continuous labeling with L-[methyl-3H1methionine is shown in Table II. 83 .m onswfim 8H no cosqmnwoum8onno 88m Hoomaw ocoaznuo no: 8H mm.: mm .opmanom 88H8o88m z 2.0 82 uo>Hoann .conaafinaoma .Aoumuoom ssfinossmv x<3lxonofiaaom 808m nousao mm: m moo mo coflumomflo Pm omooaosn soon nocfiopno ooHnooHosn 8:2 one .oqmm mI< xocfi8< 22 n am now uoaonoa 4228 no oufimooaosc 8:2 no mHn2Hmno sowsonfinunflo ovamooaoscaznuoz .o onswam .OZ ZO_._.U<~I 8A 0.? CW O 100 s s s s a EU E< EU ED 2.0.; .33 I00— Figure 6 W03 H8 85 .oonmowone nofluanoq Honsuosnnm onn ca nnomona 2na>wpomoaumn HmnOp on» no omonnoosoq o no connomona we memo one .m onsmflm an no nonmaonm cnm oomnmnamonnonea onenooaosn noes oopmomfiu ono: monznosnnm e moo com ocfinooaoznomHHo onoo .m onswwm 82 no confia Inouoo no: 8:2 8H nonwmooaozn no nownsnfinumflo one .Aononooo amzv n8saoo x<3IxonoHHaom o no vo>Homon onoz onemooaosn 8:2 one ocfinooaosnomfiao :onoo: on» .onmoaozn aswaaaoflnoq no“: monsnosnnm qmo no nofiumomflo mnflzoaaom .Aoon: z ev xoomnqomumnma no oononmqom onoz monsnoznnn m goo one e goo “onmumna Imona onwamxao new we oomzm nun: nonnoweu no: 2228 wnflnamnnooIA 8H mouHmooHosnahnnoemmMH no nofinsnenumflo HH oanme 87 DISCUSSION In earlier studies (Desrosiers EEHEl-I 1975) we examined the distribution of methylnucleosides in Novikoff mRNA after continuously labeling the cells for 13 h with L-[methyl-BHJmethionine. Analysis of mRNA methylnucleoside composition indicated that an average Novikoff mRNA contained 5.7 methyl groups, 3 of which were present 7G and 1.7 Nm's. The non-integral as internal m6A, 1 terminal m number of Nm's could result from an unequal distribution of cap 1 and cap 2 structures as well as m°Am in the N'm position. An explanation for the necessity of two different types of cap structure in mRNA was not readily apparent. However, the stabilizing effect that 2'582methyl groups have on RNA molecules in the presence of specific nucleases (Stuart and Rottman, 1973) led us to suggest that the ratio of these two cap structures might reflect multiple classes or fractions of mRNA with different cellular stabilities (Desrosiers 88,81., 1975). If this were true, one might expect the relative levels of these two structures to change as a function of labeling time. Previous determinations of cap 1 to cap 2 ratios utilized mRNA that had been treated by periodate oxidation and B-elimination to remove terminal m7G prior to alkaline digestion. Since the elimina- 7 tion of m G by periodate oxidation of mRNA was not reproducible in our laboratory, we explored the possibility of using the enzyme nucleotide pyrophosphatase on intact mRNA molecules and found this method of In7 G removal to be superior. It should perhaps be noted ‘that the susceptability of the 5'-terminal m7G-containing cap to nucleotide pyrophosphatase implies that it must be open and accessible 88 to the enzyme and not buried within the folded structure of the mRNA molecules. Following removal of terminal phosphate and subse- quent alkaline digestion, the N'mpr and N'mpN"mpr obtained from cap 1 and cap 2 structures, respectively, were separated by HSLC on Pellionex-WAX. Poly (A)-containing mRNA was isolated from Novikoff cells that had been labeled with L-[methyl-BHImethionine for different periods of time. Analysis of cap structures obtained from these mRNA samples indeed indicated that the relative labeling of cap 1 and cap 2 changed with time (Figure 2). At 20 min there was much more L-[methyl-BH] label in cap 2, reflecting enhanced labeling at the N"m position of cap structures that were earlier methylated in the nucleus at m7G and N'm from cold methyl precursors. After 5 h the ratios were reversed (Figures 1 and 3), and after longer periods (2A h) the ratio approaches a "steady-state" level of cap 2 to cap 1 of approx- imately two. These studies employed continuous labeling with L- [methyl-3H]methionine to permit continued formation of newly methylated mRNA sequences, since the ratio of cap structures present at a particular time reflects both synthetic and degradative events. The time-dependent changes observed in the relative labeling of each cap structure prompted us to further characterize the methyl- ation occurring at each specific site within mRNA as a function of time. Such an approach necessitated the prior separation and isolation of 5'-terminal cap 1 and cap 2 structures as well as the In°A located within the mRNA. DEAE-Sephadex (7 M urea) columns provided excellent resolution of the material obtained from 20 min, 5 h and '2A h-labeled mRNA (Figure 3). Analysis of the material present 89 in the mononucleotide fractions by subsequent HSLC on Aminex A—5 showed it to be mainly N°-methyladenosine (data not shown). Since there have been several reports of 5-methylcytidine (mSC) in viral and cellular mRNA (Dubin and Stoller, 1975; Salditt-Georgieff 88 81., 1976) the mononucleotide fraction was further analyzed. HSLC 6 5 on Aminex A-5 provided separation of m A and m C and indicated a 5 small peak of radioactive material eluting with m C at later labeling 5C was detected. times (data not shown). After 20 min very little m The maximum levels at 5 h and 2A h were 2% and At of the total internal methylnucleoside, respectively, indicating a low but possibly signifi- cant amount of this methylnucleoside that accumulates with time. Another interesting time-dependent comparison involves the internal methylnucleoside m°A and 5'-terminal m7G. With longer labeling 6 7 times the amount of m A relative to m G decreases (Table I). This could be due to a selective time-dependent loss of mRNAs rich in m°A or an increase in the number of 5'-terminal m7G residues on mRNA. Determination of the average size of Novikoff mRNA sequences on denaturing sucrose gradients indicates a distinct reduction in mRNA size as a function of labeling time (Figure A). At 20 min most of the mRNA sediments in MeZSO gradients at approximately 208. At 5 h a biomodal distribtuion is obtained with components sedimenting at 208 and 158 while after 2A h most of the mRNA sediments at 158. This indicates a reduction in the average size of mRNA sequences with time accompanied by a loss of m6 A which may be distributed at approximately equal intervals throughout the mRNA. Similar results have been obtained in duplicate experiments in which cells were grown in the presence or absence of adenosine, guanosine and sodium 90 formate, which were added to suppress purine ring labeling. Also it should be noted that these results on mRNA are in essential agree- ment with recent studies on HeLa hnRNA (Salditt-Georgieff 22.22-I 1976) and L-cell mRNA (Perry and Kelly, 1976). However, the HeLa and L-cell studies concentrated on the relative reduction in the number of internal m°A residues with reduction in size of RNA at a fixed labeling time. Thus in Novikoff cells it can be stated that with increased labeling time the average size of methyl-labeled mRNA gets shorter and the content of internal m°A is reduced. The isolation of labeled cap 1 and cap 2 permits a systematic analysis of specific methylated positions within these oliognucleo- tides as a function of labeling time. Comparisons can be made between cap 1 from different labeling times as well as between cap 1 and the analogous "core" structure, In7 GpppN'm, from cap 2. Earlier analyses were performed on either mixtures of cap 1 and cap 2 (Des- rosiers 88,81., 1975) or on cap structures obtained at a single fixed time of labeling (Perry SE.§l" 1975b). The methylnucleoside composition of cap 1, particularly at position N'm, appears to change with time. Although the relative labeling of Um, Gm and Am is nearly equivalent at 5 and 2A h (Figure 5 and Table II) there is a significant increase in Cm with time. This increase probably reflects a time- dependent enrichment of a sub-class of mRNA sequences with enhanced cytoplasmic stability. Also, as shown in Figure 5, virtually all of the Am present in the N'm position is found as the doubly methylated nucleoside, N6, 2'egemethyladenosine. At 5 h the amount of radio- activity recovered as N6, 2'gggmethyladenine is consistent with 6 the presence of only m Am and no Am, while at 2A h the distribution 91 is 87% m°Am, 13% Am. This result differs from comparable analyses on L-cell mRNA which contained larger amounts of singly methylated Am in the N'm position (Perry and Kelley, 1976). In an attempt to determine if the flux in cap 1 and cap 2 labeling was primarily due to labeling of a specific site within the cap structure or, alternatively, general labeling at all positions, the N"m position was selectively removed and separately analyzed. After 20 min, greater than 80% of the methyl label is in the N"m position (Table II). The distribution of methylnucleosides in the N"m position at 2A h labeling (Figure 6) indicates a high level of both Um and Am. None of the Am appears as the doubly methylated nucleoside m°Am. Removal of N"m from cap 2 results in the production of "core" structures that can be easily separated from N"m and recovered on Pellionex-WAX in the presence of ammonium acetate. Following complete degradation with nucleotide pyrophosphatase and alkaline phosphatase the methylnucleosides in the N'm position and m7 G can be analyzed on Aminex A-5. These data enable one to make an interesting com- parison between the methylnucleoside composition of cap 1 and that of the analogous "core" structure derived from cap 2. The results summarized in Table II indicate that at a given labeling time the distribution between these two structures is remarkably similar. This correspondence in methylnucleoside distribution even extends to the relative increase in Cm observed at later labeling times. Data obtained with L-cell cap 1 and cap 2 "core" structures showed differences