J. Mol. Biol. (1967) 28, 479-490

Recognition of tRNA by Aminoacyl tRNA Synthetases

i~. YARUS t AND P. BERG

Del~artment of Biochemistry, Stanford University School of Medicine Palo Alto, California, U.S.A.

(Received 5 June 1967)

A new method is described for the detection of tRNA in complex with an amino. acyl tRNA synthetase. Escherichia coli isoleucyl tRNA synthetase complexes only with tRNA TM and tyrosyl tRNA synthetase complexes exclusively with tRNA TM. ATP and amino acid are not required to establish the complex, which also forms equally well whether the tRi~A is acylated or not. Quantitative analysis of the binding curves suggests a single binding site per enzyme particle which binds either aminoaeylated or unacylated tRNA, both tRNA's having an association constant of approximately 108 liters]mole. The binding technique also provides a short and direct procedure for obtaining tRNA highly enriched for a selected acceptor.

1. Introduction Unambiguous translation of genetic structural hfformation depends equally on the fidehty of two processes: first, amino acids must be ]inked to their cognate transfer RNA's and secondly, aminoacyl tRNA's must be matched to the sequence of the messenger.

In all likelihood specificity is preserved by different means at each of these steps. In the charging reaction a single protein (Baldwin & Berg, 1966a; Calendar & Berg, 1966; Muench & Berg, 1966b; Fangman, Nass & Neidhardt, 1965; Yamane & Sueoka, 1964) selects 1 of 20 amino acids, activates it in an enzyme-bound amlnoacyl adenylate (Norris & Berg, 1964; Allende, A]lende, Gatiea & Matamala, 1964; Lagerkvist & Waldenstrom, 1965), and transfers the aminoacyl group to one, or the few, tRNA's specific for that amino acid. By contrast, matching of aminoacyl tRNA's to the structural message occurs on a ribosome, which is potentially able to bind every aminoacyl tRNA, but binds effectively only the few called for by the codon to be translated (Nirenberg & Leder, 1964). In the first case, the specificity of the charging reaction depends on the selective affinity of each enzyme for an amino acid and tRNA polynucleotide chains. In the second, specificity in translation requires correct base pairing between tRNA and message. Whereas specificity is a famillar property of base pairing interactions, the way a protein--the amlnoacyl tRNA synthetase-- distinguishes among tRNA chains is lmknown.

~or some time we have been interested in the problem posed by recognition of appropriate tRNA's by aminoacyl tRNA synthetases. ~minoacylation of tRNA, while it requires recognition, is not necessarily an assay of recognition. I t is easy to conceive, for example, of alterations of the enzyme or tRNA structure which in- activate the charging reaction without impairing recognition, that is, without impair-

Present address: Department of Chemistry, University of Colorado, Boulder, Colorado 80302, U.S.A.

479

480 M. YARUS AND P. BERG

ing the initial association of protein and t R N A (Torres-Gallardo & Kern, 196, Ve have, therefore, sought and developed a method which allows rapid, specific assay for the recognition step. Our approach utilizes the properties of nitrocellulose filters, which bind aminoacyl t R N A synthetases but not free tRNA's : under suitable con- ditions we follow the formation of specific complexes between the enzyme and its cognate tRNA by measuring the retention of t R N A on filters in the presence of enzyme. The method may be used to evaluate the requirements for the specificity and strength of the binding and has also been used to prepare highly purified tRNA.

Detection of aminoacyl t R N A synthetase- tRNA complex has been recently reported by Lagerkvist, Rymo & Waldenstrom (1966), who used filtration through Sephadex gel to separate complexed from free tRNA.

2. Materials and Methods (a) Enzymes

All the aminoacyl tRNA synthetases were from Escherichia cell: isoleucyl tRNA synthetase was purified as by Baldwin & Berg (1966a); the tyrosyl tRNA synthetase by Calendar & Berg (1966), the leueyl tRNA synthetase was about 100-fold purified according to Bergmann, Berg & Dieckmarm (1961), alanyl tRNA synthetase and tRNA.CCA pyrophosphorylase (Preiss, Dieckmann & Berg, 1961), when required, were provided from a DEAE-cellulose fraction made accord~g to Muench & Berg (1966b) and glycyl tRNA synthetase was a sonic extract. Protein concentration was determined according to Lowry, Rosebrough, Fan" & Randall (1951) with bovine serum albumin as standard. For some experiments the molar concentration of isoleucyl tRNA synthctase was obtained by titration of the enzyme with isotopically labeled isoleueine and ATP. Norris & Berg (1964) have shown that enzyme is quantitatively recovered as an enzyme-isoleucyl AMP complex after exposure to sufficient concentrations of substrates and filtration through Sephadex gels. The amount of isoleucine and ATP, found complexed with the enzyme in equimolar quantities, was taken to be the number of moles of active enzyme.

