THE ,JOURNAL OF BIO1,OGICAI. CHEMISTRY

Printed [n C’. S. A Vol. 257, No. 9, Iswe of May 10, pp. 4796-4Ho5, 19x2

The Effect of Bisulfite-induced C + U Transitions on Aminoacylation of Escherichia coli Glycine tRNA*

(Received for publication, July 13, 1981, and in revised form, December 28, 1981)

Esther L. SabbanS and Opinder S. Bhanot From the Department of Biochemistry, New York University School of Medicine, New York, New York 10016

The effect of bisulfite-induced C + U transitions on the aminoacylation of highly purified Escherichia coli tRNAt,’& has been studied. On treatment with 3.2 M NaHS03 at 25 “C and pH 5.8, C + U transitions occurred at 4 of the 21 cytosine residues during the first 16 h. After about 12 h, 2 additional unidentified residues begin to react. From these data and theoretical consid- eration, we conclude: 1) E. coli tRNA:’&, exists in the native conformation under the conditions of the reac- tion; 2) its ordered structure is similar to yeast tRNAPhe; 3) its anticodon loop is flexible in solution.

During the reaction, loss of glycine acceptor activity followed first order kinetics with a t lIz = 1.8 h. After modification for 1.25 h and aminoacylation (61% glycine acceptor activity remaining), the aminoacylated frac- tion was isolated. The fractional change at each of the reactive residues in the unfractionated mixture ( f ) and the aminoacylated fraction (p) was measured. The re- sults were: C35, f = 0.33, p = 0.19; C36, f = 0.25, p = 0.10; C74, f = 0.28,~ = 0.26; C75, f = 0.25 ,~ = 0.20. From these data, the Modulation Constant for each reactive residue was calculated from the equation M = (f - p)/ f(l - p): M35 (anticodon) = 0.52; M36 (anticodon) = 0.67; M74 (CCA end) = 0.10; M75 (CCA end) = 0.25. These values, which are based on the assumptions that the C + U reactions occur independently and that the effect of each change on the acceptor activity is an independent event, express the fractional loss in activ- ity that would occur from a C + U change at the residue in question by itself. From these results, we conclude: 1) -80% of the observed inactivation was due to changes in the anticodon; 2) neither of these anticodon residues (C35 and C36) is essential for aminoacylation; 3) a C -+ U change at C75 (CCA end) has a small effect on aminoacylation; 4) a C + U change at C74 (CCA end) has little or no effect on aminoacylation.

Several studies have explored the effects of bisulfite-in- duced C + U changes in the anticodon on the amino acid acceptor activity of tRNA. Such a change in the first (wobble) position in Escherichia coli tRNA? (1, 2), the second posi- tion of E. coli tRNApg (3) or the third position of yeast tRNA$‘ (4) destroys a l l , or almost all, of the acceptor activity.’

* This investigation was supported by Grant 5R01 GM07262-18 to Dr. H. W. Chambers from the National Institute for General Medical Science, Department of Health, Education and Welfare. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduer- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Present address, Neurochemistry Laboratories, Department of Psychiatry and Cell Biology, New York University School of Medi- cine, New York, NY 10016.

’ Throughout this paper, we will read the anticodon in a 5‘ + 9’

In E. coli tRNATrp, a C + U change in the second position of the anticodon changes the aminoacylation specificity from Trp to Gln (5).

Clearly, the anticodon plays an important role in aminoac- ylation of certain tRNAs (7, 8). E. coli tRNAG’” is a particu- larly interesting case because a series of missense suppressor tRNAs having changes in the anticodon, modification adja- cent to the anticodon, and markedly reduced rates of amino- acylation have been reported (9, 10). For example, the mis- sense suppressor tRNA that suppresses the trpA36 mutation in E. coli is produced from the tRNA;::, gene (glyT) by a C -+ U change at the third position of the anticodon and a “carbamoylthreonine-like” modification on the adenine resi- due adjacent to the anticodon (11, 12). Its rate of aminoacyl- ation is reduced to 10-4-fold compared to the wild type tRNA (13). A similar effect on aminoacylation has been reported for the trpA78 missense suppressor, tRNA:& which is produced from tRNA$& gene (glyW) by a C - A transversion at the 3’-end of the anticodon and a 2-methylthioisopentenyl modi- fication of the adjacent adenine residue (14). In contrast, the suppressor, tRNA:::,, which is derived from the tRNA:‘;‘?., gene (glyU) by a C + U change at the second position of the anticodon, has no other modification; and its rate of amino- acylation is only slightly reduced compared to normal tRNA (15). Since all three of these suppressor tRNAs, as well as the parent isoacceptors are aminoacylated by the same glycyl- tRNA synthetase (lo), some interesting questions arise. Does the third position of the anticodon play an essential role in aminoacylation? What role does the modification of the ade- nine residue adjacent to the anticodon play in the loss of acceptor activity? It is difficult to answer these questions by studying the suppressor tRNAs themselves because in those cases where a very large decrease in activity has been ob- served, each molecule has two changes; one in the anticodon and one adjacent to it. Target Distribution Analysis of bisul- fite-induced C + U changes in tRNAGIy provides an approach for examining these questions since the extra modification that complicates interpretation of the data from the missense suppressor tRNAs will not be present.

This paper deals with E. coli tRNA$&. We chose this for our initial studies with tRNAG’” because it is the major isoac- ceptor, so it is relatively easy to isolate homogeneous material. In addition, the anticodon, G-C-C- contains a RNase TI cleav- age site that we thought would simplify quantitative analysis of the changes in the anticodon. We will present evidence that a C --* U change at either the second or the third position of the anticodon (GCC + GUC or GCC + GCU) reduces the rate of aminoacylation compared to unmodified tRNA::,,

direction. Thus, the first position corresponds to the “wobble posi- tion” and to the third position of the corresponding codon, also read 5’ -+ 3’. We will use the abbreviation suggested earlier for isoacceptor tRNAs (5). Thus, tRNA:g,, is the third isoacceptor and it carries the anticodon, G-C-C-.

