Photoaffinity polyamines: interactions with AcPhe-tRNA free in solution or bound at the P-site of Escherichia coli ribosomes

Abstract

Two photoreactive derivatives of spermine, azidobenzamidino (ABA)-spermine and azidonitrobenzoyl (ANB)-spermine, were used for mapping of polyamine binding sites in AcPhe-tRNA free in solution or bound at the P-site of Escherichia coli poly(U)-programmed ribosomes. Partial nuclease digestion indicated that the deep pocket formed by nucleosides of the D-stem and the variable loop, as well as the anticodon stem, are preferable polyamine binding sites for AcPhe-tRNA in the free state. ABA-spermine was a stronger cross-linker than ANB-spermine. Both photoprobes were linked to AcPhe-tRNA with higher affinity when the latter was non-enzymatically bound to poly(U)-programmed ribosomes. In particular, the cross-linking at the TψC stem and acceptor stem was substantially promoted. The photolabeled AcPhe-tRNA·poly(U)·ribosome complex exhibited moderate reactivity towards puromycin. The attachment of photoprobes to AcPhe-tRNA was mainly responsible for this defect. A more complicated situation was revealed when the AcPhe-tRNA·poly(U)·ribosome complex was formed in the presence of translation factors; the reactivity towards puromycin was stimulated by irradiating such a complex in the presence of photoprobes at 50 µM, with higher concentrations being inhibitory. The stimulatory effect was closely related with the binding of photoprobes to ribosomes. The results are discussed on the basis of possible AcPhe-tRNA conformational changes induced by the incorporation of photoprobes.

INTRODUCTION

Previous studies on the interaction of polyamines and polyamine analogs with tRNA aimed to determine the number and the affinity of binding sites (1–6), as well as the effect of polyamines on tRNA structure (7–10). As a result of these studies, it appears that tRNA^Phe, in its crystalline L-shaped conformation, strongly binds three molecules of spermine. One spermine is located at the major groove of the double-helical region in the anticodon stem, while the second spermine is found in the deep pocket formed by bending of the D-stem nucleosides towards those of the variable loop; the third spermine is apparently wrapped in the major groove of the helix formed by the acceptor stem and the TψC stem. Two additional weak binding sites in the major grooves of the TψC stem and the anticodon stem have also been predicted theoretically (11). In solution, two classes of spermine binding sites in tRNA^Phe have been determined by equilibrium dialysis or gel filtration techniques; two to five spermine molecules bind very strongly and in a cooperative manner, while many others bind with lower affinity and non-cooperatively (9). In contrast, ¹H-NMR analysis showed that only one spermine molecule binds tRNA^Phe (6); the binding site was localized at the TψC stem. This variance can be explained by the fact that the binding of polyamines to tRNA depends on buffer components such as Mg²⁺, intercalators, temperature and pH.

Despite the fact that much has been accomplished in the elucidation of polyamine interactions with tRNA in its crystalline conformation or in solution, interactions between polyamines and tRNA bound to ribosomes have never been investigated. Transfer RNA interactions with ribosome are central to the translational process. The ribosome must position the peptidyl or initiator tRNA in such a way that the ester bond linking the peptidyl moiety of the tRNA is chemically accessible to the free aminoacyl group of the incoming aminoacyl-tRNA. Polyamines, Mg²⁺ and other buffer components influence the biological function and positioning of tRNA on the ribosome. For instance, E-site binding of tRNA on the ribosome has been visualized by cryoelectron microscopy only under conditions of a polyamine buffer (12). The use of conventional buffer conditions prevents a stable occupation of the E-site and orientates deacylated tRNA binding in another site, termed E2-site. Polyamines, except the tRNA positioning and binding to ribosomes, also regulate the in vitro translation process at several levels (9,13). We have previously demonstrated that spermine in a cell-free system derived from Escherichia coli, not only promotes the formation and stabilization of the initiator ribosomal ternary complex at 6 mM Mg²⁺, but also has a sparing effect on Mg²⁺ requirements (14). Furthermore, we have established that spermine affects the extent of puromycin reaction and influences the activity status of peptidyltransferase, in a manner depending on the experimental conditions used (15). In the absence of translation factors, spermine inhibits the puromycin reaction (partial non-competitive inhibition), whereas in the presence of translation factors there is a biphasic dose response. However, the precise and specific interactions between spermine and AcPhe-tRNA bound to ribosomes remain unknown, and the mechanism of spermine action has never been completely elucidated.

To gain insight into tRNA conformational changes induced by the binding of polyamines, footprinting analysis could be a useful strategy. However, several practical difficulties have to be overcome, when such an analysis has to be performed. These include the sensitivity of many chemical probes to compounds containing amino residues, and the necessity of using such probes under conditions which can be damaging to nucleic acids or ribosomal proteins (16). On the other hand, the astonishing resistance of Phe-tRNA, either free (17) or complexed with ribosomes (18), to chemical probing is a major problem. Moreover, polyamines are small molecules and exchange rapidly compared to the time scale of analytical techniques.