Purified, concentrated enzyme was diluted in a solution of 0-01 ~-potassium phosphate buffer, pH 7, 50 ~g bovine serum albumin[ml, and 0.010 ~-mercaptoethanol.

(b) Transfer R N A

Unfraetionated tRNA from E. coli B was prepared by a modification of the method of Zubay (1962). This material gives a single mode and symmetrical peals of absorbance and isoleucine accepter activity when filtered through Sephadex G75 or G200. Fractions enriched in various Am~uo acid accepter activities were obtained by gradient partition chromatography as described by Mueneh & Berg (1966a).

Determination of total accepter was carried out on 1 to 10 A260 units of tRNA using the reaction mixture of Berg, Bergmann, Ofengand & Dieckmann (1961), and 20- to 100- fold excess enzyme. At the end of the reaction, 2 to 3 ml. 2 ~-HC1 was added (along with carrier RNA or DNA if less than 3 A2e0 units tRNA were used) and the precipitate was collected on a Whatman GF/C glass-paper filter and washed 6 times with 3 ml. of cold 2 ~r-HC1, then 3 ml. cold 95% ethanol. After drying under a heat lamp, the filters were counted under 5 ml. of toluene-based PP0-POPOP scintillator fluid.

Large amounts of charged tRNA were prepared in 0-5 rot. to 2.0 ml. of the usual charging reaction mixture using I00 to 400 A2so units of tRNA. (In the preparation of alanyl tRNA, the other nineteen common 12C-A.m~nO acids were also present at 0.2 tanole/ml.) Sufficient enzyme activity was used to completely esterify the tRl~A with amino acid in 30 to 60 mill at 37°C. When precipitation of small samples indicated that acceptance had reached its l~m~ting value, reaction mixtures were chilled in ice and an equal volume of phenol, pre-equilibrated with the reaction mixture, was added and the two phases shaken vigorously. The phases were separated by centrifugation at 0°C, the aqueous phase was made 0-5 M-NaC1, and 2 vol. of cold absolute ethanol were added to precipitate the aminoacyl tRNA. The precipitate was collected by centrifugation, dissolved in 0.5 ~-

RECOGNITION OF tRNA 481

NaC1, and then either dia lyzed extensively against water or passed over Sephadex G25 equi l ibrated wi th 0.001 ~oKH2PO ~. The result ing aminoacy l t R N A was stable for several months i f s tored frozen (--15°C) a t 20 to 100 A260 units/ml. Concentrat ion of t R N A is expressed as A2eo, measured in 0.01 N-NaOH: ~(P ) in this solvent is 8.8 × 108 (Ofengand, Dieckmann & Berg, 1961).

(c) ~il tera

Although m a n y types of nitrocellulose filters were found usable for sequestrat ion of e n z y m e - t R N A complexes from the f i l tered react ion miYture, Millipore (Millipore F i l t e r Corp., Bedford, Mass.) filters, t ype HA, 24 m m diameter , give var ied results, some detect- ing complex efficiently and some not a t all, wi th reasonably consistent behavior among filters in the same box. We have rout ine ly used Schleicher & Schuell (Carl Schleicher & Schuell Co., Keene, N. H.) filters, whose behavior is reasonably uniform from ba tch to ba tch and reproducible wi th in a ba tch (a single lot number) . Types MC4, C2, C6, B6 and B13 Schleicher & Schuell filters have been t r ied and the B6, 24 m m diameter have been found to be general ly most efficient. Different lots va ry less than twofold in the efficiency with which they detect the isoleucyl t R N A synthetase- isoleucyl t R N A complex. Much of the work repor ted here was done using lot no. 220/5411. After learning of the detergent content of nitrocellulose filters (Cahn, 1967), filters were ext rac ted using boiling water and thorough washing; these t rea tments had li t t le effect on the efficiency of the binding assay.

(d) B i n d i n g assay

Under the condit ions described below, about 100/zg or less of isoleucyl t R N A synthetase is quant i ta t ive ly re ta ined b y a nitrocellulose filter. Transfer R N A or isoleucyl t R N A ( tRNA charged with isoleucine), under identical conditions, passes through the filter, being quant i ta t ive ly recovered in the f i l t rate and wash. When the enzyme, and either [14Clisoleueyl t R N A or unacyla ted, uniformly [14C]base-labeled t R N A or unacyla ted t R N A labeled in the terminal CpA moie ty wi th 14C, are mixed and then filtered, radio- act ive label is re ta ined on the filter: the amount re ta ined serves as a measure of complex formation. The details of the assay are as follows: the react ion mixture , 0.20 ml., contains 0-044 ~-KH2P04, 0.006 ~-K2HPO4, 0.010 ~-MgC12, 0.010 ~-2-mercaptoethanol , 10 ~g of bovine serum albumiu, 0 to 100 ~g of enzyme prote in and 0 to 400 ~/zmoles of t R N A TM

