Modified nucleosides and the chromatographic and aminoacylation behavior of tRNAIle from Escherichia coli C6

Biochimica et Biophysica Acta, 950 (1988) 172-181 Elsevier

172

BBA 91828

Modified nucleosides and the chromatographic and aminoacylation behavior o f t R N A ne f r o m E s c h e r i c h i a c o l i C 6 C h a r l e s L. H a r r i s , F a r h a d M a r a s h i

* and Sameer Sakallah * *

Department of Biochemistry, West Virginia University, School of Medicine, Morgantown, WV (U.S.A.) (Received 11 January 1988)

Key words: Modified nucleoside; tRNAne; Aminoacylation; Nucleic acid-protein interaction; (E. coli)

Transfer RNA from Escherichia coli C6, a M e t -, Cys -, relA - mutant, was previously shown to contain an altered tRNA ne which accumulates during cysteine starvation (Harris, C.L., Lui, L., Sakallah, S. and DeVore, R. (1983) J. Biol. Chem. 258, 7676-7683). We now report the purification of this altered tRNA ne and a comparison of its aminoacylation and chromatographic behavior and modified nucleoside content to that of tRNA ne purified from cells of the same strain grown in the presence of cysteine. Sulfur-deficient tRNA ne (from cysteine-starved cells) was found to have a 5-fold increased Vmax in aminoacylation compared to the normal isoacceptor. However, rates or extents of transfer of isoleucine from the [isoleucyi - A M P Ile-tRNA synthetase] complex were identical with these two tRNAs. Nitrocellulose binding studies suggested that the sulfur-deficient tRNA n~ bound more efficiently to its synthetase compared to normal tRNAnL Modified nudeoside analysis showed that these tRNAs contained identical amounts of all modified bases except for dihydrouridine and 4-thiouridine. Normal tRNA ne contains 1 mol 4-thiouridine and dihydrouridine per mol tRNA, while cysteine-starved t R N A Ile contains 2 mol dihydrouridine per mol tRNA and is devoid of 4-thiouridine. Several lines of evidence are presented which show that 4-thiouridine can be removed or lost from normal tRNA n~ without a change in aminoacylation properties. Further, tRNA isolated from E. coli C6 grown with glutathione instead of cysteine has a normal content of 4-thiouridine, but its tRNA ne has an increased rate of aminoacylation. We conclude that the low content of dihydrouridine in tRNA n~ from E. coli cells grown in cysteine-containing medium is most likely responsible for the slow aminoacylation kinetics observed with this tRNA. The possibility that specific dihydrouridine residues in this tRNA might be necessary in establishing the correct conformation of tRNA ne for aminoacylation is discussed. Introduction During amino-acid starvation of E. coil, accumulation of altered tRNA isoacceptors has been * Phillips Petroleum Corp., Bartlesville, OK, U.S.A. * * Chemistry Dept. Florida State University Tallahassee, FL, U.S.A. Abbreviations: s4U, 4-thiouridine; D, dihydrouridine; acp3U, 3-(3-amino-3-carboxypropyl)uridine.

Correspondence: C.L. Harris, Department of Biochemistry, West Virginia University, School of Medicine, Morgantown, WV 26506, U.S.A.

reported [1-4]. These changes are seen in both relA + and relA- organisms, and are not necessarily limited to starvation of cells auxotrophic for amino acids needed for tRNA modification. We previously reported the presence of an altered tRNA lIe in cysteine-starved cells of E. coli strain C6 [5,6]. This organism is a relA-, M e t - , C y s mutant of E. coli which produces tRNA deficient in thionucleotides during cysteine starvation [7]. During starvation, RPC-5 chromatography revealed that the normal tRNA ne is gradually replaced by an altered species, which elutes at a higher salt concentration [6]. If cysteine is added to the starved culture, one observes the apparent

0167-4781/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)

173 conversion of the altered tRNA to the normal species over a 2 h recovery period. These chromatographic changes are associated with two functional changes, aminoacylation and regulation of the ilvGEDA operon [6]. With crude tRNA, sulfur-deficient tRNAne was shown to have a 4-5-fold increased rate of aminoacylation, compared to normal tRNAm from the same organism. The in vivo aminoacylation ratio (isoleucyltRNAn~/total tRNA n~) was increased for cysteine-starved ceils as compared with that observed with cultures grown with cysteine. This was correlated with increased repression of the ilvGEDA operon in starved cells. Hence, the change in aminoacylation properties after cysteine starvation result in a change in the ratio of isoleucyl-tRNATM to free tRNAn~, with regulatory effects on the operon controlled in part by this ratio. We now report the purification of tRNA ne from exponentially grown and cysteine-starved cells of E. coli C6 and a comparison of their kinetic behavior and state of base modification. Materials and Methods