in composition at the N'm position (Perry and Kelley, 1976). Aside from different cell types used in both experiments 92 it should be noted that studies with L-cells were performed under pulse-chase labeling conditions while for Novikoff cells continuous labeling was employed. The data obtained with Novikoff mRNA do not suggest a selective methylation of certain mRNAs containing a unique cap 1 Nm content to which a N"m is added to form cap 2 structures. The presence of altered methylnucleoside compositions in mRNA's with different cytoplasmic half lives could produce the differences observed in pulse chase vs. continuous labeling. Earlier data on hnRNA methylation patterns indicated the pre- sence of 5'-terminal cap 1 structures and internal m°A but no cap 2 (Perry EEHEl-v 1975b). These data, in combination with the present study, are consistent with a model in which m7 m°A are products of nuclear methylation events. After exit of mRNA G, N'm and internal sequences containing these modifications into the cytoplasm, there is a cytoplasmic methylation event at the N"m position, forming cap 2. The selective enrichment, with time, of mRNA sequences con- taining a relatively high Cm composition suggests an enhances stability of a subclass of mRNAs that happen to have this altered methylnucleo- side composition. The complexities of this system in which thousands of mRNA sequences are being simultaneously synthesized and degraded are indeed enormous. Examination of methylnucleoside composition in mRNAs collectively grouped by cap 1 and cap 2 content is only a first step in following these mRNA modifications as a function of time. Of real interest will be those studies dealing with a specific homogeneous eukaryotic mRNA sequence. REFERENCES Adams, J.M. and Cory, S. (1975), Nature (London) 222, 28. DeKloet, S. and Andrean, B.A.G. (1976) Biochim. Biophys. Acta A25, A01. Desrosiers, R., Friderici, K. and Rottman, F. (1975), Biochemistry ‘12, A367. Desrosiers, R., Friderici, K. and Rottman, F. M. (197A), Proc. Nat. Acad. Sci. USA 11, 3971. Dubin, D.T. and Stollar, V. (1975), Biochem. Biophys. Res. Commun. éé. 1373. Furuichi, Y. and Miura, K. (1975), Nature (London) 25 . 37A. Furuichi, Y., Morgan, M., Shatkin, A.J., Jelinek, W., Salditt-Georgeiff, M. and Darnell, J.E. (1975a), Proc. Nat. Acad. Sci. USA 12, 190A. Furuichi, Y., Muthukrishnan, S. and Shatkin, A.J. (1975b), Proc. Nat. Acad. Sci. USA 12, 7A2. Furuichi, Y., Shatkin, A.J., Stravnezer, E. and Bishop, J.M. (19750), Nature (London) 257, 618. Keith, J. and Fraenkel-Conrat, H. (1975) FEBS Letters 21, 31. Moyer, 8., Abraham, G., Adler, R. and Banerjee, A.K. (1975), Cell 2, 59. Munns, T., Podratz, K. and Katzman, P. (197A), Biochemistry 12, AA09. Perry, R.P. and Kelley, D.E. (1976), Cell, 8, A33. Perry, R.P., Kelley, D.E., Friderici, K. and Rottman, F. (1975a), Cell 2, 387. Perry, R.P., Kelley, D.E., Friderici, K.H. and Rottman, F.M. (1975b), Cell 8, 13. Perry, R.P. and Scherrer, K. (1975), FEBS Letters 21, 73. 93 9A Rottman, F., Shatkin, A., and Perry R. (197A), Cell 2, 197. Salditt-Georgieff, M., Jelinek, W., Darnell, J.E., Furuichi, Y., Morgan, M. and Shatkin, A. (1976), Cell 1, 227. Sripati, C.E., Groner, Y. and Warner, J.R. (1976), 8, Biol. Chem., 251. 2898. Stuart, S.E. and Rottman, F.M. (1973), Biochem. Biophys. Res. Commun. 22, 1001. Tener, G.M. (1976), Methods 18_Enzymology XII, 398. Wei, C.M., Gershowitz, A. and Moss, B. (1975). Cell 2, 379. Wei, G.M. and Moss, B. (1975), Proc. Nat. Acad. Sci. USA 12, 318. PART III 12 VIVO INHIBITION OF NOVIKOFF CYTOPLASMIC MESSENGER RNA METHYLATION BY S-TUBERCIDINYLHOMOCYSTEINE 95 96 ABSTRACT The analogue S—tubercidinylhomocysteine (STH) has been used to study the methylation of mRNA 32.2212- Partial inhibition of cytoplasmic poly (A)-containing RNA methylation was observed using a level of inhibitor which still permitted cell growth. Character- ization of the partially methylated mRNA indicated the presence of cap structures lacking 2'-Q¢methylnucleosides, m7GpppN, which are normally not present in mammalian mRNA. Inhibition of additional methylated sites in mRNA at the second 2'5Qemethylnucleoside, and at internal N6-methyladenosine was also observed. Methylation of 7-methylguanosine was not affected under the conditions used in these experiments. The methylnucleoside composition of cap structures differed in STH-inhibited and uninhibited cells. These results indicate that a completely methylated cap is not required for trans- port of mRNA into the cytoplasm. Furthermore, it may now be possible to assess 18 vivo the sequential nature of mRNA methylation and its potential role in mRNA processing. INTRODUCTION Most eukaryotic mRNA molecules are now known to contain blocked and methylated 5'-terminal structures called "caps" (of. reviews Shatkin, 1976; Rottman, 1976). These structures are commonly found in two forms, m7GpppN'mpN (cap 1) and m7GpppN'mpN"mpN (cap 2). In addition, some mRNA molecules contain internally located N6-methyl- adenosine (m6A) (Desrosiers, ££.él-’ 197A; Adams and Cory, 1975; Perry 88 81., 1975a) and occasionally 5-methylcytosine (mSC) (Dubin and Stollar, 1975). 97 Identification of the cap structures, together with evidence suggesting that these methylation events may occur at specific times and separate cellular locations (Perry and Kelley, 1976; Friderici, £2.2l-I 1976), has intensified the need to address the question of mRNA methylation function 22.2212- Previous studies on the function of mRNA methylation have focused on the role of the cap structures EBHYEEEQ at the translational level, using cell-free protein-synthe- sizing systems (Muthukrishnan, 22.22-1 1975; Both, 22.2l-I 1975a), ribosome binding studies (Dasgupta, 2£.§l°: 1975, Both, ££.§l°9 19750; Kozak and Shatkin, 1976) and cap analogues as specific inhibitors of binding and/or translational activity (Hickey, 22.2l-I 1976; Weber, 22.22-» 1976). However, the possible relationship between methylation events and processing of mRNA remains to be established. Precedent for such a relationship exists in studies on rRNA maturation in HeLa cells grown under conditions of limiting methionine (Vaughn, 88,21., 1967). In preliminary studies performed in our laboratory, Novikoff hepatoma cells were deprived of methionine in an attempt to reduce methylation of mRNA, and thereby perturb mRNA processing. Under conditions which inhibited the appearance of cytoplasmic rRNA through reduced methylation of the nuclear precursor, mRNA methylation was maintained at the normal level. This result might reflect an "internal priOrity system" for the utilization of available methyl donors, the net result of which is to protect methylation of mRNA and ensure its complete complement of methyl groups. We concluded that methylation of mRNA could not be affected by depletion of the common amino acid methionine. If an evaluation of the interaction between methylation 98 and processing of mRNA was to be feasible, a specific 12.2122 inhi- bitor of mRNA methylation would be necessary. S-tubercidinylhomocysteine (STH), the 7-deaza analogue of S- adenosylhomocysteine (SAH), has been synthesized in one of our labor- atories (Coward, 88_81., 197A) and has been shown to function as an effective 22.2122 inhibitor of tRNA methylation in cultured stimu- lated rat lymphocytes (Chang and Coward, 1975), and of dopamine methylation in murine neuroblastoma cells (Michelot, 22.22-: 1977) and human fibroblasts (X.O. Breakefield and J.K. Coward, unpublished results). The use of STH as a methylation inhibitor offers several advantages: it is able to permeate cell membranes, and it is not susceptible to the enzymes responsible for SAH metabolism in mammalian cells (Crooks, 22.21-2 1977). We have investigated the effect of STH on the methylation of Novikoff RNA, and have presented below the characterization of the methyl distribution in the poly (A)-containing cytoplasmic RNA. The results of this study have indicated that STH inhibits mRNA methylation and that the use of this inhibitor may permit a qualita- tive and quantitative evaluation of the effects of methylation on mRNA processing. It may also enable determination of the relative time sequence of methylation 18 vivo at specific sites within mRNA. MATERIALS AND METHODS Cell Culture and Labeling Conditions Novikoff hepatoma cells (N1S1 strain) were grown in Swim's S-77 medium (GIBCO) supplemented with 10% calf serum (Desrosiers, ££Hé£-: 197A). Cells in midlogarithmic growth were harvested asceptically 99 and resuspended at a concentration of 1.5 x 106/ml in fresh warm medium containing 20 pM methionine (one-fifth the normal concentration) for labeling experiments. The cells were equilibrated in the medium for 1 h; a portion of the cells was exposed to 250 uM STH for the last A0 min of the pre-equilibration period. L-[methyl-3H1methionine (Amersham/Searle, 1A Ci/mmol) and [U-1uC1uridine (Amersham/Searle, 537 mCi/mmol) were added simultaneously at concentrations of 0.1 mCi/ml and 0.18 uCi/ml, respectively. Cells were incubated with radioisotope for 1 h. S-tubercidinyl-D, L-homocysteine (STH) was prepared by alkylation of 5'-chloro, 5'-deoxytubercidin (Coward, 88,81., 1977) with D,L- homocysteine thiolactone, according to the general procedure of Legraverend and Michelot (1976). The crude product was purified by ion-exchange chromatography on Dowex 50 X-8 (H+), the desired material being eluted with 1N NHuOH, following a water wash of the column to remove impurities. The peak tubes were pooled, and the contents lyophilized to give a white, fluffy powder in 83% yield. This material was identical by chromatographic and spectral compari- sons with that prepared previously (Coward, 22.