An Efficient mRNA-Dependent Translation System from Reticulocyte Lysates

Eur. J . Biochem. 67, 247-256 (1976)

An Efficient mRNA-Dependent Translation System from Reticulocyte Lysates Hugh R. B. PELHAM and Richard J . JACKSON Department of Biochemistry, University of Cambridge (Received April 1, 1976)

A simple method is described for converting a standard rabbit reticulocyte cell-free extract (lysate) into an mRNA-dependent protein synthesis system. The lysate is preincubated with CaCI, and micrococcal nuclease, and then excess ethyleneglycol-bis(2-aminoethylether)-N,N’-tetraacetic acid is added to chelate the Ca2 and inactivate the nuclease. Lysates treated in this way have negligible endogenous amino acid incorporation activity, but 75% of the activity of the original lysate can be recovered by the addition of globin mRNA. The efficiency of utilisation of added mRNA and the sensitivity of the system are both very high. N o residual nuclease activity could be detected, and the tRNA is functionally unimpaired. Several different species of mRNA have been shown to be translated efficiently into full-sized products of the expected molecular weight up to about 200000, and there is no detectable accumulation of incomplete protein products. The efficient translation of RNA from two plant viruses (tobacco mosaic virus and cowpea mosaic virus) required heterologous tRNA. +

The most efficient eukaryotic cell-free protein synthesis system available is the unfractionated reticulocyte lysate [l]. Exogenous mRNA is also translatable in this system, but the fact that it is already saturated with endogenous mRNA means that the added mRNA is only translated to the extent that it can compete with the endogenous mRNA. A further disadvantage is that the added mRNA cannot be assayed by simple radioactive amino acid incorporation methods, but only by analysis of the radioactive protein products synthesised, and this in turn demands that the mRNA under test codes for a protein which is easily resolved from the proteins synthesised on endogenous reticulocyte mRNA. It would clearly be an advantage if the reticulocyte lysate could be treated in some way to remove or inactivate the endogenous mRNA whilst retaining the full activity of all the other components required for protein synthesis. Attempts to achieve this aim by fractionating the lysate to remove endogenous mRNA have resulted in systems which, even when saturated with exogenous mRNA, are considerably less active than the original unfractionated lysate [2]. Two attempts have been made to exploit the greater nuclease sensitivity of mRNA as opposed to tRNA and ribosomal RNA by using low concentrations of pancreatic ribonuclease to destroy endogenous mRNA

[3,4], but removal of the residual nuclease has necessitated fractionation of the lysate with all the disadvantages that this entails. It occurred to us that it might be advantageous to use micrococcal nuclease rather than pancreatic ribonuclease to destroy endogenous mRNA since micrococcal nuclease requires Ca2+ [5] and can therefore be inactivated by the addition of selective chelating agents such as ethyleneglycol-bis(2-aminoethylether)-N,N’-tetraacetic acid (EGTA). In this paper we show that reticulocyte lysates which have been preincubated with CaCl, and micrococcal nuclease, and then supplemented with excess EGTA have negligible endogenous protein synthesis activity yet translate added mRNA with an efficiency approaching that of the untreated lysate.

Abbreuiatioizs. TMV, tobacco mosaic virus: EGTA, ethyleneglycol-bis(2-aminoethylether)-N,N’-tetraaceticacid. Enzymes. Creatine kinase (EC 2.7.3.2); micrococcal nuclease (EC 3.1.4.7); pancreatic ribonuclease (EC 3.1.4.22).

Immediately on thawing the lysate was made 50 pg/ml in creatine kinase and 25 pM in haemin, using stock solutions of 5 mg/ml creatine kinase in 50%

MATERIALS AND METHODS Preparation of Rabbit Reticulocyte Lysates

Lysates were prepared from rabbits which had been made anaemic by acetylphenylhydrazine injection as described previously [l, 61. The lysates were stored under liquid nitrogen in I-ml aliquots. Lysate Incubationsf o r Cell-Free Protein Synthesis

248

aqueous glycerol, and 1 mM heamin in 90% ethylene glycol, 20 mM Tris-HC1 (pH 8.2), 50 mM KC1. The standard system for assay of protein synthesis contained per ml final volume: 0.8 ml lysate (supplemented with creatine kinase and heamin), 0.15 ml of master mix, and either 0.05 ml r H ] leucine (1.6 mM, 206 Ci/mol) or 0.02-0.05 ml [35S] methionine (approx. 10' counts/min). The master mix contained per 0.15 ml: 0.05 ml creatine phosphate (0.2 M), 0.05 ml KCl/MgCl, mixture ( 2 M and 10 mM respectively), and 0.05 ml of a mixture of 19 unlabelled amino acids at concentrations related to their frequency of occurrence in rabbit globin as described in detail previously [l]. Incubation was at 30 "C. To assay protein synthesis, 5-pl samples were withdrawn and delivered into 1 ml glass-distilled water. To this was added 0.5 ml 1 M NaOH containing 0.5 M H,O, (to decolourise the sample) and 1 mg/ml of unlabelled amino acid corresponding to the radioactive amino acid used. After incubation at 37 "C for 15 min, protein was precipitated by the addition of 1 ml 25% trichloroacetic acid. The precipitate was collected by filtration on glass fibre discs (Whatman GF/C), washed with 8% trichloroacetic acid and dried with a warm air blower. The dried filters were counted using a toluene scintillant (5 g/l 2,5diphenyloxazole) in a Packard model 3330 counter at an efficiency of approximately 20% for 3H and 70% for 3sS.