Bacterial strains and cell growth E. coli HfrC, relA-, Met-, Cys- (C6 mutant) was grown in M9 medium [8] in the presence of 0.16 mM cysteine and 0.13 mM methionine, as previously described [5]. Cells were maintained at 37 °C in either a gyratory shaking bath or a New Brunswick Microferm fermentor, with growth being monitored at 650 nm. Large-scale cultures were grown in a 350 liter fermentor at the Biology Division of Oak Ridge National Laboratory, with the cooperation of Dr. G. David Novelli and Mr. Ed. Phares. In all cases, cysteine starvation was carried out for 6 h in M9 medium supplemented with 0.016 mM cysteine and 0.13 mM methionine. Under these conditions growth ceased at 0.5-0.7 A650, due to depletion of cysteine. Cultures of E. coli Q13 were grown in M9 medium containing 0.13 mM methionine and tyrosine. E. coli B, used for enzyme preparations, was obtained from Grain Processing Corp., Muscatine, IA. Transfer RNA and aminoacylation Transfer RNA was prepared as described previously [5] using isopropanol fractionation as the

final step in purification. An aminoacyl-tRNA synthetase mixture was prepared from E. coli Q13 using the method of Kelmers et al. [9]. Isoleucylt R N A synthetase was purified from E. coli B using a modification of the procedure of Baldwin and Berg [10] and has been described previously [11]. Aminoacylation assays [5] were carried out in a final volume of 0.5 ml and contained: 20 mM Tris-HC1 (pH 7.3), 10 mM magnesium acetate, 2 mM ATP (pH 7), 0.2-1.1 mM L-[14C]- or L[3H]isoleucine, up to 1 A260 unit of tRNA and various amounts of isoleucyl-tRNA synthetase. Initial velocity determinations were carried out with from 0.07 to 0.15/~g of isoleucyl-tRNA synthetase, while up to 25/~g of this enzyme was used when total aminoacylation of tRNA was measured. Incubations were at 37 o C, samples of 50 #1 being taken at appropriate times and applied to 2.4 cm Whatman 3MM paper discs. The discs were added to 5 % trichloroacetic acid, washed and counted by the scintillation method as previously described [5]. Determination of the kinetic constants for tRNAs was carried out using the above assay, the data being analyzed by the curve-fitting program of Knack and Rohm [12].

tRNA-synthetase binding Normal or sulfur-deficient t R N A was aminoacylated with [14C]isoleucine as described above, and [14C]isoleucyl-tRNA was isolated from the reaction mixture by phenol extraction and ethanol precipitation [5]. Binding of isoleucylt R N A to purified isoleucyl-tRNA synthetase was measured using the method of Yarus and Berg [13]. Reaction mixtures of 200 #1 contained 8.8 /~mol MgC12, 2 /~mol 2-mercaptoethanol, 10 #g bovine serum albumin, 15 /~g purified isoleucyltRNA synthetase (150 pmol) and variable amounts of [14C]isoleucyl normal or sulfur-deficient tRNAs. The mixtures was incubated for 15 s at 17 °C and applied to 2.4 cm nitrocellulose filters (BA-85, Schleicher & SchueU). The filters were presoaked and washed with 44 mM KH2PO4/6 mM K 2 H P O J 5 0 mM MgC12 (pH 5.5) and dried, and radioactivity was measured by the scintillation method as above. Transfer reaction The rate of isoleucine transfer from [isoleucyl - AMP. Ile-tRNA synthetase] to tRNA was mea-

174 sured by the method of Eldred and Schimmel [14]. The complex was isolated from the following 300 #1 reaction mixture: 4/~mol K2HPO 4 (pH 8), 0.4 /~mol bovine serum albumin, 2 #mol 2mercaptoethanol, 32 #Ci [3H]isoleucine (40 Ci/mmol) and 20 #g Ile-tRNA synthetase. After incubation at 25 °C for 5 rain the mixture was applied to a Sephadex G-50 column (1 x 33 cm) equilibrated with 10 mM sodium cacodylate buffer (pH 6)/0.5 mM EDTA/50 mM KC1. Elution was with this same buffer, the complex eluting with the void volume, far from free amino acid. Transfer reactions were carried out by incubating 0 to 1.0 nmol of tRNA with 1 pmol of the complex, in a reaction mixture containing 10 mM sodium cacodylate (pH 6), 5 mM 2-mercaptoethanol, 0.5 mM EDTA, 50 mM KC1, and 1 nmol of purified normal or sulfur-deficient tRNA "e. After 15 s at 3°C the reaction mixture was filtered through Whatman 3MM filters. The filters were washed with 5% trichloroacetic acid and counted as above.

Chromatographic procedures BD-cellulose chromatography was carried out according to the method of Gillam et al. [15]. DEAE-Sephadex A-50 and RPC-5 chromatography were carried out using the procedure of Ohashi et al. [16] and Kelmers and Heatherly [17], respectively. Separation of tRNA on Bio-Gel A-5m using reverse ammonium sulfate gradients followed the procedure of Holmes et al. [18]. All chromatographic procedures were carried out at 4 ° C, with the exception of RPC-5 separations, which were done at room temperature. Nucleoside analysis Nucleoside digests were prepared by incubating 0.1-1 A260 unit of tRNA in 0.2 M sodium acetate with 0.21 /~g of snake venom phosphodiesterase and 8 units of E. coli alkaline phosphatase [19]. Incubation for 4 h at 37 °C was sufficient for full hydrolysis of tRNA to the nucleoside level. Nucleoside digests were analyzed by HPLC using a Zorbax ODC C18 column (DuPont) attached to a Perkin-Elmer Model 75 Chromatograph, using the procedure of Gehrke et al. [20]. The eluate was monitored at 254 nm and nucleosides were identified by their elution position in comparison with known standards. Areas of the nucleoside peaks

were determined by planimetry and corrected for differences in absorption coefficient at 254 nm. This column was particularly useful in determining the amounts of dihydrouridine which was well separated from other nucleosides and could be directly detected at 230 nm. Nucleotide analysis was determined by the method of Harada et al. [21]. 2-4 A260 units of tRNA were digested with 1-2 units of RNAase T2, respectively, in 0.2 ml of 0.05 M potassium acetate buffer (pH 4.7) for 30 h at 37°C. The hydrolysates were directly applied to 20 x 20 cm cellulose plates and developed in the first dimension with isobutyric acid/0.5 M NH4OH (5:3, v/v). Separation in the second dimension utilized 2-propanol/conc. HC1/water (70 : 15 : 15, v/v). Spots were located using a UV lamp, scraped and extracted with 0.1 M HC1 for 2 h. The eluted nucleotides were analyzed using a Cary 17 spectrophotometer, their UV absorption spectra being determined at pH 1, 7 and 12. The spectra were used for identification and quantitation of each nucleotide [22].