4796

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Specific C + U Transitions in Glycine tRNA 4797

without inactivating the molecule. The tRNA containing either of these changes, by itself, can still be aminoacylated to about 40% the steady state level of starting material under standard assay conditions. Taking all the data together, it appears that none of the anticodon residues in E. coli

are essential for aminoacylation, but that the cyto- sine residues in the second and third position of the anticodon play some role in the reaction. It also appears that the modi- fication of the adenine residue adjacent to the anticodon in the trpA36 and trpA78 suppressor tRNAs decrease the glycine acceptor activity markedly. We interpret this to mean that the glycyl-tRNA synthetase binds to the cytosine residues in the anticodon, and modification of the residue adjacent to the anticodon with bulky groups that are “abnormal” for tRNA”” interferes with this binding.

MATERIALS AND METHODS

Most of the materials and methods used have been described previously (16-18). Additional information may be found in the min- iprint supplement to this paper.’

RESULTS

The Experimental Approach-From theoretical and prac- tical considerations, we envisage the following ideal protocol for studying the effect of bisulfite-catalyzed C -+ U changes on aminoacylation of tRNA: 1) purify the tRNA to homoge- neity; 2) measure the reaction kinetics for HS03--induced C + U changes; 3 ) establish standard conditions for aminoac- ylation of unmodified, homogeneous starting material; 4) measure the loss of amino acid acceptor activity as a function of C + U changes; 5) select a reaction time that will give a preselected loss of activity; 6) modify the tRNA on a prepar- ative scale under the selected conditions; 7 ) identify the re- active residues and measure the C + U changes quantita- tively; 8) separate the tRNA into an R-fraction and a P- fraction”; 9) analyze the C + U changes in the P-fraction; 10) calculate the fractional loss in activity (one-hit modulation constant) at each reactive residue; 11) repeat Steps 5 to 10 at different reaction times to see if the modulation constants change, indicating that cooperative effects are present. We have carried out Steps 1 through 10 in the experiments reported here.

Purification of Starting Material-Preparation of homo- geneous E . coli tRNA& free of denatured material was achieved in five steps as described under “Materials and Methods.” These procedures are routine and only the deac- ylation step requires further comment.

The usual conditions for removal of the O-phenoxyacetyl- glycyl group (0.5 M Tris-HC1, pH 9.0 at 37 “C for 30 min) caused a 30% loss of glycine acceptor activity. Part of this could be restored by increasing the concentration of enzyme in the assay. These results suggested that a conformation change was occurring, giving rise to material that aminoacy- lates more slowly than native tRNA?IY and altering the steady

Portions of this paper (including “Materials and Methods,” part of “Results,” Tables I and 11, and Figs. 1-3, separate Footnotes 1 and 2, and additional references) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magni- fying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 RockviUe Pike, Bethesda, MD 20814. Request Document No. 81M-1877, cite authors, and include a check or money order for $7.20 per set of photocopies. Full size photocopies are also included in the microfdm edition of the Journal that is available from Waverly Press.

R-fraction stands for “reactant fraction.” This contains the tRNA molecules that do not aminoacylate under standard conditions. P- fraction stands for “product fraction.” This fraction contains the

state level of aminoacylation (19). Therefore, a series of ex- periments were conducted to find conditions for removing the 0-phenoxyacetylglycyl group without loss of activity. The pseudo-first order rate constants calculated from these exper- iments and the loss of amino acid acceptor activity measured under standard conditions are summarized in Table 111. These results are important since the conditions that have been used by others to remove the 0-phenoxyacetylglycyl residue causes significant loss of acceptor activity and can confuse the inter- pretation of experiments dealing with this biological activity. From these experiments, we selected 0.5 M Tris-HC1, pH 7.0, 37 “C for 7 h for removal of the 0-phenoxyacetylglycyl group. Chromatography of the free tRNA on BD-cellulose gave essentially pure, native tRNA$lY.

The purified tRNA accepted 1840 pmol of glycine/AZm unit under standard assay conditions. This agrees well with 1846 pmol of adenosinelA2m unit released fnm the 3‘-end following RNase A treatment of tRNA$lY. No other isoacceptors were detectable in this purified tRNA. The base composition (data not shown) was in good agreement with that reported previ- ously (20). The tRNA was digested with RNase TI and each oligonucleotide from the digest was isolated and analyzed for its nucleoside composition. The results were in complete agreement with the reported sequence (20); no extra peaks were found. Thus, this sample of E . coli tRNA?” appears to be homogeneous and free of denatured material.

Kinetics for C+ U Transitions-The kinetics of the bisul- fite-induced C + U transisitions in tRNAPY was studied in a mixture of 25 ,UM tRNA, 2 mM MgS04, and 3.2 M NaHSO:], a t pH 5.8 and 25 “C, exactly as described previously (16). The results shown in Fig. 4 indicate that 6 out of the 21 cytidine residues in the tRNA react under these conditions. However, there is a break in the curve indicating that 4 of the 6 residues react rapidly. Two other residues may either be reacting slowly or they may become reative a t later times due to unfolding the tertiary structure of the tRNA. Using data reported in Table I1 to obtain the pseudo-fist order rate constant for the 4 most reactive residues, a difference curve can be constructed as shown in Fig. 4. This suggests that the second of the two possibilities considered above seems to be correct.