mRNA-Dependent Translation System from Reticulocytes

of 0.1 M CaClz and 1 mg/ml nuclease (dissolved in glass-distilled water, and stored at -20 "C). After thorough mixing it was transferred from ice to a water bath at 20 "C and incubated for 15 min. Then sufficient 1 M or 0.1 M stock solution of EGTA (neutralised with KOH) was added to give a final concentration of 2 mM, and, after mixing, the material was divided into small aliquots which were frozen in liquid nitrogen. To assay mRNA, aliquots of nuclease-treated lysate were taken, supplemented with radioactive amino acid and mRNA (where appropriate), incubated at 30 "C and sampled as for the standard reticulocyte lysate system. Radioactive amino acid and mRNA may be added in a combined volume of up to 0.15 ml per ml of final incubation volume. In some of the trial experiments reported here, smaller volumes of lysate were treated in the same way except that (a) it was convenient to use less concentrated stock solutions of CaC1, and EGTA, (b) an incubation time of 10 min at 20 "C was found more suitable (see Results), and (c) the material was not frozen prior to use. Sometimes the master mix was freeze-dried before adding the lysate: this allowed larger volumes of mRNA to be added later without altering the final composition of the protein synthesis assay. In a few cases the material was preincubated with master mix and 150 pg/ml poly (U) for 5 min at 30 "C before the addition of CaCl, and nuclease.

Wheat Germ Cell-Free System

Dodecylsulphate- Polyacrylamide Gel Electrophoresis

Wheat germ cell-free extracts were prepared by the procedure of Roberts and Paterson [7]. The assay system contained per ml final volume: 0.5 ml wheat germ extract, 0.05 ml of a mixture of KCl (0.6 M), spermidine (16 mM) and dithiothreitol (20 mM), 0.05 ml of ATP/GTP mix (20 mM and 2 mM respectively), 0.05 ml creatine phosphate (0.2 M), 0.05 ml of a mixture of 19 unlabelled amino acids (as used for reticulocyte lysate cell-free system), 0.02 -0.05 ml [35S]methionine (approx. lo7counts/min), and mRNA and glass-distilled water to give the required volume. Incubation was at 30 "C and samples were taken for assay of radioactive acid-precipitable protein as described for the reticulocyte lysate system except that HzOzwas unnecessary, and 0.25 mg bovine serum albumin was added as carrier.

Samples (10 pl) of cell-free protein synthesis assays were analysed on slab gels of the type described by Studier 181, using the gel system of Laemmli [9]. After electrophoresis the gels were dried down on to Whatman 3 MM paper. Radioactive protein were located by autoradiography using Kodak RP Royal X-omat film exposed for 18 h to 6 days, or by fluorography as described by Laskey and Mills [lo], in which case exposure was for 7 - 24 h.

m RNA- Dependent Ret iculocyte System The standard procedure for batch preparation was as follows. To 0.8-ml lysate supplementedwith haemin and creatine kinase as given above, 0.15 ml master mix was added but no labelled amino acids were added at this stage. This was then made 1 mM in CaCl, and 10 pg/ml in micrococcal nuclease using stock solutions

Preparation of Ribosomes by Sepharose 6 B Gel-Filtration

Reticulocyte ribosomes were isolated by gelfiltration of lysate on a column of Sepharose 6 B equilibrated at 2 "C with buffer A (25 mM KCI, 10 mM NaC1, 1 mM MgCl,, 10 mM Tris-HC1 buffer pH 7.2, 0.2 mM dithiothreitol) as described previously [6]. Sucrose Gradient Analysis of Ribosomes

Samples (50 pl) of incubation mixtures were diluted with 250 pl ice-cold buffer A and layered on to linear 15 - 40% sucrose gradients made up in the same