Chemicals and radioisotopes Chromatographic materials were purchased from the following sources: BD-cellulose, Schwarz/Mann Bioresearch; DEAE-Sephadex A50, Pharmacia Fine Chemicals, Bio-Gel A-5m, Bio-Rad Laboratories, cellulose TLC plates (500 micron), E. Merck. RPC-5 components were generously donated by G.. Novelli of Oak Ridge National Laboratories. Nucleosides were obtained from Schwarz/Mann or PL Laboratories. RNAase T1 and T2 and E. coli alkaline phosphatase were obtained from Sigma Chemical Co. Snake venom phosphodiesterase was purchased from Worthington. [14C]Isoleucine (325-360 mCi/mmol) and [3H]isoleucine (40-100 Ci/mmol) were obtained from New England Nuclear. All other chemicals were of the highest quality available commercially. Results

Purification of tRNAlle We previously observed that isoleucine acceptance occurred at a faster rate with tRNA from cysteine-starved cells of E. coli C6, a relA -, Met -, Cys- strain [6,7]. This is illustrated by the data of

175 Table I, where it is shown that crude t R N A f r o m cysteine-starved cells has a 3.5-fold greater rate of isoleucine acceptance (line 5) than t R N A f r o m cells grown with cysteine (line 1). However, the t R N A TM level in cells supplemented with cysteine is the same as in cells starved for this a m i n o acid. Since this rate returned to n o r m a l during recovery f r o m cysteine starvation [6], it appears that the aminoacylation change is related to the state of t R N A modification. To investigate this further, we have purified t R N A Il~ f r o m exponential and cysteine-starved cells in an attempt to learn the reason for the increased aminoacylation. Two species of t R N A n~ were observed in exponential cultures (cysteine present) of E. coli C6, the purity and kinetic properties of these have been published [11] and are included for comparison in Table I, lines 2 and 3. The isoacceptor n u m b e r was assigned on the basis of their elution positions on methylated albumin kieselguhr chrom a t o g r a p h y [5]. The data show that isoacceptor 2 contains 1 mol s4U per mol t R N A and has a slow rate of acylation c o m p a r e d to peak 3, which has no s4U. F o r comparison, purified t R N A s with 1 mol s 4 U / m o l t R N A had A 3 3 5 / / A 2 6 0 percentage ratios of 1.9-2.3% [23]. We have also isolated t R N A TM f r o m exponential cells of strain C6 cells b y aminoacylation followed b y n a p h t h o x y a c t y l a tion and BD-cellulose c h r o m a t o g r a p h y using the m e t h o d of Tener et al. [15]. The final p r o d u c t in this case was a single t R N A II~, with a slow rate of aminoacylation and very little s4U (Table I, line 4). Control experiments showed that the derivatization used to purify this t R N A did not alter its aminoacylation kinetics but did lead to the loss of s4U from the general t R N A population. I n other experiments, we have also noted the loss of s4U f r o m lyophilized, crude t R N A . The t R N A TM purified from this stored t R N A was devoid of s4U (A335//h260=0.5%) and had kinetic properties similar to the preparations listed in Table I, lines 2, 4 and 6 (data not shown). The experiments reported above and previously [6,7] might seem to indicate a role for s n u in the aminoacylation of t R N A n L Several pieces of evidence argue against this hypothesis. First, removal of s4U using C N B r had no effect on the rate of aminoacylation of t R N A n~ [7]. Secondly, the data of Table I show that t R N A n~ f r o m exponential

TABLE I KINETIC PROPERTIES OF tRNAne FROM EXPONENTIAL AND CYSTEINE-STARVED CELLS OF E. COLI C6 tRNAs were prepared from E. coli C6 grown with cysteine (exponential cells, lines 1-4), glutathione (line 9) or starved for cysteine (lines 5-8). Isoleucine accepting tRNAs purified as previously described [11]; tRNAI~e and tRNAg~, lines 2 and 3 and tRNAl~ajor and tRNAl~nor (lines 6 and 7). The major tRNAne from exponential cells (line 4) was purified by aminoacylation naphthoxyacetylation and BD-cellulose chromatography [15]. The remaining tRNAI~ajor from starved cells was purified by chromatography in unacylated form using the column procedures described in Materials and Methods. In the case of 'crude' samples, kinetic data were obtained using unfractionated tRNA from fed, cysteine-starved or cells grown with GSH in place of cysteine using 20 #g of a mixture of aminoacyl-tRNA synthetases from E. coli strain Q13. The tRNA concentration was varied from 0 to 20/tM, with 50/~1 samples being taken from the standard reaction mixture. The rate of isoleucyl-tRNA formation under all conditions was linear for 5 min. Total isoleucine acceptance was determined using 0.2 mg of the synthetase mixture to ensure complete aminoacylation. Kinetic data for purified tRNA were collected as above but using from 0 to 2 /~M tRNA and 0.16 /tg of purified isoleucyl-tRNA synthetase. Data from two separate experiments are combined, Vma~ being expressed per #g of purified isoleucyl tRNA synthetase. Total acceptance was measured as above, using 7 /xg of purified synthetase, n.d., not determined. Transfer RNA a Isoleucine acceptance (pmol/A260) Exponential cells 1. Crude 85 2. tRNA~2~e 1511 3. tRNA~e 1475 4. tRNAne 1143

