Ribosome formation from subunits studied by stopped-flow and Rayleigh light scattering.

Light scattering and standard stopped-flow techniques were used to monitor rapid association of ribosomal subunits during initiation of eubacterial protein synthesis. The effects of the initiation factors IF1, IF2, IF3 and buffer conditions on subunit association were studied along with the role of GTP in this process. The part of light scattering theory that is essential for kinetic measurements is high-lighted in the main text and a more general treatment of Rayleigh scattering from macromolecules is given in an appendix.

Introduction

In eubacteria, association of ribosomal subunits and initiation of protein synthesis require the three initiation factors IF1, IF2 and IF3 (1-3). In eukaryotes, subunit association and initiation of translation are more complex and require at least twelve initiation factors (2). All three prokaryotic initiation factors have their corresponding functional homologues in eukaryotes. Initiation factors IF1 and IF2 are close sequence and functional homologues of the eukaryotic initiation factors eIF1A (4) and eIF5B (3), respectively. Initiation factor IF3 has no sequence homology with any of the eukaryotic initiation factors (2). It has, however, several functions in common with eukaryotic eIF3 as well as with eukaryotic eIF1. The latter factor associates with eIF3 in mammals and is one of the subunits of eIF3 in yeast (5).

Termination of protein synthesis in eubacteria is carried out by either one of the class-1 peptide release factors RF1 or RF2 in a stop codon dependent way (6). After peptide release, rapid dissociation of the class-1 release factor is accomplished by the GTP-dependent action of the class-2 release factor RF3 (7-9). Subsequently, the ribosome is split by the combined activities of RRF, EF-G and IF3 (10), making the ribosomal 30S and 50S subunits ready for a new round of initiation of protein synthesis. Here, the 30S subunit, in complex with IF3, binds a messenger RNA, IF1, IF2:GTP and initiator tRNA (fMet-tRNAfMet) in a 30S pre-initiation complex (1), which rapidly recruits the 50S subunit in the formation of a 70S initiation complex. After GTP hydrolysis, IF2 rapidly dissociates from the 70S initiation complex, thereby making the ribosome ready to form the first peptide bond in a nascent protein (11). Subunit joining is an essential step in initiation of protein synthesis, but has in the past received comparatively little attention.

Subunit association or dissociation can be directly monitored by light scattering (12, 13) or ultracentrifugation (14) methods. More recently, Rayleigh light scattering, in combination with stopped-flow techniques, was used to study rapid formation of translation competent ribosomes from different pre-initiation 30S complexes and the 50S subunit (11).

In contrast to the rapid and non-invasive light scattering techniques, ultra centrifugation methods provide little kinetic information on subunit association or dissociation, but have been used to monitor the extent of eukaryotic 80S assembly (15). Ultracentrifugation methods are of non-equilibrium type and subunits, originally in ribosome complexes, may become separated during a centrifugation run. This potential problem is aggravated by the high pressure that develops in the rotor during a centrifuge run which further promotes subunit dissociation (16).

In this work, we describe the principles of Rayleigh light scattering and explain how this method can be combined with stopped-flow techniques to monitor the kinetics of formation or disruption of macromolecular complexes.

We apply the method, using a standard stopped-flow instrument (11), to initiation of protein synthesis in eubacteria, and present novel experiments that high-light the roles of IF3 and GTP on IF2 for selective and rapid 70S initiation complex formation.

The method of light scattering is general and can also be used to study the formation or disruption of macromolecular complexes other than the ribosome.

Materials and Methods

Chemicals and buffers

Nucleoside triphosphates (ATP, UTP, and GTP), radioactive amino acids and unlabelled nucleotides were from Amersham (USA). Non-hydrolysable GTP analogue GDPNP (GMPPNP), CTP, phosphoenolpyruvate (PEP), myokinase (MK), pyruvate kinase (PK), putrescine, spermidine, puromycin dihydrochloride, and non-radioactive amino acids were from Sigma (USA). All other chemicals were of analytical grade from Merck (Germany). Before use in binding and exchange assays, the guanine nucleotides GTP and GDP were further purified as described (8). All experiments were carried out in polymix buffer (17) which has the following final composition: [95 mM KCl, 5 mM NH4Cl, 5 mM Mg(OAc)2, 0.5 mM CaCl2, 8 mM putrescine, 1 mM spermidine, 5 mM potassium phosphate (KP) (pH 7.5) and 1 mM DTE]. One ml of this buffer is prepared by adding 0.1 ml of 10 times polymix, 0.05 ml of 20 times KP and 0.02 ml of 50 mM DTE to 0.83 ml of water. Preparation of 10 times polymix buffer is described in the protocol section. It contains the components of the polymix buffer at 10 times concentration but does not contain KP and DTE to avoid precipitation of calcium phosphate.

