Purines in tRNAs Required for Recognition by ATP/CTP:tRNA Nucleotidyltransferase from Rabbit Liver Peter Spacciapolif and David L. Thurlow* Department of Chemistry, Clark University, 950 Main Street, Worcester, MA 01610, USA Recognition of tRNA by the enzyme ATP/CTPtRNA nucleotidyltransferase from rabbit liver was studied using 12 tRNAs, previously treated with the chemical modifier diethylpyrocarbonate(DEP). Such chemically modified tRNAs as a cosubstrate. A carbethoxylated purine at were labeled with 3zPby nucleotidyltransferase, using U-[~~P]ATP position 57 in the Y-loop interfered with recognition of the tRNA in Srll instances. DEP-modified purines at other positions (58 in the Y-loop, 52 or 53 in the Y-stem, and 71-73 in the acceptor stem), also interfered with the interaction, but in only a few tRNAs. The mammalian enzyme was more similar to the homologous enzyme from yeast than that from bacteria, in its requirements for chemically unmodified purines. The extent of exclusion of modified bases from 3zP-labeledmaterial diminished as the concentration of enzyme increased, demonstrating that interference was not due to the inability of the chemically altered tRNA to refold into a recognizable conformation. The degree of purification of the enzyme did not affect the identity of bases that inhibited the reaction when modified.
INTRODUCTION The enzyme ATP/CTP:tRNA nucleotidyltransferase (EC 2.7.7.25), found in all living systems examined to date (reviewed by Deutscher, 1982), catalyses ligation of CMP and AMP onto the 3'-end of tRNA molecules to restore the 3'-terminal sequence CpCpA. It is an unusual enzyme in that it requires no nucleic acid template, yet is highly specific for tRNAs and catalyses the synthesis of the CpCpA sequence with impressive fidelity. Nucleotidyltransferases have been isolated from several sources, with that from rabbit liver being especially well characterized in terms of its physical and kinetic properties (Deutscher, 1982). Its active site is thought to contain binding sites for ATP, CTP and sites near the 3'-end of its tRNA substrate (Masiakowski and Deutscher, 1980). In addition, one set of experiments suggested the possibility of additional sites of interaction on the tRNA molecule (Masiakowski and Deutscher, 1979). We wish to identify all regions of the tRNA substrate that are required in chemically unaltered form for interaction with this enzyme. We have previously reported the identification of nucleotides involved in recognition by nucleotidyltransferases from Escherichiu coli and the yeast Saccharomyces cerevisiue (Spacciapoli et ul., 1989). Our experimental approach was to treat the tRNA with the chemical modifiers diethylpyrocarbonate (DEP) or hydrazine, and then label the modified tRNA with 32Pusing nucleotidyltransferase and C~-[~~P]ATP. tRNAs containing a chemically altered base at a site crucial to the interaction, did not become labeled. Hence, characterizing the content of modified bases in the labeled tRNAs. tPresent address: Laboratory of Biochemical Pharmacology, National Institute of Diabetes and Digestive and Kidney Diseases, NIH, Building 8, Room 2A-15, Bethesda, M D 20982, USA. *Author to whom correspondence should be addressed.
8 1990 by John Wiley & Sons, Ltd.
permitted identification of sites within the tRNA molecule at which chemically altered nucleotides interfered with the recognition process (Spacciapoli et ul., 1989). In this report we utilize the same approach and the chemical probe DEP to define regions of 12 tRNAs required for interaction with nucleotidyltransferase from rabbit liver.
EXPERIMENTAL
All chemicals used in this study were Reagent grade. Buffers were prepared using deionized distilled water (Barnstead NANOpure 11, Dubuque, IA, USA) and were stored at 4 "C. Purified isoacceptor species of tRNA were purchased from Boehringer-Mannheim Biochemicals (Indianapolis, IN, USA) or Subriden RNA (Rolling Bay, WA, USA). Some contaminating species were resolved on gels following labeling with 32P;they were recovered and analyzed separately. E-[~~P]ATP ( ~ 3 0 0 0 Ci/mmol) was purchased from ICN Radiochemicals (Irvine, CA, USA). Snake venom phosphodiesterase (29.1 units/mg) was obtained from Worthington Biochemicals (Freehold, NJ, USA). Frozen rabbit liver was obtained from Pel Freeze Biologicals (Rogers, AR, USA). The enzyme ATP/CTP:tRNA nucleotidyltransferase was prepared from rabbit liver as described by Deutscher (Deutscher and Masiakowski, 1978). Briefly, 250 g of liver was homogenized; following centrifugation at 13 000 x g and ammonium sulfate fractionation of the supernatant, the material was chromatographed on columns of DEAE cellulose (Whatman, Hillsboro, OR, USA), Affi Gel blue (Bio Rad, Rockville Centre, NY, USA), and P11 phosphocellulose (Whatman). Eluants from each chromatographic step were used in damage selection experiments. The homologous enzyme from Escherichiu coli, prepared according to Carre et ul. (1970), was a generous gift by Dr D. C . Fritzinger
0952-3499/9O/O 14-1 55 $05.00 JOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990 149
(Georgetown Medical Center, Washington, DC, USA). Nucleotidyltransferase from S. cerevisiae was kindly provided by Dr Jacek Wower (University of Massachusetts, Amherst, MA, USA), as the eluant from the DEAE cellulose column step in the procedure of Sternbach et al. (1971). To remove the 3'-terminal A (or CpA) residue(s), aliquots of individual tRNAs (50 pg) were treated with 1 pL snake venom phosphodiesterase (1 mg/mL) in 50 mM Tris+ acetate (pH 8.0), 10 mM magnesium acetate for 20 min at 23 "C. A tenfold concentrated solution was added to final concentrations of 10 mM Tris+acetate (pH 7.6) + 100 mM LiCl+ 2.5 mM disodium ethylenediaminetetraacetate (Na,EDTA) + 0.5% sodium dodecyl sulfate (SDS), and the solution was extracted with an equal volume of phenol, saturated in the LiCl+SDS buffer. Following precipitation with 4 volumes of ethanol, the tRNA, lacking its terminal A (or CpA), was dissolved in 200 pL 0.3 M sodium acetate, reprecipitated, rinsed with 95% ethanol, dried and resuspended in water. Treatment of tRNA prepared in this manner with DEP was carried out as described by Peattie (1979), except that the length of incubation was shortened to 1.5 min and 23s rRNA (10 pg) was used as carrier. The extent of modification under these conditions was less than one residue/tRNA, as verified by most of the material remaining intact following cleavage, by treatment with aniline, at modified bases. Prior to restoration of AMP, 5 pg of tRNA substrate was heated to 70 "C in 20 pL 50 mM Tris+ glycine (pH 8.5), 5 mM MgCl,, 8 mM dithiothreitol, 25 p~ CTP (CCA buffer), and cooled slowly to 37 "C. 50 pCi a[32P]ATP(to a final concentration of 0.7 p ~ and ) 34 punit enzyme (1 unit is defined as the amount of to tRNA enzyme required to add 1 pmol of CW-[~~P]ATP in 1 h at 37 "C, under the assay conditions of Deutscher, 1972a) were added (enzyme/tRNA = 0.6-0.8 unit/g) and incubation at 37 "C was carried out for 30 min. The samples were dried to 5 pL; an equal volume of formamide, 0.02% xylene cyanole and 0.02% bromophenol blue were added and the labeled tRNA was purified on a 0.4 mm x 18 cm x 35 cm 20% polyacrylamide 'sequencing' gel (acrylamide + N,N'-methylenebisacrylamide = 19:l) in 100 mM Tris+borate (pH8.3)+2.5 mM Na, EDTA + 8 M urea. Gels were run at 2&25 mA constant current until the xylene cyanole tracking dye had migrated 26 cm. Labeled material was eluted from the gel slice overnight at 3 "C in 200 pL buffer (0.5 M ammonium acetate + 0.5% SDS + 0.1 mM Na,EDTA) containing 10 pg carrier tRNA, precipitated with ethanol, washed with sodium acetate, and rinsed with ethanol. Control 32P-labeledmaterial, that had not previously been chemically modified, was treated with DEP as described above, and then repurified on a sequencing gel. In some instances, as indicated in the Fig. legends, control material was used without being repurified. Cleavage of the tRNA at modified bases was achieved by incubation with 1 M aniline + acetate (pH 4.5) at 65 "C for 5 min in the dark, followed by drying and washing twice with water as described by Peattie (1979). Aniline cleavage products were analyzed on 20% polyacrylamide sequencing gels in 100 mM Tris + borate (PH 8.3) 2.5 mM Na,EDTA 8 M urea. The radio-
+
+
active contents of the samples were adjusted to be approximately equal ( f 15%) using a Bioscan QC 2000 radiation detector (Bioscan Inc., Washington, DC, USA). Autoradiograms were made using Kodak X-Omat AR film and were exposed at - 70 "C for 12-60 h.