(0 to 10 A2eo units of unfract lonated tRNA). The react ion miYture, af ter addi t ion of the enzyme, is passed, af ter 15 sec incubation, through a nitrocellulose filter which has been presoaked in wash fluid. The filter is mounted on a perfora ted meta l p la te in a stainless- steel holder through which water a t 17°C is circulated. A 0.2-ml. react ion mixture is drawn through the filter using sufficient suction to give a flow ra te of 1 to 2 ml./min. The react ion tube and p ipe t te are r insed wi th about 50/zl. of wash fluid (containing 0.044 M-KHuPO4, 0-006 ~ -K2HP04 and 0.05 M-MgC12, p H 5.5) and this is passed through the filter a t the same f i l t rat ion rate. Suction is then increased (about 15 ml./min) and 3 ml. (more can be used) of wash fluid (17°C) is drawn through the filter to remove unbound tRNA. After washing, the filters are dr ied under a hea t l amp and counted wi th 5 ml. of a to luene- P P 0 - P O P O P scinti l lator solution, or a l ternat ively , the damp filter is dissolved in 5 ml. of Bray ' s scinti l lator solution (Bray, 1960). Rout ine controls show tha t added aminoacyl t R N A remains fully charged during the binding assay procedure.

Comments on the binding assay: (a) 2-Mercaptoethanol is added to the react ion m~Yture to stabilize the enzyme bu t is not essential for formation of the complex or i ts detect ion on the filter. The concentrat ions of potass ium phosphate and magnesium chloride, and the p H in the react ion and wash are op t imum for the detect ion of E. coli isoleucyl tRI~A s y n t h e t a s e - t R N A TM (E. co~i) complexes. The efficiency of the assay falls rap id ly with increasing pH, or decreasing magnesium chloride and potass ium phosphate, and slowly with increasing potass ium phosphate and magnesium chloride concentration. Mg 2+ is required for detect ion of e n z y m e - t R N A complexes; in contrast , i t is apparen t ly not required for t ransfer of isoleucine from enzyme-isoleucyl AMP to t R N A (Norris & Berg, 1964). Mg 2+, then, m a y be required for preservat ion of the complex on the filter and not for recognition, per Be. (b) Serum albumin in the react ion m ~ t u r e , or a l ternat ively, filtra- t ion of sertun albumin (10 ~g/filter) before f i l t rat ion of the react ion increases the efficiency of detect ing the complex, possibly b y neutral izing or removing something from the filter

~82 M. Y A R U S AND P. B E R G

which disrupts the e n z y m e - t P ~ A complex. (c) The amount of enzyme which can be used is l~m~ted by the binding capacity of the filter. Usually we use about 5 ~g, which is far below capacity and, therefore, is quantitatively retained. (d) The emciency of detedtion of the complex is a function o f ~emperature; 17°C is optimum, so tha t the incubation, filtration, and subsequent wash are usua~y done at this temperature. Maximum binding is achieved virtually instantaneously at any temperature f rom 0 to 45°C. This is in accord with our expectation, sbace the over-all u.m~noacylation reaction ~ vitro under these conditions proceeds at several cycles/see, and binding can be no slower than this. (e) Detec- tion of the complex is not critically dependent on the filtration rate, but slow filtration of the reaction m~x~ure and the use of smaller volumes of wash buffer give slightly more sensitive assays. Once at tached to the filter, the enzyme- tRNA complex is not freely dissociable, since washing of the filter with large amounts of buffer reduces the amount of t R N A retained only slightly. (f) The blank is a reaction in which an equal weight of serum albumin replaces enzyme; these values are usually 0.1 to 1% of those found at enzyme saturation although at lower levels of binding the b]u.n1~ may be 10~o of the experimental value (see legend to Fig. 1). (g) I f ATP and s,mino acid are added to the wash fluid, no transfer of the amino acid to previously bound, unacylated t R N A is detected. Sizn~larly, use of wash fluid containing AMP and pyrophosphate does not remove lsoleucine from previously bound isoleucyl tRl~TA. Finally, pyrophosphate ex- change into ATP is not detected when enzyme bound to the filter is incubated with ATP and o~PP,, even though all these reactions go freely in solution in wash fluid. I t seems likely, therefore, tha t a t tachment to the filter traps t R ~ A in complex with enzyme, but tha t at tachment distorts the complex so tha t normal catalytic activity is not possible.