Here, we follow another approach in an attempt to circumvent these problems. Two photoactivatable analogs of spermine, synthesized by linking an arylazido group at one of the terminal amino groups of spermine, are used for mapping of spermine binding sites in AcPhe-tRNA, free or bound at the P-site of E.coli poly(U)-programmed ribosomes. Under mild irradiation conditions, the photoprobes preferentially and covalently link in regions of AcPhe-tRNA where spermine specifically binds. The amount of probe in excess is then removed and the sites of covalent binding are localized by conventional techniques. Furthermore, the functionality of photolabeled AcPhe-tRNA in peptide bond formation is evaluated and compared to that obtained under conditions where spermine, free in solution, reacts reversibly with the initiator ribosomal complex.

MATERIALS AND METHODS

Materials

Poly(U), puromycin dihydrochloride, heterogenous tRNA from E.coli W, spermine tetrahydrochloride and N-(5-azido-2-nitrobenzoyl)-N-oxysuccinimide were purchased from Sigma and 4-aminobenzonitrile was obtained from Sigma-Aldrich. [¹⁴C]spermine tetrahydrochloride and l-phenyl-[2,3-³H]alanine were obtained from Amersham. Cellulose nitrate filters (type HA, 24 mm diameter, 0.45 µm pore size) were from Millipore. Ribonucleases T₁, U₂, PhyM and RNase from Bacillus cereus were from Pharmacia.

Synthesis of photoprobes

Azidonitrobenzoyl-spermine (ANB-spermine) was synthesized from N-(5-azido-2-nitrobenzoyl)-N-oxysuccinimide and spermine and purified according to Morgan et al. (19). Azidobenzamidino-spermine (ABA-spermine) was synthesized from methyl 4-azidobenzoimidate and spermine and purified on a sulfopropyl-Sephadex column (20). Methyl 4-azidobenzoimidate was obtained from 4-azidobenzonitrile and methyl alcohol in anhydrous ether (21). When desired, the azido group of ABA- and ANB-spermine was reduced with dithiothreitol to an amino group (rANB-spermine and rABA-spermine) by using a procedure described by Morgan et al. (19).

Biochemical preparations

Partially purified translation factors, salt-washed (0.5 M NH₄Cl) and polyamine-depleted ribosomes from E.coli cells and crude Ac[³H]Phe-tRNA charged with 15.8 pmol of [³H]Phe per A₂₆₀ unit were prepared as described previously (22). Purified Ac[³H]Phe-rRNA was prepared essentially according to Rheinberger et al. (23). Uncharged tRNA molecules in the AcPhe-tRNA preparation were oxidized at their free 3′-terminal hydroxyl groups by sodium periodate (24). Initiation ribosomal complex, i.e. the Ac[³H]Phe-tRNA·poly(U)·ribosome complex (complex C), was prepared and purified through adsorption on cellulose nitrate filters by established procedures (14). The filters were washed quickly with three 2 ml portions of reaction buffer (100 mM Tris–HCl pH 7.2, 6 mM magnesium acetate, 100 mM NH₄Cl, 6 mM 2-mercaptoethanol) and the trapped radioactivity was measured in a liquid scintillation spectrometer.

Peptide bond formation assay and first-order analysis

The peptidyltransferase activity of ribosomes was assessed by the puromycin reaction performed at 25°C in the presence of 6 mM Mg²⁺ (25). Under these conditions, the reaction between complex C and excess puromycin (S) proceeds as an irreversible pseudo-first-order reaction.

C + S C′ + P

At each concentration of puromycin, the relationship

1

holds, where k_obs is the first-order rate constant, t is the reaction time, and x′ is the percent of the bound Ac[³H]Phe-tRNA that is converted to product. It should be mentioned that the value of x′ is corrected, taking into account the parallel inactivation of complex C during the puromycin reaction and the inter-vention of other species, except for complex C. In the presence of reduced photoprobes or when the puromycin reaction is carried out with photolabeled complex C, the first-order rate constant obeys the equation

2

The term k_max represents the catalytic rate constant of peptidyltransferase which is a function of photoprobe concentration (25), while K_s is the dissociation constant of the encounter complex between puromycin and complex C. The values of k_max and K_s were determined from the double-reciprocal plot of equation 2 by linear regression.

Photoaffinity labeling of AcPhe-tRNA, free or complexed with poly(U)-programmed ribosomes

Photoaffinity labeling of AcPhe-tRNA with ANB-[¹⁴C]spermine or ABA-[¹⁴C]spermine was performed by using a 366 nm (low pressure mercury tube 8W, CAMAG UV Cabinet) or a 300 nm (Ultraviolet-B, LL 40W/12RS, Philips) light source, respectively. In the latter case, a filter was used to cut off shorter wavelengths. The light source was placed ∼5 cm over a microtiter tray containing the sample and placed on an ice-water bath. In a typical experiment, the irradiated mixture (15 µl) contained 50 mM HEPES–KOH pH 7.2, 6 mM magnesium acetate, 100 mM NH₄Cl, 13.44 A₂₆₀ units/ml AcPhe-tRNA and ANB-[¹⁴C]spermine or ABA-[¹⁴C]spermine at final concentrations ranging from 0.05 to 2 mM (total radioactivity: 80 000 c.p.m.). Before irradiation the mixture was incubated for 10 min at 25°C, in the dark. When complex C was irradiated, 32 A₂₆₀ units/ml ribosomes, 320 µg/ml poly(U) and 0.4 mM GTP were added to the above mixture. When desired, partially purified translation factors (400 µg/ml protein) were also included. After irradiation within specified time intervals, 100 µM dithiothreitol was added to scavenge any unreacted photoprobe. The photolabeled AcPhe-tRNA was discharged from non-covalently bound photoproducts by gel-filtration over a Sephadex G50 column. When complex C was irradiated, the photolabeled product was filtered through cellulose nitrate filters and washed extensively with cold buffer containing 50 mM HEPES–KOH pH 7.2, 100 mM NH₄Cl, 6 mM MgCl₂, 6 mM 2-mercaptoethanol and 10 mM spermine. The adsorbed radioactivity was measured in a liquid scintillation spectrometer.