RESULTS Modified bases were excluded from position 57 in all tRNAs and from other sites in some tRNAs. Twelve individually purified
tRNAs from E. coli were examined to determine if carbethoxylation of purines with DEP interfered with recognition of the modified tRNA by the enzyme ATP/ CTP:tRNA nucleotidyltransferase from rabbit liver. The extent to which such modified bases interfered with the interaction was assessed qualitatively by the degree of exclusion of carbethoxylated nucleotides from tRNAs that were recognized as substrate by this enzyme. Chemically treated tRNAs, lacking their 3'-terminal A (or CpA), were incubated together with R-[~,P]ATPand nucleotidyltransferase. 32P-labeledtRNA was recovered on denaturing polyacrylamide gels and the material was then treated with aniline to induce breakage of the chain at modified bases. Control tRNA, not treated with DEP, was labeled with 32Pusing nucleotidyltransferase, then modified with DEP, purified on a gel, and finally treated with aniline. Aniline induced fragments were then separated by size on sequencing gels, and the content of modified bases was assessed by the presence or absence of bands on an autoradiogram of the gel (see Fig. 1 in Spacciapoli et a!., 1989). The length of a labeled fragment generated in this manner, was determined by the distance of the modified nucleotide from the labeled (3') end of the tRNA. Consequently, the content of modified bases in an entire tRNA could be determined by analysis of one gel, on which fragments extending from mononucleotides, to full length tRNAs were resolved using multiple loadings of the samples. Autoradiograms presented in Fig. 1 illustrate separation of aniline-induced fraghments from three tRNAs, each of which had been labeled with 32Pusing nucleotidyltransferase from rabbit liver, either before (lanes 1,3,5) or after (lanes 2,4,6) treatment of the tRNA with DEP. Positions at which bands were reproducibly missing, or present at reduced intensity, in material labeled following modification (indicated by arrows in Fig. 1) identify sites where modified bases had interfered with the labeling reaction. Occasionally, reduced intensities were observed in one experiment, but were not apparent in others (Fig. l(a), position 71). We do not know the cause for this variation; we identified exclusions as instances of interference with enzyme activity only when comparable effects were seen in all experiments. Accordingly, our data represent a minimum set of bases capable of inhibiting the enzyme. A total of twelve tRNAs were examined in this fashion and a summary of sites from which carbethoxylated purines were excluded is presented in Table 1. As previously observed for nucleotidyltransferases from E. coli and yeast (Spacciapoli et al., 1989), a modified purine at position 57, a universally conserved purine located in the
RECOGNITION OF tRNA BY M A M M A L I A N NUCLEOTIDYLTRANSFERASE 12
56
34
12
56
34
12
56
34
4
4 71
4 4
4 19
53
4
4
4
4 4 58
4
4 57
4 73
(a)
(b)
(c)
Figure 1. Autoradiograms showing exclusions of carbethoxylated purines from 32P-labeled tRNAs by rabbit liver nucleotidyltransferase. tRNAfMe' (lanes 1,2); tRNAMe' (lanes 3,4); and tRNAG"' (NUG) (lanes 5,6) were labeled with [32P]AMP using rabbit liver nucleotidyltransferase either before (lanes 1, 3 and 5) or after (lanes 2, 4 and 6) treatment of the tRNA with DEP. Following treatment of the modified, gel-purified material with aniline, the resulting fragments were resolved on sequencing gels run at 20 mA constant current for: A, 2 h; 6 ,5 h; C, 8 h. Arrows indicate exclusions of modified bases from tRNAs labeled after DEP modification. Nucleotide positions are numbered according to Sprinzl eta/. (1987).
Table 1. DEP-modified nucleotides that interfered with nucleotidyltransferase from rabbit liver Nucleotide positions 19
52
53
57
-
+++
+++ +++
+++ +++ +++
+
-
-
+++
+++ +++
-
-
-
-
-
-
+
-
-
+
+++
-
++
+
-
-
-
+
+
-
+ -
+ +++
71-73 -
72, 73, 72, 73,
+++
++
++ ++
-
+++ ++ +++ +++
+++ +++ +++ +++
++
71,
+++ -
++ +++ +
71, 72. 71,
-
+++
-
Enzyme from rabbit liver, both the DEAE cellulose and Affi Gel blue eluants, were tested for their ability to label DEP-treated tRNAs. The final concentration of enzyme was 0.15 4 . 2 0 units/mL. Nucleotide positions are numbered according to Sprinzl e t a / . (1987): + + +, total ( > 90%); + +, substantial (50-90%); and +, partial (20--50%) extents of exclusion, inferred from reduction of band intensities on autoradiograms as judged by visual comparsion with control bands.
0 1990 by John Wiley & Sons, Ltd.
Y-loop (numbered according to Sprinzl et al., 1987), dramatically interfered with the labeling reaction in all tRNAs examined. Exclusion of a modified base from this site was independent of whether the nucleotide base was an adenine (lane 2, Fig. 1) or a guanine (lanes 4 and 6 , Fig. I). In addition, exclusion of modified nucleotides from positions 19 (D-loop) or 58 (Y-loop), located in a stacked array near the corner of the tRNA's tertiary structure, was identified in nine tRNAs. Interference by modified bases was also observed in the Y-stem, positions 52 or 53, in six tRNAs; and near the end of the acceptor stem, sites 71-73, in seven tRNAs. There was no apparent correlation among exclusions from positions 19,58,52-53 and 71-73. Nucleotidyltransferasefrom rabbit Liver was more similar to that in yeast than its counterpart in bacteria. The enzyme from
rabbit liver was similar to both those from E. coli and yeast in that a modified purine at position 57 interfered with enzymatic activity in almost all instances. Similar to the yeast enzyme, the rabbit liver protein was inhibited by chemically altered purines at positions 19 and 58 in some, but not all tRNAs. As specific examples, significant exclusions from positions 19 and 58 were evident with both the yeast and rabbit liver enzymes in JOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990
151
PETER SPACCIAPOLI A N D DAVID L. THURLOW 1 2 3 4 5 6 7 8
I 2 3 4 5 6 7 8
12345678
4 19
4 19
4 53
4 57
4 58
4 65
4 53
4 73
4 57 4 58
(a) (b) (C) Figure 2. Autoradiograms showing exclusions by f. coli, yeast, and various preparations of rabbit liver nucleotidyltransferases.tRNAPhefrom E. coli was labeled with [32P]AMP either before (lanes 1, 2) or after (lanes 3-8) treatment with DEP. Control material in lane 1 was not gel purified following modification with DEP, whereas that in lane 2 was. Enzyme preparations were from f. coli (lane 3). yeast (lane 4). or rabbit liver (lanes 5-8).The rabbit liver enzyme was obtained from the DEAE cellulose eluant (lane 5). the Affi Gel blue eluant (lane 6). peak 1 of the phosphocellulose column (lane 7) or peak 2 of the phosphocellulose column (lane 8). All enzymes were present at a final concentration of 0.02 units/mL. Following treatment of the modified material with aniline, the resulting fragments were resolved on sequencing gels run at 20 mA constant current. Assignment as for Fig. 1.