él°: 197A). Isolation and Characterization of Poly (A)-ContainingRNA Total cytoplasmic RNA fractions were isolated essentially as described (Desrosiers, 22.2l-2 197A). Poly (A)-containing RNA from each fraction was isolated by oligo (dT)-cellulose chromatography, including a heat step prior to a second passage over the column to eliminate ribosomal RNA contamination (Desrosiers, £2.2l-I 1975). RNA was analyzed by sedimentation through denaturing gradients (99% 100 Me 80 :5-201 sucrose) as described (Friderici, ££.él°’ 1975) except 2 that sedimentation was for 17 h. Egzymatic and Acid Degradation of Poly (A)-Containing RNA The procedures for enzymatic degradation of poly (A)-containing RNA, and for high speed liquid chromatography (HSLC) methodology employed for analyses of cap structures, are detailed in previous publications (Friderici, 22.22:: 1975; Rottman, 88_81., 1976) and will be briefly outlined below. Cytoplasmic poly (A)-containing RNA was digested with RNase T2 (Sigma; Calbiochem) and alkaline phosphatase (Worthington) to produce nucleosides and intact 5'-terminal cap structures. After resolution of the reaction products on DEAE-Sephadex (7 M urea), intact cap 1 (m7GpppN'mpN) and cap 2 (m7GpppN'mpN"mpN) structures were separately isolated and desalted on Biogel P-2 (100-200 mesh); mononucleoside fractions were adsorbed on charcoal and eluted with 20% pyridine. Further digestion of the desalted cap structures with nuclease P1 (Yamasa Shoyl Co. Ltd.) and alkaline phosphatase yielded methyl- 7GpppN'm (core cap labeled m7GpppN'm from cap 1, and a mixture of m 2) plus released nucleosides (N"m) from cap 2. The nucleosides were resolved from the core cap on Biogel P-2 (100-200 mesh). Alternatively, cap 1 and core cap 2 structures were degraded to 2'222methylnucleosides and 7-methylguanosine with nucleotide pyro— phosphatase and alkaline phosphatase. Analysis of these nucleosides on Aminex A-5 HSLC (see below) permitted an independent determination of methylnucleoside distribution. Acid hydrolysis of intact mRNA 101 and of cap structures was performed in concentrated formic acid in sealed tubes at 100°C for 1 h. Distribution Analysis of Methy1ation The distribution of base-methylated nucleosides was determined using Aminex A-5 chromatogrpahy in 0.A M ammonium formate, pH A.55. 2'-8:Methylnucleosides and 7-methylguanosine were separated using the same chromatographic system at pH A.15 and in the presence of A05 ethylene glycol. Released purine bases in the acid hydrolysate were resolved on Aminex A-5 using 0.A M ammonium formate at pH 5.30. The distribution of cap species obtained from P1/alkaline phos- phatase-digested cap 1 and cap 2 was determined by Partisil-SAX HSLC (Whatman) using a gradient of 0.1-0.3 M KHZPOA’ pH 3.A5. Aliquots of the oligonucleotide eluting with a -2.5 charge (m7GpppN) were injected directly and resolved by gradient elution from 0.1 M KH2POA’ pH 3.50 to 0.3 M KHZPOA’ pH 3.90. Cap 1 and cap zero standards (m7Gppme and m7GpppN, respectively; PL Biochemical Co.) were moni- tored by A260 absorbance. RESULTS Preliminary studies on Novikoff poly (A)-containing RNA following exposure of the cells to STH, suggested that the compound effectively altered mRNA synthesis after relatively short time periods. No detect- able decrease in cell viability, as monitored by vital staining, was observed up to 2A h of exposure to STH. These studies also indi- cated that STH concentration, exposure time to inhibitor, and labeling time influenced the extent of radioactivity incorporated. Conditions were chosen which resulted in partial inhibition of mRNA methylation. 102 Presumably, higher concentrations of STH might further inhibit mRNA methylation, although cell viability may be decreased as well. RNA was labeled for comparison of both RNA synthesis and methyl- ation in the presence and absence of inhibitor. Since 1uC-uridine 3 incorporation is indicative of total RNA synthesis and H-methyl incorporation reflects the extent to which RNA was methylated, 3H cpm/17C cpm ratios for each RNA fraction can be employed for internal comparison of normal and inhibited samples. Table I summarizes the incorporation data obtained from parallel labeling of Novikoff cells in the presence (STH) or absence (normal) of inhibitor. Incorporation of both 1nC-uridine and 3H-methyl groups into total cytoplasmic RNA was reduced to approximately the same extent in the presence of STH. However, incorporation of methyl groups into poly (A)-containing RNA was reduced (32% of the normal level) to a greater extent than incorporation of uridine (73% of the normal level). These results, coupled with the methyl distribution analysis presented below, indi- cated that S-tubercidinylhomocysteine functions as an 22.2212 inhi- bitor of mRNA methylation. The size of poly (A)-containing RNA from STH-inhibited and normal cells was found to be nearly identical by denaturing sucrose gradient sedimentation (data not shown). No degradation was apparent in either sample and the size distribution was characteristic of cytoplasmic poly (A)-containing RNA obtained from these cells (Des- rosiers, 2£.§l" 197A). The poly (A)-containing RNA from normal and STH-treated cells was then further characterized to determine if the decreased incorporation of 3H-methyl groups reflected site- specific or overall inhibition of mRNA methylation. 103 .0onmofionn no moor no none» pounce onu .monan30 Hosnon 8on2 mnoanomne <22 wnflonoamonnoo onna nononoanoone 800 no owonnoonoq o no monsnaso mew once mononoqnoona 800 noonaxo nanonnnonmn 0H nonswfimo Ammon AnoOPV nmmmo 00.0 me.m eoex<22 Honon < nHHoo AaNmV 2200.0 Aumev 2200.0 nos <22 mcacamscoo o...m ome.m. om...P oo~.m. Iaev anon an 2am ee.00 em.mm m.e0 e.00 <22 wnflnamnnooIA<0 meoQInon <22 02.0 m0._ mm.0 m:.0 0.NF m.PP weanfiounoOIA<0 mace no 220 no a 22000 “noopv Aum0v oAuoopv nHHoo 0.0m m.em 00mem.m 00me0.m aopxpe.0 mopxmm.F e0P\<22 Honou 8H :20 mew n<2202 mew n<2202 mew n<2202 820 0:F\8qo mm noflnonoqnoonw HmnnozImm noenonoqnoonw onH0H8010=F <22 2o nofinoonm <22 ofismoaaoneo ounfl coflnmnoqnoonH onflnofinuozlmamnnossmmHIn 00m onwuwnqun— no mew 2o noowum H oanme 1ou CytOplasmic poly (A)-containing RNA was digested with RNAse T2 and alkaline phosphatase, and the reaction products were separated on DEAE-Sephadex to resolve internal N6-methyladenosine and the 5'- terminal cap structures. The radioactive profiles of both normal and STH-treated cytoplasmic poly (A)-containing RNA digests are pre- sented in Figure 1; the percentage of total radioactivity eluting with each peak was as indicated in the figure. The most striking difference between the two samples was the appearance of a new oligo- nucleotide eluting as peak III in RNA isolated from STH-inhibited cells. This oligonucleotide, bearing a charge of approximately -2,5 was not a result of incomplete enzymatic digestion. Re-digestion and re-chromatography of a portion of the desalted peak fractions resulted in more than 92% of the material eluting identically to that shown in Figure 1. In addition, similar digestion and chromato- graphy of nuclear poly (A)-containing RNA fractions also indicated the presence of this oligonucleotide only in the RNA obtained from STH-treated cells (data not shown). The -2.5 charge oligonucleotide, found only in the inhibited RNA sample, was further analyzed to determine its structure. An aliquot of the desalted peak fractions was acid hydrolyzed and the released purine bases separated by HSLC on Aminex A-S resin. Greater than 951 of the base-methyl radioactivity migrated with 7-methy1- guanine and guanine, the latter representing 15% of the total cpm and arising from ring-label, which was not suppressed during the labeling period (Friderici, gt al., 1975). Less than 3% of the total cpm eluted in the solvent front with degradation products arising from Z'ggrmethylnucleosides. These results suggested that this 105 Figure 1. DEAE-Sephadex column separation of RNase T2 and alkaline phosphatase digestion products from whole mRNA. The reaction mixture was diluted with H volumes of 7 M urea, 0.02 g Tris-HC1, pH 7.0, and applied to a 0.9 X 25 cm DEAE-Sephadex column. The mononucleosides were eluted with 0.05 5 NaCl, 7 M urea, 0.02 §_Tris-HC1, pH 7.0. To resolve oligonucleotides a 180 m1 gradient of 0.1 N to 0.“ g NaCl in 7 M urea, 0.02 §_Tris-HC1, pH 7.0 was used at a flow rate of -8 ml/h; 2 ml fractions were collected, and 50111 aliquots were removed from scintillation counting. Standard oligonucleotides (pUm)1_5 were included to indicate approximate charge. (A) Digest of poly (A)-containing RNA isolated from normal cells; (B) Digest of poly (A)-containing RNA isolated from STH-treated cells. Numbers in parentheses indicate the percentage distribution among the labeled fractions: Peak I - mononucleosides mN; Peak II - dinucleotides, NmpN: Peak III - oligonucleotide bearing -2.5 charge identified as cap zero, m7GpppN§ Peak IV - cap 1, m7GpppN'pmpN; and Peak V - cap 2, m7GpppN'mpN"mpN. 106 3H CPM |25 75 50 ¥ *T“\l fl II(|.5) -A. I (65.0) END) 25 -2 -3 -4 -5 $ (r <1 «1 IV(|7.9) . 