H. R. B. Pelham and R. J. Jackson

buffer. The gradients were centrifuged for 45 min at 50000 rev./min and at 2 "C in the Beckman Spinco SW 50.1 rotor. The gradients were analysed by pumping the contents through a flow cell in a recording spectrophotometer to yield a continuous trace of the absorbance profile at 260 nm. Preparation of mRNA Globin mRNA was isolated by chromatography on oligo(dT)-cellulose according to the rapid method described by Krystosek et al. [4], except that reticulocyte polysomes were obtained by Sepharose 6 B gel-filtration, and no carrier tRNA was added during the ethanol-precipitation step. RNA was prepared from rat liver ribosomes by phenol extraction as described previously [6], and mRNA was likewise purified by chromatography on oligo(dT)-cellulose. Other RNAs were generously provided by the following : tobacco mosaic virus (TMV) RNA by T. M. A. Wilson, fowl plague virus RNA by S. C. Inglis, immunoglobulin mRNA by Dr D.Secher, mouse liver tRNA by Dr T. Hunt, and cowpea mosaic virus RNA by Dr T. Hunt after partial fractionation of material provided by Dr A. Van Kammen. Sucrose density gradient centrifugation by the methods described previously [6], yielded two main fractions each of which was enriched for one of the two components of cowpea mosaic virus RNA [ll], but complete separation was not achieved.

249 Table 1. Ejfhct of CaC12 and EGTA on protein synthmis Standard assays with ['HI leucine were set up with the additions indicated at the final concentrations stated. Incubation was for 20 min Additions

3H incorporation

None 1 m M CaCl, 1 m M EGTA 1 m M CaCI, and 2 mM EGTA

counts/min 10500 2 500 11 800 9 500

Table 2. Efleci ofmicrococcal nucleuse in the uhsmce of added Cci2' Standard assays with [3H] leucine were set up with micrococcal nuclease at the final concentration stated, and 2 mM EGTA where indicated. Incubation was for 20 min. The 'H incorporation in each assay is given as a percentage of that obtained in the absence of nuclease and EGTA. Different lysates and different batches of micrococcal nuclease were used for the two experiments: 100% corresponds to an incorporation of 10500 counts/min in experiment 1. and 8600 counts/min in experiment 2 Micrococcal nuclease concn

3H incorporation in Expt 1

0 0.1 1 10 100

-EGTA

-EGTA

+EGTA

100 106 102 69 11

100 96 94 77 49

111 108 103 107 101

Materials Micrococcal nuclease (8000units/mg) and poly (U) were obtained from Boehringer Mannheim GmbH (Mannheim, West Germany), and EGTA from Sigma Chemical Co. (St. Louis, Mo, U.S.A.). Radioisotopes were obtained from the Radiochemical Centre (Amersham, UK). RESULTS Preparation of mRNA- Dependen t System Since the use of micrococcal nuclease to destroy endogenous mRNA would require the addition of Ca2+and later EGTA to the lysate, we first tested the effects of these compounds on protein synthesis in the absence of nuclease. A concentration of 1 mM CaCI, was chosen as this is about the minimum required for full activation of micrococcal nuclease [5]. This concentration was found to inhibit the rate of protein synthesis quite severely (Table l), probably because the rate of elongation of nascent chains or the termination step was inhibited as judged by the build-up of polysomes seen under these conditions (results not shown). However 1.5-2.0mM EGTA was by itself non-inhibitory to protein synthesis, and completely

prevented the inhibition seen in the presence of 1 mM CaCl, (Table 1). We next examined the effects of various concentrations of micrococcal nuclease on the activity of lysates in the absence of added CaCl,. The extent to which the nuclease inhibited incorporation varied somewhat between different lysates, but partial inhibition invariably occurred at 10 pg/ml and more marked inhibition at 100 lg/ml (Table 2). The fact that 2 mM EGTA completely overcomes the inhibitory effects of even 100 pg/ml micrococcal nuclease (Table 2) implies that endogenous Ca2 in the lysate is responsible for activation of the nuclease in these conditions, and hence that the variability between lysates is probably due to different levels of endogenous Ca2+. These experiments indicated that if a lysate was first treated with 1 mM CaCI, and any level of micrococcal nuclease up to 100 pg/ml, subsequent addition of 2 mM EGTA would completely suppress the potential inhibitory effects of both Ca2' and nuclease. We therefore tested the effects of such a pretreatment at different concentrations of nuclease, different temperatures and different times with the aim of finding preincubation conditions leading to a system which had the minimum amino acid incorporation activity +

250

mRNA-Dependent Translation System from Reticulocytes

Table 3. Effect of nuclease dose and digestion time mRNA-dependent lysate was prepared using the concentrations of micrococcal nuclease and the digestion times given below. In experiment 2 the lysate was preincuhated with poly(U) as described in Materials and Methods. Assays were done with ["S] methionine and an incubation time of 20 min, with 58 kg/ml added mouse liver tRNA, and 200 pg/ml TMV RNA where indicated Expt