A335 /

Vmax

Km

A260(%) (pmol/min) (/~M)

1.70 1.90 0.50 0.70

2.6 145.0 1045.0 150.0

7.39 0.72 1.01 0.53

Cysteine-starved cells 5. Crude 90 6. tRNAI~nor 1495 7. tRNAI~a,or 1375 Ile ~J 8. tRNAmajo r 1192

0.70 2.00 0.47 0.70

9.1 143.1 877.0 783.0

3.95 0.64 1.00 0.36

GSH-grown cells 9. Crude n.d.

1.70

6.1

3.28

cells has the same kinetic behavior whether s4U is present or n o t ( c o m p a r e lines 2 and 4). The exception to this is tRNAI~e, line 3, which was devoid of s4U b u t had a fast rate of aminoacylation. Since this t R N A was o b t a i n e d in low yield it is possible that this t R N A represents a minor, undermodified f o r m of t R N A ne similar to that isolated from

176

cysteine-starved cells (see below). Finally, we found that crude tRNA lost s 4 U upon storage, yet the kinetic properties of this tRNA were identical to those of freshly isolated tRNA.

tRNA "e from cysteine-starved cells Purification of tRNA n~ from cysteine-starved cells was carried out using two separate approaches, with both aminoacylated and unacylated t R N A as starting material. When [14C]isoleucyl-tRNA was chromatographed on DEAE-Sephadex A-50, we observed a minor and a major species. These isoacceptors were separately purified exactly as reported previously for normal tRNA n~ [11]. Table I shows that the minor tRNA ll~ (line 6) had nearly the same kinetic behavior as the slowly acylated tRNA TM from cells grown in the presence of cysteine (lines 2 and 4). It also contained 1 mol of saU, similar to tRNAI1L Since cysteine starvation is established by growing E. coli C6 on a suboptimal level of this amino acid (0.016 M), it is likely that this tRNA represents residual normal tRNA II~, present in these cells prior to the onset of cysteine starvation. The major tRNA n~ present in cysteine-starved cells has been purified using two separate strategies: by successive chromatographies of isoleucyltRNA (line 7) or unacylated tRNA (line 8). Regardless of the method of purification, the major tRNA n~ purified from cysteine-starved cells had a 5-fold-increased rate of aminoacylation, compared to normal tRNA. This indicates that the aminoacylation behavior observed with crude tRNAs is maintained with purified isoacceptors isolated from exponential and cysteine-starved E. coli C6. Chromatographic properties of purified tRNAs We previously showed that tRNA TM isolated from cysteine-starved cells of E. coli C6 eluted at a higher salt concentration on RPC-5 chromatography than normal tRNA n~ [6]. On this matrix one observes only a single major tRNA n~ with either fed or starved cells. To determine whether the purified tRNAs has this same behavior, we aminoacylated each tRNA and co-chromatographed them on RPC-5 as shown in Fig. 1. Sulfur-deficient tRNA m elutes at a higher salt concentration on this column relative to normal tRNA n~, just as was the case for the crude tRNAs.

i

,

i

80C

800

",T-

--

60o

~

600

n c.)

6

~t n 0

400

400

Ioo

~,

I

0

~

70

80

. . . . . ,n

90

0

I00

Fraction

Fig. 1. Analysis of purified normal and sulfur-deficient tRNA ne by RPC-5 Chromatography. 0.1 A260 unit of purified normal or sulfur-deficient tRNA TM was aminoacylated with [14C]- and [3H]isoleucine, respectively, reisolated from the reaction mixture by phenol extraction and ethanol precipitation and cochromatographed on a 0.6 × 45 cm RPC-5 column equilibrated with 10 m M sodium acetate (pH 4.5)/10 mM M g C I 2 / 2 mM 2-mercaptoethanol/0.4 M NaCl. The column was washed with 28 ml of this buffer at 0.8 m l / m i n and eluted with a 150 ml gradient from 0.4 to 0.8 M NaCl in the same buffer. 1 ml fractions were collected, mixed with 3 ml of ACS (Amersham) and counted by the scintillation method under double label conditions. It should be noted that the small amount of tritium which appears under the major normal tRNA TM peak was not observed when sulfur-deficient tRNA nc was chromatographed separately. This peak is due to spillover of 3H into the 14C channel, i.e., is the result of a counting artifact. O, normal [14C]isoleucyl-tRNAnC; O, sulfur-deficient [3H]isoleucyltRNA lIe.

In addition, sulfur-deficient tRNA TM chromatographs distinctly from freshly isolated normal tRNA ne on methylated albumin kieselguhr [5], BD-cellulose, agarose and DEAE-Sephadex columns (data not shown). Since we find a difference in elution position for sulfur-deficient tRNA Ile compared to a s4U-free normal tRNA n~ (Fig. 1; Table I, line 4), there must be some other differences in these tRNAs which give rise to altered chromatographic elution and rather large differences in aminoacylation.