'V(3.0) IV (27. I) III (5.3) V(|.5) 50 Fraction Number Figure 1 75 107 oligonucleotide (peak III in Figure 1) was in fact cap "zero", i.e. m7GpppN', in which the 5'-5' pyrophosphate linkage between 7-methyl- guanosine and the N' nucleoside is intact, but the 2'-Qymethyl ribose group is absent. Verification that the material eluting with a —2.5 charge from DEAE-Sephadex (Figure 1) was in fact cap zero was obtained by injection of an aliquot of the peak III fraction onto Partisil-SAX. The profile shown in Figure 2 demonstrates the resolution of peak III material into fractions coincident with authentic cap zero standards. Chromato- graphy of a second aliquot of the -2.5 charged oligonucleotide under conditions employed for cap 1 separation resolved the sample into peaks which were not coincident with cap 1 standards (data not shown). All four cap zero structures are present, with predominately purines in the N' position. Since m7Gpppm6A was not available as a standard, its presence could not be directly verified. It is possible, however, that the small peak present after m7GpppA (Figure 2) is in fact 6A. m7Gpppm The distributional analysis data for cap zero, and for both cap 1 and cap 2 from inhibited and normal cytoplasmic poly (A)-con- taining RNA fractions, are summarized in Table II. In contrast to cap zero analysis, caps 1 and 2 were subjected to nuclease P1/a1kaline phosphatase digestion prior to analysis. Following this treatment, m7GpppN'm structures obtained from either cap 1 or cap 2 were analyzed directly on Partisil—SAX. If the base distributions at N' of cap zero and cap 1 are compared, similar distributional data are obtained. This indicates that inhibition of N' 2'sgymethy1ation to produce cap zero is not base specific. A comparison of cap 2 distribution 108 .mocwanouoo no: u .0.20 .coflumowflo onmummomoso cowamxam\—2 Louwm ago 8:2 acne Boo 0200 N one no cofiumgmoom noumofloca emancoonomo .uxmu mow “odomH2m>mcs cnmocmpn oumanoonoo< .cowumomeo onmumsomoco ocwflmxam\em Lmuem muosoono czooxmonn wceosaocfio 0302023 .2aco mono cw “cor on coueamenoo mommucooaoo nogwm xHumuHucwso Lou Boo acoaoaeesmcwv 02300 N now no «000 a + 0.0.2 + + + <22 mew + no N omo chemo N one no «N00 um um_ «o no. “on mam <22 Hmanoz co m amo .unm II. a II. mm. mm. me so a<0 a< 0 0 monouosnuw N moo Bonn monenooaosz a..2 Aeao N emu co ammv 0m 0: 0 P 0 28.2ooooeav onoo N omo 0m Nm 0 P w 28.2oooueav P mac 3 sagas... g 3 a <22 Hmanoz “ago m amo co nomv Acoflumcwanoumo o>HumuHucmsu Low Boo pcoeoeuesmofiv 28.2aqoueav choc N omo Pm op em 0 ow Aa.zaqaoea0 F amo “em nAucv nmm “0 nNN A.2ooooeav ones can Aavoqqqoea Aav<0aqqqcea wafimmmmwm havmqqqcea Aavuqqqoea <22 mew .mnHHoo ewoxe>o2 oopmonulmew 0cm Hmanoz some <22 wcficempcooua<0 2H02 oesmmHQOpho no monsuosnuw omu :« 220 H2nuoanmm no coHusoHLumHQ HH wanme 109 .ooeanom 2uH>HuomoHumL kuop on» no n02 ems» mmoa ucowmnoon new .mwmxmmno amo no o>emeHUCe ohm mangoes no gnome CH weepsao nxmoo o3» mce .5: 00m um 0022002008 202 oaoemm on» cue: cocoonce who: monmccmun oLoN qmo cegonuczw .0muooeaoo mew: mcoeuomne as 0.0 “22mm omNez v cHe\Ha e zeopmsexonoam mm: mama 30am .mmeooow one on» mo 80532 2: 83mm: 3 82. mm: om.m ma ..ioammx 0. m5 3 omnm mg 30.30. m To sot #5320 E 2: m “canao 0» cmwoo mangoes amps: 00.0 mm .aomNmM m.P.0 sue: 002mm: mm: csoaoo one .csseoo xo ocm .moomnmz m.—o.o sue: Godusam 20 cesaoo N2 HmUIOem Bo Hocmsuo nufl3 “0N mums mm: mssao> 020> mm» CH Hmenmumz 0m x m.e m so vmuammou mum: AF .022 CH HHH xmoav m.Nu no ownmno mumsexonoqm am am senaoo xocmcqmw|2HuomovaLImm no soapsnwnuwen .N mnswwm 110 50:52 02.68... 0m 00 _ — 4 a 088 028%. Om * 0282.. am On we was Hg Figure 2 111 in mRNAs from STH-inhibited and normal cells (Table II) indicates that less base methylation at N' is occurring in the presence of STH. In addition, the data shows that the composition of N'(m) in the cap structures from STH-treated and normal RNA is considerably different. The low amount of radioactivity in cap 2 from mRNA of inhibited cells permitted only a limited qualitative analysis of N"m. It appears, however, that all four bases are present in N"m, and that Um is the predominant species, as is the case for normal cellular mRNA. The mononucleoside fraction from DEAE-Sephadex (Peak I in Figure 1), representing internal base-methylated nucleosides, was also analyzed. The nucleosides were desalted by charcoal adsorption and resolved on Aminex A-S by HSLC as described. In both the normal and inhibited samples, more than 98% of the 3H-methyl cpm was present as N6-methyladenosine. If S-methylcytosine was present in either sample, it consisted of less than 1.5% of the total radioactivity. Identical results were obtained by chromatography of the purine bases released by acid hydrolysis of the mononucleoside fraction (data not shown). The effect of STH on the level of internal base methylation can be assessed by comparison of the labeling of terminal 7-methyl- guanine with that of N6-methyladenine. In order to avoid including 6A was determined the m6A present in cap structures, the amount of m from acid hydrolysis data of the mononucleoside fraction eluting from DEAE-Sephadex. Total 7-methylguanine levels were determined by acid hydrolysis of intact RNA. In the normal sample, per 107 cells, 1038 3H—cpm was incorporated as m7G, and 10,822 3H-cpm as 112 m6A in poly (A)-containing cytoplasmic RNA, resulting in an m6A/m7G ratio of 10.”. Corresponding incorporation in the STH-treated sample 6 was 782 3H-cpm as m7G, and 2,h32 3H-cpm in m6A; the m A/m7G ratio was 3.1. The ratio of m6A to m7 G provides a measure of internal methylation in messenger RNA, relative to the amount of cap present. Since partial ring-opening of 7—methylguanine in cap structures can occur during the numerous analytical procedures performed over a period of several weeks, alternative analysis of methyl distribu- tion in caps was determined from acid hydrolysis and nucleotide pyro- phosphatase degradations, as outlined in MATERIALS AND METHODS. For illustrative purposes, Figure 3 displays the profiles obtained from HSLC anslysis of cap 1 from STH-inhibited cells. Resolution of intact cap 1 species on Partisil SAX (Figure 3A), of the nucleotide pyrophosphatase digest on Aminex A-S (Figure 3B), and of the acid hydrolysate on Aminex A-5 (Figure 3C) are presented. The results for cap structures from the three types of analyses were internally consistent. DISCUSSION Several in zitgg systems, particularly those using viral sources (Rhodes, 32 al., 197R; Shatkin, 197“; Martin and Moss, 1975, 1976), have provided information on the enzymatic generation of cap struc- tures in mRNA. Analysis of isolated eukaryotic mRNA has produced a clearer understanding of the kinds of mRNA methylation which occur in eukaryotes, including 5'-termina1 and internal methylation of both cytoplasmic mRNA and the corresponding nuclear precursors (Perry, gt gl., 1975b; Salditt-Georgieff, gt_al., 1976). Studies on the 113 Figure 3. Analysis of cap 1 structures obtained from mRNA of STH-treated cells. A) Resolution of cap 1 structures by HSLC on Partisil-SAX. Cap 1 fractions eluted from a DEAE-Sephadex column (peak IV in Fig. 1) were desalted on Bio-Gel P-2 as described in Fig. 2. The cap structures were digested with P1 nuclease and T2 RNase in the presence of 1A260 unit carrier rRNA. The digest was injected onto a Partisil PXS 1025-SAX column and cap 1 species were resolved as described in Fig. 2, except that the gradient elution was with 0.1 M KHZPOH’ pH 3.u5 to 0.3 M KHZPOH’ pH 3.U5. Synthetic cap 1 structures were used as standards. B) Distribution of methyl- nucleosides in cap 1 structures isolated from STH-treated cells. The desalted material was digested with nucleotide pyrophosphatase and alkaline phosphatase, dried with N2 and resuspended in 125 pl column buffer. HSLC on Aminex A-5 (1/8 in X 90 cm) was performed in 0.fl M ammonium formate, pH ”.15, "0% ethylene glycol at “0°C. Flow rate was ~10 ml/h (#000 psi); fraction size was ~0.8 ml until Cm had eluted, after which the volume was increased to ~1.2 ml/frac- tion. C) Resolution of methylated bases released from cap 1 struc- tures by acid hydrolysis. A portion of the sample was hydrolyzed in concentrated formic acid by heating at 1000 for 1 h. The hydroly- sate "33 dried With N2 and dissolved in column buffer. The released bases were resolved by HSLC on Aminex A-5 (1/8 in. X 90 cm) in 0.A M ammonium formate at pH 5.3, “0°C. 11" 100 50 300 200 3H CPM IOO ISO 100 50 7 m Gppme A 1 6 ” m7GpppLUm ”TGprip'“ Am 1 D 7 7 m GpripCm m Gppp Am 1 60 90 m o— 0" v'v'v Fraction Number Figure 3 115 function(s) of mRNA methylation using these approaches, however, are not directed at the possible role of methylation in processing of mRNA. What is required for these studies is an in vizg system in which the role of methylation and its relationship to mRNA pro- cessing can be assessed. Since SAH is known to be a very effective inhibitor of mRNA methylases ingitgg (Both, £2 21., 1975a, 1975b; Tonguzzo and Ghosh, 1976), and since STH, an analogue of SAH, has previously been shown to inhibit RNA methylation (Chang and Coward, 1975; Michelot, §t_§l., 1977), the effect of STH on mRNA methylation in 1112 was studied. Labeling of RNA with both PnCIuridine andlSH-methlemethionine permits an evaluation of overall methylation relative to the total amount of RNA synthesized. The data presented in Table I provides a comparative overview of this relationship. The reduction of total radioactivity in RNA from STH-inhibited cells is nearly equivalent for both 3H-methyl and 1” C-uridine incorporation. However, the methylation of poly (A)-containing RNA is significantly reduced, as reflected by the 3H/1uc values for each poly (A)-containing frac- tion. Perhaps the most significant result of this study is the detec- tion of cap zero structures in poly (A)-containing RNA isolated from the cytoplasm of cells exposed to STH. The appearance of peak III (Fig. 1) led to extensive characterization of the -2.5 charged mate- rial. The oligonucleotide was not a result of incomplete digestion since redigestion with alkaline phosphatase and RNase T2 did not alter its retention on DEAE-Sephadex. Acid hydrolysis indicated that nearly all the radioactivity was present as m7G, suggesting 116 the structure was m7GpppN. Verification of the identity of this oligonucleotide as cap zero was obtained by HSLC on Partisil-SAX. As shown in Figure 2, the oligonucleotide fractions co-eluted with actual cap zero standards. The observation of a large amount of cap zero, which previously