Micrococcal nuclease concn

Digestion time

x [35Ss] Methionine incorporation

Stimulation by

-TMV RNA

TMV RNA

+TMV RNA

counts/min

1

10

173 147 125

102 73 83

10

10 15 20

0

10

19

127

7

1 1

10 20

21 15

163 122

6 8

5 5

5 10 20

19 7.5 2.7

165 168 130

9 22 48

1.5 5 10

24 6.2 2.8

164 163 165

7 26 59

I0

2

I .7 2.0 1.5

-fold

5 10 10 10

in the absence of added mRNA and the maximum activity when a standard amount of globin mRNA or TMVRNAwasadded. Preliminary experiments showed that preincubation at 0 'C would entail a rather long preincubation time unless very high nuclease levels were used. On the other hand preincubation with 10 pg/ml nuclease for 5 min at 30 "C or higher, although satisfactory for the majority of experiments, introduced the risk of activation of the translational repressor: when mRNA was added to such systems the initial rate of incorporation was always high, but after 20 - 30 min there was sometimes a decrease which we ascribe to the effect of the translational repressor since the decrease was prevented by high concentrations of 2-amino-purine or cyclic 3' : 5'-AMP, two compounds which antagonise the action of this repressor [12,13]. To overcome these difficulties the compromise temperature of 18 - 20 "C was chosen, and it was found that when small samples (0.1-0.2 ml) of lysate were processed preincubation with 10 pg/ml micrococcal nuclease and 1 mM CaCl, for 10 min was sufficient to reduce the endogenous activity to a very low level and yet retain high activity when mRNA was added (Table 3). With batch preparations (1 -2 ml) the optimum preincubation time was longer (15 min) possibly because of slower temperature equilibration with these larger volumes of material. Preincubation for periods longer than the optimum time resulted in a reduced activity in the presence of added mRNA (Table 3), and it was our experience that a doubling of the preincubation time was more deleterious than the

use of twice the normal level of nuclease (i. e. 20 pg/ml instead of 10 pg/ml). When the same batch of micrococcal nuclease was tested with different batches of lysate the optimum preincubation conditions were found to be independent of the lysate used. When the optimum concentration of nuclease was determined with different batches of enzyme using the standard preincubation time, slight differences were noted but these were not so serious as to make recalibration of each batch an absolute necessity, although such standardisation is very easy to do. Preliminary screening tests suggest that the method described here should be satisfactory when micrococcal nuclease from other commercial sources is used. It may be noted that the preincubations with CaCl, and micrococcal nuclease were routinely supplemented with the master mix (KCI, MgCl,, creatine phosphate and unlabelled amino acids) required for the actual protein synthesis assay. This was done as a matter of convenience, and in fact it proved immaterial to the success of the preincubation step whether the master mix was added to the lysate before or after nuclease treatment (Table 4). The preincubation with nuclease would be expected to convert polysomes into monomeric ribosomes carrying nascent protein chains and attached to fragments of mRNA. It seems likely that such ribosomes would be unable to participate in protein synthesis when mRNA is added at a later stage, and so we considered whether the system might not be improved if the lysate was first incubated with an inhibitor of initiation to

H. R . B. Pelham and R . J . Jackson

251

Table 4. Recovery qfprorein synfhesis on uddition of globin m R N A to mRNA-dependent l,v.sute 0.1 ml aliquots of mRNA-dependent lysate were prepared using a preincubation time of 10 min, and with the master mix added before o r after the digestion with 10 pg/ml micrococcal nuclease. Assays were carried out using either [j5S] methionine or r H ] leucine, with or without globin mRNA at 33 pg/nil. Incubation was for 40 min: zero-time samples were taken whilst the incubations were still on ice Incubation

x

[35S]Methionine incorporation

zero time counts/min Untreated lysate (control) Treated +master mix Treated -master mix

-m R N A

fmRNA

204 8.0 2.4

198 (100) 153 (77) 153 (77)

(xcontrol)

1.5

1 0 ~ x~ [‘HI ’ Leucine incorporation

zero time

- mRN A

fmRNA

21.4 1.o

21.3 (100) 15.8 (74)

counts/min (% control) Untreated lysate (control) Treated +master mix

0.4

allow ribosome run-off before the treatment with micrococcal nuclease. We therefore tried the effect of incubating the lysate with poly(U), an inhibitor of initiation [14] which should be destroyed during the subsequent treatment with CaCl, and micrococcal nuclease. It was found that this incubation with poly(U) offered no advantages over the simpler procedure (Table 3), nor did alternative methods of allowing ribosome run-off using initiation inhibitors which do not have irreversible effects : incubation with sodium fluoride which was later removed by gelfiltration using Sephadex G-50, or incubation with methioninol-AMP, an inhibitor of the formation of methionyl-tRNA which is competitive with respect to methionine and can therefore be overcome by the addition of excess methionine. Sensitivity and Dependence on Added m R N A