tRNAPhe(Fig. 2, lanes 4-8), yet were not observed in tRNAGIY with either enzyme (Fig. 3, lanes 3 4 ) . The general patterns of exclusions by yeast and rabbit liver nucleotidyltransferases were similar for each individual tRNA (Table 1 and Spacciapoli et al., 1989). Relative to the rabbit liver enzyme, and in accord with previous observations with that from yeast (Spacciapoli et al., 1989), nucleotidyltransferase from E. coli appeared to have much more stringent reguirements for unmodified nucleotides near the corner of the tRNA's tertiary structure, that is at positions 19, 57, and 58 (Fig. 2, lane 3 and Fig. 3, lane 2). In contrast, near the 3'-end the E. coli enzyme did not generally discriminate against modified bases to the extent that either the yeast or rabbit liver forms did (see Fig. 2(a), position 73). Extent of exclusion of modified purines was dependent on the concentration of enzyme. To assess effects of changing the
concentration of enzyme, the final concentration of rabbit liver nucleotidyltransferase (Affi Gel blue eluant) in the labeling reaction was varied from 0.0074.54 unit/mL. Exclusions of modified purines from labeled tRNAPhewere followed in the usual manner (Fig. 4). In all instances the extents of exclusion decreased as the concentration of enzyme increased. Similar results were 152 JOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990
obtained with tRNA'" (data not shown) and have previously been reported for the bacterial and yeast enzymes (Spacciapoli et al., 1989). These data demonstrate that exclusions did not result from inability of the modified tRNA to refold into a form recognizable by the enzyme (see Discussion). In addition, the possibility that additional instances in which a modified base interferes with labeling might be found at still lower concentrations of enzyme means that our results consititue a minimum set of bases required for interaction with nucleotidyltransferase from rabbit liver. Extent of purification of the enzyme did not affect pattern of exclusions. The initial survey of tRNAs described above
was carried out using two preparations of the enzyme; a 30 mM phosphate eluant from the DEAE cellulose column, and material released by tRNA from the Affi Gel blue column. The latter was approximately 150 fold purified relative to the DEAE cellulose eluant. In all cases the pattern of excluded bases was the same for both preparations of the enzyme. It was previously reported that two distinct forms of the enzyme could be separated by chromatography on phosphocellulose (Deutscher, 1972a). They appeared to have nearly identical physical properties and were
0 1990 by John Wiley & Sons,Ltd.
RECOGNITION OF tRNA BY M A M M A L I A N NUCLEOTIDYLTRANSFERASE
I 2 3 4
I 2 3 4
1234
4 19 4 19
4 71
4 57 4 58
(a)
(b)
(C)
Figure 3. Autoradiograms showing exclusions by E. coli, yeast, and rabbit liver nucleotidyltransferases. tRNAG'y from E. coli was labeled with [32P]AMP either before (lane 1) or after (lanes 2 4 ) treatment with DEP. Control material (lane 1) was gel purified following modification with DEP. Enzyme preparations were from E. coli (lane 2), yeast (lane 3), or rabbit liver, the DEAE cellulose eluant (lane 4). All enzymes were present at a final concentration of 0.1-0.2 units/mL. Following treatment of the modified material with aniline, the resulting fragments were resolved on sequencing gels run at 20 mA constant current. Assignments as for Fig. 1.
similar, but not identical, as judged by kinetic measurements and substrate specificity (Deutscher, 1972b). Accordingly, a column of phosophocellulose was used to separate enzymatic activity into two peaks, a further purification of approximately fivefold over the Affi Gel blue eluant. All four preparations were used to examine whether the purity of the enzyme affected the pattern of excluded bases in damage-selection experiments. Three tRNAs were tested in this manner and the purity of the enzyme did not significantly affect either the identity of bases excluded in modified form or the degree to which they were excluded (see Fig. 2, lanes 5-8).
restoration of [32P]AMP(Table I). It was somewhat puzzling that chemically altered purines in this region of the molecule did not inhibit the enzymatic reaction in all cases. In tRNAs where modified purines near the 3'-end did not interfere with the interaction, binding of the enzyme to cytidines 74 and 75 must have been sufficient to permit catalytic activity. In other instances, additional interactions may have occurred, which varied depending on differences among conformations of the 3'-terminal region. As a specific example, tRNAme' exhibited total exclusion of a modified purine from position 72 (Fig. l(a)); it was only one of two tRNAs in which a modified purine was totally excluded from the 3'-terminal region (Table 1). From analysis of the crystal structure of this tRNA it is clear that the adenine (A) at residue 72 is highly exposed as the sugar/phosphate backbone makes a sharp turn (Woo et al., 1980), and could thereby interact with nucleotidyltransferase in a totally different fashion than the homologous base in tRNAphe,for example, which is in a highly dissimilar environment in the crystal structure (Woo et al., 1980). Another set of previous experiments suggested that the rabbit liver enzyme may interact with a region of the tRNA at some distance from the 3'-end, and that this interaction may be required for stimulation of enzymatic activity (Masiakowski and Deutscher, 1979). Accordingly we note that modified bases located at the corner of the tRNA's tertiary structure, site 57 in all cases and positions 19 or 58 in several, had pronounced effects on the reaction (Table 1). We propose that rabbit liver nucleotidyltransferase interacts with its tRNA substrate by attachment at one or both 3'-terminal cytidines, as previously suggested (Masiakowski and Deutscher, 1980), contacts bases near the 3'-end depending on the particular tRNA, and, in all cases, simultaneously extends to the corner of the tRNA's three-dimensional structure, where the Y-and D-loops are juxtaposed. Interaction with tRNA by the rabbit liver enzyme is therefore very similar to that by the homologous enzyme from yeast, as demonstrated by the similarities between bases excluded in modified form by these two enzymes (Figs 2 and 3, Table I , and Spacciapoli et al., 1989). Recognition by the E. coli enzyme also involves the same regions of the tRNA, but appears to be more stringent in requiring unaltered bases near the corner of the molecule, where the D- and Y-loops join (Fig. 3 and Spacciapoli et al., 1989). Nucleotides at positions 52 and 53, located in the Y-stem near the top of the molecule (Woo rf al., 1980), were also required in chemically unaltered form in several instances (Table I). This agrees with our proposed model in that one would expect minor points of contact along the top of the tRNA as the enzyme extends from the 3'-end across the molecule towards the corner.
DISCUSSION The active site of nucleotidyltransferase from rabbit liver has been characterized extensively and has been shown to have binding sites for ATP, one or two CTPs and the 3'-end of its tRNA substrate (Masiakowski and Deutscher, 1980). In accord with these findings, we observed that carbethoxylated purines near the 3'-end of seven out of twelve tRNAs examined. interfered with
01990 by John Wiley & Sons, Ltd.