had not been identified as a component in any mammalian system, demon- strates that N' methylation has been inhibited by the presence of STH. The similarity of the base composition at N' positions in cap zero and cap 1 structures (Table II) suggests that inhibition of methylation within this group of mRNAs is not selective for a particu- lar base at N'. Instead an overall inhibition of N' methylation appears to have occurred. It is possible, however, that base methyla- tion at the 6-position of adenine may be inhibited to a greater extent 6-adenine is that the ribose methylation at N', since much less N present in cap zero structures. The relative decrease observed in the amount of cap 2 from RNA labeled in the presence of STH (Fig. 1) indicates that Z'ggymethylation at N" is also being inhibited. The data presented in Table II indicate that the composition of N'(m) in cap structures from STH-treated samples is considerably different from the base distribution at N'm in normal RNA. The increased frequency of pyrimidines at N'(m) in STH-inhibited poly (A)-containing RNA is consistent with preferential inhibition of a subpopulation of mRNA. Inasmuch as RNA synthesis is initiated with a 5'-terminal purine triphoisphate (Chambon, 197R; Schibler and Perry, 1976; Schmincke, gt al., 1976) and since cap 1 structures (m7GpppN'mpN) have been found in hnRNA (Perry, g 3;” 1975; Salditt- Georgieff, gt al., 1976), it has been postulated that 5'-terminal 117 cap structures of mRNA might arise in two ways. Nascent 5'-termini of nuclear precursors would contribute to the mRNA species bearing caps of the type m7GpppPu, or internal cleavage of hnRNA would generate RNA species terminated with either a purine or a pyrimidine, which would then be capped and processed (Furuichi, §t_§l., 1975; Ensinger, gt_§l., 1975; Schibler and Perry, 1976). The apparent enrichment of mRNAs containing a pyrimidine at the N' position in cells exposed :- to STH raises the possibility that the presence of STH favors pro- cessing of mRNA bearing caps generated internally rather than from the 5'-terminus of a precursor molecule. E Although previous studies of nuclear poly (A)-containing RNA have suggested that methylation at the N' site occurs in the nucleus, (Perry, 23 al., 1975; Salditt-Georgieff, 23 al., 1976), finding cap zero in poly (A)-containing cytoplasmic RNA indicated that methylation at the N' position in mammalian mRNA is not a prerequisite for inter- nal cleavage prior to capping. These results also suggest that trans- port of nuclear mRNA precursors does not require a fully methylated cap structure. Currently, investigations are in progress to determine if cap zero-containing mRNAs are located on polysomes, in an attempt to determine if N' methylation is required for ribosomal binding of those mRNAs containing a capped 5'-terminus. The data indicates that the presence of STH did alter the pro- 6 portion of internal m A present in the RNA fraction relative to the 6A/m7G is much lower for mRNA amount of m7G observed. The ratio of m derived from STH-treated cells, indicating that internal methylation is inhibited significantly in the presence of STH. This interpreta- tion is supported by the fact that when expressed as absolute cpm 118 incorporated as m6A, only 23% of the normal incorporation level is observed. It would be of interest to ascertain if this inhibition is expressed uniformly at all m6A sites in mRNA molecules (Wei, gt 31., 1976; Dimock and Stoltzfus, 1977). In contrast, incorporation of 3H-methyl as m7G in mRNA from STH-treated cells was decreased to a lesser extent. The observed incorporation, 75% of the normal level, is comparable to the overall 1"C-uridine incorpor- decrease in mRNA synthesis (73%, as measured by ation; cf. Table I). It thus appears that under the conditions used in these experiments, in which partial methylation is occurring, little or no inhibtion of methylation at the 7-position of guanine is observed in the cytoplasmic poly (A)-containing RNA. These results may indicate a cellular response in terms of priority for methylation at the 7-methylguanosine site over base methylation, possibly reflec- ting the functional or sequential nature of methyaltion events during processing. Alternatively, RNA molecules which do not contain m7G may not be transported into the cytoplasm, and thus would not be observed. Experiments using 32 P-labeled RNA are currently being pur- sued in our laboratory in order to investigate further the possible inhibition of m7G formation. The data presented above indicates that STH inhibits methylation in gigg at several sites. Base methylation to generate m6A is affected both at internal sites and at the N‘ position of the caps. In addi- tion, the presence of cap zero and the relative decrease in cap 2 structures in mRNA exposed to STH indicates inhibition of 2'1Qrmethyla- tion at N' and N", respectively. This lack of complete selectivity for individual methylation sites by STH is in accord with previous 119 data obtained using several isolated methylases (Coward, gt al., 197“; Borchardt, gt_§l., 1976). The fact that STH affects both 2'- Qymethylation and base methylation indicates that this inhibitor may prove to be useful in examining the role of internal N6-methyl- adenosine as well as the function of the cap structure. In an earlier report describing the 5'-terminal cap structure (Rottman, gt al., 197“), a possible function of the cap in mRNA pro- cessing was proposed. The use of S-tubercidinylhomocysteine as an in vivg methylation inhibitor should provide a useful approach to this complex question. Although it can be stated that decreased and altered patterns of methylation are observed in the presence of STH, more specific questions can be asked if the system employed for this analysis permits evaluation of a single mRNA species. Studies of the effect of STH on methylation of a specific mRNA se- quence are in progress. The analyses on poly (A)-containing RNA presented here indicate that STH does affect messenger RNA methyla- tion in 1339, and that this in givg approach may be useful in assessing the role of mRNA methylation and its relationship to mRNA processing. REFERENCES Adams, J.M. and Cory, S. (1975), Nature (London) 255, 28. Borchardt, R.T. (1976) in The Biochemistry of Adenosylmethionine (Salvatore, F. and Borek, E., eds.) Columbia University: New York, p.151. Both, G.W., Banerjee, A.K. and Shatkin, A.J. (1975a), Proc. Natl. Acad. Sci. USA 12, 1189. Both, G.W., Furuichi, Y., Muthukrishnan, S. and Shatkin, A.J. (1975b), Cell g, 185. Chang, C.D. and Coward, J.K. (1975), Mol. Pharm. 11, 701. Chambon, P. (197“) in The Enzymes 19 (Boyer, P., ed.) Academic Press, New York, 261. Coward, J.K., Bussolotti, D.L. and Chang, C.-D. (197“), g, Reg. Chem. _1_7, 1286. Coward, J.K., Motola, N.C. and Moyer, J.D. (1977), g, Med. Chem. g9, 500. Crooks, P.A., Dreyer, R.N.and Coward, J.K., in preparation. Dasgupta, R., Shih, D.S., Saris, C. and Kaesberg, P. (1975), Nature (London) 256, 62“. Desrosiers, R., Friderici, K. and Rottman, F. (197“), Proc. Natl. Acad. Sci. USA 7_1_, 3971. Dimock, K. and Stoltzfus, G.M. (1977), Biochemistry 1Q, “71. Dubin, D.T. and Stollar, V. (1975), Biochem. Biophys. Res. Comm. .§§. 1373- Ensinger, M.J., Martin, S.A., Paoletti, E. and Moss, B. (1975), Proc. Natl. Acad. _Sgi. U_SA_'_72, 2525. Friderici, K., Kaehler, M. and Rottman, F. (1976), Biochemistry 15, 523“. 120 121 Furuichi, Y., Morgan, M., Muthukrishnan, S. and Shatkin, A.J. (1975), Proc. Natl. Acad. Sci. USA 1g, 362. Hickey, E.D., Weber, L.A. and Baglioni, C. (1976), Proc. Natl. Acad. Sci. USA 1;, 19. Kozak, M. and Shatkin, A.J. (1976), g. Biol. Chem. 251, “259. Legraverend, M. and Michelot, R. (1976), Biochimie 58, 723. Martin, S.A. and Moss, B. (1975), g, Biol. Chem. 250, 9330. Martin, S.A. and Moss, B. (1976), J. Biol. Chem. 251, 7313. Michelot, R.J., Lasks, N., Stout, R.N. and Coward, J.K. (1977), Mol. Pharmacol. 13, 368. Muthukrishman, 8., Both, G.W., Furuchi, Y. and Shatkin, A.J. (1975), Nature (London) 255, 33. Perry, R.P. and Kelley, D.E. (197“), Cell. 1, 37. Perry, R.P. and Kelley, D.E. (1976), Cell. 8, “33. Perry, R.P., Kelley, D.E., Friderici, K.H. and Rottman, F.M. (1975a), Cell 3, 387. Perry, R.P., Kelley, D.E., Friderici, K.H. and Rottman, F.M. (1975b), Cell 6, 13. Rhodes, D.P., Moyer, S.A. and Banerjee, A.K. (197“), Cell. 3, 327. Rottman, F.M., Desrosiers, R.C. and Friderici, K. (1976), Progg, Nucleic Acid figs. 12, 21. Rottman, F.M., Shatkin, A.J. and Perry, R.P. (197“), Cell. 3, 197. Salditt-Georgieff, M., Jelinek, N., Darnell, J.E., Furuichi, Y., Morgan, M. and Shatkin, A.K. (9176), Cell, 1, 227. Schibler, U. and Perry, R.P. (1976), Cell. 9, 121. Schmincke, C.D., Herrmann, K. and Hansen, P. (1976), Proc. Natl. Acad. Sci. USA 13, 199“. Shatkin, A.J. (1975), Cell. 2, 6“5. Shatkin, A.J. (197“), Proc. Natl. Acad. Sci. USA 11, 320“. Toneguzzo, F. and Ghosh, R.P. (1976), g, Virology 11, “77. 