The background amino acid incorporation activity of the mRNA-dependent lysate is at most 1 -2% of that of control lysates and usually ceases completely after about 20-min incubation. Some of this background activity seems to be due to a process which is independent of ribosomes since it is not inhibited by sparsomycin. When globin mRNA was added at saturating levels the nuclease-treated lysate synthesised protein at a rate about 75% of that seen in the corresponding untreated control lysate (Table 4). Saturation was achieved at a concentration of 20 pg/ml added globin mRNA (Fig. 1). Although this is approximately twice the level of globin mRNA which we calculate to be present as endogenous mRNA in our lysates, the fact that our globin mRNA preparations are certainly not

150

1

0’

0

20

40

60

80

100

1x)

140

160

[mRNAl ( w l m l )

Fig. 1. Dose-reponse of’mRNA-dependent lysafr with glohin tnRNA. Standard incubations of mRNA-dependent lysate were set up with [“‘S] methioninc and various concentrations of globin mRNA. After 60-inin incubation samples were taken for assay of 35S incorporation, which is plotted against globin mRNA concentrdtion

loo:/, pure shows that added globin mRNA must be utilised by the nuclease-treated lysate with an efficiency not very different from the utilisation of endogenous mRNA in control lysates. At lower doses of globin mRNA, the incorporation was linearly dependent on mRNA concentration (Fig. 1). The lowest level tested (0.3 pg/ml) gave a 2.5-fold stimulation over the background incorporation after an incubation of 1 h. Thus with a 10-p1 incubation volume it is possible to assay amounts of the order of 1 - 10 ng mRNA simply by using amino acid incorporation methods. It should be possible to assay even lower levels of a single species of mRNA if some method such as polyacrylamide gelelectrophoresis is used to distinguish the protein

mRNA-Dependent Translation System from Reticulocytes

252 Table 5. Stimtilation of p h m virul RNA translation by I R N A . Aliquots of mRNA-dependent lysate were incubated with [3sS] methionine at 2.5 times the concentration used in the experiments given in Tables 3 and 4, and with viral RNA at the concentrations stated. Where indicated, mouse liver tRNA was present at 58 pg/ml. Incubation was for 60 min CPMV = cowpea mosaic virus RNA added

Amount

x 35S incorporation

-tRNA pg/ml None TMV RNA 1 TMV RNA 100 CPMV RNAlargecomponent 65 CPMV RNA smallcomponent 65

+tRNA

counts/min 5.1 10.3 119 49 78

9.6 578 296 159

specified by this mRNA from the background incorporation. Even higher incorporation of [3’S] methionine was obtained with added TMV RNA (provided tRNA was also added), presumably because this RNA is richer in methionine codons than is globin mRNA. The stimulation by TMV RNA was usually more than 100-fold over the background level seen in the absence of added mRNA (Tables 3 and 5). Even this high figure may be an underestimate since over 60% of the added [3sS] methionine was incorporated into protein after an incubation time of 1 h, and it seems likely that incorporation was limited by the availability of amino acids. In addition, the assays of background incorporation are probably an overestimate, since the zerotime incorporation is a significant fraction of the total observed after an incubation period of 1 h (Table 4).

Fig. 2. Time course ofprotein synthesis. Control lysate and mRNAdependent lysate, prepared by the batch procedure, were incubated with the additions given below, and with [35S]methionine supplemented with 7 pM L-methionine to obviate depletion of the pool of radioactive methionine. [35S]Methionine incorporation was assayed by taking samples at various times as shown. Control lysate ( 0 ) : mRNA-dependent lysate with no further additions (A);mRNAdependent lysate with 8 pg/ml globin mRNA (O), 32 pg/ml globin mRNA (O), and 32 pg/ml globin mRNA plus 58 pg/ml mouse liver tRNA (m)

Lack ojDamage to tRNA

Lack of Residual Nuclease Actiuity

Protein synthesis in the mRNA-dependent lysate is considerably stimulated by the addition of mouse liver tRNA when TMV RNA or cowpea mosaic virus RNA are used as mRNAs (Table 5). This stimulation by tRNA should not be taken to mean that the preincubation with micrococcal nuclease has damaged and inactivated some reticulocyte tRNAs, since mouse liver tRNA also strongly stimulates the translation of TMV RNA in untreated reticulocyte lysates (T. Hunt, personal communication; see also Fig. 5). The supposition that the reticulocyte tRNAs have not been damaged by the nuclease treatment is reinforced by the observation that the translation of globin mRNA in the mRNA-dependent lysate is uninfluenced by the addition of tRNA even when saturating levels of globin mRNA are used (Fig. 2). The stimulation of the translation of certain viral mRNAs by the addition of tRNA (mouse liver) seems

A critical point to be established with any cell-free system used for the assay of mRNA is the level of nuclease activity. By fragmenting the added mRNA, residual nuclease activity can lead to the synthesis of large amounts of incomplete polypeptide chains ; this complicates the analysis of the products synthesised by the mRNA under test, especially if a mixture of mRNA species is being tested or if there is no prior knowledge as to the products coded by the mRNA in question. Three results show that the system described here has negligible nuclease activity after the addition of EGTA. The first is the time course of protein synthesis, which shows linear incorporation for as long as the control untreated lysate even at low doses of mRNA (Fig. 2). There is no sign of the gradual reduction in rate which would be expected if the mRNA were being degraded by residual nuclease activity.