Kinetic analysis of tRNA lie isoacceptors from E. coil C6 The data of Table I show that the maximal rate of isoleucylation of sulfur-deficient tRNA was about 3.5-times greater than in the case of normal tRNA, using a crude mixture of tRNA and syn-

177

thetases. A 5-fold greater Vmax was seen with purified sulfur-deficient tRNA he, as compared with the purified normal species. In the latter case, both the tRNAs and the isoleucyl-tRNA synthetase were purified, demonstrating that the kinetic effects are not due to the presence of inhibitors in the crude preparations. While there was a clear difference in Vmax between the normal and cysteine-starved tRNAs, comparison of K m values is not conclusive. With unpurified tRNAs and enzyme, the K m value observed for cysteinestarved tRNA n~ was half that of the normal tRNA. It is likely that other components of the tRNA mixture may have influenced these values, since the K m values of the purified tRNAs were not that different. If we average the K m va!ues for the three slowly acylating, purified species (Table I, lines 2, 4 and 6) we obtain 0.63 #M. The corresponding value for the 'fast' tRNAs (lines 3, 7 and 8) is 0.79 #M. Because of the variability in K m values for purified tRNAs, we are unable to reach a conclusion as to whether the increased rate of aminoacylation is due to a change in the binding properties of cysteine-starved tRNA n~.

tRNA-synthetase binding To see if tRNA binding differences could be demonstrated, we also measured the direct bind-

A tn O

2,0 E3 Z

o m z

o

oo

I

W

o

o

g

1.0

O

g

O

0 I

10

20

30

40

50

ISOLEUCYL -tRNA ADDED (pmoles)

Fig. 2. Binding of normal and sulfur-deficient isoleucyl-tRNA to isoleucyl-tRNA synthetase. Various amounts of [14C]isoleucyl-tRNA was added to isoleucyl-tRNA synthetase and binding was measured as described in Materials and Methods. n, normal isoleucyl-tRNA; ©, sulfur-deficient isoleucyl-tRNA.

I

I

I

I

0.75

,i0

C

~

O. 5

v

f

0.10

0.05 ILl

n-

0

0.25

0 5

ADDED tRNA (nmoles)

Fig. 3. Rate of isoleucine transfer from the [isoleucyl - AMP. isoleucyl-tRNA synthetase] complex to normal and sulfur-deficient tRNAne. The complex was prepared as described above and 1 pmol was reacted with 0 to 1.0 nmol of each tRNAne at 4 ° C for 15 s. The reaction mixtures were then filtered through Whatman 3MM paper discs, washed with 5% trichloroacetic acid, dried and counted by the scintillation method. Composite data from three separate experiments is shown here. D, normal tRNAne; O, sulfur-deficient tRNAIle.

ing of t R N A TM to isoleucyl-tRNA synthetase, using the method of Yarus and Berg [12]. These experiments were carried out at p H 5.5 in the absence of the other aminoacylation substrates, using normal and sulfur-deficient [14C]isoleucyl-tRNAn¢, and purified isoleucyl-tRNA synthetase. The data of Fig. 2 show a clear difference between normal and sulfur-deficient tRNA Ile in the amount of isoleucyl-tRNATM bound to isoleucyl-tRNA synthetase. In addition, the binding efficiency observed for sulfur-deficient tRNA TM is higher than observed for normal tRNAnL We also determined binding constants for these two tRNAs, using a curve-fitting program [13]: K d for sulfur-deficient tRNA n¢ = 41.5 + 23.0 nM (n = 3); K d for normal tRNA n~ = 117 + 37 n M (n = 3). Taken together, these data demonstrate an increased binding affinity of sulfur-deficient isoleucyl-tRNA TM for isoleucyl-tRNA synthetase.

Rate of transfer The rate of transfer of isoleucine from [ I l e A M P . isoleucyl-tRNA synthetase] to normal and sulfur-deficient t R N A TM was determined using the method of Eldred and Schimmel [154]. The data (Fig. 3) show the relationship between the rate of isoleucyl-tRNA formation and the amount of ad-

178

ded tRNA n~, indicating no differences in transfer between normal and sulfur-deficient tRNA TM. In addition, initial velocity determinations were also carried out in the presence of 1 nmol of each tRNA n~ and 2 pmol of [ I l e - AMP-isoleucyltRNA synthetase], wherein rate values of 0.1 p m o l / m i n were observed for both tRNAs. Hence, the difference in the overall rate of aminoacylation observed above was not the result of a more efficient transfer reaction for the sulfur-deficient tRNA.

Modified nucleoside analysis Previous work suggested that the tRNA ne which accumulates in cysteine-starved cells of E. coli C6 is undermodified, and is converted to normal tRNA upon addition of cysteine [5]. To determine which nucleoside is altered we analyzed normal and sulfur-deficient tRNA ne digests using cellulose T L C [21] and H P L C on Zorbax ODC [20]. A composite of these studies is given in Table II. The major differences between these isoacceptors is the absence of s a u (see Table I) and one more residue of D comparing sulfur-deficient to normal tRNAnL We did not obtain a clear separation of acp3U and Gp on TLC; hence, an absolute value for this modified nucleoside is not indicated. We did observe equal amounts of an altered tRNA TM on chromatography of either normal or sulfur-deficient tRNA n~ purified by naphthoxyacetylation. Since this procedure is known to result in derivatization of acp3U in tRNA [24], we conclude that these tRNAs probably do not differ with regard to this modified nucleoside. The amounts of t6A, mTG, T and ~k were also analyzed by H P L C as described in Materials and Methods, with results in complete agreement with the data of Table II (not shown). Effect of glutathione as a sulfur source Since the nucleoside analyses indicated that normal tRNA TM had a low level of dihydrouridine coupled with a slower rate of aminoacylation, we considered the possibility that these properties are the result of growth in cysteine-containing medium. Indeed, cysteine has been known to inhibit growth in E. coli [25], due to inhibition of isoleucine synthesis [26]. We also showed that glutathione (GSH) could satisfy the requirement