122 Vaughan, M.H., Soeiro, R., Warner, J. and Darnell, J.S. (1967), Proc. Natl. Acad. Sci. M31: 1527. — Weber, L.A., Feman, E.R., Hickey, E.D., Williams, M.C. and Baglioni, C. (1976), g, Biol. Chem. 251, 5657. Wei, C.M., Gershowitz, A. and Moss, B. (1976), Biochemistry 15, 397. PART IV CYTOPLASMIC LOCATION OF UNDERMETHYLATED MESSENGER RNA FROM NOVIKOFF CELLS 123 12“ ABSTRACT Novikoff cells in culture were simultaneously labeled with L-[methyl-3H]methionine and [32P]orthophosphate in the presence or absence of S-tubercidinylhomocysteine, an inhibitor of RNA methyla- tion. Total cytoplasmic, polysomal and monosomal poly (A)-containing RNAs from both normal and inhibited cells were analyzed to ascertain the subcellular location of undermethylated mRNA. Both monosomal and polysomal mRNA fractions from S-tubercidinylhomocysteine-treated cells contain partially methylated cap structures, suggesting that 2'fQ—methylation of the nucleoside adjacent to the pyrophosphate linkage in caps is not required either for ribosomal binding or for translation. These partially methylated cap structures are also present in nuclear poly (A)-containing RNA isolated from cells labeled in the presence of S-tubercidinylhomocysteine. Totally unmethylated cap structures, if present at all, are not accumulating in the cells. INTRODUCTION Methylated cap structures (cap 1, m7GpppN'mpN and cap 2, m7GpppN'mpN"mpN) have been identified at the 5'-terminus of a variety of viral and eukaryotic messenger RNAs (of. reviews: Shatkin, 1976; Rottman, 1976, 1978). This posttranscriptional modification of mRNA has been the subject of a number of investigations which have attempted to elucidate cap function(s) within the cell. These efforts have generally focused on the role of methylated cap structures during translation (Muthukrishnan, gt_§l. 1975; Kozak and Shatkin, 1977; HiCKGY. gfiflgl., 1977) and have indicated that cap structures 125 may facilitate translation, but are not absolutely required. However, the presence of cap structures on hnRNA molecules (Perry, 92 al., 1975; Salditt-Georgieff, gt al., 1976) and the kinetics of methylation at specific sites within cap structures (Perry and Kelley, 1976; Friderici, §t_al., 1976; Rottman, gt al., 1976), suggest that capping and methylation may be important processing events in the generation of mature mRNA. The discovery of intervening sequences within the coding region of unique genes (Jeffreys and Flavell, 1977; Tilghman, gt al., 1977; Breathnach, gt al., 1977; Brack and Tonegawa, 1977), and the demonstration that transcription of the intervening sequences into primary RNA transcripts occurs during expression of Bpglobin genes (Tilghman, gt_§l,, 1978), suggests a model for mRNA processing in which both termini of an hnRNA molecule are conserved during mRNA biogenesis. In order to assess the possible role of methylation in mRNA processing, we have perturbed methylation in vivg and thereby have generated undermethylated mRNA (Kaehler, gt al., 1977). These studies demonstrated that S-tubercidinylhomocysteine (STH), the 7-deaza analogue of S-adenosylhomocysteine (SAH), inhibited mRNA methylation in viable Novikoff cells. The cytoplasmic presence of "cap zero" structures (m7GpppN'), in which 2'figemethylation was absent, indicated that ribose methylation of the N' nucleoside was not required for nuclear processing and transport of mRNA in viable Novikoff cells. We report here the results of experiments designed to establish whether cap zero-bearing mRNA was associated with polysomes in Novikoff cells. Such an association would imply that these 126 undermethylated mRNA molecules can be translated in 1222- In addition, we have investigated the possibility that totally unmethylated cap structures were generated in cells with STH present. MATERIALS AND METHODS Cell Culture and Labeling Conditions. Novikoff hepatoma cells (N181 strain) were grown in Swim's S-77 medium (GIBCO) supplemented with 10% calf serum (Desrosiers, 33 al., 197“). Cells in midlogarithmic growth were harvested and resuspended at 1.5 x 106/ml for labeling in fresh medium containing no phosphate and 20 pM methionine (one-fifth normal concentration). Cells were equilibrated for 3 h; a portion of the culture was exposed to 500 uM STH for the final 50 min of the equilibration period. L-[methyl-Bfilmethionine (Amersham, 8.8 Ci/mmole) and [32P]orthophos- phate (Amersham, 127 Ci/mg P) were added simultaneously at concen- trations of 0.1 mCi/ml and 0.15 mCi/ml, respectively. Cells were labeled for 2 h. STH was synthesized and characterized as previously described (Coward, gt al., 1976). Isolation of Poly (A)-Containing RNA. Cells were poured over frozen crushed saline and harvested by centrifugation at 1500 xg for 5 min. The washed cells were resuspended and allowed to swell in hypotonic buffer (10 mM NaCl, 1.5 mM MgC12, 10 mM Tris, pH 7.0). Cells were lysed by dounce homogenization in the presence of 1.5 mg/ml cycloheximide. Nuclei were pelleted by centrifugation at 1000 xg for 5 min. Total cyto- plasmic RNA was isolated from the postnuclear supernatant as pre— viously described (Desrosiers, gt al., 197“). 127 Polysomal and monosomal fractions were separated by sucrose gradient sedimentation. A portion of the postnuclear supernatant was made 0.5% in both sodium deoxycholate and Triton X-100 (Eschenfeldt and Patterson, 1975) and layered onto 11 m1 gradients of 10-“O% sucrose in 100 mM NaCl, 1.5 mM MgCl 10 mM Tris, pH 7.0. The 2. gradients were overlaid on a 0.5 ml cushion of 60% sucrose in the same buffer. Sedimentation was for 75 min at “0,000 rpm in a Beckman SW “1 rotor. Gradients were analyzed using a Gilford model 2“8O gradient scanner, and aliquots equivalent to 5% of each fraction were removed and analyzed for radioactivity. Appropriate fractions were pooled, brought to 0.2 M NaCl, 10 mM EDTA, 0.2% SDS and 20 mM Tris, pH 7.0, and digested with 200 ug/ml proteinase K (EM Biochemicals) for 10 min at 37°C. RNA was precipitated in 67% ethanol. The RNA was pelleted by centrifugation at 27,000 xg for 15 min, dried under nitrogen and dissolved in 0.1 M NaCl, 10 mM EDTA, 0.2% SDS, 10 mM Tris, pH 7.0 for redigestion with proteinase K. The RNA was extracted with phenol:chloroform:isoamyl alcohol (50:“9:1) and reprecipitated in ethanol. Washed nuclei were resuspended in hypotonic buffer and vortexed for 3 sec in the presence of 1% sodium deoxycholate and 2% Tween“0 prior to re-centrifugation at 1500 xg for 5 min. Following resuspen- sion of the nuclear pellet into anti-RNase buffer (3 mM MgC12, 3 mM each 2',3'AMP, 2',3'UMP and 2',3'CMP, 30 pg/ml polyvinyl sulfate, 200 ug/ml heparin, 10 mM sodium acetate, pH 5.2; (Kwan, gt al., 1977)), “O ug/ml of DNase I (Worthington) was added and the solution was incubated at room temperature for approximately 10 min. Pro- teinase K digestion and extraction of nuclear RNA were performed 128 as described above for cytoplasmic RNA. Poly (A)-containing RNA was isolated from each RNA fraction by repeated binding to oligo(dT)-cellulose (Desrosiers, gt al., 1975). RNA was dissolved in high salt buffer (0.12 M NaCl, 1 mM EDTA, 0.2% SDS, 10 mM Tris, pH 7.0) and applied to a 2.5 ml oligo(dT)- cellulose column. Non-poly (A)-containing RNA fractions were pooled and precipitated by the addition of two volumes of ethanol. RNA which bound to the column was eluted with low salt buffer (1 mM EDTA, 0.2% SDS, 10 mM Tris, pH 7.0), made 0.2 M in NaCl, and pre- cipitated in ethanol. The second chromatographic passage of RNA on oligo(dT)-cellulose was preceded by heat denaturation of the RNA in the presence of 90% DMSO, as described by Bantle, gt al. (1976). Poly (A)-containing RNA was precipitated in the presence of unlabeled Novikoff rRNA by addition of two volumes of ethanol. Enzymatic Digestion and Analysis of Poly (A)-Containing RNA. Poly (A)-containing RNA was digested with approximately 2 units RNase T2 per A unit of RNA in 10 mM ammonium acetate, 260 pH “.5. for 2 h at 37°C. The digest was then diluted with four volumes of 7 M urea, 10 mM Tris, pH 7.0, and the digestion products were resolved on DEAE-Sephadex (7 M urea). A 200 ml gradient from 0.1 M to O.“ M NaCl in 7 M urea, 10 mM Tris, pH 7.0 was used to elute the digestion products; (pUm)1_5 were employed as absorbance markers. Alternatively, when cap structures were to be isolated for further analysis, an acetylated DBAE-cellulose column was used to separate mononucleotides from cap structures. The RNase T2 digest was diluted with six volumes application buffer (0.6 M KCl, 129 10 mM MgC12, 50 mM Tris pH 7.7, 20% ethanol; (Thomason, gt al., 1978)) and applied to a 1 ml DBAE-cellulose column. The column was washed with application buffer until no further radioactivity was detectable. The cap structures were eluted with 0.2 M NaCl, 1 M sorbitol, 50 mM sodium acetate, pH 5.0 (Thomason, gt al., 1978), diluted to approximately 0.05 M NaCl with 7 M urea, 10 mM Tris, pH 7.0 and chromatographed on DEAE-Sephadex (7 M urea) as described above. Fractions containing cap structures were pooled, diluted to approximately 0.1 M in NaCl, and passed over a 1 ml DBAE-cellulose column to remove urea and salt. The column was washed with 0.02 M ammonium acetate, and cap structures were eluted with 0.5 M ammonium acetate. The ammonium acetate was subsequently removed by repeated evaporation and lyophilization. Cap structures were analyzed by Partisil-SAX (Whatman) high speed liquid chromatography (HSLC). Cap 1 structures were first digested with P1 nuclease in 10 mM ammonium acetate, pH 5.3 for 1 h at 37°C. The mixture was adjusted to 10 mM Tris, pH 8.3 and 10 mM magnesium acetate prior to digestion with alkaline phosphatase at 37°C for “5 min. The digestion mixture was evaporated under a gentle stream of nitrogen and resuspended in 100 pl 0.01 M KH2P0“’ pH 3.33 for injection onto Partisil-SAX. Cap zero structures were digested with alkaline phosphatase prior to injection. Resolution of cap 1 species was achieved using a 100 ml gradient of 0.1 M to 0-3 M KHZPOA! pH 3.55; cap zero species were resolved with a similar gradient using 0.1 M KH2P0“’ pH 3.6“ to 0.3 M KHZPO“’ pH 3.90. Cap standards (P-L Biochemicals) were monitored at 25“ nm. 130 Due to low levels of radioactivity in cap 2 species, the P1 nuclease-digested caps were chromatographed on DBAE-cellulose as described above, and fractions were analyzed for radioactivity. The 3H-methyl distribution between core cap structures (m7GpppN'mp) and N"mp was determined by the radioactivity in bound and unbound fractions, respectively. Acid hydrolysis in concentrated formic acid, followed by chroma- tography