Components of the translation system

Synthetic mMFTI mRNA, encoding the tetra-peptide Met-Phe-Thr-Ile, was prepared according to (18). 70S ribosomes, 50S and 30S subunits were prepared from the E. coli strain MRE 600, using sucrose gradient zonal ultracentrifugation according to (19). Initiation factors were purified from overproducing strains according to (20). [3H]fMet-tRNAfMet and Phe-tRNA synthetase (PheRS) were prepared according to (7). Elongation factors EF-Tu, EF-Ts and tRNAPhe were purified according to (21).

Kinetics of macromolecular complex formation analyzed by stopped-flow and light scattering

In a typical light scattering experiment to monitor a binary complex formation between particles of type A and B, a solution containing particles A is rapidly mixed with a solution containing particles B and the intensity of light scattered perpendicular to the beam of illuminating light is recorded as a function of time. Initially, the mixture contains particles A and B at concentrations a(0) and b(0), respectively, while the concentration, c(0), of complex C is zero. The scattering intensity, I(t), at a time t after the mixing is the sum of the scattering intensities from free A-particles, free B-particles and C-complexes:

a(t) and b(t) are the concentrations of free particles A and B, c(t) is the concentration of complexes, C. I, I and I are the scattering intensities per unit concentration for the corresponding particles and complexes. Since, for every C complex that is formed, one particle A and one particle B are consumed, a(t) and b(t) are related to the initial concentrations a(0) and b(0) as: a(t) = a(0) – c(t) and b(t) = b(0) – c(t). Introducing these mass relations in expression [1] gives

where ΔI is the increase in light scattering intensity when the two particles A and B form a complex C. For particles with dimensions much smaller than the wave length of the illuminating light, the scattering intensity of a particle is proportional to the square of its molecular mass and does not depend on particle shape (12) (see also Appendix 1). Since the complex C between particles A and B is just a bigger particle, ΔI can be estimated as:

M and Mare the molecular masses of the A and B particles, respectively, and Z is a proportionality coefficient for the particular experimental set up. Ribosomes and their subunits do not extend more than 30 nm, which is less than one tenth of the wave length (λ= 430 nm) of illuminating light used in most light scattering experiments on these particles (11-13). Therefore, relation [3] holds very well and it follows from this expression that, with x=M, the scattering intensity increases by a factor of 1+2x/(1+x when molecules A and B form a complex. The largest relative increase is by a factor of two, when M so that x=1.

Relation [2] shows that the increase, I(t)-I(0), in scattering intensity with time is directly proportional to the concentration c(t) of formed complexes. When complex formation has reached equilibrium, the plateau value, I, of the scattered intensity is given by

Combining the experimentally measured parameters I(0), I(t) and I, and using the relations [2] and [4] one gets:

The time evolution of the function f, which is one at time zero and zero at infinite time, contains all kinetic information about the complex formation. The ratio f is the difference between the current and the equilibrium concentration of the complex C, normalized to the value of this difference at time zero. Notice that f can be obtained from the experimentally measured intensities I(0), I(t) and I without knowledge of the absolute value of ΔI. Therefore, kinetic experiments can be interpreted without knowledge of the coefficient Z in relation [3], which depends on the experimental set up.

To exemplify the kind of kinetic information one can get from light scattering experiments, we consider the irreversible formation of a complex C from A- and B-particles that have the same initial concentration [A]0

The corresponding rate equation is:

Its analytical solution is

as can be verified by substituting c(t) in [7] with the expression for c(t) in [8]. After a long time all particles A and B will eventually end up in complexes C and an equilibrium concentration of complexes c will be reached.