Standard Conditions for Aminoacylation-E. coli tRNA$IY shows an enzyme-dependent steady state level of aminoacylation similar to those described for yeast tRNA:”’ (19). This is shown in Fig. 5. At low [Elo, aminoacylation of

TABLE 111 Deacylation kinetics and the effect of p H on glycine acceptor

activity

PH Incubationh

k” Activity remaining Temperature Time

min” “ C h B

9.0 0.115 25 0.5 70 8.5 0.072 25 1.0 73 8.0 0.037 25 2.0 72 7.5 0.018 25 3.0 86 7.0 0.009 37 7.0 100

The pseudo-fist order rate constants ( k ) were calculated from the kinetics of deacylation which were found to be linear under the conditions used (see below). ’ Phenoxyacetyl [‘4C]Gly-tRNA?‘Y was incubated in 0.5 M Tris-HC1 at the specified pH for the time necessary for complete deacylation. After incubation, the pH was brought to 7.0 with HCl and the mixture was dialyzed against 1 m~ MgS04. The tRNA was assayed as de- scribed under “Materials and Methods” except that the reaction mixture contained 10 mM KC1, 0.4 mg/ml of bovine serum albumin, and the r’4Clglycine concentration was 0.1 mM. Partially purified Gly-

molecules that are aminoacylated under standard conditions. tRNA synthetase (0.7 milliunit/ml) was used. ”

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

4798 Specific C += U Transitions in Glycine tRNA

Reaction t ime ( h )

FIG. 4. Kinetics for C + U transitions in E. coli tRNAYLY. The reaction conditions and the analytical procedures are described under “Materials and Methods.” M, the experimental curve for C + U transitions induced by 3.2 M NaHS03 at pH 5.8 and 25 “C; - - -, the theoretical curve for 4 residues obeying the rate law for independ- ent reactions occurring in parallel and based on fmst order rate constant evaluated from the data shown in Table 11; -.-, the difference curve calculated from the experimental and theoretical curves.

200 10 30 [E], ( rnUn l t s / rn l I

50

I I I

10 30 50 Incubat ion time (min)

FIG. 5. Determination of “standard conditions” for steady state aminoacylation of E. coli tRNA$’Y. The assay was carried out as described under “Materials and Methods” except that various enzyme concentrations were used. 0, 2; 0, 4; A, 10; A, 20; and 0, 50 milliunits/ml. At the times indicated, 25-pl aliquots were removed and [14C]Gly-tRNA estimated by acid precipitation (5). The data plotted in the inset were obtained from 30-min incubation points. The arrow represents the enzyme concentrations selected (20 milliunits/ ml) for [tRNA10 = 1.5 X M.

tRNA$’Y reaches a steady state plateau. Increasing the incu- bation time does not effect this level. This has been explained for tRNA?’ (19) by a steady state between ATP-dependent enzymatic aminoacylation, and deacylation (both enzymatic and nonenzymatic). It is important to find a value for the initial enzyme concentration, [Elo, that is just sufficient to aminoacylate modified tRNA under specific assay conditions and to use these conditions both for measuring the inactiva- tion kinetics and for separating the R-fraction and P-fraction tRNA. Thus, we selected conditions that would give 98% aminoacylation of the unmodified tRNA in 40 min. At [tRNA10 = 1.5 X M, 20 milliunits/d (5 X lo-’ M) of highly purified synthetase were required (arrow in the inset of Fig. 5). Unless otherwise stated, these conditions were used as “standard” conditions for all of the aminoacylation reac- tions described in this paper, including that used for separa- tion of R- and P-fraction.

Modulation Kinetics-Starting material was treated with 3.2 M NaHS03 at pH 5.8 and 25 “C for various times and then worked up as described previously (16,18). An aliquot of each time sample was assayed for glycine acceptor activity under standard conditions described above. The results are shown in Fig. 6. The data fit pseudo-fist order kinetics with a half-

time of 1.8 h. This inactivation rate is somewhat faster than has been observed with other tRNAs (1, 3, 4, 6, 16). At half- inactivation, an average of 1.4 C + U transitions have oc- curred.

There are several ways by which C + U transitions can reduce the ability of a tRNA to be aminoacylated. Since the steady state level of aminoacylation of any tRNA is resultant of the rate of aminoacylation and the competing rate of hydrolysis (both enzymatic and nonenzymatic) of the amino- acyl-tRNA, it follows that when a C -+ U change reduces the rate of aminoacylation significantly without affecting the rate of hydrolysis, a new steady state level will be established. If the same conditions are used for measuring this activity with the modified tRNA as are used for the unmodified material, then the new steady state wiU be less than the original value. If the molecules that contain this change are not really inac- tive, the steady state level of aminoacylation should increase when [El0 is increased while keeping [tRNA10 constant. We have done such an experiment. The tRNA?Y was modified with NaHS03 for 2.5 h. The steady state level of aminoacyl- ation of modified tRNA was 40% of that obtained with the unmodified starting material under standard assay conditions. When [El0 was increased, the steady state level increased reaching 60% of the theoretical value when [E30 has been increased 40-fold over its standard concentration. For techni- cal reasons, we were unable to increase the enzyme concen-

IO0 90 80 70 60

50

40

- 30 Y 8

2 20 Q)

c 0

0 0 0 0

0,

0 h

n

.- = IO - (3

Reaction time (hrs) FIG. 6. Modulation kinetics. E . coli tRNAj;lY was treated with

3.2 M NaHSOR as described under “Materials and Methods” and C + U-dependent loss of glycine acceptor activity was determined under standard conditions of aminoacylation. The results are nor- malized to a sham reaction substituting NaCl for NaHS03.