of the hydrolysate on an Aminex A-5 column, was performed as previously described (Rottman, gt al., 1976). Both mononucleotides and cap structures were analyzed. This method permits determination of 3H-methyl groups present in methylated purine bases and of incorporation into purine ring structures through gg'ggvg synthesis (Munns gt al., 197“). RESULTS Previous studies in our laboratory have shown that STH is an effective inhibitor of mRNA methylation in gizg in Novikoff hepatoma cells (Kaehler, gt 3;., 1977). In order to further establish the specificity of STH as a methylation inhibitor, the following experiment was performed. Cultured Novikoff cells were subdivided into five separate cultures and labeled with L—[methyl-BH]methionine and [U-1uC]uridine in the presence of (a) 250 uM STH, (b) 250 uM SAH, (c) 250 uM homocysteine, (d) 250 uM tuberdicin, or (e) without any additions (control culture). The labeling conditions and analytical methods were identical with those previously described (Kaehler, gt al., 1977). Analysis of the methyl distribution in poly (A)-containing RNA from cultures treated with SAH and homo- cysteine gave virtually identical results when compared to the 131 control (Table I). As predicted, tubercidin was toxic to the cells and both RNA synthesis and methylation sharply decreased. Only in cultures labeled in the presence of STH were cap zero structures observed to represent a significant portion of [3H]methyl radio- activity. These results are summarized in Table I. To determine if cap zero-bearing mRNA was functional in these cells we studied the cytoplasmic location of undermethylated mRNA 7' molecules. In these experiments, simultaneous labeling with L-[3H- methyl]methionine and [32P]orthophosphate permitted evaluation of the effect of 500 pM STH on mRNA synthesis, methylation and transport during the labeling period. In agreement with our earlier E observations (Kaehler, 32 al., 1977), 3H-methyl incorporation was 32P-incorporation, i.e. to 18% inhibited to a greater extent than and 70%, respectively, of the corresponding incorporation levels observed in the control cultures. This inhibition of methylation appears to depend upon the concentration of the inhibitor, since doubling the concentration of STH (from 250 pH to 500 uM) decreased the ratio of 3H-methyl incorporation in STH-treated vs. normal cultures from .32 to .18. In contrast, mRNA synthesis (monitored 1“ 32 C-uridine or P-orthophosphate incorporation) was decreased by to approximately 70% of normal levels at either STH concentration. In order to establish the cytoplasmic location of cap zero- bearing mRNA molecules, monosomal and polysomal RNAs were separated by sedimentation of the postnuclear supernatant through sucrose gradients. Figure 1 shows the absorbance profiles obtained from both normal and STH-treated cells, and indicates the regions pooled .<22 wcflcemacooflona mm onwr moonuoa Hmoapeamcm 0cm nmoeueocoo wceaonmqm 0: a z N m A N emu me 0_ e. em 0, F emu 0 P P m 0 onmu emu m _ P P F eeeeoeaoseee m: 0e 0e mo 0e ocanooaoscocoe mm a "ecoeesnneemne Heseee-mm N0.m 0N.P e0._ 00.0 00.2 <22 weanemunoonA. I .— DJ 2 500'- v—I— I (45.0) 3L. H) 2 50 "’ I(77.3) l00 50 1) ’ L 1 --AJ. A A A- I (pUm). (pUm)5 III (24.7) N (3.7) II (20.7) III (19.7) Blue) I (23.5) III (20.3) 40 D e 1102.)) I III (6.0) 80 120 FRACTION NUMBER Figure 2 138 the 32P-radioactivity from the sample prior to resolution of the cap structures by DEAE-Sephadex (7 M urea). As shown in Figure 3, more than 99% of the 32P-radioactivity in total cytoplasmic mRNA from STH-treated Novikoff cells did not bind to DBAE-cellulose after RNase T2 digestion. The bound fraction, containing 55% of the 3H-methyl radioactivity, and 0.9% of the total 32P-cpm, was chromatographed on DEAE-Sephadex (7 M urea) (Figure “A). The distri- bution of 3H-methyl cpm in caps zero, 1, and 2 (Figure “A, peaks II, III and IV, respectively) was virtually the same as the distribu- tion observed in the DEAE-Sephadex (7 M urea) profile shown in |.'m’.__‘ _v_ I Figure 28 (peaks II, III and IV, respectively). Most of the 32P-cpm which bound to DBAE-cellulose eluted as mononucleotides (Figure “A, peak I), presumably due to trailing of the unbound fraction. The remainder of the 32P-cpm, representing 0.17% of the total radio- activity incorporated into STH-treated cytoplasmic mRNA, was distri- buted as follows: 0.10% in cap zero (Figure “A, peak II), 0.0“% in cap 1 structures (Figure “A, peak III), and 0.03% eluting as peak V in Figure “A. Unmethylated cap structures would be expected to elute from DEAE-Sephadex (7 M urea) columns between caps zero and 1, since the unmethylated caps would not carry the partial positive charge of 7-methylguanosine. There was no detectable radioactivity eluting in this position, but rather peak V in Figure “A was observed, at a position indicating an approximate charge of -8. The fractions containing this radioactive material were pooled and desalted for further analysis. 139 Figure 3. Acetylated DBAE-cellulose chromatography of RNase T2 digestion products from STH-inhibited cytoplasmic poly (A)-containing RNA. The digest was diluted with application buffer and chromatographed on a 1 m1 acetylated DBAE-cellulose column as described in MATERIALS AND METHODS. Thirty drop fractions were collected, corresponding to a fraction size of 0.95 ml during appli- cation, and 1.6 ml during elution. Radioactivity present in each fraction was determined from 25 pl aliquots. Circles (-o-o—) indicate 3H-methyl cpm; triangles (++-) denote 32P-cpm. 1“0 A I [3H] METHYL CPM x Io'ZH DO I 5 IO 5 EIuIion buffer I h- ---- FRACTION NUMBER Figure 3 O) 42 [3213] CPM x10"3 ~—-- DO 1111 Figure “. DEAE-Sephadex (7 M urea) elution profile of the DBAE-cellulose-bound fraction from RNase T2 digestion of poly (A)-containing RNA. The radioactive material released from acetylated DBAE-cellulose (cf. Figure 3) was chromatographed on DEAE-Sephadex (7 M urea) as described in MATERIALS AND METHODS. Aliquots from each fraction were analyzed for radioactivity levels. Profiles are shown for (A) total cytoplasmic poly (A)-containing RNA, and (B) nuclear poly (A)-containing RNA, both from STH-inhibited cultures. Radioactive material was present in peak I, mononucleotides, (Np and mNp); peak II, cap zero, peak III, cap 1; peak IV, cap 2 and peak V. Circles (-o—o—) denote 3H-methyl cpm, and triangles (H) indicate 32P-radioactivity. 1“2 I I I I I 300 pUm (pUm)2 (pUmI3 (pUm)4 (pUm)5 I I I I A II 200 ~lOO I III IOO -‘ 50 I . = I a. o 4 I z ' II o. ;‘ A“; ........................ w J fl . ........ 0 I I )— _ u.) a ’ B 33.. T II 0 _J .. I .... 4) 25- " IOO III 4 m l 20 BO IOO Figure “ FRACTION NUMBER 1“3 Although no absorbance standards were available for chromato- graphy, it was possible that the anomalous elution behavior of the material contained in peak V is consistent with a GpppN'p structure. The desalted oligonucleotide eluting as peak V in Figure “A was therefore treated with alkaline phosphatase and rechromatographed on DBAE-cellulose. Sixty per cent of the radioactive material bound to the column (Figure 5). This result is consistent with the predicted behavior of unmethylated cap structures only if 20% of the pyrophosphate linkages had been broken during desalting and/or phosphatase digestion. The low level of radioactivity in this highly negatively charged compound prevented further analysis. Nuclear poly (A)-containing RNA isolated from STH-inhibited cultures was analyzed to determine if unmethylated cap structures were present in the nucleus. The RNA was digested with RNase T2, and the digestion products chromatographed on DBAE-cellulose. The DBAE-bound fraction was eluted on DEAE-Sephadex (7 M urea) and the resultant profile is shown in Figure “B. 32P-Radioactivity was detected only in the mononucleotide and possibly in cap zero fractions (Figure “B, peaks I and II, respectively). It should be noted, however, that only 0.02% of the total 32P-cpm incorporated into poly (A)-containing nuclear RNA eluted with cap zero structures. This level approached the limits of detection, and thus unmethylated cap structures would not be detected if present at a level below that of cap zero. Parallel analysis of normal nuclear poly (A)- containing RNA revealed no cap zero structures (data not shown). mu .2eneneem Iowoonum 202 mepooneu oonnmooe ono nouooHHoo onoz onowuoonm .nEBHoo oooasaHoonm<00 oonoaanooo no no Nm eoneeemoemeoeno we: eeemne one .moomemz 02¢ ms xooov onoeuoone oowuooaosnoweao oownmno manwfin one .<228 oHEnoHnouao oonoonuumew some nonoeomfi Hoenonos oownono 2Hnwfln on» mo oafieona oooasaaoorm on onwm m< xonfls< an oonoonwouoaonno new nonmaonomn 020m onoz monsuoonno onou moo OHEmMquuho .oanmaeo>o ono: nonoonopn oonmnnonno noenz Lou nonsuosnum onou moo noon on» Bone maunonoeeeo oonmnwea mue>eaoooeoon on oonen nonnoene no: < 5222023 20 oonoono oneo 0 m2 um m2 Pm u. emmeoee =2 2. 22 em e_ enammaeoueo . 