Substituting the above expressions for c(t) and c into [5] one obtains a very simple expression for f:

Accordingly, a plot of 1/f versus time gives a straight line with slope k, from which the association rate constant k can be obtained from linear regression and knowledge of the initial concentrations [A]. In general, however, it is better to use non-linear regression methods to obtain k along with normalization parameters. For this, the scattered intensity I(t) can be written as:

This relation follows from Eq. [5], with a equal to I and a equal to I - I(0). A best fit of this theoretical expression for I(t) to its experimental counterpart by variation of the parameters k, a and a , e.g. with the Marquardt algorithm (22), gives an estimate of k along with the expected error (standard deviation) of this estimate. The content of this section is all that is required to apply Rayleigh light scattering to the kinetic analysis of macromolecular complex formation. An extended and more detailed description of light scattering theory and its experimental applications can be found in Appendix 1.

Light scattering experiments

Association of ribosomal subunits was monitored with light scattering after their rapid mixing in an SX-18MV stopped-flow instrument (Bio-sequential SX-18MV, Applied Photophysics, Leatherhead, UK) equipped with Xenon arc light source. To study the role of IF3 in subunit association, two mixtures, A and B, were prepared. Mixture A contained 4µM mMFTI mRNA, 0.5 mM ATP, 0.5 mM of GTP, 2 µM 30S and either (i) no additional components, (ii) 4 µM IF3, (iii) 2 µM IF1, 4µM IF2 or (iv) 4 µM IF3, 2 µM IF1, 4µM IF2 and 4 µM [3H]fMet-tRNAfMet as indicated. Mixture B contained 0.5 mM ATP, 0.5 mM GTP and 2 µM 50S. To remove dust particles, the mixtures were spun for 3 min at 14000 rpm in an Eppendorf centrifuge before they were loaded into the syringes of the stopped-flow instrument and pre-incubated at 37°C for at least 5 minutes. After rapid mixing, light scattering at 436 nm at right angle to the illuminating light was recorded as a function of time. The instrument was in emission detection mode, the photomultiplier voltage set to 520 V, and the time constant of the noise reduction filter to 5 ms. The volume of mixes (around 1 ml each) loaded into the syringes of the stopped-flow instrument was sufficient for at least ten independent time traces. Kinetic parameters were obtained for each individual trace by non-linear regression (see the section “Curve fitting” below) and used to obtain average estimates of rate constants and the standard deviations of these estimates.

To study the effects of GTP and GDP on the kinetics of subunit association in the presence of IF2, two mixtures, A and B, were prepared. Mixture A contained 0.5 mM ATP, 1 mM PEP, 0.5 mM of either GTP, GDPNP or GDP, 4 µM [3H]fMet-tRNAfMet, 2 µM 30S, 4µM mMFTI mRNA, 2 µM IF1, 4 µM IF2 and 4 µM IF3. Mixture B contained 0.5 mM ATP, 1 mM PEP, 2 µM 50S. Both mixtures were centrifuged for 3 min at 14000 rpm, loaded into the syringes of the stopped-flow instrument and pre-incubated at 37°C for at least 5 min before fast mixing.

Effects of buffer composition on the kinetics of 70S initiation complex formation were studied as follows. Two mixtures, A and B, were prepared and loaded into the syringes of the stopped-flow instrument. Mixture A contained 2 µM 70S ribosomes, 4µM mMFTI mRNA, 4 µM IF1, 4 µM IF3, 0.5 mM ATP, 0.5 mM GTP and indicated concentrations of PEP and Mg (OAc)2 in polymix buffer. Mixture B contained 0.5 mM ATP, 0.5 mM GTP, 4 µM [3H]fMet-tRNAfMet, 4 µM IF2 and the same concentrations of PEP and Mg(OAc)2 in polymix buffer as in mixture A. Mixture A was pre-incubated for at least 10 min at 37°C to ensure ribosome dissociation into 30S and 50S subunits. The formation of 70S initiation complexes was then initiated by mixing the mixtures A and B in a stopped-flow instrument as described above.