Kinetics of photolabeling

At each concentration of free photoprobe (F), the amount of covalently linked photoprobe (B) was calculated from the adsorbed radioactivity into the target molecule. The value of B was corrected by subtracting the non-specific binding. The time of irradiation was sufficient to assure that the incorporation of photoprobe had reached the maximum level. Since the photoprobe was present in excess, the reaction with the target molecule proceeded as an irreversible pseudo-first-order reaction

nF + R (Step 1) RF_n RF_n* (Step 2)

where R is the target molecule, RF_n is the encounter complex between the target molecule and photoprobe (non-covalently attached) and RF_n* represents the irreversible complex between R and F (covalently attached). Step 1 represents the equilibrium attainment during preincubation in the dark, whereas step 2 describes the irradiation process. The equation that relates B and [F] is:

3

where [F] is the concentration of free photoprobe, B_max is the maximum level of covalently linked photoprobe, and K_D is the overall dissociation constant of RF_n complex. The [F] value was calculated by subtracting the concentration of bound photoprobe from the total concentration. The values of B_max, K_D and n were obtained by non-linear regression fitting of data to equation 3, using the Hill hyperbola function of the Microcal origin 4.0 program (Microcal software, Inc.). The number of binding sites on AcPhe-tRNA was estimated by dividing the B_max values by the amount of the target molecule.

Mapping of cross-linking sites of photoprobes in AcPhe-tRNA

AcPhe-tRNA photolabeled with non-radioactive photoprobe was deacylated (24), 3′-end labeled with [5′-³²P]pCp and T4 RNA ligase (26), purified by gel electrophoresis on 8% polyacrylamide/7 M urea gel, excised and ethanol precipitated. The pellets were washed twice with cold ethanol, dissolved in 10 µl of appropriate buffer and digested with RNases T₁ (guanine specific), U₂ (adenine specific), PhyM (uracil/adenine specific) and RNase from B.cereus (cytosine/uracil specific). The solutions were quenched on ice with 3 µl of loading buffer (5 mM Tris, 5 mM boric acid, 1 mM EDTA, 0.01% xylene cyanol F, 0.01% bromophenol blue and 10 M urea), heated to 95°C for 2 min and analyzed on 8% polyacrylamide/7 M urea gel. The gels were subjected to autoradiography, and the intensity of bands was quantified by image analysis. When complex C was photolabeled, the photoproduct was adsorbed on a cellulose nitrate filter, washed and then treated with 2 mM puromycin for 10 min at 25°C. This treatment specifically deacylates P-site bound AcPhe-tRNA. Total tRNA was eluted from the filters as described by Peattie and Herr (17), phenol extracted and ethanol precipitated. Subsequently, the tRNA was analyzed as described above.

Kethoxal probing

Photolabeled AcPhe-tRNA, free or bound at the P-site of poly(U)-programmed ribosomes, was treated with kethoxal for 50 min at 37°C, under native conditions (24). The AcPhe-tRNA was isolated as described previously, deacylated and 3′-end-[³²P]labeled. Parallel experiments with non-photolabeled AcPhe-tRNA, but treated as above, were also carried out. The modification sites were analyzed by partial T₁ digestion. For the calculation of relative intensities, the data were processed according to Dabrowski et al. (27). In this set of experiments, purified AcPhe-tRNA was used (23) and its specific binding at the P-site was performed according to Dabrowski et al. (27).

Gel mobility assay of photolabeled tRNA

tRNA^Phe was isolated from complex C photolabeled with 300 µM ANB- or ABA-spermine and was 3′-end-[³²P]labeled. One-half of the labeled tRNA^Phe was incubated at 37°C for 20 min in buffer containing 50 mM HEPES–KOH pH 7.2, 20 mM magnesium acetate, 100 mM NH₄Cl and 6 mM 2-mercaptoethanol, whereas the second half was denatured by heating to 90°C for 15 min, followed by cooling in an ice bath. The Mg²⁺ concentration was then normalized at 6 mM and the solution was resolved on an 8% non-denaturing polyacrylamide gel, according to Friederich et al. (28).

RESULTS

Covalent binding of ANB- and ABA-spermine to the AcPhe-tRNA

The chemical structures of photoprobes are given in Figure 1. The photoreactive azido groups of ANB- and ABA-spermine are positioned ∼8 and 9 Å from the N¹-amino group of spermine, respectively. In the dark, the photoprobe interacts in a reversible fashion with AcPhe-tRNA. Appropriate irradiation of the arylazido group triggers the formation of nitrene, a highly reactive species, which reacts immediately and covalently with a neighboring group of the target molecule (19). Control experiments demonstrated that both AcPhe-tRNA and complex C remain functional when irradiation is carried out in the absence of photoprobes. Neither AcPhe-tRNA nor complex C covalently bind photoprobes in the dark, during the time required for a typical experiment.