0

20

40

60

Time (rnin)

to reflect some deficiency in the tRNA complement of reticulocytes and is not merely a consequence of the treatment with micrococcal nuclease.

H. R. B. Pelham and R . J. Jackson

253

0.1

0

w

7

0 A

i

D

Fig. 3. Luck ofnuclruse uctiuity in mRNA-deyendn?t f.wrrte. Reticulocyte polysomes (from Sepharose 6B columns) were incubated at 30 C with an equal volume of mRNA-dependent lysate (A,B), or of wheat germ incubation mix (C,D) in the presence of 75 pM sparsomycin. Samples were taken for sucrose density gradient analysis either prior to incubation (A, C) or after 80 rnin (B, D). The profile of absorbance at 260 nm is shown with a 2.5-fold reduction i n scale where 80-S ribosomes were found in (A) & (B), and a &fold scale reduction in (C)& (D) where the 80-S ribosomes and the ribosomal subunits occurred. Sedimentation direction is from left to right

Secondly, as we shall show in the following section, the products of translation directed by various mRNA species were of discrete size and co-electrophoresed with appropriate markers during dodecylsulphate-polyacrylamide gel electrophoresis. In particular, the products translated from TMV RNA were similar to those observed with the same RNA in untreated control lysate (Fig. 5), and there was no indication that the mRNA-dependent lysate system synthesised a large number of unusually short protein products. Finally, we could find no evidence for residual nuclease activity when we carried out a direct test in which normal reticulocyte polysomes obtained by Sepharose 6B gel-filtration were incubated with an equal volume of mRNA-dependent lysate in the presence of sparsomycin to prevent any breakdown of polysomes as a result of translation. After 80-min incubation at 30 "C there was no sign of any polysome degradation by residual nuclease activity, and the only changes were a slight dissociation of the single ribosomes and a very slight increase in the average size of the polysomes, presumably arising from binding of ribosomes to vacant initiation sites (Fig. 3A, 3B). In contrast, in a similar experiment in which wheatgerm extract was used in place of mRNA-dependent lysate, there was a shift in the polysome distribution from larger polysomes to smaller size classes, which is indicative of nuclease activity in the wheat germ extract (Fig. 3C, 3D).

Analysis of Products Synthesised in the mRNA-Dependent Lysate The products synthesised by the mRNA-dependent lysate with different exogenous mRNAs were analysed by dodecylsulphate-polyacrylamide gel electrophoresis followed by autoradiography. The results are shown in Fig. 4 and 5. Analysis of the radioactive proteins formed in the absence of added mRNA required fluorographic detection methods because the incorporation was very low. One type of product made in this condition is very small peptides which migrate close to the blue marker dye, and which are not labelled in the presence of sparsomycin (Fig. 4). Another protein which is labelled only in the absence of sparsomycin is a component of M , about 22000. It corresponds to a minor radioactive protein seen when incubations of control lysates are analysed, and it therefore seems to be the product of an mRNA which is insensitive to (or protected from) nuclease action. Finally, there is a protein of M , about 42000 which is labelled even in the presence of sparsomycin and is the only product to show increased labelling after about 20-inin incubation (Fig.4). The incorporation of [35S] methionine into this protein is presumably the result of some ribosomeindependent process. The addition of reticulocyte mRNA ('globin mRNA') promoted not only the extensive synthesis of globin but also the synthesis of the same minor products as are made in the control lysate (Fig. 4).

mRNA-Dependent Translation System from Reticulocytes

254 1

220

2

3

4

5

6

7

8

9

10

11

-

130 -

68

-

50

-

H

4036

-

23.5

-

L

-

Fig. 4. Analysis of products synthesised in mRNA-dependent lysure. Assays using [35S]methionine were analysed by polyacrylamide gel of markers run electrophoresis. The figure shows selected tracks from three different slab gels: the positions and molecular weights ( x on each gel are shown. Tracks 1 - 4 are analyses of mRNA-dependent lysate incubated without added mRNA for 20 min (tracks 2844) or 50 min (tracks 1 & 3 ) , in the presence (tracks 3&4) or absence (tracks 1 & 2 ) of 100 pM sparsomycin, using fluorography (24-11 exposure). Tracks 5 - 11 show autoradiograms exposed for 1 day (tracks 5 & 8) or for 7 days, and represent analyses of 60-min incubations of control lysate (track 8), or mRNA-dependent lysate incubated without added mRNA (tracks 6 & 1l), with rat liver mRNA (track 5), with reticulocyte niRNA (track 9),with immunoglobulin heavy chain inRNA (track 1 0), or with 400 pg, ml total R N A from fo\l.l-pla@ue-virus-infected cells (track 7). In track 7 the bands marked are: P protein (P), putative haemagglutinin precursor (HA), nucleoprotein (NP), membrane protein (M) and a probable non-structural protein (NS). H refers to immunoglobulin heavy chain and L to immunoglobulin light chain precursor visible in track 10