The kinetics of association of naked ribosomal subunits was studied in the following way. Two mixtures, A and B, were prepared and loaded into the syringes of the stopped-flow instrument. Mixture A contained 2 µM 30S ribosomal subunits, 0.5 mM ATP, 0.5 mM GTP, 1.5 mM PEP and indicated concentrations of Mg(OAc)2 in polymix buffer. Mixture B contained 2 µM 50S ribosomal subunits, 0.5 mM ATP, 0.5 mM GTP and the same concentrations of PEP and Mg(OAc)2 in polymix buffer as in mixture A. The formation of 70S initiation complexes was then initiated by mixing mixtures A and B in a stopped-flow instrument as described above.

Dipeptide-Formation assay

The effect of IF3 on the formation of translation-competent 70S initiation complexes was also studied with a dipeptide formation assay. To this end, two mixtures, A and B, were first prepared. Mixture A contained 0.5 mM ATP, 2 mM PEP, 0.5 mM GTP, 1.5 µM 30S, 2.5 µM mMFTI mRNA, 2.5 µM IF1, 2.5 µM IF2. Mixture B contained 0.5 mM ATP, 2 mM PEP, 2 µM 50S, 3 µM EF-Tu, 5 µM tRNAPhe, 30 µM phenylalanine, 1 µg/ml PK, 0.1 µg/ml MK and 10 U/ml PheRS (1 U of PheRS aminoacylates one pmol of tRNA per second). Then, 2.5 µM [3H]fMet-tRNAfMet was added either to mixture A or to mixture B. After pre-incubation for 10 min at 37°C the mixtures A (0.025 ml) and B (0.025 ml) were loaded into a quench flow instrument (KinTech, USA), mixed and quenched after the indicated times by 50% formic acid. The samples were centrifuged and the amount of formed fMet-Phe-tRNAPhe in the pellet was determined by HPLC as described previously (18).

Curve fitting

The association rate constant (ka) for subunit association in the stopped-flow light scattering experiments was estimated by non-linear regression (22) or by the Origin Program, using the three-parameter relation Eq. [10].

Results

IF3 as anti-association factor

In the absence of initiation factors, the 30S:mRNA complex associated with the 50S subunit with an association rate constant ka=1.2 μM-1 s-1 (Fig. 1A). Fig. 1C shows that the addition of IF1, IF2 and GTP to the 30S:mRNA complex resulted in a faster association of 30S with 50S (ka=4.1 μM-1 s-1). In the presence of only IF3, there was no complex formation between 50S and 30S:mRNA alone (Fig. 1B), or together with IF1, IF2 and GTP (not shown).

Fig. 1

The anti-association activity of IF3.

The extent of 70S initiation complex formation was monitored as a function of time by light scattering after rapid mixing in the stopped-flow instrument of a volume containing 30S subunits, mRNA, GTP and initiation factors as indicated with a volume containing 50S subunits. Time traces obtained with no initiation factors added to the 30S subunits (Panel A), with only IF3 added (Panel B), with IF1 and IF2 added (Panel C) and with IF1, IF2, IF3 and [3H]fMet-tRNAfMet added (Panel D).

The results of these experiments, summarized in Table 1, demonstrate the ability of IF3 to block subunit association when the 30S pre-initiation complex lacks initiator tRNA. When, however, fMet-tRNAfMet was present in the pre-initiation 30S:mRNA complex together with IF1, IF2 and GTP, the block was removed and the ribosomal subunits joined with an association rate constant k = 12 μM-1s-1 (Fig. 1D).

Table 1

Association rate constants of 50S subunits with 30S:mRNA in the presence of different combinations of Initiation Factors and fMet-tRNA.

Factors added	kass(μM-1s-1)
none	1.2 ± 0.13
+IF3	<0.0
+IF1+IF2:GTP	4.1 ± 0.3
+IF1+IF2:GTP+IF3	< 0.01
+IF1+IF2:GTP+IF3+fMet-tRNA	12.2 ± 1.6
+IF1+IF2:GTP+IF3+fMet-tRNA (*)	8.6 ± 1.2
+IF1+IF2:GDP+IF3+fMet-tRNA (*)	0.13 ± 0.03
The association rate constant kass was calculated as an average (kav ) over rate constants (ki ) obtained from n (between six and eight) individual time traces of light scattering in a stopped-flow experiment. The error is the standard deviation σ obtained from the variance, estimated as
(*) The reaction buffer contained 1 mM phosphoenolpyruvate (PEP).

