Eur. J. Biochem. 210,451 -454 (1992) 0FEBS 1992

Effect of inorganic pyrophosphate on the pretransfer proofreading in the isoleucyl-tRNA synthetase from Escherichia coli R. Kalervo AIRAS Department o f Biochemistry, Univcrsity of Turku, SF-20500 Turku. Finland (Received June Y/August 17, 1992)

-

EJB 92 0809

A total rate equation was used to calculate the discrimination of valine by the isoleucyl-tRNA synthetase from Escherichia coli. The PPipresent in the cell makes the backward reaction or the pyrophosphorolysis of the E . possible. If the E . Ile-AMP has been corrected for wrong aminoacyl adenylation by the pretransfer proofreading, the pyrophosphorolysis rapidly equilibrates the corrected E . Ile-AMp with E . and thus spoils the effect of the proofreading. The loss of the corrected species is avoided if there is a barrier (perhaps conformational) formed by a slow reaction step between the noncorrected E . l'e-AMP and the corrected *E$&Mp. If such a slow conformational change exists, the increase in accuracy from the pretransfer proofreading would be benefitial, and, in addition, the PPi increases the accuracy by optimizing the initial discrimination of the wrong amino acid.

The activation of amino acids by the aminoacyl-tRNA synthetases produces PP,, which is able to bind to the aminoacyl adenylate in the reversc reaction [I]. Both the activation and the pyrophosphorolysis are rapid reactions [2] which allow an equilibrium at this step in the presence of PPi. The PPi strongly shifts the equilibrium from the aminoacyl adenylate intermediates towards the start of the total reaction. PP, is decomposed by the inorganic pyrophosphatase present in all living cells. The pyrophosphatase, however, does not destroy all PP,, but a moderate concentration is maintained. The PP, concentration in Escherichia roli cells, thoroughly studied in this laboratory [3 - 51, is about 0.5 mM and is not changed very much by varying the conditions. The amino acid specificity of the aminoacyl-tRNA synthetases has been described to be increased in four steps [6].Two initial discrimination steps occur before the activation reaction, the pretransfer proofreading works at the aminoacyl adenylate step and the posttransfer proofreading at the aminoacyl-tRNA step (Eqn 1).

E

The effect of PP, on the aminoacyl-tRNA synthetase reaction has already been considered in two earlier papers on this work [7, 81. The present work uses the kinetic equations derived in the preceding paper [9]. The roles of PP, in the initial discriminations and pretransfer proofreading were calculated.

MATERIALS AND METHODS Equations The rate equations were derived as suggested by Cha [lo] and as described in detail in the preceding paper [9]. The reaction scheme was divided in separate rapid equilibrium segments. Since the amino acid specificity is to be examined, the binding reaction for the amino acid makes one boundary of the segments, and the different amino acids, isoleucine and valine, have their own segments (Eqn 2). iRN.4

+ aa + ATP-

ETTp-

"EYTpr> pp,

Initial discrimination 11

Initial discrimination I2

Correspondence to K. Airas, Department of Biochemistry, University o f Turku, SF-20500 Turku, Finland Enzyme. Isoleucyl-LRNA synthctase (EC 6.1.1 3).

Eaa-iRNA +E

Eaa-AMP+>

1

E + aa + A M P

AMP

4

+ aa-tRNA.

+ +tRNA

E aa

Pretransfer proofreading

Posttransfer proofreading

Zl

712

Each box contains the re1evant of the given intermediate and ATP, tRNA or PPi. Eqn (3) specifically shows the segments containing the E . Val-AMP intermediates.

452

/

tRNA

ValAMP

*

ValAMP Eppi IRW

EPPI

Ksp

E

IRNA

K4R

ValAMP

tRNA

ValAMP ppi

h

PP,

EVaUMP

~

E

IRNA

K,

tRNA

_ _ \_

The possible existence of segments of the conformationally changed *E intermediates in Eqn (2) is considered in the Results. The last E . aa-tKNA segments in Eqn (2) were omitted when deriving the rate equations ; their role is insignificant under conditions where the dissociation of E . aa-tRNA is not rate limiting. In [9], such conditions were, e.g. [Mg"] = [ATP] + [PPj] + 1 mM and [spermidine] = 1 mM. The activators, Mg" or polyamines, were not taken into account in the rate equations. This requires conditions where high enough activator concentrations are present to keep the reaction components mainly as MgATP, MgPPi, MgE and (spermidine), . tRNA. The conditions in the cell apparently fulfil this requirement [9]. The rate equations are best derived using determinants in an analogous way as for normal steady-state rate equations Ill]. The discrimination equation of valine (vlle/vVal) and the C terms contained in it are expressed by Eqn (4).