3. R ~ u l t s

(a) .Formation of isoleucyl t RNA synthetase-4 RNA TM complex

W h e n isoleucyl t R N A synthe tase and [14C]isoleucyl t R N A are rn~×ed and passed th rough a nitrocellulose filter as described above, increasing amoun t s o f 14C are retained b y the filter unti l a p la teau is reached at a molar rat io of added t R N A TM to enzyme of a p p r o ~ m a t e l y two (Fig. 1). The bound ~4C is still insoluble in cold 2 ~-HC1 and, as shown later, in the fo rm of arn~noacyl t R N A . W h e n enzyme is present in excess (molar rat io o f enzyme to t R N A TM ~ 8), the a m o u n t o f label re ta ined is

20 i 18

16

"~ t.4 0

< 12 z

3"

0

2

I I I 1

I ] I I I I0 20 30 40 50

I I i. I J

! q O "~

I I [ 60 70 80 90

/~/~moles Ile-tRNA added

Fro. 1. Binding of isoleucyl tRNA to isoleueyl tRNA synthetase. 6-5 ~g of isoleucyl tRNA synthetase w ~ combined with the indicated amounts of [~4C]isoleueyl

tRNA prepared from Mueneh & Berg (1966a) fraction no. 168 (see Fig. 4), and the mixture assayed for binding. Using bovine serum album~ instead of enzYme , filtration of 7.5 ~maoles of [l~C]isoleueyl tRNA leaves 0.04 ~mole on the filter; filtration of 75/~moles leaves 0'07 ~mole.

R E C O G N I T I O N OF t R N A 483

proportional to the amount of [z4C]isoleucyl tRNA added (Fig. 5). Since the reaction is specific for tRNA TM (see below), this measurement assays the relative amount of tRNA TM (or other tRNA bound by the enzyme). Similarly, with an excess of tRNA TM,

the amount of tRNA bound is linearly related to the amount of enzyme added, and the amount of enzyme capable of the recognition process may be determined.

(b) Spe~'ficity of the binding assay Figure 2(a) shows that neither labeled tyrosyl, glycyl, leuoyl, nor alanyl tRNA is

bound to the filter when mixed with isoleueyl tRNA synthctase. Since these charged tRNA preparations also contain unesterified tRNA TM, which is efficiently bound by the enzyme (see below), it seems that at high concentrations other tRNA's do not compete to an appreciable extent with tRNAn% In addition, no binding occurs at low levels of the added labeled arn~noaeyl tRNA's, where there is enzyme free of tRNA~% We estimate that complexes between unlike partners, ff detected with normal efficiency, would have been observed ff their association constants were greater than or equal to 1/100 that of like partners.

.~ 16[- ~ - t

I I 81- I - I

• ~ ' . ~ c ~ ' ~ I le~_¢16.._L¢ ! I :~e.._l.o__J_e_l 8 l I l e 1 I ©l I e I I¢ I I . I I:,_.1_, 0 10 20 3 0 4 0 50 60 70 80 9 0 1 0 0 1 2 0 1 4 0 1 6 0 1 8 0 0 10 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 1 2 0 1 3 0 1 4 0 1 6 0 1 8 0

/#.zmoles Amino acid-tP,.NA odded (a) (b)

FzG. 2. Binding of various aminoaeyl tRNA's to isoleucyl tRi~A synthetese (a) and to tyrosyl tRNA synthe~ase (b).

6.5 ~g of ~soleucyl tRl~A synthetase (a) or approx. 5 ~g tyrosyl tRNA synthetese (b) were mixed with heterologous amlnoacyl tRNA's prepared from unfraetionated tRNA (E. cell) and binding assayed as described in Materials and Methods. ( e ) Ile--tRNA; (O) Tyr--tRNA; (O) Gly--tRNA; (~)) Leu--tRl~A; (~) Ala--tRNA.

Although the conditions used for detecting complexes between enzyme and tRNA were optimized for the reaction between isoleucyl tRNA synthetase and isoleucyl tRNA, interaction between tyrosyl tRNA synthetase and tyrosyl tRNA can also be demonstrated (Fig. 2(b)). The assay is less efficient, but as with isoleucyl tRNA synthetase, only the homologous combination is retained on the filter: in particular, isoleucyl tRNA, bound efficiently in Fig. 2(a), is not bound with tyrosyl tRNA synthctase.

Conceivably, complexes between other tRNA's and isoleucyl tRNA synthetase do form but are disrupted during binding to the filter and are not detected. This may be tested by measuring the ability of different tRNA's to compete with isoleucyl tRNA for binding to the enzyme. In this ease, only one type of complex, that involving matching partners, is measured throughout, and the event signaling interaction

32

484 M. Y A R U S AND P. B E R G

between isoleucyl t R N A synthe tase and o ther t R N A ' s , t h a t is, competi t ion, will, i f i t occurs, occur in solution.

I n order to consider the compet i t ion wi th unlil~e t R N A ' s , we first show t h a t compet i t ion between t R N A TM a nd [14C]isoleucyl t R N A can provide a quant i ta t ive assay for recognizable R N A in an ~ml~nown mix ture o f t R N A ' s . I f isoleueyl t R N A synthetase is rn~xcd wi th a constant , sa tura t ing a m o u n t of [14C]isoleucyl t R N A and increasing quanti t ies o f unesterified t R N A TM a nd then filtered, using the usual assay procedure, the a m o u n t o f [14C~isoleucyl t R N A reta ined b y the filter is progressively decreased (Fig. 3). I f ( t R N A • ENZ)~ax is the amoun t o f complex detected in the

"6 i " i i .... I .......... i j

I 5 T

~ 4

T z z .k.<3 z z

0 I 2 3 4 5 6 (tRNA i) (tRNA i) ( tRNA) ' ( I l e - tRNA)

~IG. 3. Competition between [14C]isoleucyl tRNA and uncharged tRNA TM for binding to isoleueyl tRNA synthetase.