The vital importance of IF3 for proper initiation of protein synthesis in eubacteria is further illustrated by quench-flow experiments that monitored the rate of dipeptide formation in the absence of IF3 (Fig. 2). In one experiment, pre-initiation 30S:mRNA complex together with fMet-tRNAfMet, IF1 and IF2 was mixed with 50S complex and all factors needed for peptide-bond formation (Fig. 2A). In an otherwise identical parallel experiment, fMet-tRNAfMet was present in the 50S, rather than in the 30S, mixture when the rate of peptide bond formation was followed (Fig. 2B). The rate of peptidyl-transfer was fast in the former (Fig. 2A), but virtually zero in the latter experiment (Fig. 2B). The absence of dipeptide formation in the second experiment reflects the rapid formation of a translationally inactive 70S complex lacking initiator tRNA. This complex is of the type seen with light scattering in Fig. 1C. This inactive ribosome complex was, in other words, formed before initiator tRNA had time to bind to the 30S:mRNA pre-initiation complex.

Fig. 2

Rate of initiation in the absence of IF3 monitored by di-peptide formation.

The extent of dipeptide formation was monitored as a function of time after rapid mixing in a quench flow instrument of a volume containing 30S subunits, mRNA, GTP, IF1 and IF2 with an equal volume containing 50S subunits. Initiator tRNA was present either in the 30S or in the 50S mix. Time curves obtained with [3H]fMet-tRNAfMet added with 30S (Panel A), [3H]fMet-tRNAfMet added with 50S (panel B).

The role of GTP in subunit association

It was recently shown that GTP on IF2 is important for fast association of a 30S pre-initiation complex with the 50S subunit (11). Those experiments were carried out with the functionally active β-form of IF2, lacking part of the N-terminal domain of the α-form of the factor (23, 24). Since differences in the GTP dependency of these factors cannot be excluded, we present here a similar study, but with a his-tagged version of the full-length α-form of IF2. Pre-initiation 30S complexes were formed with IF1, IF2, IF3, mRNA and fMet-tRNA in the presence of GTP, GDP or the GTP analogue GDPNP. Subsequently, these were rapidly mixed with 50S subunits in the stopped-flow instrument and the intensity of the scattered light was recorded. The rate constant for subunit association was around 8.5 μM-1 s-1 with GTP (Fig. 3A) or GDPNP (not shown) and 0.13 μM-1 s-1 with GDP (Fig. 3B). This means that GTP accelerated subunit formation sixty-fold compared to the rate obtained with GDP, in line with our previous results with the β-form of IF2 (11), but in contrast to results obtained by others (25).

Fig. 3

The effects of G-nucleotides on the association of 30S pre-initiation complex with 50S subunits.

The extent of 70S initiation complex formation was monitored as a function of time by light scattering after rapid mixing of pre-initiation 30S complexes with 50S subunits in a stopped-flow instrument. Traces obtained with GTP (Panel A) and GDP (Panel B).

Dependence of the subunit association rate constant on buffer composition

It is well known that the concentration of magnesium ions, ionic strength and composition of the buffer have a profound effect on the rate and accuracy of protein synthesis (21). It has also been demonstrated that the association rate constant of ‘empty’ ribosomal subunits increases by almost an order of magnitude when the Mg2+ concentration in the buffer increases from 4 to 8 mM (13). It was therefore of considerable interest to study the effect of Mg2+ and other components, like phospho-enolpyruvate (PEP), usually included in buffers for in vitro translation on the rate of formation of ‘real’ 70S initiation complexes. In the light scattering experiments described below ribosomes were first dissociated into their subunits in the presence of mRNA and initiation factors IF1 and IF3 in a buffer of indicated composition and then initiation factor IF2:GTP was added together with fMet-tRNA to dissociated 70S ribosomes in the stopped-flow instrument.

Fig. 4 shows that the rate constant of subunit association increased by only 50% from 7 µM-1s-1 to about 10 µM-1s-1 when the free Mg2+ concentration increased from 3 to 7 mM in our standard polymix buffer (compare Fig. 4A and 4B). At the same time, the addition of 10 mM PEP to polymix buffer (Fig. 4C) resulted in a three-fold decrease in the rate constant for formation of the 70S initiation complex to 2.3 µM-1s-1.

Fig. 4

The effects of buffer composition on the rate of 70S initiation complex formation.