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Specific C -+ U Transitions in Glycine tRNA 4799

TABLE IV Target Dis tr ibut ion Analys is of the C + U changes i n E . c o l i CK:dA;ly produced by a 1 .25 h r e a c t i o n w i t h 3 . 2 M NaHS03 a t pH 5.8. 25’.

Fraction changed at residue

c - 35 C - 36 c - 74 c - 75

f a

b P

MC

0 . 3 3 0 . 2 5 0 . 2 8 0 . 2 5

0.19 0.10 0 . 2 6 0.20

0.52 0.67 0.10 0 . 2 5

tration further, but there was no indication of “leveling off’ (data not shown). At this stage, we are uncertain if more of the inactivation could be reversed by the use of increasingly higher enzyme levels.

Separation of Bisulfite-modified tRNA into R- and P- fractions-A reaction time of 1.25 h (Fig. 6) was selected so that sufficient amount of P-fraction would be available for quantitative analysis of C + U changes. After modification of 74 nmol of tRNA as described under “Materials and Meth- ods,’’ the mixture was aminoacylated under standard condi- tions. There was 39% loss of activity relative to a sham reaction using NaCl instead of NaHS03. Our initial attempts to isolate P-fraction via its phenoxy-

acetylglycyl derivative were unsatisfactory. The derivatized tRNA isolated from the BD-cellulose column invariably had a low specific activity (1600 pmol/Azm unit as measured directly from radioactivity and absorbance), presumably be- cause of contamination with R-fraction. Since it is essential for experiments reported here that the specific activity of the P-fraction be >95% of theoretical, we modified our usual procedure by using the napthoxyacetyl derivative instead of phenoxyacetyl derivative (21). This modification was satisfac- tory. The P-fraction eluting with 19% ethanol and 1 M NaCl (Fig. 2) had a specific activity of 1810 pmol/A2m (measured directly from radioactivity and absorbance) as compared to 1840 pmol/AzW for unmodified starting material.

P-fraction was hydrolyzed with 0.5 M Tris-HC1, pH 7.0 for 7 h at 37 “C. Under these conditions, the denaturation of

is minimum. The free tRNA was isolated by chro- matography on BD-cellulose. A single peak (data not shown) was obtained. This material was used for quantitative analysis of C + U changes present in the P-fraction (see Miniprint). Since analysis of R-fraction is not mandatory for calculation of modulation constants, it was not investigated further.

Calculations of the One-Hit Modulation Constants-Since partial reaction at four different sites have occurred under conditions used for modification of tRNAflY, Z4 = 16 different molecular species are present in the reaction mixtures. In order to investigate the biological effect of the individual changes and to determine their effect on one another, we have used an approach called Target Distribution Analysis.

aFraction C -+ U change a t re s idue ind ica ted in the unfrac t ionated reac t ion mix ture .

bFraction C -+ U change a t r e s i d u e i n d i c a t e d i n t h e P - f r a c t i o n .

CModulation constant = ( f - p ) / [ f ( l - p)].

Target Distribution Analysis is based on a simple idea. Suppose exactly 2 residues in a given molecule react inde- pendently under a given set of conditions. Suppose one of these changes inactivates the molecule and the other change has no effect. One can determine which change inactivates the molecule by separating active and inactive molecules after partial modification (22). A change that causes inactivation will never be found in the active fraction. A residue having this effect is defined as an inactivation target. A change that has no effect on the activity will be found in both the active and inactive fractions, and the extent of this change will be the same in both fractions. A residue with this property is called a no effect target. The terms “active” and “inactive” imply all-or-none activity effects.

During our detailed study of the effects bisulfite-induced C + U transitions have on aminoacylation of yeast tRNA::’ (7, 15, 23)4 and yeast tRNAik (10, 24); we observed that the loss of activity could not be explained by all-or-none effects. For example, only one inactivation target was found in tRNA”“‘ and this was insufficient to explain the total observed inactivation (7, 23). The results with yeast tRNATv (17) and tRNAit (24) were even more striking. No inactivation target was found, yet a C + U-dependent loss of activity was observed.

A priori, a change in the primary structure of a macromol- ecule may produce an activity effect that falls on a continuum ranging from no effect to complete inactivation. Therefore, it seems more appropriate to use the term “modulation” rather than “inactivation” to describe the activity change. Ideally, we would like to know exactly where the modulation produced by each individual modification falls on the activity contin- uum. In order to calculate the modulation constant, we have used the following equation:

M = - f - P f(1 - P) (1)

Where M = modulation constant. This is the fractional loss in

0. S. Bhanot, H. Zawadzka, Y. Furukawa, S. Aoyagi, R. W. Chambers (1981) J. Mol. Biol., manuscript submitted for publication. ’ 0. S. Bhsnot and R. W. Chambers (1981) J. Mol. Biol., manuscript submitted for publication.

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

4800 Specific C + U Transitions in Glycine tRNA

activity that occurs from a C + U change at the reactive residue. f = fraction C "* U change at the reactive residue in the unfractionated reaction mixture;^, fraction C + U change at the reactive residue in the P-fraction.

This general equation is valid for each reactive residue provided there are no cooperative effects operating between the targets. The derivation of the basic equation and a detailed discussion of Target Distribution Analysis is described else- where.6

The calculated M values are given in Table IV. The data show that the reactive residues at the second and third posi- tion of the anticodon (C35 and C36) do not fall into either of the two extreme categories of a no effect (M = 0) or an inactivation target (M = 1.0). The modulation constant (C35 = 0.52; C36 = 0.67) indicate that the molecules modified a t one of the residues alone can be aminoacylated to about one- half (for C35) and one-third (for C36) of the steady state level.