Figure 1.

Chemical structures of ANB-spermine and ABA-spermine.

The kinetics of the AcPhe-tRNA photoreaction with 0.6 mM photoprobe, is presented in Figure 2. The photoincorporation of ANB- and ABA-spermine levels off after 2 and 10 min irradiation periods, respectively. The incorporation of photoprobes into AcPhe-tRNA is prevented up to 80% by the addition of a 250-fold excess of spermine in the reaction mixture. Reduced labels at the azido group (rANB-spermine, rABA-spermine) were more potent competitors than spermine, hindering labeling at only 10-fold excess (data not shown). The level of photoincorporation in the presence of excess spermine was used to measure non-specific binding and the values obtained were subtracted. The amount of specifically bound photoprobe to AcPhe-tRNA was used to estimate the ratio v/F, where v is the pmoles of photoprobe bound per pmol of AcPhe-tRNA and F is the concentration of free photoprobe. Figure 3 shows the plots of v/F versus v (Scatchard plot analysis) for the photoincorporation of ANB- or ABA-spermine to free AcPhe-tRNA. In both plots a pronounced ‘hump’ appears, which is characteristic of strong cooperativity between the binding sites (1). Therefore, the data were further analyzed using equation 3 (see Materials and Methods). This equation is described by a sigmoidal hyperbola (Fig. 4) and is mathematically equivalent to the Hill equation of enzyme kinetics (29). Table 1 summarizes the results obtained from this analysis in terms of the number of cooperatively interacting sites, the overall binding constant associated with these sites and the molecular interaction coefficient for photoprobe binding to AcPhe-tRNA or complex C.

Figure 2.

Kinetics of photoincorporation of ANB-spermine and ABA-spermine into AcPhe-tRNA. Purified AcPhe-tRNA (363 pmol) was incubated in the dark with 0.6 mM ABA-spermine (open circles) or ANB-spermine (solid circles) in irradiation buffer (50 mM HEPES–KOH pH 7.2, 6 mM magnesium acetate, 100 mm NH₄Cl) at 25°C for 10 min, and then was irradiated at 300 and 366 nm, respectively. Dashed lines, photoprobe incorporation in the presence of excess spermine (non-specific binding).

Figure 3.

Scatchard plot of ANB-spermine or ABA-spermine photoincorporation into AcPhe-tRNA. Purified AcPhe-tRNA was photolabeled with varying concentrations of ANB-spermine (circles) or ABA-spermine (squares), as indicated in Figure 2. F, the concentration of free photoprobe; v, pmoles of photoprobe bound per pmol of AcPhe-tRNA.

Figure 4.

ANB-spermine photoincorporation into AcPhe-tRNA at various concentrations of photoprobe. Circles, specific binding; squares, photoincorporation in the simultaneous presence of excess spermine.

Table 1. Hill-plot analysis for photoincorporation of (a) ANB-spermine or (b) ABA-spermine into AcPhe-tRNA or complex C.

Target molecule	Nc	KDc (mMn)	nc
(a) Photolabeling with ANB-spermine
AcPhe-tRNA	2.8 ± 0.1	1.02 ± 0.05	2.46 ± 0.25
Complex Ca	396.0 ± 5.0	0.93 ± 0.03	1.32 ± 0.02
Complex C*b	131.0 ± 4.0	0.60 ± 0.04	1.75 ± 0.02
(b) Photolabeling with ABA-spermine
AcPhe-tRNA	3.9 ± 0.1	0.59 ± 0.02	2.92 ± 0.35
Complex Ca	111.4 ± 2.3	0.77 ± 0.01	1.48 ± 0.02
Complex C*b	66.9 ± 1.0	0.39 ± 0.01	1.93 ± 0.03

^aAcPhe-tRNA·poly(U)·ribosome complex prepared in the absence of translation factors.

^bAcPhe-tRNA·poly(U)·ribosome complex prepared in the presence of translation factors.

^cN, K_D and n represent the number of binding sites, the overall binding constant and the Hill coefficient, respectively.

Biochemical properties of photoprobes

Acylation of spermine at the N¹ position results in a derivative affecting peptidyltransferase by the same mechanism, but less active than the native counterpart (25). Therefore, the activity of photoprobes was tested and compared to that of spermine. In these tests, instead of native photoprobes, we used rANB- and rABA-spermine. These compounds fully retain all the biochemical properties of photoprobes and allow the experiments to be conducted in the light. In summary (data not shown), the reduced photoprobes exhibit the following behavior. First, both rANB- and rABA-spermine promote the binding of Ac[³H]Phe-tRNA to poly(U)-programmed ribosomes. Moreover, rABA-spermine exhibits a sparing effect on the Mg²⁺ requirements for AcPhe-tRNA binding; addition of rABA-spermine in the binding buffer results in a decrease of the Mg²⁺ optimum (10 mM) to ∼6 mM. Secondly, rANB- and rABA-spermine significantly enhance the stability of complex C in a dose-dependent manner. This effect is more pronounced in the absence of translation factors. Last, the reduced photoprobes display both inhibitory and stimulatory effects on peptidyltransferase activity, depending on the absence (partial non-competitive inhibition) or the presence of translation factors (non-essential activation in concert with partial non-competitive inhibition). Although weaker, these properties are reminiscent of the behavior of spermine (15).