TMV RNA directed the synthesis of the same large products ( M , approximately 140000 and 165000) as are seen when this RNA is added to untreated reticulocyte lysates (Fig. 5). In the mRNA-dependent lysate, as in the control lysate, these large products are only obtained in high yield if mouse liver tRNA was added, or if low levels of TMV RNA (1 pg/ml) are used (Fig. 5). The wheat germ system synthesised relatively few high-molecular-weight products and a much higher proportion of apparently incomplete chains even when tRNA was added (Fig. 5). Cowpea mosaic virus RNA also required exogenous tRNA for maximum activity (Table 5). It directed the synthesis of a pair of products of M , about 220000 and another pair of M , about 120000 (Fig. 5). The former seem to be coded by the large RNA component ( M , = 2.02 x lo6),and the latter by the small

component ( M , = 1.37 x lo6), as judged from experiments in which RNA enriched for one or other of these two cowpea mosaic virus RNA species [ll] was tested (Fig. 5). The molecular weights of the products implies that they represent approximately the total coding capacity of the RNAs. It is not clear whether the observed doublets are due to alternative intitiation sites, or some other phenomenon such as read-through of a termination site, heterogeneity of the RNA, or post-translational cleavage. There is similar uncertainty as to the origin of the product of M , about 30000 which seems to be coded by the large RNA species (Fig. 5). When total RNA from fowl-plague-infected chick embryo fibroblasts was tested, the only products detected were viral proteins [15] identified by coelectrophoresis with the appropriate markers (Fig. 4).

H. R . B. Pelham and R . J. Jackson 1

2

3

4

5

6

7

8

9

10

11

12

13

-220 - 130

-68 --5o

- 40 - 36

-23.5

Fig. 5. Trunslution products qfplunt viral R N A s with and vvithout added t R N A . Assays using [35S]methionine and a 60-min incubation time were analysed by polyacrylamide gcl electrophoresis and autoradiography. Selecled tracks from two slab gels are shown with the molecular weights and markers indicated as in Fig.4. Control lysate was used in the case of track 5 , mRNA-dependent lysate for tracks 1- 4 and 6 - 10, and wheat germ system for tracks 1I - 13. Mouse liver tRNA was added at 58pg!mI to the assays shown in tracks 2,4,5,7,9,12.Viral RNA was present as follows: 65 pg/ml cowpea mosaic virus R N A large component (tracks 1 & 2), 65 pg/ml cowpea mosaic virus R N A small component (tracks 3&4), 1 pg/ml TMV R N A (tracks 7 & 8), 100 p'g/ml TMV R N A (tracks 5 , 9 - 12), and n o added viral R N A (tracks 6 & 13)

The haemagglutinin precursor protein does not appear to be processed (cleaved) in this system (Fig. 4). Total poly(A)-containing RNA from rat liver ribosomes stimulated amino acid incorporation in the mRNA-dependent lysate about 20-fold over background, yielding many discrete products most of which were smaller than 65000 M , (Fig.4). Purified immunoglobulin heavy chain mRNA from a mutant line of MOPC 21 (P3K) cells which produce a heavy chain slightly shorter than normal [16], directed the synthesis of a protein product of the expected size (Fig. 4). The main general conclusion to be drawn from these experiments with different mRNA species is that the mRNA-dependent lysate synthesises discrete products of the expected size and does not accumulate incomplete polypeptide products, as appears to happen in the wheat germ system (Fig. 5).

DISCUSSION The mRNA-dependent cell-free system which we have developed seems to have significant advantages over other mRNA assay systems in current use. It is quickly and easily prepared From a standard reticulocyte lysate using commercially available materials. It has very low background incorporation activity in the absence of added mRNA, and yet is very sensitive to quite low concentrations of mRNA. Its utilisation of exogenous mRNA is very efficient, and is comparable with that of the parent reticulocyte lysate, which means that protein synthesis proceeds at a rate close to that observed in whole reticulocyte cells [l].It has negligible nuclease activity and is capable of synthesising very large protein products in high yield with correspondingly low yields of incomplete protein chains. In comparison, the Krebs I1 ascites cell-free system typically has a higher background incorporation