TABLE II NUCLEOTIDE COMPOSITION OF PURIFIED tRNA ne ISOACCEPTORS With exceptions quoted in the footnotes, these nucleotide data were obtained by two-dimensional thin-layer cellulose chromatography of RNAase T2 digests of normal tRNA Ilc (1511 pmol/A26o) and cysteine-starved tRNA nc (1375 pmol/A26o). Nucleotide

Ap Cp Up Gp +p m7Gp Tp t6A a Dp b s4Up c acp3Up

E. coli B

tRNA

Normal tRNAI~e

Cysteinestarved tRNAi~ajor

14 19 10 24 2 1 1 1 2-3 0 1

14.0 20.2 10.4 24.2 2.1 1.1 1.1 0.9 0.8 1.0 1.0

11.9 19.5 10.0 28.0 1.7 1.0 0.7 0.6 2.1 0 + d

The amount of t6A shown for cysteine-starved tRNA ne was obtained by nucleoside analysis on a Zorbax ODC colunm, using tRNA with a purity of 1192 pmol/A26o. Normal tRNA ne (1143 pmol/A260) gave a value of 0.9 residues tOA/mol tRNA using this analytical method. b The amount of D was determined at 230 nm by Zorbax chromatography of the nucleoside digests described in a. c s4U was estimated from the A335/A260 values obtained with unhydrolyzed tRNAs. a This nucleotide was incompletely separated from Gp in the chromatogram for cysteine-starved tRNAt~ajor. a

for cysteine in strain C6, with no inhibition of growth even at high concentrations of GSH. Transfer R N A was isolated from E. coli C6 grown in M9 medium containing 0.21 mM GSH in place of cysteine. The A335/A260 ratio of this tRNA was 1.7% (Table I), indicating a normal complement of s4U. Hence, GSH can apparently provide sufficient cysteine to meet the needs for both protein biosynthesis and thionucleotide synthesis. Aminoacylation data (Table I) show that tRNA isolated from ceils grown in GSH-containing medium and sulfur-deficient tRNA are similar in terms of Vmax and apparent g m values, and clearly different from normal tRNA. Other experiments show that tRNA ~le from cells grown with GSH is not only normal in its s a u content but also coelutes with normal tRNA ne on RPC-5 columns. The data of Fig. 4 show that tRNA ~le from cells grown

179 !

i

|

i

|

30

Oo ta o-

20

"' >

/

\

.

W

Q:

IC

N

50

60

70

80

90

FRACTION

Fig. 4. RPC-5 chromatography of cysteine-starved tRNA and tRNA isolated from E. coli C6 grown in 0.21 mM glutathione (GSH) in place of cysteine. In the latter instance, cells grew with a doubling time of 130 rain, about half as fast as observed with exponential cultures of strain C6 in cysteine-containing medium. Sulfur-deficient and GSH tRNAs were aminoacylated with [14C]- or [3H]isoleucine, respectively, as above, obtained from reaction mixtures by DEAE-cellulose chromatography, precipitated with ethanol and co-chromatographed on RPC-5 as described in the legend to Fig. 1. The gradient volume was 200 ml, with 2 ml fractions being collected and radioactivity measured as above. The symbols are: ©, [14C]isoleucyl-tRNA from cells grown in GSH; o, [3H]isoleucyl-tRNA from cysteine-starved cells.

in G S H medium elutes earlier than the cysteinestarved species. Therefore, we have uncoupled the kinetic and chromatographic properties of t R N A Ile in the case of GSH-grown cells. The kinetic behavior of tRNA n~ in GSH-grown cells is similar to that seen for cysteine-starved tRNA TM, while the chromatographic elution of this tRNA on RPC-5 is identical to normal tRNAII9 Discussion

A chromatographically altered tRNA n~ was observed to accumulate during cysteine starvation of E. coli C6, a relA-, Me t - , Cys-, mutant [5]. When starved cells were allowed to recover in the presence of cysteine, restoration of the normal isoacceptor pattern occurred within 2 h. This suggests that the cysteine-starved tRNA ne may be related to its normal counterpart through base modification. The data reported here show that the cysteine-starved tRNA is undermodified with respect to s4U, as expected, since cysteine is the sulfur donor for thionucleotide synthesis in E. coli [27,28]. While our results show that the lack of