Effects of increasing the concentration of enzyme. The observation that extents of exclusion of modified bases were dependent on enzyme concentration (Fig. 4) is significant for two reasons. These data clearly demonstrate that exclusion did not result from the inability of the modified tRNA to refold into a conformation capable of being recognized by the enzyme. That is, at elevated concentrations of enzyme, no carbethoxylated purine had sigJOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990
153
PETER SPACCIAPOLI AND DAVID L. THURLOW I23456
I 2 3 4 5 6
I 2 3 4 5 6
4 19
4 19 4 53 4 57
4 50
4 65
4 53
4 57 4 58
4 73
(a)
(b)
(C)
Figure 4. Titration of the rabbit liver enzyme. tRNAPhe from E. coliwas labeled with [32P]AMP either before (lanes 1, 2 ) or after (lanes 3-6) treatment with DEP using rabbit liver nucleotidyltransferase,the Aff i Gel blue eluant. Control material in lane 1 was not gel purified following modification with DEP, whereas that in lane 2 was. Final concentrations of enzyme were: lane 3,0.54; lane 4, 0.18; lane 5, 0.018; lane 6, 0.007 units/mL. Following treatment of the modified material with aniline, the resulting fragments were resolved on sequencing gels run at 20 mA constant current. Assignments as for Fig. 1.
nificant effects on the labeling reaction (Fig. 4, lane 3), regardless of its location in the tRNA’s tertiary structure. We conclude that the chemically modified bases present in our experiments did not result in gross distortions of the tRNA’s structure following renaturation. However, based on these data alone, we cannot exclude the possibility that the tRNA refolded into a structure slightly different from that of native tRNA and that the rate at which the enzyme utilized the chemically altered substrate was lowered. If these were the case, increasing the concentration of enzyme could compensate for the reduced rate of reaction. Furthermore, the effects of enzyme concentration signify that the sites we have identified as being important in the interaction (Table 1) represent a minimum set of such bases. At lower concentrations of enzyme there may be additional instances where effects would not have been detected in our survey. Using our approach alone, we cannot distinguish effects of modified bases on the affinity between substrate and enzyme from those on the rate of the reaction. It was not possible to judge effects on KM or V,,, by simply measuring initial rates as a function of substrate concentrations, because the population of chemically modified substrate molecules was heterogeneous. For this reason we are currently using site-directed mutagenesis to pre154 JOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990
pare a homogeneous population of tRNAs, containing one altered nucleotide at a position we have identified as being required in chemically unmodified form. The absence of modified bases can also result from purification of chemically modified material on polyacrylamide gels. Instances
in which chemically altered bases were excluded as a result of purification of modified material on polyacrylamide gels were occasionally observed, most notably in tRNAPhe(Figs 2 and 4). Such exclusions occurred reproducibly from positions 69-71, 44, and 27-31 in tRNAPhe.We made similar observations in our previous set of damage-selection experiments (Spacciapoli et af., 1989) and have undertaken a more thorough investigation of this phenomenon (Hegg and Thurlow, 1990). From these studies it is clear that substantial secondary structure persists during purification of some tRNAs on 8 M urea gels run at 45°C and that the electrophoretic mobility of chemically modified tRNAs can vary, depending on which base had been altered. Accordingly, all instances of exclusion we identify as resulting from interference with enzymatic activity were determined relative to control material that had been purified on the same purification gel as the experimental sample.
01990 by John Wiley & Sons, Ltd.
RECOGNITION OF tRNA BY M A M M A L I A N NUCLEOTIDYLTRANSFERASE
Extent of purification of the enzyme did not affect exclusion of modified bases. Given the highly similar nature of modified
bases excluded by nucleotidyltransferases from yeast and rabbit liver, it was not suprising to observe that the two chromatographically distinct forms of activity from rabbit liver were indistinguishable by this criterion (Fig. 2, lanes 7, 8). Previous work has also shown them to be very similar in their physical and kinetic properties (Deutscher, 1972b). It is important to note that even relatively crude preparations of the enzyme exhibited exactly the same behavior as the highly purified forms (Fig. 2). There may have been several proteins in the crude preparation which would, under appropriate circumstances, be capable of competing with nucleotidyltransferase for binding to tRNAs. However, under our set of experimental conditions most of these interactions would have been unlikely to be significant. For instance, because the tRNAs were fully processed, they would not have been approriate substrates for tRNA processing enzymes. In addition, since the 3'-terminal A had been removed and because we did not include amino acids in our incubations, interactions with aminoacyl-tRNA synthetases would have been unlikely. Also, because the tRNAs were
not aminoacylated, they would not be expected to undergo complex formation with elongation or initiation factors.
CONCLUSION The interaction between ATP/CTP:tRNA nucleotidyltransferase from rabbit liver and DEP-modified tRNAs was inhibited when tRNAs contained carbethoxylated purines near the 3'-end of the tRNA or near the corner of the molecule where the D- and Y-loops are juxtaposed. Nucleotide positions 19, 52, 53, 57 and 58 most frequently required chemically unmodified bases for proper recognition. The extent of interference was dependent on the concentration of enzymatic activity, but was not affected by the purity of the enzyme preparation.
Acknowledgements We gratefully acknowledge Dr Murray Deutscher (University of Connecticut Health Center, Farmington. CT, USA) for his help and advice in the purification of nucieotidyitransferase from rabbit liver.
REFERENCES Carre, D. S., Litvak, S., and Chapeville, F. (1970). Purification and properties of Escherichia coli CTP/ATP:tRNA nucleotidyltransferase. Biochim. Biophys. Acta 224, 371-381, Deutscher, M. P. (1972a). Reactions a t the 3'-terminus of transfer ribonucleic acid: purification and physical and chemical properties of rabbit liver transfer ribonucleic acid nucleotidyltransferase. J . Biol. Chem. 247. 450-458. Deutscher, M. P. (1972b) Reactions at the 3'-terminus of transfer ribonucleic acid: catalytic properties of two purified rabbit liver transfer ribonucleic acid nucleotidyltransferases. J. Biol. Chem. 247.459468. Deutscher, M P. (1 982). tRNA nucleotidyltransferase. The Enzymes 15, 183-21 5. Deutscher, M. P., and Masiakowski, P. (1978). Binding of tRNA nucleotidyltransferase to Affi Gal blue: rapid purification of the enzyme and binding studies. Nucleic Acids Res. 5, 1947-1 954. Hegg, L. A., and Thurlow, D. L. (1990). Residual tRNA secondary structure in 'denaturing' 8 M urea/TBE polyacrylamide gels: effects on electrophoretic mobility and dependency on prior chemical modification of the tRNA. Nucleic Acids Res. (in press) Masiakowski, P., and Deutscher, M. P. (1979). Separation of functionally distinct regions of a macromolecular substrate: stimulation of tRNA nucleotidyltransferase by a nonreacting fragment of tRNA. J. Biol. Chem. 254, 2585-2587.
0 1990 by John Wiley & Sons, Ltd.
Masiakowski, P., and Deutscher, M. P. (1980). Dissection of the active site of rabbit liver tRNA nucleotidyltransferase: specificity and properties of subsites for donor nucleoside triphosphates. J. Biol. Chem. 255, 11240-1 1246. Peattie, D. A. (1979). Direct chemical method for sequencing RNA. Proc. Natl Acad. Sci. USA 76, 1760-1 764. Spacciapoli, P., Doviken, L., Mulero, J. J., and Thurlow, D. L. (1 989). Recognition of tRNA by the enzyme ATP/CTP:tRNA nucleotidyltransferase: interference by nucleotides modified with diethyl pyrocarbonate or hydrazine. J. Biol. Chem. 264, 3799-3805. Sprinzl, M., Hartmann, T., Meissner, F., Moll, J., and Vorderwulbecke, T. (1987). Compilation of tRNA sequences and sequences of tRNA genes. Nucleic Acids Res. 1 5 Suppl., r53-rl88. Sternbach, H., von der Haar, F., Schlimme, E., Gaertner, E., and Cramer, F. (1971). Isolation and properties of tRNA nucleotidyltransferase from yeast. Eur. J. Biochem. 22, 166-1 72. Woo, N. H., Roe, 6. A,, and Rich, A. (1980). Three-dimensional structure of Escherichia coli initiator tRNAfMe' Nature 286, 346-351.