256

H. R. B. Pelham and R. J. Jackson: mRNA-Dependent Translation System from Reticulocytes

activity, a much poorer response to added mRNA, and is also known to contain nucleases which degrade the added mRNA fairly rapidly [17]. Whilst the wheat germ system has a background activity similar to that of the inRNA-dependent lysate described here, it appears less sensitive to added mRNA and its utilisation of added mRNA seems less efficient. Incorporation per unit volume of mRNA-dependent lysate is about five times that obtained under comparable conditions with the wheat germ system in this laboratory. In addition the wheat germ cell-free system has higher nuclease activity (Fig. 3), and produces a much higher proportion of incomplete translation products (Fig. 5). An advantage over frog oocytes [I81 (apart from the ease of mRNA addition) is the very low background activity, which allows rapid assay of even very low amounts of mRNA without the need to isolate the product specified by the mRNA concerned. Although low background activity is a feature of other, highly fractionated, systems derived from reticulocytes [2],such systems are considerably more trouble to prepare, and they seem to be significantly less sensitive and less efficient. Whilst they are extremely useful, if not essential, for the study of the molecular mechanisms of protein synthesis, these fractionated systems are less suitable for the assay of mRNA. Moreover, lack of fractionation leaves the system described in this paper sensitive to regulation by haemin [19], double-stranded RNA [20], oxidised glutathione [12], and depletion of certain metabolites [21], thus facilitating the study of the effects of these regulatory mechanisms on the translation of different types of mRNA. Finally, there seems no reason why the method of preparation of the reticulocyte mRNA-dependent lysate should not be extended to other cell-free systems where removal of endogenous template activity is desirable. We are very grateful to Dr Tim Hunt for valuable advice and for gifts of RNAs, to Paul Farrell for providing [35S]methionine,

and to the many donors of RNAs. This work was supported by grants from the Medical Research Council and the Cancer Research Campaign. H.R.B.P. is the recipient of an M.R.C. Research Studentship.

REFERENCES 1. Hunt, T. & Jackson, R. J. (1974) in Modern Trends in Human Leukuemiu (Neth, R., Gallo, R. C., Spiegelrnan, S. & Stohlman, F., eds) pp. 300-307, J. F. Lehrnanns Verlag, Munich. 2. Schreier, M. H. & Staehelin, T. (1973) J . Mol. Biol. 73, 329 - 349. 3. Crystal, R. G., Nienhuis, A. W., Elson, N. A. & Anderson, W. F. (1972) J . Bid. Chem. 247, 5357-5368. 4. Krystosek, A., Cawthon, M. L. & Kabat, D. (1974) J . Bid. Chtnz. 250, 6077 - 6084. 5. Cuatrecasas, P., Fuchs, S. & Anfinsen, C. B. (1967) J . Biol. Cliem. 242, I541 - 1547. 6. Darnbrough, C. H., Legon, S., Hunt,T. &Jackson,R. J. (1973) J . Mol. Bid. 76, 379-403. 7. Roberts, B. E. & Paterson, B. M. (1973) Proc. Natl Acud. Sci. U.S.A. 70, 2330-2334. 8. Studier, F. W. (1973) J . Mol. Bid. 79, 237-248. 9. Laernrnli, U. K. (1970) Nurure (Lond.) 227, 680-685. 10. Laskey, R. A. & Mills, A. D. (1975) Euu. J. Biochenz. 56, 335-341. 11. Reijnders. L., Aalbers, A. M. J., Van Kammen, A. & Thuring, R. W. J . (1974) Virologj,,60, 515-521. 12. Legon, S., Brayley, A,, Hunt, T. & Jackson, R. J. (1974) Biochem. Biophys. Res. Cornmuit. 56, 745 - 752. 13. Balkow, K., Hunt, T. & Jackson, R. J. (1975) Biochem. Bio-phys. Res. Cominun. 67, 366 - 375. 14. Hardesty, B., Miller, R. & Schweet, R. (1963), Proc. Nut1 Acud. Sci. U.S.A. SO, 924-931. 15. Skehel, J. J . (1972) Virology, 49, 23-36. 16. Cowan, N. J., Secher, D. S. & Milstein, C. (1976) Eur. J . Biochem. 61, 355 - 368. 17. Mohier, E., Hirth, L., Le Meur, M.-A. & Gerlinger, P. (1975) Virolo2y, 68, 349 - 359. 38. Lane, C . D., Marbaix, G. & Gurdon, J. B. (1971) J . Mol. B i d . 61.73-91. 19. Legon, S., Jackson, R. J. & Hunt, T. (1973) Nar. Nen. Biol. 241, 150-152. 20. Hunter, T., Hunt, T., Jackson, R. J. & Robertson, H. D . (1975) . I . Biol. Chem. 250, 409-417. 21. Giloh (Freudenberg), H. & Mager, J. (1975) Biochirn. Bioplzys. Acts, 414, 293 -308.

H. R. B. Pelham* and R. J. Jackson, Department of Biochemistry, University of Cambridge, Tennis Court Road, Cambridge, Great Britain, CB2 1QW

* To whom all correspondence should be addressed.