s4U may contribute to the shift in chromatographic elution of t R N A lIe on several chromatographic matrices, it is clear that the increased rate of aminoacylation is not due to the loss of this thionucleoside. Previous experiments showed that removal of s4U from normal t R N A by CNBr treatment did not alter the rate of isoleucyl-tRNA formation [11]. Further, the loss of s4U in tRNA ne during storage failed to alter the aminoacylation properties of normal tRNA. Indeed, s4U can be lost by simply repeated scanning of tRNA I1~ in a Cary 17 spectrophotometer [29]. Finally, crude tRNA isolated from E. coli C6 cells grown with G S H in place of cysteine was normal in terms of s4U content, but contained a tRNA ne which had a fast rate of acylation. This conclusion is consistent with our previous observation that the chromatographic shift and s4U content both return to normal during recovery from cysteine starvation, before the complete recovery of slow aminoacylation kinetics. Some other modified nucleoside or combination thereof must account for the altered aminoacylation behavior. Modified nucleoside analysis showed that the normal and cysteine-starved tRNAs contain similar amounts of ~b, mTG, T and t6A, but differed in both s4U and D. The purified tRNA n~ from starved cells lacked s4U and contained twice the D level seen with purified tRNA TM from cells grown with cysteine. Hence, the only observed modified nucleoside difference between normal and cysteine-starved tRNA, other than s4U, is D. The cysteine-starved tRNA n~ contains the amount of D predicted from the published sequence of this tRNA in E. coli B [30]. The sequence data were obtained with a mixture of two isoacceptors specific for isoleucine, but differing in their content of D. Considering the above findings, one might be justified in concluding that the tRNA II~ present in cysteine-starved cells is the 'normal' tRNA he. It has a high Vm~x in the aminoacylation reaction and a normal complement of D. In contrast, cells grown in the presence of cysteine contain a tRNA lIe which is deficient in D and has a much lower rate of aminoacylation. We would predict that the 'fast-acylating' tRNA H~ we observed in E. coli C6 cells grown with G S H contains a higher D level than the corresponding isoacceptor in cells grown with cysteine. Proof of

180 this awaits the future purification and characterization of tRNA ne from GSH-supplemented cells of strain C6. We also attempted to learn which step in the overall aminoacylation reaction is altered with cysteine-starved tRNAnL No difference in K m or in the rate of transfer from i s o l e u c y l - AMP to normal and cysteine-starved tRNA n~ could be demonstrated. We observed nearly a 3-fold lower dissociation constant for the cysteine-starved, isoleucyl-tRNA, isoleucyl-tRNA synthetase complex, as compared with the K d measured for normal isoleucyl-tRNA. It is difficult to interpret how this increased binding would influence the overall rate of aminoacylation, since these measurements were made in the absence of the other aminoacylation substrates. If the binding of the aminoacylated and free forms of these tRNAIl~s to isoleucyl-tRNA synthetase are similar, then enhanced binding might be the reason for the increased rate seen for cysteine-starved tRNA. It is possible that equilibrium of binding is reached more rapidly with cysteine-starved tRNA n~, thus resulting in a faster overall rate of aminoacylation than observed with normal tRNAnL It is interesting to note that Schimmel proposed the formation of a covalent link between tRNA ne and its synthetase, the interaction occurring between the enzyme and position 8 of the tRNA [31]. Perhaps the presence of D or the conformation of tRNA in this region is important for the formation of this aminoacylation intermediate. Growth of E. coli strain C6 with G S H in place of cysteine is slower than with cysteine itself, presumably due to the need to transport and metabolize this peptide. Cells grown in the presence of G S H contain a tRNA ne with increased aminoacylation kinetics compared with tRNA ne from C6 cells grown in cysteine. The tRNA n~ from GSH-grown cells contains normal levels of s4U and is indistinguishable from normal tRNA TM on RPC-5 columns (Fig. 4). We would predict that the altered aminoacylation kinetics of tRNA ne from cysteine-starved and GSH-grown cells of strain C6 may be due to an increased dihydrouridine levels in these tRNAs as compared to that of normal tRNA. Since some D is found in normal tRNA TM it will be important to know which residue remains and how the modification of the

other uridines so profoundly alters the aminoacylation behavior of this tRNA. There are some similarities and differences between the current results and previous studies of altered tRNAs in amino-acid-starved relAmutants of E. coli. Fournier and his co-workers described the accumulation of several modification-deficient tRNAs after amino acid starvation [1-4]. In the case of an altered tRNA Phe, they observed a decreased rate of aminoacylation and translation in an in vitro poly(U)-directed system [3] associated with a deficiency of D at position 16. This altered t R N A was also missing several other modified bases, making assignment of the altered functions to a specific modified nucleoside impossible [4]. In addition, the altered tRNAs which accumulated were either converted slowly to the normal species during recovery from starvation, or were not changed at all. In this study, we find that the 'normal' t R N A lie is missing D and is slow in aminoacylation, in substantial agreement with the observations of Fournier cited above. However, the sulfur-deficient tRNA TM seen here is converted back to the normal species during recovery, both in vivo [6] and in vitro [5]. Our data also agree with those of Waters et al., who found a chromatographically altered tRNA II~ in chloramphenicol-treated E. coli [32]. Although the bulk tRNA was found to be deficient in D and s4U and judged normal in total amino-acid acceptance, the reaction kinetics were not studied. Their results differ from those presented here in that the deficiency in D is associated with tRNA II~ in exponential cultures, instead of conditions where protein synthesis is blocked by amino-acid starvation or chloramphenicol treatment. Fast and slowly acylated forms of tRNA I1~ were previously noted in E. coli by Yegian and Stent [33] with chromatographic properties similar to those reported by our laboratory [5]. Future experiments will determine the amount of D in the corresponding positions of normal and sulfur-deficient tRNA ne purified here. The present results show the importance of certain base modifications to tRNA function. At the outset, we expected that any functional changes in tRNA isolated from cysteine-starved cells would involve thionucleotides. However, our results show that another modified nucleoside (perhaps D) is likely to be responsible for the altered aminoacyla-

181

tion kinetics of cysteine-starved tRNATM. Finally, the recent review by BjiSrk [34] emphasizes that modified nucleosides are likely to be involved in fine tuning the reactions of tRNA. However, examples of large functional changes between tRNAs related through modification are also known. Roe et al. [35] observed dramatically lower aminoacylation rates for tRNAPhe not methylated at G10. In the current study, a large aminoacylation effect is correlated with the D content of E. coli tRNAnL It will be important to find out whether this effect is direct, or is due to an altered tRNA conformation [34] when D is missing.