Received 1 December 1989; accepted (revised) 30 January 1990.
JOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990
155
INTRODUCTION The enzyme ATP/CTP:tRNA nucleotidyltransferase (EC 2.7.7.25), found in all living systems examined to date (reviewed by Deutscher, 1982), catalyses ligation of CMP and AMP onto the 3'-end of tRNA molecules to restore the 3'-terminal sequence CpCpA. It is an unusual enzyme in that it requires no nucleic acid template, yet is highly specific for tRNAs and catalyses the synthesis of the CpCpA sequence with impressive fidelity. Nucleotidyltransferases have been isolated from several sources, with that from rabbit liver being especially well characterized in terms of its physical and kinetic properties (Deutscher, 1982). Its active site is thought to contain binding sites for ATP, CTP and sites near the 3'-end of its tRNA substrate (Masiakowski and Deutscher, 1980). In addition, one set of experiments suggested the possibility of additional sites of interaction on the tRNA molecule (Masiakowski and Deutscher, 1979). We wish to identify all regions of the tRNA substrate that are required in chemically unaltered form for interaction with this enzyme. We have previously reported the identification of nucleotides involved in recognition by nucleotidyltransferases from Escherichiu coli and the yeast Saccharomyces cerevisiue (Spacciapoli et ul., 1989). Our experimental approach was to treat the tRNA with the chemical modifiers diethylpyrocarbonate (DEP) or hydrazine, and then label the modified tRNA with 32Pusing nucleotidyltransferase and C~-[~~P]ATP. tRNAs containing a chemically altered base at a site crucial to the interaction, did not become labeled. Hence, characterizing the content of modified bases in the labeled tRNAs. tPresent address: Laboratory of Biochemical Pharmacology, National Institute of Diabetes and Digestive and Kidney Diseases, NIH, Building 8, Room 2A-15, Bethesda, M D 20982, USA. *Author to whom correspondence should be addressed.
8 1990 by John Wiley & Sons, Ltd.
permitted identification of sites within the tRNA molecule at which chemically altered nucleotides interfered with the recognition process (Spacciapoli et ul., 1989). In this report we utilize the same approach and the chemical probe DEP to define regions of 12 tRNAs required for interaction with nucleotidyltransferase from rabbit liver.
EXPERIMENTAL
All chemicals used in this study were Reagent grade. Buffers were prepared using deionized distilled water (Barnstead NANOpure 11, Dubuque, IA, USA) and were stored at 4 "C. Purified isoacceptor species of tRNA were purchased from Boehringer-Mannheim Biochemicals (Indianapolis, IN, USA) or Subriden RNA (Rolling Bay, WA, USA). Some contaminating species were resolved on gels following labeling with 32P;they were recovered and analyzed separately. E-[~~P]ATP ( ~ 3 0 0 0 Ci/mmol) was purchased from ICN Radiochemicals (Irvine, CA, USA). Snake venom phosphodiesterase (29.1 units/mg) was obtained from Worthington Biochemicals (Freehold, NJ, USA). Frozen rabbit liver was obtained from Pel Freeze Biologicals (Rogers, AR, USA). The enzyme ATP/CTP:tRNA nucleotidyltransferase was prepared from rabbit liver as described by Deutscher (Deutscher and Masiakowski, 1978). Briefly, 250 g of liver was homogenized; following centrifugation at 13 000 x g and ammonium sulfate fractionation of the supernatant, the material was chromatographed on columns of DEAE cellulose (Whatman, Hillsboro, OR, USA), Affi Gel blue (Bio Rad, Rockville Centre, NY, USA), and P11 phosphocellulose (Whatman). Eluants from each chromatographic step were used in damage selection experiments. The homologous enzyme from Escherichiu coli, prepared according to Carre et ul. (1970), was a generous gift by Dr D. C . Fritzinger
0952-3499/9O/O 14-1 55 $05.00 JOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990 149
(Georgetown Medical Center, Washington, DC, USA). Nucleotidyltransferase from S. cerevisiae was kindly provided by Dr Jacek Wower (University of Massachusetts, Amherst, MA, USA), as the eluant from the DEAE cellulose column step in the procedure of Sternbach et al. (1971). To remove the 3'-terminal A (or CpA) residue(s), aliquots of individual tRNAs (50 pg) were treated with 1 pL snake venom phosphodiesterase (1 mg/mL) in 50 mM Tris+ acetate (pH 8.0), 10 mM magnesium acetate for 20 min at 23 "C. A tenfold concentrated solution was added to final concentrations of 10 mM Tris+acetate (pH 7.6) + 100 mM LiCl+ 2.5 mM disodium ethylenediaminetetraacetate (Na,EDTA) + 0.5% sodium dodecyl sulfate (SDS), and the solution was extracted with an equal volume of phenol, saturated in the LiCl+SDS buffer. Following precipitation with 4 volumes of ethanol, the tRNA, lacking its terminal A (or CpA), was dissolved in 200 pL 0.3 M sodium acetate, reprecipitated, rinsed with 95% ethanol, dried and resuspended in water. Treatment of tRNA prepared in this manner with DEP was carried out as described by Peattie (1979), except that the length of incubation was shortened to 1.5 min and 23s rRNA (10 pg) was used as carrier. The extent of modification under these conditions was less than one residue/tRNA, as verified by most of the material remaining intact following cleavage, by treatment with aniline, at modified bases. Prior to restoration of AMP, 5 pg of tRNA substrate was heated to 70 "C in 20 pL 50 mM Tris+ glycine (pH 8.5), 5 mM MgCl,, 8 mM dithiothreitol, 25 p~ CTP (CCA buffer), and cooled slowly to 37 "C. 50 pCi a[32P]ATP(to a final concentration of 0.7 p ~ and ) 34 punit enzyme (1 unit is defined as the amount of to tRNA enzyme required to add 1 pmol of CW-[~~P]ATP in 1 h at 37 "C, under the assay conditions of Deutscher, 1972a) were added (enzyme/tRNA = 0.6-0.8 unit/g) and incubation at 37 "C was carried out for 30 min. The samples were dried to 5 pL; an equal volume of formamide, 0.02% xylene cyanole and 0.02% bromophenol blue were added and the labeled tRNA was purified on a 0.4 mm x 18 cm x 35 cm 20% polyacrylamide 'sequencing' gel (acrylamide + N,N'-methylenebisacrylamide = 19:l) in 100 mM Tris+borate (pH8.3)+2.5 mM Na, EDTA + 8 M urea. Gels were run at 2&25 mA constant current until the xylene cyanole tracking dye had migrated 26 cm. Labeled material was eluted from the gel slice overnight at 3 "C in 200 pL buffer (0.5 M ammonium acetate + 0.5% SDS + 0.1 mM Na,EDTA) containing 10 pg carrier tRNA, precipitated with ethanol, washed with sodium acetate, and rinsed with ethanol. Control 32P-labeledmaterial, that had not previously been chemically modified, was treated with DEP as described above, and then repurified on a sequencing gel. In some instances, as indicated in the Fig. legends, control material was used without being repurified. Cleavage of the tRNA at modified bases was achieved by incubation with 1 M aniline + acetate (pH 4.5) at 65 "C for 5 min in the dark, followed by drying and washing twice with water as described by Peattie (1979). Aniline cleavage products were analyzed on 20% polyacrylamide sequencing gels in 100 mM Tris + borate (PH 8.3) 2.5 mM Na,EDTA 8 M urea. The radio-
+
+
active contents of the samples were adjusted to be approximately equal ( f 15%) using a Bioscan QC 2000 radiation detector (Bioscan Inc., Washington, DC, USA). Autoradiograms were made using Kodak X-Omat AR film and were exposed at - 70 "C for 12-60 h.