Effect of AcPhe-tRNA and complex C photolabeling on the catalytic properties of peptidyltransferase

Photo-incorporation of ANB- or ABA-spermine into complex C formed in the absence of translation factors, inhibited the activity of peptidyltransferase by decreasing the catalytic rate constant (k_max), while retaining the affinity constant (K_s) value (Table 2). Experiments with complex C prelabeled in each one of its components, revealed that sizeable inhibition occurs, particularly when AcPhe-tRNA is pretreated with the photoprobes (Table 2).

Table 2. Kinetic parameters of AcPhe-puromycin synthesis carried out with complex C formed in the absence of translation factors and partially or totally photolabeled with ANB- or ABA-spermine^a.

Complex C species	Photoprobe concentration (µM)	kmax (min–1)
Not labeled (control)		1.60 ± 0.05
Totally labeled with ANB-spermine	50	1.41 ± 0.04
	300	1.15 ± 0.04
Totally labeled with ABA-spermine	50	1.45 ± 0.03
	300	1.29 ± 0.04
Labeled in AcPhe-tRNA with ANB-spermine	50	1.50 ± 0.03
	300	1.20 ± 0.03
Labeled in AcPhe-tRNA with ABA-spermine	50	1.52 ± 0.04
	300	1.32 ± 0.04
Labeled in poly(U) with ANB-spermine	300	1.68 ± 0.05
Labeled in poly(U) with ABA-spermine	300	1.67 ± 0.05
Labeled in ribosome with ANB-spermine	300	1.55 ± 0.05
Labeled in ribosome with ABA-spermine	300	1.58 ± 0.02

^aThe k_max values were obtained from the corresponding reciprocal plots of equation 2. The K_s value was invariable and equal to 665 ± 31 µM.

In experiments where complex C was formed in the presence of translation factors, photolabeling of complex C at 50 µM ANB- or ABA-spermine stimulated the peptidyltransferase activity, while photolabeling at 300 µM decreased the k_max value without affecting the K_s value (Table 3). Pre-photolabeling of ribosomes with ANB- or ABA-spermine at 50 µM, resulted in a highly active form of complex C. In contrast, pre-photolabeling of translation factors or poly(U) did not affect the catalytic properties of peptidyltransferase. Furthermore, pretreatment of AcPhe-tRNA with photoprobes caused only inhibition (Table 3).

Table 3. Kinetic parameters of AcPhe-puromycin synthesis carried out with complex C formed in the presence of translation factors and totally or partially photolabeled with ANB- or ABA-spermine^a.

Complex C species	Photoprobe concentration (µM)	kmax (min–1)
Not labeled (control)		2.22 ± 0.06
Totally labeled with ANB-spermine	50	3.26 ± 0.02
	300	1.40 ± 0.08
Totally labeled with ABA-spermine	50	3.46 ± 0.03
	300	1.55 ± 0.05
Labeled in AcPhe-tRNA with ANB-spermine	50	2.31 ± 0.04
	300	1.58 ± 0.03
Labeled in AcPhe-tRNA with ABA-spermine	50	2.42 ± 0.05
	300	1.68 ± 0.03
Labeled in ribosome with ANB-spermine	50	2.71 ± 0.03
	300	2.12 ± 0.02
Labeled in ribosome with ABA-spermine	50	2.95 ± 0.04
	300	2.20 ± 0.03

^aThe k_max values were estimated as indicated in Table 2. The K_s value was again constant and equal to 665 ± 31 µM. Prelabeling of poly(U) or translation factors with photoprobes had no effect on the catalytic properties of peptidyltransferase.

Localization of the ANB- and ABA-spermine binding sites on AcPhe-tRNA free in solution or bound at the P-site of poly(U)-programmed ribosomes

In order to perform meaningful comparisons with previous kinetic results (15), the present experiments had to be carried out under strictly identical conditions. Consequently, a crude AcPhe-tRNA preparation was used. Prior to use of this preparation, uncharged tRNA molecules were oxidized at their 3′-terminal hydroxyl groups. Oxidation inhibited the ligation with [5′-³²P]pCp, so that uncharged tRNAs could not interfere in the subsequent gel analysis. On the other hand, <85% of the AcPhe-tRNA bound to ribosomes reacted with puromycin. This finding revealed that >15% of the bound AcPhe-tRNA was localized at the A-site. Thus, photolabeled complex C was treated with 2 mM puromycin for 5 min, at 25°C. This treatment specifically deacylated P-site bound AcPhe-tRNA, so that only P-site bound tRNA could be isolated in a form able to interact with [5′-³²P]pCp. The distribution of AcPhe-tRNA between P- and A-site did not change during the time-course of photolabeling or puromycin reaction. This precludes any spontaneous translocation of AcPhe-tRNA from the A- to the P-site. In addition, the reactivity of A-site bound AcPhe-tRNA towards puromycin is very slow and even at 25°C this reaction can be considered negligible (30).