The extent of 70S complex formation was monitored as a function of time by light scattering after rapid mixing of mixture A containing dissociated 70S ribosomes together with IF1, IF3 and mRNA with mixture B containing IF2:GTP together with fMet-tRNA. Complex formation in polymix buffer (PM) with 3 mM of free Mg2+ (Panel A), in PM buffer with 7 mM of free Mg2+ (Panel B) and in PM buffer with 3 mM free Mg2+ plus 10 mM of PEP (Panel C).

An unexpectedly modest effect of Mg2+ on the association rate of 50S subunits with pre-initiated 30S complexes (see Table 2) compared to the large effects seen for association of “naked” 30S and 50S subunits observed by Wishnia et al. (13) may be due to the presence of mRNA, fMet-tRNA or initiation factors in the 30S pre-initiation complex. Alternatively, the difference could be due to the different ionic milieu in the polymix buffer compared to that in the buffer used in (13). The latter work employed a simple TMN buffer containing 10 mM Tris pH 7.5, 50 mM NH4Cl, 7 mM β-mercaptoethanol and different concentrations of MgCl2.

Table 2

Dependence of association rate constants of 70S initiation complex (IC) or naked 70S ribosomes on buffer conditions. 70S initiation complexes were formed from 30S and 50S subunits in the presence of mRNA, IF1, IF2:GTP, IF3 and fMet-tRNA.

Free Mg++	PEP	kass (μM-1 s-1)	70S type
3	0	6.9± 0.7	IC
7	0	9.6± 1.5	IC
3	10	2.3± 0.4	IC
3	1.5	10.1± 0.9	“naked”
7	1.5	18.3± 1.8	“naked”
The association rate constant kass and errors were calculated as for Table 1.

To discriminate between these possibilities we have measured the association rate of naked 30S and 50S subunits in polymix buffer containing either 3 or 7 mM of free Mg2+. The results shown in Fig. 5 clearly demonstrates that the increase in Mg2+ concentration in polymix buffer from 3 to 7 mM results in about 80% increase in the subunit association rate from approximately 10 µM-1s-1 to 18 µM-1s-1 which is comparable to the 50% increase observed for pre-initiated 30S complexes and 50 subunits (see Table 2). Thus, the different effect of Mg2+ on subunit association in the two buffer systems is probably due to the difference in composition of the buffers and not to the presence or absence of initiation factors, mRNA or fMet-tRNA.

Fig. 5

The effects of the level of magnesium on the rate of naked 70S formation formation from its subunits.

The extent of 70S complex formation was monitored as a function of time by light scattering after rapid mixing of mixture A containing 30S ribosomes with mixture B containing 50S subunits. 70S formation in polymix buffer (PM) with 3 mM of free Mg2+ (Panel A), in PM buffer with 7 mM of free Mg2+ (Panel B).

Discussion

This work demonstrates the power of combining stopped-flow and light scattering techniques for experimental studies of how ribosomal subunits join during initiation of protein synthesis. Light scattering techniques were used in early experiments to determine how the equilibrium constant for subunit association depends on translation initiation factors (12). The kinetics of association of naked 30S and 50S subunits and its dependence on buffer conditions have previously been studied with stopped-flow techniques (13) and the effect of IF3 on the rate of ribosome splitting has been addressed with light scattering and manual mixing (12). However, under near-physiological conditions used in our in vitro experiments (21), subunit association catalyzed by initiation factors (Figs. 1 and 2; Antoun et al. (11)) and ribosome splitting, catalyzed by EF-G, RRF and IF3 (10), are rapid processes and their study therefore requires the combination of stopped-flow techniques and light-scattering. Here, we used stopped-flow with light scattering techniques to demonstrate the anti-association property of IF3 in the absence of initiator tRNA (Fig. 1), and complemented these measurements with quench-flow experiments, performed under similar conditions, to follow the rate of formation of the first peptide bond after initiation of protein synthesis (Fig. 2). The IF3 dependent block in the association of ribosomal subunits can be removed by the presence of initiator tRNA, IF1 and IF2. In cases when the subunits associate in the absence of IF3 and initiator tRNA the formed 70S ribosomes are unable to participate in protein synthesis. This suggests that IF3 plays a fundamental role in preventing premature ribosome formation in the absence of initiator tRNA.