RESULTS Modified bases were excluded from position 57 in all tRNAs and from other sites in some tRNAs. Twelve individually purified
tRNAs from E. coli were examined to determine if carbethoxylation of purines with DEP interfered with recognition of the modified tRNA by the enzyme ATP/ CTP:tRNA nucleotidyltransferase from rabbit liver. The extent to which such modified bases interfered with the interaction was assessed qualitatively by the degree of exclusion of carbethoxylated nucleotides from tRNAs that were recognized as substrate by this enzyme. Chemically treated tRNAs, lacking their 3'-terminal A (or CpA), were incubated together with R-[~,P]ATPand nucleotidyltransferase. 32P-labeledtRNA was recovered on denaturing polyacrylamide gels and the material was then treated with aniline to induce breakage of the chain at modified bases. Control tRNA, not treated with DEP, was labeled with 32Pusing nucleotidyltransferase, then modified with DEP, purified on a gel, and finally treated with aniline. Aniline induced fragments were then separated by size on sequencing gels, and the content of modified bases was assessed by the presence or absence of bands on an autoradiogram of the gel (see Fig. 1 in Spacciapoli et a!., 1989). The length of a labeled fragment generated in this manner, was determined by the distance of the modified nucleotide from the labeled (3') end of the tRNA. Consequently, the content of modified bases in an entire tRNA could be determined by analysis of one gel, on which fragments extending from mononucleotides, to full length tRNAs were resolved using multiple loadings of the samples. Autoradiograms presented in Fig. 1 illustrate separation of aniline-induced fraghments from three tRNAs, each of which had been labeled with 32Pusing nucleotidyltransferase from rabbit liver, either before (lanes 1,3,5) or after (lanes 2,4,6) treatment of the tRNA with DEP. Positions at which bands were reproducibly missing, or present at reduced intensity, in material labeled following modification (indicated by arrows in Fig. 1) identify sites where modified bases had interfered with the labeling reaction. Occasionally, reduced intensities were observed in one experiment, but were not apparent in others (Fig. l(a), position 71). We do not know the cause for this variation; we identified exclusions as instances of interference with enzyme activity only when comparable effects were seen in all experiments. Accordingly, our data represent a minimum set of bases capable of inhibiting the enzyme. A total of twelve tRNAs were examined in this fashion and a summary of sites from which carbethoxylated purines were excluded is presented in Table 1. As previously observed for nucleotidyltransferases from E. coli and yeast (Spacciapoli et al., 1989), a modified purine at position 57, a universally conserved purine located in the
RECOGNITION OF tRNA BY M A M M A L I A N NUCLEOTIDYLTRANSFERASE 12
56
34
12
56
34
12
56
34
4
4 71
4 4
4 19
53
4
4
4
4 4 58
4
4 57
4 73
(a)
(b)
(c)
Figure 1. Autoradiograms showing exclusions of carbethoxylated purines from 32P-labeled tRNAs by rabbit liver nucleotidyltransferase. tRNAfMe' (lanes 1,2); tRNAMe' (lanes 3,4); and tRNAG"' (NUG) (lanes 5,6) were labeled with [32P]AMP using rabbit liver nucleotidyltransferase either before (lanes 1, 3 and 5) or after (lanes 2, 4 and 6) treatment of the tRNA with DEP. Following treatment of the modified, gel-purified material with aniline, the resulting fragments were resolved on sequencing gels run at 20 mA constant current for: A, 2 h; 6 ,5 h; C, 8 h. Arrows indicate exclusions of modified bases from tRNAs labeled after DEP modification. Nucleotide positions are numbered according to Sprinzl eta/. (1987).
Table 1. DEP-modified nucleotides that interfered with nucleotidyltransferase from rabbit liver Nucleotide positions 19
52
53
57
-
+++
+++ +++
+++ +++ +++
+
-
-
+++
+++ +++
-
-
-
-
-
-
+
-
-
+
+++
-
++
+
-
-
-
+
+
-
+ -
+ +++
71-73 -
72, 73, 72, 73,
+++
++
++ ++
-
+++ ++ +++ +++
+++ +++ +++ +++
++
71,
+++ -
++ +++ +
71, 72. 71,
-
+++
-
Enzyme from rabbit liver, both the DEAE cellulose and Affi Gel blue eluants, were tested for their ability to label DEP-treated tRNAs. The final concentration of enzyme was 0.15 4 . 2 0 units/mL. Nucleotide positions are numbered according to Sprinzl e t a / . (1987): + + +, total ( > 90%); + +, substantial (50-90%); and +, partial (20--50%) extents of exclusion, inferred from reduction of band intensities on autoradiograms as judged by visual comparsion with control bands.
0 1990 by John Wiley & Sons, Ltd.
Y-loop (numbered according to Sprinzl et al., 1987), dramatically interfered with the labeling reaction in all tRNAs examined. Exclusion of a modified base from this site was independent of whether the nucleotide base was an adenine (lane 2, Fig. 1) or a guanine (lanes 4 and 6 , Fig. I). In addition, exclusion of modified nucleotides from positions 19 (D-loop) or 58 (Y-loop), located in a stacked array near the corner of the tRNA's tertiary structure, was identified in nine tRNAs. Interference by modified bases was also observed in the Y-stem, positions 52 or 53, in six tRNAs; and near the end of the acceptor stem, sites 71-73, in seven tRNAs. There was no apparent correlation among exclusions from positions 19,58,52-53 and 71-73. Nucleotidyltransferasefrom rabbit Liver was more similar to that in yeast than its counterpart in bacteria. The enzyme from
rabbit liver was similar to both those from E. coli and yeast in that a modified purine at position 57 interfered with enzymatic activity in almost all instances. Similar to the yeast enzyme, the rabbit liver protein was inhibited by chemically altered purines at positions 19 and 58 in some, but not all tRNAs. As specific examples, significant exclusions from positions 19 and 58 were evident with both the yeast and rabbit liver enzymes in JOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990
151
PETER SPACCIAPOLI A N D DAVID L. THURLOW 1 2 3 4 5 6 7 8
I 2 3 4 5 6 7 8
12345678
4 19
4 19
4 53
4 57
4 58
4 65
4 53
4 73
4 57 4 58
(a) (b) (C) Figure 2. Autoradiograms showing exclusions by f. coli, yeast, and various preparations of rabbit liver nucleotidyltransferases.tRNAPhefrom E. coli was labeled with [32P]AMP either before (lanes 1, 2) or after (lanes 3-8) treatment with DEP. Control material in lane 1 was not gel purified following modification with DEP, whereas that in lane 2 was. Enzyme preparations were from f. coli (lane 3). yeast (lane 4). or rabbit liver (lanes 5-8).The rabbit liver enzyme was obtained from the DEAE cellulose eluant (lane 5). the Affi Gel blue eluant (lane 6). peak 1 of the phosphocellulose column (lane 7) or peak 2 of the phosphocellulose column (lane 8). All enzymes were present at a final concentration of 0.02 units/mL. Following treatment of the modified material with aniline, the resulting fragments were resolved on sequencing gels run at 20 mA constant current. Assignment as for Fig. 1.