2 emu N: 200 N: m :e anodes: N: ofiov 0m 2 22 oesnoaaopmo I onoN moo "wonsuaso confinencw I mew NF 0N Nm 0 me LmoHos: mp em 2N me e2 oeanoaqoumo I e moo ”nonsuaso Hoanoz flavoqqquea Asv<0amqqcea Aeveuoooeoo2 ahnquImm mo nownsnanunea HH oHnoe 1“8 DISCUSSION Determination of the role of mRNA methylation in eukaryotes is complicated by the inability to isolate a subclass of undermethylated mRNA molecules from cell cultures. We approached this problem indirectly by studying the intracellular location of cap zero-bearing mRNAs, which were formed in the presence of S-tubercidinylhomocysteine. Cap zero was the predominant cap structure of both polysomal and monosomal mRNAs (Figure 2C and D, respectively). The location of these mRNAs on monosomes implies that ribosomal recognition of the undermethylated mRNAs was not prevented by lack of 2'29: methylation at the N' position. Furthermore, identification of cap zero structures in the polysomes suggests that these under- methylated structures are translated in viva. Although the results presented here indicate that ribose methylation at N' was not requisite for either ribosomal binding or subsequent translation, these data do not exclude the possibility that 2'7ggmethylation facilitates these processes in 2232- It should be noted that these conclusions have been drawn from experiments using viable cell cultures. Both 33 §l° (1975) and Muthrikrishnan, 32 al. (1975) have previously reported that 2'292methylation at N' was not necessary for ribosome binding and cell-free translation of capped reovirus mRNA in wheat germ extracts. Subsequent studies, however, indicated that ribosomes preferentially bound viral RNA and synthetic ribo- polymers terminated by cap structures which contained ribose-methylated nucleosides at the N' position (Both, 22 al., 1976; Muthukrishnan, £2 21., 1976). Ribose methylation at N' does not appear to 1“9 significantly affect ribosomal binding of vaccinia mRNA (Muthukrishnan, gt al., 1978), unless high concentrations of mRNA are used to enhance competition for binding. Sedimentation analysis of the postnuclear supernatant of cells labeled in the absence and presence of STH (Figure 1A and B, res- pectively) also permitted evaluation of the inhibitor's effect on polysome distribution. Polysomal profiles have been recognized as a criterion of the specificity of various metabolic inhibitors of RNA synthesis (Craig, N., 1973). The similar monosome/polysome ratio observed for both control and STH-inhibited cells indicated that the inhibitor had not significantly altered the polysome levels, implying that protein synthesis has remained relatively unperturbed by the presence of STH. This nontoxicity of STH is highly desirable for studies involving RNA processing, since cell viability must remain high in order to maintain normal regulatory functions in the cells. [32P]Orthophosphate was used as a radioactive precursor in these studies to detect the possible presence of completely unmethyl- ated cap structures. No detectable levels of 32P-radioactivity were found in inhibited mRNA digests that chromatographed at the expected elution position of unmethylated caps (Figure 2B, C and D; Figure “A and B). Total 32P-incorporation into cap structures of both normal and inhibited cytoplasmic mRNAs was slightly lower than predicted. Assuming an average mRNA length of 2000 nucleotides, cap structures should have contained approximately 0.20-0.25% of the total 32P-cpm (depending on whether cap zero or cap 1 structures are used for calculation). Cap 1 structures derived from normal 150 cytoplasmic mRNA contained 0.20% of the 32 P-cpm incorporated into the sample. Approximately 0.17% of the total 32P-radioactivity in STH-inhibited cytoplasmic mRNA eluted from DEAE-Sephadex in three distinct peaks (Figure “A): 0.10% was in cap zero structures (peak II), 0.0“% in cap 1 structures (peak III) and 0.03% in the late eluting fractions of peak V in Figure “A. We estimate that the lower limit of detection of unmethylated caps would have enabled us to see one-fifth of the cap zero level in this sample, or 0.02% of the total 32P-cpm incorporated into mRNA. The small peak of 32P-labeled material referred to as peak V in Figure “A represents approximately one-third the level of incorporation observed into cap zero structures. Analysis of the material was limited by the low amount of radioactivity present. The structure possessed the following chracteristics: 1) it contains no radioactive methyl groups, but is resistant to RNase T2 digestion; 2) it binds to acetylated DBAE-cellulose, and therefore must contain gigghydroxyl groups; 3) alkaline phosphatase digestion does not eliminate its ability to bind the substituted borate column, although “0% of the 32P-cpm is released into the unbound fraction; “) it does not appear in the DEAE-Sephadex (7 M urea) profiles of STH- inhibited nuclear mRNA digests (although its detection would require this structure to be present in levels comparable to cap zero levels); and 5) it elutes from DEAE-Sephadex (7 M urea) columns as if it contained a charge of approximately -8. With the exception of its apparent excessive charge, this material exhibited behavior consistent with unmethylated cap structures. The absence of this material in nuclear RNA, however, indicated that it is not stable 151 and therefore does not accumulate in the nucleus. Pugh, et_§l. (1978) have demonstrated that STH is a very potent inhibitor of mRNA (guanine-7-)methyltransferase in vitrg. THe data presented here do not exclude the possibility that inhibition of methylation at the 7-position of guanine is occurring! 2.2212: but the resultant cap structures are rapidly degraded. This situation is consistent with the moderate decrease in RNA synthesis observed in the presence of STH. The 3H-methyl distribution among cap structures from total cyto- plasmic mRNA of both control and STH-inhibited cultures are remarkably similar to the corresponding nuclear cap distributions (Table II). This implies that transport of cap-bearing mRNA molecules is not selective relative to the presence or absence of a 2'2Qemethyl group in the N' nucleotide of the cap. Cap zero structures were observed in the nuclear and cytoplasmic fractions of only STH-exposed cells. The absence of m6A at N' of cap zero structures sharply contrasts the relative predominance of this methylated base in cap 1 structures, and suggests that 2'ggemethylation of adenosine at N' may be necessary for subsequent base methylation. This sequence of methylation events has been observed in the formation of mouse globin mRNA cap structures (Cheng and Kazazian, 1978). The predominance of uridine at N' in both nuclear and cytoplasmic cap one structures derived from inhibited cells was somewhat surprising (Table II), since m7GpppU(m) is generally observed as a minor compo- nent of cap structures from Novikoff cells (Friderici gt al., 1976; Kaehler, gt al., 1977). The presence of 500 M STH in these cultures, however, may have altered the relative stability or processing efficiency of certain RNA species which happen to be enriched in 152 uridine at the N' position of caps. An increase in the amount of pyrimidines at N' might reflect an increase in the processing and/or transport of RNA molecules whose 5'-termini are generated internally from primary RNA transcripts. Alternatively, STH may selectively inhibit certain methylases to a greater extent than others. Such selectivity could account for the observation that cap zero structures from STH-treated cultures contain low levels of uridine at N'. In an earlier report, the presence of cap zero-bearing mRNAs in the cytoplasm of cells labeled in the presence of STH had indicated that 2'ggemethylation was not required for processing and transport of mature mRNA (Kaehler, §t_al., 1977). The present study indicates that cap zero-containing mRNAs are present on both monosomes and polysomes of cells which remained viable in the presence of 500 pM STH. We conclude that ribose methylation at N' does not appear to be requisite for ribosomal binding and subsequent translation of the undermethylated mRNA. REFERENCES Bantle, J., Maxwell, I. and Hahn, W. (1976) Analytical Biochem. 12, u13. Both, G., Furuichi, Y., Muthukrishnan, s. and Shatkin, A. (1975) Cell 6, 185. Both, G., Furuichi, Y., Muthukrishnan, S. and Shatkin, A. (1976) J. Mol. Biol. 10“, 637. Brack, C. and Tonegawa, S. (1977) Proc. Nat'l. Acad. Sci. USA 1“, 5652. Breathnach, R., Mandel, J. and Chambon, P. (1977) Nature (London) 270, 31“. Cheng, T.-C. and Kazazian, H. (1978) J. Biol. Chem. 253, 2“6. Coward, J., Motola, N. and Moyer, J. (1977) J. Med. Chem. 20, 500. Craig, N. (1973) J. Cell. Physiol. 82, 133. Desrosiers, R., Friderici, K. and Rottman, F. (197“) Proc. Nat'l. Acad. Sci. USA 11, 3971. Desrosiers, R., Friderici, K. and Rottman, F. (1975) Biochemistry 13. “367. Eschenfeldt, W. and Patterson, R. (1975) Prep. Biochem. 5, 2“7. Friderici, K., Kaehler, M. and Rottman, F. (1976) Biochemistry 15, 523“. Hickey, E., Weber, L. Baglioni, C., Kim, C. and Sarma, R. (1977) J. Mol. Biol. 109, 173. Jeffreys, A. and Flavell, R. (1977) Cell 12, 1097. Kaehler, M., Coward, J. and Rottman, F. (1977) Biochemistry 16, 5770. 153 15“ Kozak, M. and Shatkin, A. (1977) J. Mol. Biol. 112, 75. Kwan, S.-P., Wood, T. and Lingrel, J. (1977) Proc. Nat'l. Acad. Sci. USA 18, 178. Munns, T., Podratz, K., and Katzman, P. (197“) Biochemistry 13, ““09. Muthukrishnan, 5., Both, G., Furuichi, Y. and Shatkin, A. (1975) Nature (London) 255, 33. Muthukrishnan, S., Morgan, M., Banerjee, A. and Shatkin, A. (1976) Biochemistry 18, 5761. Muthukrishnan, 8., Moss, B., Cooper, J. and Maxwell, E. (1978) J. Biol. Chem. 253, 1710. Perry, R. and Kelley, D. (1976) Cell 8, “33. Perry, R., Kelley, D., Friderici, K. and Rottman, F. (1975) Cell 51, 13. Rottman, F. (1976) Trends Biochem. Sci. 1, 1217. Rottman, F. (1978) International Review of Biochemsitry, "Biochemistry of Nucleic Acids II" 11 (ed. B.F.C. Clarke) University Park Press, Baltimore, in press. Rottman, F., Desrosiers, R. and Friderici, K. (1976) Prog. Nucl. Acid. Res. 12, 21. Salditt-Georgieff, M., Jelinek, W., Darnell, J., Furuichi, Y., Morgan, M. and Shatkin, A. (1976) Cell 1, 227. Shatkin, A. (1976) Cell 2, 6“5. Thomason, A., Velicer, L. and Rottman, F. (1978) J. Virol., in press. Tilghman, 8., Curtis, P., Tiemeier, D., Leder, P. and Weissmann, C. (1978) Proc. Nat'l. Acad. Sci. USA 18, 1309. Tilghman, S., Tiemeier, D. Polsky, F., Edgell, M. Seidman, J., Leder, A., Enquist, L., Norman, B. and Leder, P. (1977) Proc. Nat'l. Acad. Sci. USA 1“, ““06. 2 4140 I1 193 031 31 IIIIIIIII