The cross-linking sites of spermine analogs on AcPhe-tRNA were localized by enzymatic probing with RNases T₁, U₂, PhyM and RNase from B.cereus and by gel electrophoretic fractionation of the digestion products. Decreased band intensity indicates restriction of the cleavage reaction due to the covalent attachment of a photoprobe near or at the respective nucleoside. We found that photolabeling of free AcPhe-tRNA with 50 µM ANB-spermine protects nucleosides U12, C13, G15, G19, D20, G22, A23, G24, C25, A26, G27, A31, G34 and C48. (Fig. 5A). Probing with 300 µM ANB-spermine conferred protection at the same nucleosides, but to a greater extent. Moreover, the intensity of additional bands was reduced (S⁴U8, A9, C11, G18, A21, C49, C56, G57 and A58), while only moderate protection was observed at nucleosides of the TψC and acceptor stems. The pattern of protection did not significantly change when we used ABA-spermine instead of ANB-spermine (data not shown). In general, the degree of ABA-spermine cross-linking in AcPhe-tRNA was stronger than that of ANB-spermine. Moreover, except for G27, the next three guanosines (G28, G29 and G30) were protected, whereas nucleosides G19, G34 and A58 showed faint protection. In conclusion, the deep pocket formed by nucleosides of the D-stem and the variable loop, as well as the anticodon stem and the TψC-loop, are preferable polyamine binding sites for AcPhe-tRNA free in solution.

Figure 5.

Gel electrophoretic fractionation of products from the enzymatic cleavage of AcPhe-tRNA, free or bound at the ribosomal P-site and photolabeled with ANB-spermine. (A) AcPhe-tRNA photolabeled with 50 µM (lanes 6, 7, 8 and 9) or with 300 µM ANB-spermine (lanes 10, 11, 12 and 13), was deacylated, 3′-end-labeled with [5′-³²P]pCp and digested with RNase T₁ (lanes 6 and 10), U₂ (lanes 7 and 11), PhyM (lanes 8 and 12) and RNase from B.cereus (lanes 9 and 13); lanes 1, 2, 3 and 4 correspond to AcPhe-tRNA non-photolabeled, but digested with RNases T_1, U₂, PhyM and RNase from B.cereus, respectively; lane 5, alkaline hydrolysis ladder. (B) Complex C was photolabeled with 50 µM (lanes 6, 7, 8 and 9) or with 300 µM ANB-spermine (lanes 10, 11, 12 and 13). The bound AcPhe-tRNA at the P-site was isolated and treated as described in (A). (C) Differences in the protection patterns of AcPhe-tRNA, free or bound at the P-site, and photolabeled with 300 µM ANB-spermine. The relative intensities were calculated according to Dabrowski et al. (27). Briefly, the intensities of all G bands obtained from native AcPhe-tRNA in solution were determined relative to that of G10 (vertical comparison). Likewise, C43, A14 and U50 were the reference bands for all Cs, As and Us, respectively. The intensity of a band of photolabeled AcPhe-tRNA, free or bound at the P-site, relative to the corresponding band of native tRNA (horizontal comparison) was multiplied by the value derived from the vertical scanning of native tRNA. The result of multiplication gives the intensity of this band of photolabeled AcPhe-tRNA relative to the reference band of native AcPhe-tRNA. Red, positions of P-site bound AcPhe-tRNA where the respective relative intensity was <30% that of the corresponding band of the tRNA in solution (strongly interactive nucleosides with ANB-spermine); blue, nucleosides with diminished accessibility (relative intensity >200%); yellow, positions with no difference in the accessibility; green, nucleosides in both free and P-site bound AcPhe-tRNA, non-interactive with ANB-spermine; gray, not determinable positions. The data were taken from (A) and (B).

With the exception of C49 and nucleosides localized at the anticodon loop, whose protection was ceased, the binding of AcPhe-tRNA to the P-site of 70S ribosomes facilitated the interaction with photoprobes (Fig. 5B and C). By increasing the photoprobe concentration, the intensity of bands decreased, in particular of those corresponding to the TψC stem/loop (U50–G53, C56–G65) and the 3′-strand of the acceptor stem (U66–G71). Little differences in the protection pattern were observed in the presence of translation factors (data not shown). For instance, the interaction of photoprobes, at 50 µM, with the D-loop/stem was increased, whereas the protection at the TψC stem was very low. Increase of the photoprobe concentration restored the protection at the latter region and enhanced the wrapping of photoprobes around the 3′-strand of acceptor stem.

Gel mobility shift assay and kethoxal probing of photolabeled AcPhe-tRNA

The relative electrophoretic mobility (µ_rel) of a tRNA species is a sensitive indicator of the angle between its anticodon and acceptor stem, with more obtuse interstem angles resulting in higher mobility (28). Thus, AcPhe-tRNA isolated from photolabeled complex C with 300 µM ANB- or ABA-spermine, was deacylated and labeled at the 3′-end with [5′-³²P]pCp. One-half of the preparation was incubated at 37°C for 20 min under native buffer conditions, whereas the second half was denatured by heating to 90°C, and then put on ice. Both preparations were resolved on a native 8% polyacrylamide gel at room temperature. The µ_rel value was calculated from the ratio of the mobility of the folded tRNA species to that of its linear counterpart. Compared to untreated sample (µ_rel = 0.43 ± 0.01; Fig. 6, lane 1), a mobility shift appeared at the lane of photolabeled tRNA with ABA-spermine, resulting reproducibly in a second band of µ_rel value equal to 0.47 ± 0.01 (Fig. 6, lane 2). This difference in mobility, albeit very small, is statistically significant and suggests that the labeled tRNA^Phe possesses a larger interstem angle than the corresponding angle of the reference tRNA^Phe. Similar results were obtained with AcPhe-tRNA isolated from complex C treated with ANB-spermine (data not shown).