Due to experimental errors in measuring f and p , we should emphasize that our conclusions regarding the effect of C35 and C36 are semiquantitative. A 5% error in the measurement of f and p can yield M values for C35 ranging from 0.48 to 0.65. The important conclusion for residues C35 and C36 is that both of the cytosine residues play a large role in amino- acylation, however, they are not absolutely essential.

On the other hand, C74 and C75 at the CCA end have an inactivation modulus of 0.10 and 0.25, respectively. Thus, they have little or no effect on the aminoacylation reaction.

The calculations of the modulation constants assume that neither cooperative reactivity nor cooperative modulation oc- cur in this particular modification experiment. Therefore, these assumptions are examined further in the discussion.

DISCUSSION

Ordered Structure of E. coli tRNAfb-Different tRNAs have been divided into groups based on their secondary (clov- erleaf) structure (25). It has been suggested that at least within these groups, tRNAs have similar tertiary structures (26-28). E. coli tRNA?'Y belongs to the same group (D4V5) as yeast tRNAPhe. Therefore, when we fitted the tRNA$'Y se- quence to the tRNAPhe stacking pattern (26-29), in order to predict which cytosine residues would react, we found that residues C48 and C56 are tightly stacked and form hydrogen bonds. As expected, no reaction was detected at these residues under the conditions of this experiment. The predicted and the experimental results are in agreement. The data indicate that the stacking lifetimes of residues 36,74, and 75 are similar and somewhat longer than residue 35. We can speculate that residues 48 (variable loop) and 56 (TqC loop) represent the additional residues that react after the tRNA partially unfolds at long reaction times, but this has not been investigated experimentally.

The anticodon loop of a tRNA can exist in two extreme conformations (30); conformation A, where 5 residues of the anticodon loop are stacked on the 3'-end and conformation B, where 5 residues of the loop are stacked on the 5'-end. Crys- talline yeast tRNAPh" is present in conformation A (29, 31). Recent data obtained by digestion of this tRNA with single- stranded nuclease S1 have been interpreted to mean that conformation A also exists in solution (32). However, bisulfite- catalyzed C + U changes probably cannot occur in either conformation A or B (16). Thus, other conformations must be possible in solution under the conditions used for the modifi- cation reaction. A similar conclusion has been reached for

R. W. Chambers, L. Schulman, G. B. Weiss, W. B. Gatland, K. A. Freude, 0. S. Bhanot, I. Kucan, Z. Kucan, S. Aoyagi, Y. Furukawa, H. Zawadzka, E. Sabban, and L. Morin (1981) J. Mol. Bioi., manu- script submitted for publication.

both yeast tRNAx? (18) and yeasttRNAek (16). Our modifi- cation data indicates that the anticodon loop oftRNAFiY is flexible in solution under the conditions employed for these studies. A similar conclusion was reached earlier from physi- cochemical measurements (33).

Eight tRNAs have been investigated by the bisulfite modi- fication. With one exception (34, 35), the results are in agree- ment with the suggestion that all tRNAs have similar tertiary structures (26) . These tRNAs appear to have the same struc- ture in 3.2 M NaHS03 at pH 5.8 to 6.0 as yeast tRNAPh' does in the crystals examined by x-ray diffraction. The exception of E. coli tRNAG'" may be due to the fact that it belongs to a different class (D3V4) than any other tRNAs that have been studied so far with the bisulfite method.

Target Distribution Analysis: Examination of the Basic Assumptions-The modulation constants for the reactive res- idues, using Equation 1 were 0.52,0.67,0.10, and 0.25, respec- tively, for residues 35,36, 74, and 75 (Table IV). This equation has two basic assumptions, noncooperative reactivity and noncooperative activity effects. The first assumption is rea- sonable, based on the location of the reactive residues and the chemistry involved. While we have not directly tested this assumption for tRNA$'Y, in the cases that have been examined, this has held true. First, careful kinetic studies on bisulfite- reduced C + U transition with E. coli tRNAPh' show that each reactive C residue follows pseudo-first order kinetics (36). This, by itself, is not compelling since one can argue that the 2nd residue reacts much faster than the fist, once it is exposed, giving the impression that they both react at the same first order rate. This is clearly not the case for tRNAPh' since the t1,2 values are significantly different for each reactive residue. Second, we have shown previously in the case of tRNA;/,"h' (18) and tRNAf% (16) that the distribution of oligo- nucleotide with 0, 1 and 2 C -+ U transitions is as expected for independent reactivity for the residues in single-stranded re- gions.

The second assumption of Equation 1 is noncooperative activity effects. A cooperative activity effect would be the situation, for example, in which a C + U change at positions 35 and 36 alone would not effect the activity, yet all molecules with U at both positions would be inactive. This could be envisioned if the tRNA synthetase required the presence of at least one C in the anticodon. We do not wish to imply that cooperative activity effects are not possible in tRNA?',", but our subsequent analysis will attempt to show that under the conditions and reaction time used in this work, the majority of the effects, at least in the anticodon, are due to individual effects a t residues 35 and 36. There are several arguments to support this conclusion.

The mot direct method would have been to isolate each type of modified tRNA and directly determine its activity. However, 24 = 16 products are possible. The distribution of these products, at 1.25 h reaction can be calculated from the data in Table IV and is shown in Table V. Several things are worth noting: first of all, the mixture is complex and the percentage of many of the molecular species is quite small. No one, so far as we know, has attempted to isolate individual species from such a mixture. Thus, we resorted to more indirect analysis of the results to evaluate the likelihood of noncooperative activity effects.