tRNAPhe(Fig. 2, lanes 4-8), yet were not observed in tRNAGIY with either enzyme (Fig. 3, lanes 3 4 ) . The general patterns of exclusions by yeast and rabbit liver nucleotidyltransferases were similar for each individual tRNA (Table 1 and Spacciapoli et al., 1989). Relative to the rabbit liver enzyme, and in accord with previous observations with that from yeast (Spacciapoli et al., 1989), nucleotidyltransferase from E. coli appeared to have much more stringent reguirements for unmodified nucleotides near the corner of the tRNA's tertiary structure, that is at positions 19, 57, and 58 (Fig. 2, lane 3 and Fig. 3, lane 2). In contrast, near the 3'-end the E. coli enzyme did not generally discriminate against modified bases to the extent that either the yeast or rabbit liver forms did (see Fig. 2(a), position 73). Extent of exclusion of modified purines was dependent on the concentration of enzyme. To assess effects of changing the
concentration of enzyme, the final concentration of rabbit liver nucleotidyltransferase (Affi Gel blue eluant) in the labeling reaction was varied from 0.0074.54 unit/mL. Exclusions of modified purines from labeled tRNAPhewere followed in the usual manner (Fig. 4). In all instances the extents of exclusion decreased as the concentration of enzyme increased. Similar results were 152 JOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990
obtained with tRNA'" (data not shown) and have previously been reported for the bacterial and yeast enzymes (Spacciapoli et al., 1989). These data demonstrate that exclusions did not result from inability of the modified tRNA to refold into a form recognizable by the enzyme (see Discussion). In addition, the possibility that additional instances in which a modified base interferes with labeling might be found at still lower concentrations of enzyme means that our results consititue a minimum set of bases required for interaction with nucleotidyltransferase from rabbit liver. Extent of purification of the enzyme did not affect pattern of exclusions. The initial survey of tRNAs described above
was carried out using two preparations of the enzyme; a 30 mM phosphate eluant from the DEAE cellulose column, and material released by tRNA from the Affi Gel blue column. The latter was approximately 150 fold purified relative to the DEAE cellulose eluant. In all cases the pattern of excluded bases was the same for both preparations of the enzyme. It was previously reported that two distinct forms of the enzyme could be separated by chromatography on phosphocellulose (Deutscher, 1972a). They appeared to have nearly identical physical properties and were
0 1990 by John Wiley & Sons,Ltd.
RECOGNITION OF tRNA BY M A M M A L I A N NUCLEOTIDYLTRANSFERASE
I 2 3 4
I 2 3 4
1234
4 19 4 19
4 71
4 57 4 58
(a)
(b)
(C)
Figure 3. Autoradiograms showing exclusions by E. coli, yeast, and rabbit liver nucleotidyltransferases. tRNAG'y from E. coli was labeled with [32P]AMP either before (lane 1) or after (lanes 2 4 ) treatment with DEP. Control material (lane 1) was gel purified following modification with DEP. Enzyme preparations were from E. coli (lane 2), yeast (lane 3), or rabbit liver, the DEAE cellulose eluant (lane 4). All enzymes were present at a final concentration of 0.1-0.2 units/mL. Following treatment of the modified material with aniline, the resulting fragments were resolved on sequencing gels run at 20 mA constant current. Assignments as for Fig. 1.
similar, but not identical, as judged by kinetic measurements and substrate specificity (Deutscher, 1972b). Accordingly, a column of phosophocellulose was used to separate enzymatic activity into two peaks, a further purification of approximately fivefold over the Affi Gel blue eluant. All four preparations were used to examine whether the purity of the enzyme affected the pattern of excluded bases in damage-selection experiments. Three tRNAs were tested in this manner and the purity of the enzyme did not significantly affect either the identity of bases excluded in modified form or the degree to which they were excluded (see Fig. 2, lanes 5-8).
restoration of [32P]AMP(Table I). It was somewhat puzzling that chemically altered purines in this region of the molecule did not inhibit the enzymatic reaction in all cases. In tRNAs where modified purines near the 3'-end did not interfere with the interaction, binding of the enzyme to cytidines 74 and 75 must have been sufficient to permit catalytic activity. In other instances, additional interactions may have occurred, which varied depending on differences among conformations of the 3'-terminal region. As a specific example, tRNAme' exhibited total exclusion of a modified purine from position 72 (Fig. l(a)); it was only one of two tRNAs in which a modified purine was totally excluded from the 3'-terminal region (Table 1). From analysis of the crystal structure of this tRNA it is clear that the adenine (A) at residue 72 is highly exposed as the sugar/phosphate backbone makes a sharp turn (Woo et al., 1980), and could thereby interact with nucleotidyltransferase in a totally different fashion than the homologous base in tRNAphe,for example, which is in a highly dissimilar environment in the crystal structure (Woo et al., 1980). Another set of previous experiments suggested that the rabbit liver enzyme may interact with a region of the tRNA at some distance from the 3'-end, and that this interaction may be required for stimulation of enzymatic activity (Masiakowski and Deutscher, 1979). Accordingly we note that modified bases located at the corner of the tRNA's tertiary structure, site 57 in all cases and positions 19 or 58 in several, had pronounced effects on the reaction (Table 1). We propose that rabbit liver nucleotidyltransferase interacts with its tRNA substrate by attachment at one or both 3'-terminal cytidines, as previously suggested (Masiakowski and Deutscher, 1980), contacts bases near the 3'-end depending on the particular tRNA, and, in all cases, simultaneously extends to the corner of the tRNA's three-dimensional structure, where the Y-and D-loops are juxtaposed. Interaction with tRNA by the rabbit liver enzyme is therefore very similar to that by the homologous enzyme from yeast, as demonstrated by the similarities between bases excluded in modified form by these two enzymes (Figs 2 and 3, Table I , and Spacciapoli et al., 1989). Recognition by the E. coli enzyme also involves the same regions of the tRNA, but appears to be more stringent in requiring unaltered bases near the corner of the molecule, where the D- and Y-loops join (Fig. 3 and Spacciapoli et al., 1989). Nucleotides at positions 52 and 53, located in the Y-stem near the top of the molecule (Woo rf al., 1980), were also required in chemically unaltered form in several instances (Table I). This agrees with our proposed model in that one would expect minor points of contact along the top of the tRNA as the enzyme extends from the 3'-end across the molecule towards the corner.
DISCUSSION The active site of nucleotidyltransferase from rabbit liver has been characterized extensively and has been shown to have binding sites for ATP, one or two CTPs and the 3'-end of its tRNA substrate (Masiakowski and Deutscher, 1980). In accord with these findings, we observed that carbethoxylated purines near the 3'-end of seven out of twelve tRNAs examined. interfered with
01990 by John Wiley & Sons, Ltd.