Figure 6.

Relative electrophoretic mobilities of tRNA^Phe species on native 8% polyacrylamide gel. Left panel: complex C was photolabeled with 300 µM ABA-spermine. AcPhe-tRNA bound at the P-site of this ribosomal complex was deacylated by puromycin, extracted and labeled at the 3′-end with [5′-³²P]pCp. One-half of the labeled tRNA^Phe was incubated at 37°C for 20 min under native buffer conditions (lane 2), whereas the second half was denatured by heating to 90°C for 15 min (lane 4) prior to gel electrophoresis. Lanes 1 and 3 correspond to AcPhe-tRNA isolated from non-photolabeled complex C, and treated as above. Right panel: bands of lanes 1 and 2, magnified.

Kethoxal is a chemical probe reacting specifically with unpaired guanosines (16). Regions of AcPhe-tRNA, free or bound at the ribosomal P-site, that were protected from kethoxal modification, were analyzed by T₁ digestion. Bar graphs indicating normalized footprinting of untreated AcPhe-tRNA (open bars) and AcPhe-tRNA photolabeled with 300 µM ABA-spermine (solid bars) are shown in Figure 7. The different protection pattern (Fig. 7B) implies that the conformations of the two tRNAs bound at the P-site are different from each other. Similar results were obtained with AcPhe-tRNA photolabeled with ANB-spermine (data not shown).

Figure 7.

Kethoxal footprinting of native or photolabeled AcPhe-tRNA with ABA-spermine. (A) Bar graphs indicating normalized kethoxal footprinting. Native AcPhe-tRNA (open bars) or photolabeled with 300 µM ABA-spermine (solid bars), free or bound at the P-site of poly(U)-programmed ribosomes, was treated with kethoxal for 50 min under native conditions. The AcPhe-tRNA was isolated, deacylated and 3′-end-labeled with [5′-³²P]pCp. The modification sites by kethoxal were analyzed by T₁ partial digestion and gel electrophoresis. The relative intensities were calculated according to Dabrowski et al. (27). Bands G27–30 could not be resolved and were treated as one band. (B) Differences in the kethoxal footprinting patterns of P-site bound AcPhe-tRNA, native or photolabeled with ABA-spermine. Red, guanosines where the respective relative intensities of the two species of AcPhe-tRNA, native or photolabeled with ABA-spermine, differed at least by a factor of two; yellow, no differences according to the above criterion; gray, positions not determinable. The data were taken from (A).

DISCUSSION

The role of tRNA in protein synthesis is intimately associated with the ability of tRNA to project multiple functional groups capable of binding other biomolecules and generating potential binding pockets for metal ions and small polycationic molecules. In the present work we have used two photoaffinity polyamine derivatives, ANB- and ABA-spermine, in an attempt to investigate possible tRNA conformational changes induced by the attachment of polyamines, especially with respect to changes that occur while the process of peptide bond formation is carried out. ANB- and ABA-spermine have been previously used to map binding sites of polyamines on several target biomolecules (19,20,31). In the ANB-derivative, one of the spermine charges is removed, resulting in a compound resembling N¹-acetylspermine or spermidine (Fig. 1). In contrast, ABA-spermine retains a charge in the vicinity of the nearest amino group more closely resembling spermine. This may explain the fact that rANB-spermine, like N¹-acetylspermine (25), does not show any sparing effect on Mg²⁺ requirements for AcPhe-tRNA binding to ribosomes, in contrast to rABA-spermine. Moreover, kinetic studies on the effect of spermine analogs on peptidyltransferase activity revealed that rANB-spermine mimics much better the behavior of N¹-acetylspermine (25), whereas rABA-spermine behaves as its parent compound (15).

Several controls were included in the affinity labeling experiments, with favorable response. Thus, when the biological preparation was photolyzed in the absence of photoprobes and tested for biological activity, no cross-linking or cleavage of macromolecules were detected by gel electrophoresis, whereas the protein-synthesizing activity of the in vitro cell-free system was perfectly preserved. Pre-photolyzed or reduced photoprobes did not covalently bind to AcPhe-tRNA or complex C. Also, the incorporation of photoprobes into AcPhe-tRNA or complex C showed saturation kinetics, while the non-specific binding was not saturable (Fig. 4). Moreover, spermine, as well as rANB- or rABA-spermine, inhibited the specific incorporation of photoprobes into AcPhe-tRNA. The high concentrations of spermine or reduced analogs, required to compete with photoprobes, are consistent with the ability of these competitors to antagonize the reversible binding of the photoprobe during the recognition phase (step 1), but not once photoincorporation has occurred (step 2). Felschow et al. (31) and Leroy et al. (32) have previously observed that high concentrations (300-fold excess) of the parent polyamine are required for competition during photoaffinity labeling of plasma membrane proteins and β-subunit of casein kinase 2, respectively. The degree of cooperativity between the binding sites as well as the affinity of the ligand towards AcPhe-tRNA or complex C appear to increase with the charge of the ligand (Table 1). The number of interacting sites of AcPhe-tRNA with ABA-spermine is similar to that reported for high-affinity binding sites of spermine in tRNA^Phe (1,9) or in ternary ribosomal complex (33). On the other hand, the kinetic analysis of ANB-spermine photoincorporation suggests the existence of three binding sites in AcPhe-tRNA, equal to the number of spermidine cooperative binding sites in yeast tRNA^Phe (9). The above results further support the specificity of photoprobe attachment to AcPhe-tRNA.