Acknowledgements The authors thank Lorena Lui, Robert Beto and Susan Capelle for technical assistance. We are also grateful to James Blair and Michael Miller for their suggestions, and to Carol Molisee for her careful and patient typing of this manuscript. This work was supported by NIH grant GM 25620 and by NIH Biomedical Research Grant 2 S07 RR5433-22.

References 1 Fournier, M.J. and Peterkofsky, A. (1975) J. Bacteriol. 122, 538-548. 2 Kitchingman, G.R. and Fournier, J.J. (1975) J. Bacteriol. 124, 1382-1394. 3 Fournier, M.J., Webb, E. and Kitchingman, G.R. (1976) Biochim. Biophys. Acta 454, 97-113. 4 Kitchingman, G.R. and Fournier, M.J. (1977) Biochemistry 16, 2213-2220. 5 Harris, C.L., Marashi, F. and Titchener, E.B. (1976) Nucleic Acids Res. 3, 2129-2142. 6 Harris, C.L., Lui, L., Sakallah, S. and DeVore, R. (1983) J. Biol. Chem. 258, 7676-7683. 7 Harris, C.U, Titchener, E.B. and Cline, A.L. (1969) J. Bacteriol. 100, 1322-1327. 8 Anderson, E.H. (1946) Proc. Natl. Acad. Sci. USA 32, 120-128. 9 Kelmers, A.D., Novelli, G.D. and Stulberg, M.P. (1965) J. Biol. Chem. 240, 3979-3983.

10 Baldwin, A.N. and Berg, P. (1966) J. Biol. Chem. 241, 831-838. 11 Harris, C.L. and Marashi, F. (1980) Nucleic Acids Res. 8, 2023-2037. 12 Knack, I. and Rohm, K.H. (1981) Hoppe Seyler's Z. Physiol. Chem. 362, 1119-1130. 13 Yams, M. and Berg, P. (1967) J. Mol. Biol. 28, 479-490. 14 Eldred, E.W. and Schimmel, P.R. (1972) Biochemistry 11, 17-23. 15 Gillam, I., Millward, S., Blew, D., Von Tigerstrom, M., Wimmer, E. and Tener, G.M. (1967) Biochemistry 6, 3043-3056. 16 Ohashi, Z., Saneyoshi, M., Harada, F., Hara, H. and Nishimura, S. (1970) Biochem. Biohys. Res. Commun. 40, 866-871. 17 Kelmers, A.D. and Heatherly, D.E. (1971) Anal. Biochem. 44, 486-495. 18 Holmes, W.H., Hurd, R.E., Reid, B.R., Rimerman, R.A. and Hatfield, G.W. (1975) Proc. Natl. Acad. Sci. USA 72, 1068-1071. 19 Sen, G.C. and Ghosh, H.P. (1974) Anal. Biochem. 58, 578-591. 20 Gehrke, C.W., Kuo, K.C. and Zumwalt, R.W. (1980) J. Chromatogr. 188, 129-147. 21 Harada, F., Kimura, F. and Nishimura, S. (1971) Biochemistry 10, 3269-3277. 22 Hall, R.H. (1970) The Modified Nucleosides in Nucleic Acids, pp. 257-280, Colombia University Press, New York. 23 Seno, T., Kobayashi, M. and Nishimura, S. (1969) Biochim. Biophys. Acta 174, 71-85. 24 Neuheimer, U. and Hedgcoth, C. (1974) Arch. Biochem. Biophys. 160, 631-642. 25 Rowley, D. (1953) J. Gen. Microbiol. 9, 37-43. 26 Harris, C.L. (1981) J. Bacteriol. 145, 1031-1035. 27 Hayward, R.S. and Weiss, S.B. (1966) Proc. Natl. Acad. Sci. USA 55, 1161-1168. 28 Lipsett, M.N. and Peterkofsky, A. (1966) Proc. Natl. Acad. Sci. USA 55, 1168-1174. 29 Marashi, F. (1977) Ph.D. Dissertation, West Virginia Univ. 30 Yams, M. and Barrell, B.G. (1971) Biochem. Biophys. Res. Commun. 43, 729-734. 31 Schimmel, P.R. (1979) Adv. Enzymol. 49, 187-222. 32 Waters, L.C., Shugart, L., Yang, W-K and Best, A. (1973) Arch. Biochem. Biophys. 156, 780-793. 33 Yegian, C.D. and Stent, G. (1969) J. Mol. Biol. 39, 59-71. 34 Bjork, G.R., Ericson, J.U., Gustafson, C.E.D., HagervaU, T.G., Jonsson, Y.H. and Wilkstrom, P.M. (1987) Annu. Rev. Biochem. 56, 263-287. 35 Roe, B., Sirover, M. and Dudock, R. (1973) Biochemistry 12, 4146-4154.