- c 2 i i r i 2 1 (c341 + c 3 2 1 + C31t~)l) (4) where u = [ATP], r = [tRNA] a n d p = [PP,]. The letters I, V and R in the subscripts refer to Ile, Val and tRNA"", respectively. The dissociation constants K1, K,, and K5 are for ATP, PPi and tRNA, respectively and K4Ris for PPi in the presence of bound tRNA, and KRI and KRvare for tRNA in the presence of the amino acids [9]. k + and k - are the rate constants of the activation reaction. knVis the proofreading rate constant for Val-AMP, and kZlis the corresponding dissociation rate constant for the correct species lle-AMP.

Rate-constant and dissociation-constant values

Most of the experimental constant values come from the works of Holler and Calvin [2], Fersht [12] and Freist et al. [13, 141. The following values were used in the calculations: K1 = 300 pM [9],k + 3 = 100 s - ' [2], k-3 = 600 S - ' [2], K4 = 0.3 pM, KRI = 250 FM [2], K ~ R 1250 pM [9], K5 = KR KRV = 5~ I(* [9]. If the binding of the amino acid is written as an one-step process, the dissociation constants are KIle = k-21/k+2;and Kval = k-2v/k+2v,where Klle = 6.8 pM, k-21 = 15 s- and k+,, = 2.2 pM-' s-' [2].The relation between the K, values is Kval = 109 x Klle [2], or Kval = 750 pM. Then k-2v = 750 x k+,,. The values of k-2Y and k + 2 Vwere vaned in the calculations according to this dependence. An increase in the Kval value compared to Krle can be made by an increase in the dissociation rate (k-21 < k - 2 V )or by a decrease in the association rate (k+,, > k+2v).In the calculated results, the COlI = (1 + a/Kl)(k+21 k+21Rr/KR)[11el, equal rate constant values of dissociation, k-2v = k - 2 1 = c o i v = (1 + a/Ki)(k+zv ~ + ~ V R ~ / K R ) W ~ ~ I , 15 s-', mean that the difference depends totally on the association rate. If k-2v = 150s-1 ( = l O ~ k - ~the ~ ) value , of CIOI= (1 + a/Ki)(k-21 ~ - ~ I R ~ / K R I ) , k +zv = k +21/11, then the difference between K1le and Kvalis CIOV = (1 + dK1) (k- 2v f k - ZVR PIKRv). contributed about equally by the increase in the dissociation rate and decrease in the association rate. c 1 2 1 = c12v = a/K1(1 f T/&)k+3, The value of k, 6 = 2 s - ' was used for a mechanism where CZII= c 2 1 v = b/K4 p / K 4 (r/Ks)lk~ 3, the transfer reaction E&iMp F? E&?!& occurs without a conformationally changed intermediate *E:{hAPP191. c231 = c Z 3 V = f P/K4R)(r/K5)k+6C'> The rate constant values for the process (conformational c 3 2 1 = CJZV = (1 P/K~K)~-~c, change) E:&$Mp + *E:&&MPwere estimated both for isoleucine and valine as: kcsn = k+,,, = k-6C1 = kMBCV = 5s-I. C 3 x 1 = (1 + P/K4R)ka1, These values have been chosen to give an approximately corc 3 x V = (1 f p/K4R)kltV, rect calculated rate of the formation of the isoleucyl-tRNA, and the values are low enough to show the expected effect on c341 = c34v = (1 -tP/K&+6, the pretransfer proofreading. The constants are equal and VflelvVal = C341C23iC121Co11[(C12V+ CIOV) ( C 2 3 ~+ C ~ I V ) written quite arbitrarily since there is as yet no basis on which estimate thcm separately. At the transfer reaction, from (c34Y + c 3 2 V f c3aV)-(c12V + clOV)c23Vc32V to *Eaa-AMP forwards, k, 6 = 20 s-' both for isoleucine and lRNA -c21Vc12v(c34V c32v + C3?IV)I/ valine. The pretransfer proofreading factor is xl = 22 1131, from which the proofreading rate constant kqV = 22 x k + 6 = (c231 + c Z l l ) { c 3 4 V c 2 3 V c 1 2 V COlV[(C121

+ + +

+

+

+

+

453

A

B

2000

c

>

'

-

1000

20 I

I

I

f

I

.

-. m

I

a

I

I

I

I

-

zz

...r t

>

I

< [Val] = 5 x [Ilel

Fig. 2. Separate proofreading scgments exist. A. The calculated effect of PP, on the discrimination when both initial discrimination and pretransfer proofreading are taken into account. k-2Y = 50 s - ' , krv = 440 s-', kZl = 0.02 s-', k + b = 20 s-'. (B) The consumption or ATP in the pretransfer proofreading. The rate constants were as in (A).

440 s-' is obtained. These values of k + 6 and knv are chosen to give reasonable calculated discrimination rates. The relative rates from and to the intermediates *E:{L%Mp correspond to the estimates of Fersht [12]; knv is fast, k+6is rather slow but k,,, is rate limiting.