Enzyme (1.1/~g) was added to reaction m~tures (17°C) containing 41/~/~moles [14C]isoleucyl tRNA and varying amounts of uncharged tRNA having 0"35 m/~mole tRNAne/10A26o units. The amount of enzyme [l~C]isoleueyl tRi~A complex was then determined. This experiment calibrates the assay (determines the slope of the curve) for use with an unknown competitor, as in the experiment of Fig. 4. Explicitly, two similar competitors, one unknown and one known, contain equivalent concentrations of recognizable tRNA when equivalent volumes give the same value of

(tRNA • ENZ)max - - 1. (tRNA • ENZ)

absence o f competi tor , using a sa tura t ing a m o u n t o f isoleucyl t R N A , denoted b y the symbol ( tRNA), and ( t R N A . ENZ) is the a m o u n t o f complex in the presence of an a m o u n t o f inhibitor, denoted ( tRNAI), then, if the sys tem is a t equilibriumS,

[ (tRN-~k ° ENZ)max --1] (~:l) (tRNA,) (t-=R-~ ENZ) - - ( k ) ( t R N A ) '

in which k I and k are the association constants o f the inhibi tor and the s t anda rd t R N A , respectively. The da ta in Fig. 3 are in accord with this simple model. Since

t Further work has shown that the conditions used here might not be suitable for the attain- ment of true equilibrium. In any case, the competition assay is empirically calibrated by the use of known amounts of tRI~TA TM, and can he used to determine the amount of competitor in other samples of tRNA by reference to this standard curve (Fig. 3). The assay used in this way yields the apparent amount of competitor in an ~ml~uown expressed as the amount of the standard competitor tRl~A TM which would have given the same competition. For this reason the assay can yield the absolute amount of unknown if it is similar to the standard competitor, as in this experiment.

RECOGNITION OF tRNA 485

each tRNA excludes the other, it is likely tha t the enzyme may bind both esterified and unesterified tRNA and tha t the binding site for charged and uncharged tRNA is the same. In addition, the slope of the line in Fig. 3 is nearly one. This implies tha t the association constants for charged and uncharged t RN A are similar, tha t is, the ~wo forms of tRNA are recognized equally well. Additional studies (Yarus & Berg, manuscript in preparation) confirm this conclusion.

To determine if tRNA chains other than tRNA n~ can compete with isoleucyl tRNA for binding to isoleucyl tRNA synthetase, t RN A taken from different fractions of a part i t ion ehromatogram (NIuench & Berg, 1966a) was tested as an inhibitor of the binding of [14C]isoleucyl tRNA to isoleucyl tRNA synthetase. Since many tRNA's are resolved by the fractionation method (l~Iuench & Berg, 1966a,b), this procedure tests the competitive ability of all aceeptors, taken a few at a t ime in partially purified form. The results (Fig. 4) show tha t primarily those fractions which contain tRNA ne chains, as measured by their acceptor act ivi ty (lower curve), com- pete with the [14C]isoleucyl tRNA for binding to the enzyme (upper curve). The two

20 I I " I '

._. 16 Competition E

~ o 1 ~

4 400

0 1 , 0 •

" 8 o

0 .:I00 200 300 400 Fraction no.

Fro. 4. Fractionation of ~RNA TM and competitor of isoleucyl tRNA by gradient partition chromatography.

tRNA fractionated by l~uench & Berg (1966a} was used as competitor with [14C]isoleucyl tRNA arid the competitor profile determined, assaying as described for Fig. 3; the acceptor profiles are determined by measuring the isoleucine-aeceptor activity; the data of Muench & Berg (lg66a) and this work are combined to yield the curve in the lower portion and the absorbance profile is that reported by Muench & Berg (1966a).

peaks of tRNA n e measured by acceptor and competitor activities are in appro~mate quanti tat ive agreement, indicating tha t the major species of unacylated tRNA ne bind equally well to the enzyme.