Complex C, formed in the absence of translation factors and photolabeled either with ANB- or with ABA-spermine, exhibited moderate reactivity towards puromycin (Table 2). It is important to notice that the attachment of photoprobes to AcPhe-tRNA is responsible for this defect. Nevertheless, a large number of photoprobe molecules were incorporated into the ternary ribosomal complex (Table 1). Consequently, we assume that these additional molecules are not directly involved in the mechanism of inhibition, but probably confer a specific conformation on complex C. When complex C was formed in the presence of translation factors, the effect of photolabeling on the catalytic properties of peptidyltransferase was more complicated. AcPhe-puromycin synthesis was markedly stimulated when complex C was prelabeled with photoprobes at 50 µM. Compared to ANB-spermine, ABA-spermine exhibited a more pronounced stimulatory effect (Table 3). Further analysis revealed that the stimulatory effect is attributed to the ligand anchoring at the ribosome. Photolabeling with photoprobe at higher concentration (300 µM) caused inhibition of peptidyltransferase activity. Although weaker, these effects are reminiscent of the effects of spermine reacted in solution (15).

Why does the incorporation of polyamines into complex C influence its reactivity towards puromycin? As a first approach to this question, the binding sites of photoprobes on AcPhe-tRNA free or bound at the P-site of ribosomes were mapped and the induction of AcPhe-tRNA conformational changes was investigated by gel mobility shift assays and chemical probing. Photolabeling of free AcPhe-tRNA with ANB- or ABA-spermine at low concentration revealed that photoprobes bind preferentially at the deep pocket formed by the D-stem and the variable loop, as well as at the anticodon stem. These are closely related to the spermine binding sites, found in yeast tRNA^Phe by X-ray crystallography and NMR studies (9). Increasing the concentration of photoprobes, new sites of protection appeared in the TψC stem/loop, probably related to weaker binding sites. ANB-spermine, compared with ABA-spermine, protects a shorter number of sites and to a lesser extent. This is in agreement with the conclusions derived by Hill plot analysis (Table 1). Also, this finding supports that the major factor determining binding of these spermine analogs to AcPhe-tRNA is the polyamine part of analogs, not the photoactivated azide.

When AcPhe-tRNA bound at the P-site was probed, the interaction of each photoprobe with the D stem/loop and the anticodon stem was elevated. Since both photoprobes behave as electropositive molecules under the experimental conditions used, it is tempting to suggest that the ribosomal environment may create negatively charged pockets, thus increasing locally the concentrations of photoprobe able to react with AcPhe-tRNA. Alternatively, conformational changes of AcPhe-tRNA occurring upon binding to ribosome could favor the photoprobe attack. The improvement of AcPhe-tRNA binding to the P-site by photolabeling, could be explained assuming that the anticodon-stem stacking is strengthened upon photoprobe incorporation. It has already been realized that the correct stacking of four base pairs, proximal to the anticodon loop, is crucial for tRNA binding at the P-site (34). On the other hand, concomitantly with the increase of photoprobe concentration, the protection at the corner of the L-shaped AcPhe-tRNA molecule increased, whereas new protection sites along the 3′-strand of the acceptor stem were raised (Fig. 5C). It is known that the angle between the anticodon and aminoacyl acceptor stem is strongly modulated by divalent cations attached at these sites (28). Therefore, it was reasonable to suggest that the photoprobe binding at the corner of the L-shaped AcPhe-tRNA may induce an alteration in the canonical interstem angle. Indeed, polyacrylamide gel analysis justified that photolabeled tRNA^Phe possesses a larger interstem angle than that of the reference tRNA^Phe (Fig. 6). Moreover, kethoxal probing of AcPhe-tRNA bound at the P-site of ribosomes, revealed that the protection pattern of photolabeled tRNA is different, compared to that of untreated tRNA (Fig. 7B). Some of the changes could be explained by the fact that kethoxal and ABA-spermine bind to mutually exclusive sites in tRNA. Nevertheless, there are differences in the protection pattern, for instance at nucleosides G15, G57, G63 and G70, which suggest the possibility of a perturbation in the three-dimensional folding of tRNA. On the basis of the above observations, we assume that spermine acting at the corner of the L-shaped AcPhe-tRNA increases the interstem angle and consequently orientates the aminoacyl group of tRNA towards a less active position of the catalytic cavity. When complex C was formed in the presence of translation factors, the strong protections in the region of the TψC stem and acceptor stem were absent at low concentration of photoprobe. This may contribute to the stimulatory effect of photoprobes observed at this concentration (Table 3). Certainly, to elucidate the molecular mechanism of peptidyltransferase stimulation by polyamines it is important to determine the binding sites of polyamine in ribosomes, and to investigate the role of translation factors.

Spermine is the most effective among polyamines and is the parent compound of a large number of synthetic polyamine analogs with antitumor and antiparasitic activity. In agreement with the present results, accumulation of excess polyamines causes a decrease in cell viability, primarily through inhibition of protein synthesis (35). Consequently, getting an insight into the regulatory role of polyamines in protein synthesis is of great significance.