RESULTS AND DISCUSSION Fig. 1 shows calculated PP, dependences for the discrimination of valine by the isoleucyl-tKNA synthetase. The proofreading occurs at the aminoacyl adenylate step, and in this model the amino acid is able to return from the adenylated to the free form through the k - 3 reaction. In Fig. 1 A, only the effect of the initial discrimination is calculated (the proofreading rate is 0). The PP, increases the accuracy so that the discrimination value approaches 11 0, which IS the relationship

between the Ks values (Kval/Klle).The value of the constant k - 2v is not known but the curves for three values are presented. k-2v is the backward rate of the conformational change at the initial discrimination step 12, For isoleucine, this value is low; k-21 = 15 s-' [2]. Jf the k-2Vvalue is 15 s-', then the PP, does not have any effect on the accuracy at the initial discrimination steps. The measured value of the initial discrimination is I1 x I2 = 53 [14]. The calculated value in the absence of PPi is close to this value when k - 2 v = 50 s-'. Fig. 1B shows the effect of PPi when both initial discrimination and pretransfer proofreading cooperate in the accuracy. The increase in accuracy, caused by the pretransfer proofreading, is lost in the presence of PPi. The limit at high PPi concentrations is the same (1 10; Kva1/Kfle) as in Fig. 1A, but it is approached from above. Both in Fig. 1A and B, the effect of PPi is almost maximal at a concentration of 100 pM,

454 which is remarkably lower than the measured PPi concentration in the E. coli cells (500 pM) [3]. In the reaction system of Fig. 2, the aminoacyl adenylate step has been divided in two parts. After the binding of tRNA, a slow conformational change occurs, and after it the pretransfer proofreading is performed, The conformational change functions as a barrier which prevents the proofread adenylate from returning to early steps. Fersht [12] has suggested this kind of model as one explanation for the rates of decomposition of correct and incorrect aminoacyl adenylates or aminoacyl-tRNA:s. According to Freist et al. [15], the AMP production is also Fast in the presence of tRNA modifications which are not able to bind the amino acid, therefore it is apparent that this proofreading really occurs before the transfer reaction. In addition, some other aminoacyl-tRNA synthetases show a ‘conformational activation’ after the binding of tRNA [16], thus the same mechanism could also work in the isoleucyl-tRNA synthetase. In Fig. 2A, the PPi increases the accuracy of the system. Almost maximal discrimination is attained at about 100 pM PPi. The value of the discrimination at [PPi] = 0 is 1020, while it was 1170 in the measurements of Freist et al. [14]. (The value of k P z Vis 50 s P 1 in Fig. 2.) Therefore, the distribution between the initial discrimination and pretransfer proofreading corresponds approximately to the measured values. As mentioned above, the system contains some estimated constants such as the rates for production of the proofreading segment and away from it. Also, the question as to whether the conformational change or the transfer reaction can occur in the presence of a bound PPiremains obscure. Improvements of these details may somewhat change the efficiency of PPi, but qualitatively the effects are as in Fig. 2A. The valine concentration in the E. coli cells is about fivetimes higher than the isoleucine concentration [17, 181. The discrimination at these concentrations is also described in Fig. 2A. The lowered consumption of ATP at the pretransfer proofreading step in the presence of PP, (Fig. 2B) is caused by the optimization of the initial discrimination, which does not consume ATP.The decrease in the requirement for ATP is not very great at these rate-constant values, but the effects of only one wrong amino acid and only the pretransfer proofreading are examined. In the discrimination of valine, posttransfer proofreading is also important [14]. Changes in the estimated constants may affect the ATP consumption remarkably.

The importance of the pretransfer proofreading is emphasized in the discrimination of amino acids other than valine [14]. According to kineticmodels, it is not to be expected that the PPi could affect the discrimination in the posttransfer proofreading. In any case, the transfer reaction acts as a barrier which does not allow a remarkable reverse reaction of the proofread aminoacyl-tRNA. In conclusion, it seems to be necessary to assume that a barrier exists between the pretransfer proofreading segment and the aminoacyl adenylate segment if the pretransfer proofreading at all is thought to be cffective in the cell. Jf all 500 pM of PPi in E. coli is not in contact with the aminoacyl-tRNA synthetases, even 10% contact causes almost the same effects. The effect of PPi in increasing the accuracy is consistent with the equations presented by Hopfield [19], where the initial discrimination is optimized if the reverse reaction of activation ( k - 3 or m in [19]) is fast. REFERENCES 1. Hoagland, M. B. (1955) Biorhim. Biophys. Acta 16, 288-289. 2. Holler, E. & Calvin, M. (1972) Biochemistry I I , 3741 -3752. 3. Kukko, E. & Heinonen, J. (1982) Bur. J . Biochem. 127,347-349. 4. Kukko-Kalske, E. & Heinonen, J. (1985) Int. J . Biochem. 17, 575 -580. 5. Kukko-Kalske, E.; Lintunen, M., Karjalainen, M.: Ldhti, R. & Hcinonen, J. (1989) J . Barteriol. 171,4498-4500. 6. Freist, W. (1989) Biochemistry 28, 6787-6795. 7. Airas, I