First, the inactivation kinetics (Fig. 6) fit a first order curve. While this is not compelling evidence for independent effects, it is consistent with this. Second, Target Distribution Analysis itself provides test data. First, we assume all inactivation events are independent and calculate the modulation constant for each reactive residue (Table IV) using Equation 1. Using these data, we can calculate the molecular distribution for the

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Specific C + U Transitions in Glycine t R N A 4801

TABLE V The calculated molecular distribution for E . coli tRNA?" having

C + U No. of

changes

the changes shown in Table IV Distribution in the Distribution in the

Molecular species unfractionated mixture" P-fraction*

C35C36C74C75 27.1 27.1

U35C86C74C73 13.4 Ca5UasC74C7,

C35C36C74U75 9.0

UJ5U96C74C75 4.5 0.7 U35C36U74C75 5.2 2.2 u:l5c36c74u,5

C:15u:l6u74c73 1.0 c,35u:16ci4u75 3.0 0.7 c35c:16u74U75 3.5. 2.4.

UaiU:I6U74Cii,

U:35C.I6U74U75 1.7 Cd5U:16U74U75 1.2 0.3

u35U:l6u;4u75 0.5 0.01

c:1sc36u74Ci5 l:::j 42.0" ;;I 3'0

25.6'

U35US6C74U75 ::;I 6,1' q 1.3'

3":;' 8.6' 24.2' "',

z = 99.9 X = 62.6

Calculated activityd = 62.2% Measured activity = 61.0%

Calculated using data from Table IV and equations of the type f a ( l - fb)(l - fc) . . ., where fa is the fractional change at a reactive residue present modified in the molecule and fb, L., . . ., are the fractional changes at reactive residues b, c, . . ., present unmodified in that molecular species.

Obtained by multiplying the values in previous column by (1 - M), where M is the modulation constant for the modified residue present in that molecular species. The M values for various reactive residues in tRNA?'' are given in Table IV. ' Sum of the molecules having the same number of C + U changes. "Sum of the P-fraction molecules.

active fraction. This is shown in the last column of Table V. Summing the amounts of each species in this calculated distribution gives 63% active molecules in the unfractionated mixture. This is very close to the experimental value of 61% obtained by direct assay of the mixture. This, too, is consistent with independent activity effects by the altered residues and suggests that if cooperative activity effects exist, they do not make a very large contribution to the loss of acceptor activity observed in this experiment. This conclusion is strengthened by further analysis of the data in Table V. Only 8.2% of the molecules are changed at both C35 and C36. The observed loss of activity in this experiment was 39%. Thus, two-hit cooperative inactivation involving these anticodon residues cannot account for the loss in activity. Even if we take the sum of all molecules hit more than once, we can only account for 75% of the observed inactivation. Thus, it is clear that one- hit molecules are making a significant contribution to the inactivation.

While cooperative activity effects cannot be ruled out, there is no evidence that they are present in this experiment. Therefore, the assumptions inherent in the equation used to calculate the modulation constants appear to be valid for these experiments.

Evaluation of the Target Distribution Analysis-The M- values shown in Table IV indicate that changes in the anti- codon residues, C35 and C36, have rather similar effects on aminoacylation and these effects are much larger (80% of the total loss of activity) than those caused by changes at the CCA end. Neither of the anticodon changes, by itself, com-

pletely inactivates the acceptor activity. Thus, C35, C36, C75, and possibly C74 are modulation targets.

The Role of the Anticodon in Aminoacylation of E . coli tRNA "lY-A major objective of this work was to collect further data on the effect that base changes in the anticodon of tRNAGly have on aminoacylation. The results of Target Dis- tribution Analysis on C -+ U changes in tRNA:$.,. combined with data from the elegant studies of Carbon and Hill and their collaborators on suppressor tRNAs derived from tRNA"IY lead to some interesting conclusions.

Three genes (glyU, glyT, and glyV) have been identified and mapped on the E . coli chromosome (10). Each of these genes codes for a different isoacceptor (10). A fourth gene, glyW, maps at a different locus from glyV, but gives the same isoacceptor (14). All three isoacceptors are aminoacylated by the same enzyme. Two forms of the enzyme are known, one with normal activity and the other derived from a mutant that aminoacylates tRNAG'? at a faster rate than the wild type protein. Specific mutations in these tRNA genes produce altered tRNAs (9, 10). Several of these tRNA have been isolated and sequenced. One of the mutations (called ins) occurs in the glyV gene and produces a G -+ U transversion in the "wobble" position of the anticodon in tRNA$$., (37). No other changes in this mutant tRNA was detected. It aminoacylates completely under the same conditions as the wild type isoacceptor from which it is derived. On the basis of this result, one can conclude that the Fist ("wobble") position of the anticodon plays no detectable role in the aminoacyla- tion.

A second missense suppressor tRNA is derived from a mutation in the glyU gene and contains a C -+ U transition at the second position of the anticodon in tRNA:%, (15). The ability of this supT mutant tRNA to be aminoacylated is somewhat reduced but not as drastically as the suppressor tRNAs involving changes at the third position (15). Although detailed kinetic studies have not been reported, this result is in agreement with our finding that similar change in tRNA$& to tRNAgg, reduces the steady state level of ami- noacylation, suggesting that this residue plays a role in ami- noacylation, but is not essential for the reaction. We empha- size that these results involve two different isoacceptors of tRNA. However, since they are both aminoacylated com- pletely by a single glycyl-tRNA synthetase, we believe that it is valid to compare these results. Another glycine suppressor (trpA58), having the change in codon specificity (Gly -+ Asp) expected for a C -+ U change in the second position of anticodon of tRNA?" has been described (10). However, the mutant tRNA has not yet been isolated and characterized (9).