We also demonstrated the fundamental role of GTP for fast subunit association catalyzed by the α-form of IF2 during initiation of eubacterial protein synthesis, in line with previous results obtained with the β-form of IF2 (11). During exponential growth of bacteria, ribosomes load on to the 5’ end of an mRNA each four seconds (26). The distance between ribosomes in a polysome is around 230 nucleotides (26). With a rate of 20 codons/s for protein elongation (27) and the need to clear the occluded ribosome binding site (1) to allow for the binding of the next ribosome, the lower limit for the initiation rate is about 0.3 s-1. This rate includes 30S docking to mRNA, fMet-tRNA and IF2 binding and subunit joining. Taking into account the results in Table 1 and that the concentrations of free ribosomal subunits in the cell are around 1 μM (1) one can conclude the rate of subunit joining catalyzed by IF2:GTP observed here is compatible with the rate of initiation in vivo.

The presented light scattering experiments show also the importance of a proper choice of buffer conditions to study the kinetics of ribosomal reactions. All experiments presented in this paper were performed in polymix buffer that mimics the ionic milieu of the bacterial cell (21). We have found, however, that some additional, supposedly “neutral,” components, often included in standard in vitro translation systems, like phosphoenolpyruvate (PEP) may result in considerable alteration in the rate of 70S complex formation (Table 2). Even an addition of small amounts of PEP (1 mM) to the reaction mixture results in a noticeable decrease of the rate of association of pre-initiated 30S complexes with 50S subunits (Table 1). The reason for this PEP effect is not clear at present. The effect of an increase of the free Mg2+ concentration from 3 to 7 mM in polymix buffer on the rate of 70S initiation complex was however much more modest, i.e. only 50%. This 50% effect is, nevertheless, quite comparable with an 80% increase in the association rate for “naked” ribosomal subunits upon the same increase in free Mg2+ concentration (see Table 2). Comparison with published data (13) on association of “naked” subunits shows, however, that the increase in free Mg2+ concentration from 3 to 7 mM results in a drastic increase in the association rate of ‘naked’ subunits from 0.63 μM-1s-1 to about 20 μM-1s-1. This discrepancy is likely to be due to the absence of organic polyamines such as putrescine and spermidine in the TNM buffer system [10 mM Tris pH 7.5, 50 mM NH4Cl, 2 to 8 mM MgCl2 and 7 mM β-mercaptoethanol] used in the previous work (13). Addition of polyamines corresponds, to a first approximation, to an effective increase of Mg2+ concentration in the buffer since polyamines mimic the most important, electrostatic, contribution of Mg2+ in shielding phosphates of rRNA and reducing the electrostatic repulsion between the subunits (13). The association rate constant of 10 μM-1s-1 at 3 mM free Mg2+ in polymix buffer is similar to the rate constant of 9.2 µM-1s-1 observed by Wishnia et al. (13) at 5.5 mM of free Mg2+. It seems therefore more appropriate to compare our results with those in TNM buffer upon the increase of Mg2+ from 5.5 to 9.5 mM. Published data (13) show that the association rate of naked subunits plateaus around 7.5 mM Mg2+ reaching 22 μM-1s-1. If this value of 22 μM-1s-1 is really a plateau, we will get a very good agreement for the effect of Mg2+ on subunit association in two different buffer systems.

The experiments described here were performed with a standard stopped-flow instrument (SX-18MV, Applied Photophysics, Leatherhead, UK) in fluorescence mode. The required ribosome concentration was in the μM range, and all solutions were centrifuged for 3 min at 14000 rpm to remove dust particles and aggregates before they were loaded into the syringes of the stopped-flow instrument. The stopped-flow measurements successfully covered a broad range of subunit association times, from 10 ms to 30 min.

As described, the scattering intensity is proportional to the molar concentration of particles and to the square of their molecular weight. Therefore, the scattering intensity from a 1 μM solution of 30S subunits (Mw 900 kD) equals the scattering intensity from a 100 μM solution of 90 kD proteins. Accordingly, light scattering methods can be used to monitor the association kinetics also of proteins with considerably smaller molecular weights than the ribosome and its subunits, albeit at higher protein concentrations and with a larger investment in the total amount of protein.