However, there is a reproducible small peak of inhibitor of the enzyme-isoleucyl tRN A interaction in the region containing tRNAAla: these fractions contain little tRN A he, as measured by isoleucine-acceptor act ivi ty with purified or cruder enzyme (the latter also contains RNA-CCA pyrophosphorylase (Preiss et al., 1961) which would restore missing terminal trinueleotide sequences). We do not believe this inhibitor is tRNA A~a chains which can interact with isoleucyl t RN A synthetase, since (a) [l~C]alanyl tRNA is not bound (Fig. 1), (b) under conditions of the binding assay

486 M. YARUS AND P. B E R G

[14C]alanyl tRNA is not broken down in the presence of the isoleucyl tRNA syn- thetase, AMP and PP, and (c) large amounts of isoleucyI tRBTA synthetase, which would be expected to sequester the inhibitor, do not slow the rate of charging of tRNA Ala by alanyl tRNA synthetase under conditions where the concentration of t R N A xl~ l imits the ra te of the charging reaction~. The compet ing mate r ia l in the

t R N A Ala region is, conceivably, a crypt ic or non-accep t ing t R N A Ile of the t ype recent ly described for other accepters b y IAndahl, Adams & Fresco (1966), Ga r t l and

& Sueoka (1966) or Muench (1966). Aside f rom this exception, the compet i tor is always associated with, and perhaps ident ical to, t R N A he.

(c) Analysis of the binding curve

Figure 1 shows t h a t i n the first por t ion of the b ind ing curve (enzyme excess), the

f rac t ion of added [zaC]isoleucyl t R N A b o u n d to the enzyme on the filter is cons tan t a n d close to 1.0. W h e n examlned in more detai l (Fig. 5) the slope, which we assume

measures the e~c iency+ + of detect ing the e n z y m e - t R N A complex on the filter, is

5

<

4 ) ~ 2

o l E

0

"i I I I I l

~ 0.76 -

I . I l ..... I I I t 2 3 4 $ 6 ./~/~moles I I e - t R N A added

:FzG. 5. Binding of [~4C]isoleucyl tRNA by an excess of isoleucyl tRNA synthetase. 13/zg of isoleucyl tRNA synthetase were added to reactions containing small amounts of

[~ 4C]isoleucyl tRNA, and the amount of complex determined.

j" The rate of [14C]alanyl tRNA formation was measured in the binding reaction mixture with the addition of ATP (0.001 ~), :L-[14C]a]anlne (0"00013 ~), tRNA Ala from Muench & Berg (1966a) fraction no. 344 (Fig. 4; there is negligible tRNA TM in this fraction), and alanyl- and isoleucyl- tRNA synthetasss (see Materials and Methods). I t may be calculated from the data in Fig. 4 that, if tRNA Ala were able to bind to isoleucyl tRNA synthetase strongly enough to compete with tRNAn% then the amount of isoleucyl tRI~A synthetase added in this experiment would have markedly slowed the charging of the tRNA Ala by the alanyl tRNA synthetase.

:~ Throughout this paper, "efficiency of the assay", then, refers to the fraction of the accepter which may be bound to the filter by a large excess of enzyme. We believe this reflects a true efficiency rather than, for example, the existence of a fraction of tRNA which is not bound because all tRI~A TM can be shown to interact with enzyme: {a) since our normal substrate is [1~C]iseleucyl tRNA, all tRNA ~e was recognized when it was charged; (b) in a preparation which is only partially bound, all isoleucyl tRNA is discharged by enzyme if AMP and pyrophosphate is added; (c) tRNA which passes through the filter in the presence of excess enzyme will be bound in about the same proportion if more enzyme is added and the solution re-filtered. After several cycles, all the iso- leucyl tRNA can be bound. Therefore, we believe that all tRNA is complexed in the presence of sufficient excess enzyme, and that some of the complexes are not detected because they are disrupted by the filtration process, releasing some tRNA. The effect of prefiltration of serum albumin (see Materials and Methods) suggests that there is a disruptive agent on filters. The linearity of the curve in Fig. 5 shows that efficiency of detection of complex is reproducible and independent of tRNA concentration at low tRNA concentratious; the observation of flat plateaus at high tRNA concentrations (Fig. 2) similarly shows that efficiency of detection is reproducible and independent of tRNA concentration at high tRNA concentrations.

R E C O G N I T I O N OF t R N A 487

0"76 (0'40 to 1"0 with different tRNA's , different filters, and variations in technique). I f we correct all points of the binding curve for the efficiency of binding and plot the reciprocal of the fraction of enzyme converted to complex versus the reciprocal of the t R N A ns concentration (Fig. 6), these two quantities should be related, for an

1 1 1 1 associative equilibrium, by the equation - = ~ d- tRNAne, where ~ is the frac-

t ion of enzyme converted to complex, n is the number of binding sites per enzyme equivalent and k is the association constant of the complex.