Results with suppressor tRNAs carrying a change in the third position of the anticodon are less clear. Two such mis- sense tRNAs have been isolated and sequenced. One of these is derived from a mutation that produces a C -+ A transversion at the third positions of the anticodon (read 5' to 3') in tRNA;$, (14). This sequence change is apparently recognized in vivo by the tRNA modification system, and a 2-methyl- thioisopentenyl group is attached to the adenine residue ad- jacent to the anticodon (14). This trpA78 suppressor tRNA is inactive when assayed under the same conditions that com- pletely aminoacylate the wild type isoacceptor. It requires a 500-fold excess of glycyl-tRNA synthetase to be aminoacyl- ated. Both K , and V,,, are effected (14). Since this suppressor tRNA contains two changes it is not possible to decide which is responsible for the observed loss of acceptor activity or whether both changes play a role.

Still another missense suppressor tRNA is derived from a mutation in the glyT gene, resulting in a C -+ U transition in the third position of the anticodon. This sequence change is

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

4802 Specific C + U Transitions in Glycine tRNA

also recognized by the modification system and a “carbamoylthreonine-like” residue is attached to the adenine residue that is adjacent to the anticodon (11, 12). This mis- sense mutant tRNA, tRNA;::,, suppresses the trpA36 locus (13). Like the trpA78 suppressor, tRNA:$,, it is not amino- acylated under standard conditions although glycine can be attached by using very high concentrations of activating en- zyme. The rate of aminoacylation is that of the normal tRNA?”. Once again, this tRNA contains two changes, one in the anticodon and one adjacent to it, so the role of the anticodon in the aminoacylation reaction cannot be evaluated from these data.

Our Target Distribution Analysis data, although derived from a different isoacceptor, shows very clearly that both the second and third positions of the anticodon play a role in aminoacylation, though neither residue is essential. The mod- ified containing a C + U change at the third position of the anticodon can still be aminoacylated to about one-third the wild type steady state level under carefully standardized conditions. Since all of these assays were done with the same enzyme, the results indicate that the C + U change in the trpA36 suppressor, tRNA;:,, and the C + A change in the trpA78 suppressor, tRNA$& probably are not responsible for the entire loss of activity that was observed. The modifications that occur on the adenine residue adjacent to the anticodon seem to play a major role in this loss of activity. It is interesting that neither of these modifications is normal for tRNAGly. These bulky groups may well interfere with the “normal” binding of the glycyl-tRNA synthetase to the suppressor tRNAs. These modifications are normal for a tRNA that inserts Cys, as the trpA78 suppressor must do, the Arg, as the trpA36 suppressor must do (10, 38). It is quite possible that these modifications play an important role in the ribosome- binding process that is involved in suppression (39). It is also interesting that these suppressor tRNAs show such low accep- tor activity in vitro while they seem to suppress rather effi- ciently in vivo and suppression requires esterification to gly- cine of the suppressor tRNA.

Roberts and Carbon (11) have argued that the base change at the third position of the anticodon rather than the modifi- cation adjacent to it is responsible for the dramatic loss of acceptor activity. A critical part of their argument depends upon the interpretation of chemical modification studies (ni- trous acid) carried out on partially purified tRNA (40). It is virtually impossible to deduce the effect produced by a single change in a multitarget molecule without a detailed Target Distribution Analysis. While one cannot strictly compare C + A and a C + U change in the same isoacceptor, tRNA?&, or the same base change (C --$ U) in different isoacceptors (tRNA:$, and tRNABtY,,), the Target Distribu- tion Analysis results suggest that the extra modification car- ried by the missense suppressor tRNAs, play an important role in the loss of acceptor activity that is observed.

The comparison of the results from these two different approaches (isolation of mutants and chemical modification) ilustrates the value of Target Distribution Analysis in resolv- ing an ambiguity that arose with some of the tRNAGly mis- sense suppressors. Taken together, the two approaches pro- vide information that neither can provide by itself in this particular case.

It is now clear that there is no simple set of rules governing the specificity of aminoacylation. As sequence changes have been studied in more tRNAs, either by isolation of mutants or by Target Distribution Analysis, some areas that are critical for a group of tRNAs have been identified. In several cases, specific residues that are essential for aminoacylation have

been identified (1-4,6,41-43). As additional tRNAs are stud- ied, other critical areas and essential residues may be revealed. We believe this information is important in trying to under- stand the chemical basis for the specificity of these important enzymatic reactions.

Acknowledgment-We thank Dr. Robert W. Chambers for his generous support and stimulating discussions throughout this work.

REFERENCES

1. Goddard, J. P., and Schulman, L. H. (1972) J. Biol. Chem. 247,

2. Schulman, L. H. (1979) in Transfer RNA: Structure, Properties and Recognition (Schimmel, P. R., SOU, D., and Abelson, J. R., eds) pp. 312-324, Cold Spring Harbor, NY

3864-3867

3. Chakraburtty, K. (1975) Nucleic Acids Res. 2, 1793-1804 4. Chambers, R. W., Aoyagi, S., Furukawa, Y., Zawadzka, H., and

Bhanot, 0. S. (1973) J. Biol. Chem. 248, 5549-5551 5. Reeves, R. H., Imura, N., Schwam, H., Weiss, G. B., Schulman, L.