Effects of increasing the concentration of enzyme. The observation that extents of exclusion of modified bases were dependent on enzyme concentration (Fig. 4) is significant for two reasons. These data clearly demonstrate that exclusion did not result from the inability of the modified tRNA to refold into a conformation capable of being recognized by the enzyme. That is, at elevated concentrations of enzyme, no carbethoxylated purine had sigJOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990
153
PETER SPACCIAPOLI AND DAVID L. THURLOW I23456
I 2 3 4 5 6
I 2 3 4 5 6
4 19
4 19 4 53 4 57
4 50
4 65
4 53
4 57 4 58
4 73
(a)
(b)
(C)
Figure 4. Titration of the rabbit liver enzyme. tRNAPhe from E. coliwas labeled with [32P]AMP either before (lanes 1, 2 ) or after (lanes 3-6) treatment with DEP using rabbit liver nucleotidyltransferase,the Aff i Gel blue eluant. Control material in lane 1 was not gel purified following modification with DEP, whereas that in lane 2 was. Final concentrations of enzyme were: lane 3,0.54; lane 4, 0.18; lane 5, 0.018; lane 6, 0.007 units/mL. Following treatment of the modified material with aniline, the resulting fragments were resolved on sequencing gels run at 20 mA constant current. Assignments as for Fig. 1.
nificant effects on the labeling reaction (Fig. 4, lane 3), regardless of its location in the tRNA’s tertiary structure. We conclude that the chemically modified bases present in our experiments did not result in gross distortions of the tRNA’s structure following renaturation. However, based on these data alone, we cannot exclude the possibility that the tRNA refolded into a structure slightly different from that of native tRNA and that the rate at which the enzyme utilized the chemically altered substrate was lowered. If these were the case, increasing the concentration of enzyme could compensate for the reduced rate of reaction. Furthermore, the effects of enzyme concentration signify that the sites we have identified as being important in the interaction (Table 1) represent a minimum set of such bases. At lower concentrations of enzyme there may be additional instances where effects would not have been detected in our survey. Using our approach alone, we cannot distinguish effects of modified bases on the affinity between substrate and enzyme from those on the rate of the reaction. It was not possible to judge effects on KM or V,,, by simply measuring initial rates as a function of substrate concentrations, because the population of chemically modified substrate molecules was heterogeneous. For this reason we are currently using site-directed mutagenesis to pre154 JOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990
pare a homogeneous population of tRNAs, containing one altered nucleotide at a position we have identified as being required in chemically unmodified form. The absence of modified bases can also result from purification of chemically modified material on polyacrylamide gels. Instances
in which chemically altered bases were excluded as a result of purification of modified material on polyacrylamide gels were occasionally observed, most notably in tRNAPhe(Figs 2 and 4). Such exclusions occurred reproducibly from positions 69-71, 44, and 27-31 in tRNAPhe.We made similar observations in our previous set of damage-selection experiments (Spacciapoli et af., 1989) and have undertaken a more thorough investigation of this phenomenon (Hegg and Thurlow, 1990). From these studies it is clear that substantial secondary structure persists during purification of some tRNAs on 8 M urea gels run at 45°C and that the electrophoretic mobility of chemically modified tRNAs can vary, depending on which base had been altered. Accordingly, all instances of exclusion we identify as resulting from interference with enzymatic activity were determined relative to control material that had been purified on the same purification gel as the experimental sample.
01990 by John Wiley & Sons, Ltd.
RECOGNITION OF tRNA BY M A M M A L I A N NUCLEOTIDYLTRANSFERASE
Extent of purification of the enzyme did not affect exclusion of modified bases. Given the highly similar nature of modified
bases excluded by nucleotidyltransferases from yeast and rabbit liver, it was not suprising to observe that the two chromatographically distinct forms of activity from rabbit liver were indistinguishable by this criterion (Fig. 2, lanes 7, 8). Previous work has also shown them to be very similar in their physical and kinetic properties (Deutscher, 1972b). It is important to note that even relatively crude preparations of the enzyme exhibited exactly the same behavior as the highly purified forms (Fig. 2). There may have been several proteins in the crude preparation which would, under appropriate circumstances, be capable of competing with nucleotidyltransferase for binding to tRNAs. However, under our set of experimental conditions most of these interactions would have been unlikely to be significant. For instance, because the tRNAs were fully processed, they would not have been approriate substrates for tRNA processing enzymes. In addition, since the 3'-terminal A had been removed and because we did not include amino acids in our incubations, interactions with aminoacyl-tRNA synthetases would have been unlikely. Also, because the tRNAs were
not aminoacylated, they would not be expected to undergo complex formation with elongation or initiation factors.
CONCLUSION The interaction between ATP/CTP:tRNA nucleotidyltransferase from rabbit liver and DEP-modified tRNAs was inhibited when tRNAs contained carbethoxylated purines near the 3'-end of the tRNA or near the corner of the molecule where the D- and Y-loops are juxtaposed. Nucleotide positions 19, 52, 53, 57 and 58 most frequently required chemically unmodified bases for proper recognition. The extent of interference was dependent on the concentration of enzymatic activity, but was not affected by the purity of the enzyme preparation.
Acknowledgements We gratefully acknowledge Dr Murray Deutscher (University of Connecticut Health Center, Farmington. CT, USA) for his help and advice in the purification of nucieotidyitransferase from rabbit liver.
REFERENCES Carre, D. S., Litvak, S., and Chapeville, F. (1970). Purification and properties of Escherichia coli CTP/ATP:tRNA nucleotidyltransferase. Biochim. Biophys. Acta 224, 371-381, Deutscher, M. P. (1972a). Reactions a t the 3'-terminus of transfer ribonucleic acid: purification and physical and chemical properties of rabbit liver transfer ribonucleic acid nucleotidyltransferase. J . Biol. Chem. 247. 450-458. Deutscher, M. P. (1972b) Reactions at the 3'-terminus of transfer ribonucleic acid: catalytic properties of two purified rabbit liver transfer ribonucleic acid nucleotidyltransferases. J. Biol. Chem. 247.459468. Deutscher, M P. (1 982). tRNA nucleotidyltransferase. The Enzymes 15, 183-21 5. Deutscher, M. P., and Masiakowski, P. (1978). Binding of tRNA nucleotidyltransferase to Affi Gal blue: rapid purification of the enzyme and binding studies. Nucleic Acids Res. 5, 1947-1 954. Hegg, L. A., and Thurlow, D. L. (1990). Residual tRNA secondary structure in 'denaturing' 8 M urea/TBE polyacrylamide gels: effects on electrophoretic mobility and dependency on prior chemical modification of the tRNA. Nucleic Acids Res. (in press) Masiakowski, P., and Deutscher, M. P. (1979). Separation of functionally distinct regions of a macromolecular substrate: stimulation of tRNA nucleotidyltransferase by a nonreacting fragment of tRNA. J. Biol. Chem. 254, 2585-2587.
0 1990 by John Wiley & Sons, Ltd.
Masiakowski, P., and Deutscher, M. P. (1980). Dissection of the active site of rabbit liver tRNA nucleotidyltransferase: specificity and properties of subsites for donor nucleoside triphosphates. J. Biol. Chem. 255, 11240-1 1246. Peattie, D. A. (1979). Direct chemical method for sequencing RNA. Proc. Natl Acad. Sci. USA 76, 1760-1 764. Spacciapoli, P., Doviken, L., Mulero, J. J., and Thurlow, D. L. (1 989). Recognition of tRNA by the enzyme ATP/CTP:tRNA nucleotidyltransferase: interference by nucleotides modified with diethyl pyrocarbonate or hydrazine. J. Biol. Chem. 264, 3799-3805. Sprinzl, M., Hartmann, T., Meissner, F., Moll, J., and Vorderwulbecke, T. (1987). Compilation of tRNA sequences and sequences of tRNA genes. Nucleic Acids Res. 1 5 Suppl., r53-rl88. Sternbach, H., von der Haar, F., Schlimme, E., Gaertner, E., and Cramer, F. (1971). Isolation and properties of tRNA nucleotidyltransferase from yeast. Eur. J. Biochem. 22, 166-1 72. Woo, N. H., Roe, 6. A,, and Rich, A. (1980). Three-dimensional structure of Escherichia coli initiator tRNAfMe' Nature 286, 346-351.
Received 1 December 1989; accepted (revised) 30 January 1990.
JOURNAL OF MOLECULAR RECOGNITION, VOL. 3, No. 4,1990
155