Examina t ion of m a n y such curves as tha t in Fig. 6 shows no systematic deviation f rom linearity; the intersection of the curves with the ordinate suggests tha t one t R N A is bound per enzyme molecule, and the slope of the reciprocally plot ted binding curve suggests an association constant of about 108 liters/mole (Table 1). This

TABLE 1

Summary of data from several binding experiments

Temperature Source of tRNA binding /¢ (°C) isoteucyl tRNA Efficiency sites per mole (1./mole)

of enzyme

17 tRNA I t 1.0 0.80 1.1 × l0 s 17 tRNA I 0.83 0.77 0.7 × l0 s 17 tRNA I, at pH 7 0-18 0-87 1.2 × l0 s 17 tRNA II~ 0.70 0.95 0.9 x l0 s 0 tRNA I 0.72 0.74 0.7 × l0 s 0 tRNA I 0.78 0.77 0.6 × l0 s

t Isoleucyl tRNA from Muench & Berg (1966a) fraction no. 168 (see Fig. 4). $ Isoleucyl tRNA from Muench & Berg (1966a) fraction no. 222 (see Fig. 4).

a t te r est imate is within a factor of two to three of the value deduced from the Km determined in the usual kinetic fashion. Table 1 shows tha t independent of the efficiency of assay (0.18 to 1.0), the apparent association constant varies only between 0.6 × 108 and 1.2 × 108 liters]mole and the number of t R N A binding sites per mole of enzyme between 0.74 and 0"95. This measurement, t aken with Norris & Berg's (1964) finding t ha t there is one site for isoleucyl AMP per mole (112,000 g) suggests that , on each enzyme particle, there is a single site for each substrate: tRNA, A T P and isoleucine. The da ta of Table 1 also confirm directly an observation implicit in the competi tor profile of unacylated t R NA' s in Fig. 5; namely, the two major species of t R N A n~ separated b y part i t ion chromatography have similar association con- stants; t ha t is, t hey are recognized equally strongly.

(d) Purification of tRNA TM by its s~ecific retention on nitrocellulose filters

The specific interaction between isoleucyl t R N A synthetase and t R N A n" and retention of the complex on nitrocellulose filters suggest a novel approach to resolu- tion of specific tRNA's . Although the complex is only slowly disrupted and t R N A

488 M. Y A R U S A N D P. B E R G

eluted when washed with the standard wash solution (pH 5.5), tRNA, but not the enzyme, is readily eluted by washing with 0.05 ~-Tris-C1, pH 8.0.

Therefore, 300/~moles ofisoleucyl tRNA synthetase were mixed with 560 ~/~moles of [14C]isoleucyl tRNA and filtered. The filter was washed as in the standard assay, and the washed filter bearing the bound [14C]isoleucyl tRNA was eluted With 1 m]. 0.05 ~-Tris-C], pH 8. The material in the eluant had an ultraviolet spectrum similar to a standard sample of tRNA, and isoleueyl tRNA accounted for about 70% of the absorbance, assuming an ~ (P) of 7950 and a chain length of 77. This represents a 22-fold purification of tRNA he.

The [14C]isoleucine recovered in the eluant was acid-preeipitable and remained so after incubation in eluant; however, just as with the starting [14C]isoleucyl tRNA, incubation in mild alkali (0.3 ~-Na2CO 8, pH 10, 15 m~nutes, 37°C) or incubation with enzyme, 1~gC12, ~ and PP in ehant converted the labeled isoleucine to a form soluble in 2 ~-HC1. Therefore, the tRNA recovered from the filter appears to be bor~ fide isoleucyl tRNA. This experiment yielded about 120 ~moles of tRNA he, but the procedure is easily scaled up by using larger filters or by using several filters at once. Our present technique sacrifices the enzyme, recovering only the tRNA, but the purification is short and convenient and is therefore useful for some purposes. Further details will be reported later.

4. D i s c u s s i o n

The association of aminoacyl tRNA synthetases with their cognate tRNA's (recognition) can be studied by measuring the retention of the complexes on nitro- cellulose filters. The strength and specificity of the interaction suggest that attach- ment to the enzyme's catalytic site is obligatory for adherence to the filter. Thus, recognition can be studied independently of activation and transamluoacylation. In fact, ATP and isoleueine are not needed for the recognition. The effects of substrates on recognition will be the subject of another publication (¥arus & Berg, manuscript in preparation).

Both chromatographically separable tRNA ne species (l~Iuench & Berg, 1966a) bind equally well to the enzyme whether or not they are esterified with isoleucine (Fig. 4, Table 1). The apparent association constant for the enzyme-tRNA complex is approximately 1 × 108 liters/mole. We are attempting to confirm these data by a method not involving binding to a solid substratum. But at the moment, we would argue that our method of analysis is likely to yield true and relevant equilibrium constants because Ca) the high specificity suggests the enzyme's catalytic RNA site is required (Figs 2 and 4), (b) a physically reasonable correction of the data yields a reproducible value for the number of RNA sites per mole (Table I), (c) the shape of the binding curve is that expected for an associative equilibrium (Figs 1 and 6) and (d) the apparent equilibrium constant agrees satisfactorily with another estimator of the same equilibrium, 1/Km (Fig. 6, Table 1). We suspect from our data that a major contribution to the specificity of amluoacyl tRNA synthetases is made by their selectivity in binding tRNA's. While isoteucyl tRNA synthetase can activate and bind both isoleucine and valine (Baldwin & Berg, 1966a)no similar ambiguity is found in its interaction with tRNA's (Figs 2 and 4). In subsequent papers we shall report on structural features of the tRNA and of the enzyme which are involved in and critical for the affinity and specificity of association.