H., and Chambers, R. W. (1968) Proc. Natl. Acad. Sci. U. S. A. 60, 1450-1457

6. Seno, T. (1975) FEBS Lett. 51, 325-329 7. Schimmel, P. R., and SOU, D. (1979) Annu. Rea Biochem. 48,

601-648 8. Schimmel, P. R. (1979) in Transfer RNA: Structure, Properties

and Recognition (Schimmel, P. R., Soll, D., and Abelson, J. R., eds) pp. 297-310, Cold Spring Harbor, NY

9. Hill, C. W. (1975) Cell 6,419-427 10. Carbon, J., and Squires, C. (1971) Cancer Res. 31,663-666 11. Roberts, J. W., and Carbon, J. (1975) Nature 250,412-414 12. Roberts, J. W., and Carbon, J. (1975) J. Biol. Chem. 250, 5530-

13. Carbon, J., Squires, C., and Hill, C. W. (1969) Cold Spring Harbor

14. Carbon, J., and Fleck, E. W. (1974) J. Mol. Biol. 85, 371-391 15. Hill, C. W., Combriato, G., and Dolph, W. (1974) J. Bacteriol.

16. Bhanot, 0. S., and Chambers, R. W. (1977) J. Biol. Chem. 252,

17. Aoyagi, S., Bhanot, 0. S., and Chambers, R. W. (1977) J. Biol.

18. Bhanot, 0. S., Aoyagi, S., and Chambers, R. W. (1977) J. Biol.

19. Bonnet, J., and Ebel, J. P. (1972) Eur. J. Biochem. 31,335-344 20. Squires, C., and Carbon, J. (1971) Nature New Biol. 233,274-277 21. Gillam, I., Blew, D., Warrington, R. C., von Tigerstrom, M., and

22. Shulman, L. H., and Chambers, R. W. (1968) Proc. Natl. Acad.

23. Chamber, R. W., Aoyagi, S., Furukawa, Y., Zawabzka, H., and Bhanot, 0. S. (1973) Fed. Proc. 32, 585

24, Bhanot, 0. S., and Chambers, R. W. (1975) Fed. Proc. 34,517 25. Levitt, M. (1969) Nature (Lond.) 224, 759-763 26. Kim, S. H., Sussman, J. L., Suddath, F. L., Quigley, G. J., Mc-

Person, A,, Wang, A. H. J., Seeman, N. C., and Rich, A. (1974) Proc. Natl. Acad. Sci. U. S. A . 71,4970-4974

27. Robertus, J. D., Ladner, J. E., Finch, J. T., Rhodes, D., Brown, R. S. Clark, B. F. C., and Klug, A. (1974) Nature 250,546-551

28. Kim, S. H. (1979) in Transfer RNA: Structure, Properties and Recognition (Schimmel, P. R., Soll, D., and Abelson, J. R., eds) pp. 83-100, Cold Spring Harbor, NY

29. Kim, S. H., Suddath, F. L., Quigley, G. J., McPherson, A., suss- man, J. L., Wang, A,, Seeman, N. C., and Rich, A. (1974) Science 185,435-440

30. Fuller, W., and Hodgeson, A. (1967) Nature (Lond.) 215,817-821 31. Klug, A., Ladner, J., and Robertus, J. D. (1974) J. Mol. Biol. 89,

32. Wrede, P., Wurst, P., Vournakis, J., and Rich, A. (1979) J. Biol.

33. Youn. K.. Turner. D. H., and Tinoco, I., Jr. (1975) J. Mol. Biol.

5541

Symp. Quant. Biol. 34,505-512

117, 351-359

2551-2559

Chem. 252,2560-2565

Chem. 252, 2566-2574

Tener, G. M. (1968) Biochemistry 7,3459-3468

Sci. U. S. A . 61, 308-315

511-516

Chem. 254,9608-9616 ~~

84,’375-388 ’

.%. Sinehal. R. P. (1971) J. Biol. Chem. 246.5848-5851 3 5 : Signhal, R. P. (1974) Biochemistry 13,2924-2932 36. Lowdon, M., and Goddard, J. P. (1976) Nucleic Acids Res. 3,

~ D ~ - ~,

3383-3396

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Specific C -+ U Transitions in Glycine tRNA 4803

37. Squires, C., and Carbon, J. (1971) Nature New Bwl. 233,274-277 Abelson, J. (1972) FEBS Lett. 22, 144-148 38. Nishimura, S. (1972) Prog. Nucleic Acid Res. Mol. Biol. 12, 49- 42. Hooper, M. L., Russell, R. L., and Smith, J. D. (1972) FEBSLett.

85 22, 149-155 39. Gefter, M. L., and Russell, R. L. (1969) J. Mol. Biol. 39, 145-157 43. Yaniv, M., Folk, W. R., Berg, P., and SOU, L. (1974) J. Mol. Biol. 40. Carbon, J., and Curry, J. B. (1968) J. Mol. Biol. 38, 201-216 86,245-260 41. Shimura, Y., Aono, H., Ozeki, H., Sarabhai, A., Lamfrom, H., and Additional references will be found on p. 4805.

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

4804 Specific C + U Transitions in Glycine tRNA

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Specific C 4 U Transitions in Glycine tRNA 4805

I I I 1 I l I I I

- 2.0

15

D G D G

G D 20

Fraction number

A O H - 0 . 2 5 U c 75 C - 0 . 2 8 U U

P G * C C - G G C7o

5G * C G - C

A.U

A 1 0 A U A

C U C G . . . a

A G A G C 25 A

C G

. U

.G

.C

u6: c u c u 60 .....

C U G C G A G T

G ;! 45

G A

G C

30C G40 A * U

C O G U A U A

G C , " , + I 0.25U

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

E L Sabban and O S Bhanotcoli glycine tRNA.

The effect of bisulfite-induced C to U transitions on aminoacylation of Escherichia

1982, 257:4796-4805.J. Biol. Chem. 

  http://www.jbc.org/content/257/9/4796Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/257/9/4796.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on October 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from