2

1

RECOGNITION OF t R N A

i I t f t 1 't _'l+'+ ++' 4 ~ i'-+o-

k ~ I x lOS:L/mole

~'I/n=1

489

I I I I I I 0 0"I 0:2 0'3 0"4 0"5 0"6 0-7

( I I I I e - t R N A )

FIG. 6. Graphic determination of association constant and tRNA sites/mole enzyme by double reciprocal plot.

5.9/~g of isoleucyl tRNA synthetase (26 ~moles of active enzyme) was mixed with varying amounts of [z4C]isoleueyl tRNA and the amount of complex determined. Observed amounts of bound isoleucy] tRNA were corrected for the efficiency (in this experiment 0.70) and the result subtracted from ]soleucyl tRNA added to yield the free isoleucyl tRNA concentration at each point. The reciprocal of the free isoleucyl tRNA concentration, 1]isoleucyl tRNA, is plotted against the reciprocal of the fraction of enzyme appearing as complex, liP. This method of analysis is analogous to the usual kinetic method, where velocity of reaction is replaced by the fraction of enzyme in complex (~) and Km by the reciprocal of the association constant(1//~).

I t seems l ike ly t h a t a d s o r p t i o n o f maeromolecu les to n i t roce l lu lose fi l ters m a y p r ov ide an a s s a y for o the r in te rac t ions . F o r example , t h e f o r m a t i o n of t h e complex be tween isoleucyl t R N A syn the t a se a n d iso leueyl A M P fo rmed f rom enzyme, A T P a n d isoleucine (Norris & Berg, 1964) is also d e t e c t e d b y i ts r e t en t i on on the fi l ters (Yarus & Berg, u n p u b l i s h e d resul ts) . Other sys t ems in which one of t he componen t s

is f i l terable suggest themselves .

One of us (M.Y.) was suppor ted during the per iod of this work b y U.S. Public Hea l th Service Fellowship 5 -F2-GM-21 , 328-01.

REFERENCES

Allende, J . E., Allende, C. C., Gatica, IVL & Matamala , M. (1964). Biochem. Biophys. Res. Comm. 16, 342.

Baldwin, A. N. & Berg, P. (1966a). J. Biol. Chem. 241, 831. Baldwin, A. N. & Berg, P. (1966b). J. Biol. Chem. 241, 839. Berg, P., Bergmarm, F. H., Ofengand, E. J . & Dieckmarm, M. (1961). J. Biol. Chem. 236,

1726. Bergmann, F. H. , Berg, P. & Dieckrnann, M. (1961). J. Biol. Chem. 236, 1735. Bray , G. A. (1960). AnaIyt. Biochem. 1, 279. Cahn, R. D. (1967). Science, 155, 195. Calendar, R. & Berg, P. (1966). Biochemistry, 5, 1681. Fangman , W. L., Nass, G. & Neidhardt , F . C. (1965). J. Mol. Biol. 13, 202. Gart land, W. J . & Sueoka, N. (1966). P~'oc. Nat. Acad. Sci., Wash. 55, 948. Lagerkvist , U., Rymo, L. & Waldenst rom, J. (1966). J. Biol. Chem. 241, 5391. Lagerkvis t , U. & Waldenst rom, J . (1965). J. Biol. Chem. 240, PC 2264. Lindahl, T., Adams, A. & Fresco, T. 1%. (1966). Prec. Nat. Acad. Sci., Wash. 55, 941.

490 M. YARUS AND P. B E R G

Lowry, O~. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951). J. Biol. Ghem. 193, 265.

Mueneh, K. H. (1966). Cold Spr. Harb. ~ymp. Quant. Biol. 31, 539. Muench, K. H. & Berg, P. (1966a). Biochemistry, 5, 970. Mueneh, K. H. & Berg, P : (1966b). In Procedures in iVucl6ic Acid Re~earch, p. 375. New

York: Harper & Row. Nirenberg, M. & Leder, P. (1964). Science, 145, 1399. Norris, A. & Berg, P. (1964). Proc. Nat. Acad. Sol., Wash. 52, 330. 0fengand, E. J., Dieckmann, J. & Berg, P. (1961). J. Biol. Ghem. 236, 1741. Preiss, J. , Dieckmann, M. & Berg, P. (1961). J. Biol. Ghem. 236, 1748. Torres-Gallardo, J. & Kern, M. (1965). Proc. Nat. Acad. Sci., Wash. 53, 91. Yamane, T. & Sueoka, N. (1964). Proc. Nat. Acad. Sci., Wash. 51, 1178. Zubay, G. (1962). J. Mol. Biol. 4, 347.