Two-step aminoacylation of tRNA without channeling in Archaea

Abstract

Catalysis of sequential reactions is often envisaged to occur by channeling of substrate between enzyme active sites without release into bulk solvent. However, while there are compelling physiological rationales for direct substrate transfer, proper experimental support for the hypothesis is often lacking, particularly for metabolic pathways involving RNA. Here, we apply transient kinetics approaches developed to study channeling in bienzyme complexes to an archaeal protein synthesis pathway featuring the misaminoacylated tRNA intermediate Glu-tRNA(Gln). Experimental and computational elucidation of a kinetic and thermodynamic framework for two-step cognate Gln-tRNA(Gln) synthesis demonstrates that the misacylating aminoacyl-tRNA synthetase (GluRS(ND)) and the tRNA-dependent amidotransferase (GatDE) function sequentially without channeling. Instead, rapid processing of the misacylated tRNA intermediate by GatDE and preferential elongation factor binding to the cognate Gln-tRNA(Gln) together permit accurate protein synthesis without formation of a binary protein-protein complex between GluRS(ND) and GatDE. These findings establish an alternate paradigm for protein quality control via two-step pathways for cognate aminoacyl-tRNA formation.

Fig. 1

Kinetics of the two-step pathway for Gln-tRNA^Gln formation in the archaeon M. thermautotrophicus. (a) Experimental design used to investigate the possibility that the intermediate Glu-tRNA^Gln is channeled from GluRS^ND to GatDE. (b) Summary of rate constants for individual steps in the two-step pathway for Gln-tRNA^Gln synthesis, derived from experiments presented in thjis manuscript.

Fig. 2

Kinetics of the tRNA^Gln aminoacylation reaction. (a) Thin-layer chromatography (TLC) separation of substrate and products from a single turnover reaction at 500 nM enzyme (quantitated in b). Plateau aminoacylation levels for the reaction are consistemtly in the 60–70% range. (b) Time traces of product formation at enzyme concentrations of 2 nM (purple), 25 nM (green), 100 nM (red), 500 nM (blue) and 2 μM (black). The tRNA^Gln concentration in each reaction is < 1 nM. Curves are fit to F = F_max(1−e^−kt), where F is the fraction Glu-tRNA^Gln formed, F_max is the maximum (plateau) value of F, k is the rate constant (sec⁻¹) and t is time (sec). (c) Plot of k_obs values obtained from the data in panel (b) as a function of GluRS^ND concentration; from this plot we derive k_max of 0.12 ± 0.01 sec⁻¹ and K_1/2 of 43 ± 18 nM. The curve is fit to: k_obs = k_obs,max[GluRS^ND]/(K_1/2 + [GluRS^ND]). (d) Pre-steady state kinetic timecourses at 500 nM GluRS^ND and either 5 μM (blue) or 10 μM tRNA^Gln (red).

Fig. 3

Binding of tRNA^Gln to GluRS^ND. (a) Time course monitoring the decrease in intrinsic Trp fluorescence upon addition of varying concentrations of tRNA^Gln using a stopped-flow instrument. A sample trace at 500 nM GluRS^ND and 500 nM tRNA^Gln is shown, from which k_obs of 50 sec⁻¹ was derived. (b) Plot of observed association constants against concentration of tRNA^Gln, from which k_on of 1.2 × 10⁸ M⁻¹sec⁻¹ is derived. Equations to fit the raw data and to derive elementary rate constants are provided in Materials and Methods (equations (1) and (2), respectively).

Fig. 4

Kinetics of the GatDE reaction. (a) Timecourse of a GatDE reaction showing TLC separation of Gln-tRNA^Gln from Glu-tRNA^Gln and tRNA^Gln. (b) Time traces for the formation of Gln-tRNA^Gln from purified Glu-tRNA^Gln under single turnover conditions. Concentrations are less than 1 nM for Glu-tRNA^Gln and as depicted for GatDE. Curves are fit to F = F_max(1−e^−kt), where F is the fraction Gln-tRNA^Gln formed, F_max is the maximum (plateau) value of F, k is the rate constant (sec⁻¹) and t is time (sec). (c) Plot of the GatDE concentration dependence of k_obs obtained from (b); from which we derive k_max of 1.2 ± 0.04 sec⁻¹. The curve is fit to: k_obs = k_obs,max[GatDE]/(K_1/2 + [GatDE]). (d) Increase in fluorescence of 2-AP labeled tRNA (average of three traces) upon binding of 2.5 μM GatDE giving k_on of 2.6 ± 0.11 × 10⁵ M⁻¹sec⁻¹. The equation used to fit the raw fluorescence data is provided in Materials and Methods (equation 3).

Fig. 5

Combined GluRS^ND and GatDE reactions. (a) TLC plate showing transient accumulation of the Glu-tRNA^Gln intermediate during conversion of tRNA^Gln to Gln-tRNA^Gln in the presence of both GluRS^ND (600 nM) and GatDE (600 nM). Plots of normalized data and kinetic simulation traces (not fits to data) for a two-step dissociative mechanism at 600 nM GluRS^ND and GatDE at 0 nM (a), 10 nM (b), 100 nM (c), 600 nM (d) and 6 μM (e). tRNA^Gln in all the reactions < 1 nM. The panels show that, at all tested concentrations of GatDE, a distributive model fits the data well. Goodness of fit values (R²) from the nonlinear regression were computed separately using GraphPad software; for formation and reaction of Glu-tRNA^Gln these values range from 0.91 to 0.97. See Supplementary Figs. S3–S5 for details of kinetic simulations and evidence that a channeling model does not fit the data.

Fig. 6

Discrimination against Glu-tRNA^Gln by archaeal EF-1α. TLC images showing spots corresponding to acylated (Gln-Ap*) and deacylated (Ap*) tRNAs (left panels) were quantified to give the fraction of (a) Gln-tRNA^Gln, (b) Glu-tRNA^Gln and (c) Glu-tRNA^Glu in the presence of 2.2 μM activated EF-1α (inverted triangles) or in the presence of activated EF-1α following proteinase K incubation for 10 minutes (triangles). When EF-1α is removed by proteolysis, the rate constants for spontaneous deacylation are 0.036 min⁻¹, 0.025 min⁻¹, and 0.025 min⁻¹ for the three aminoacyl-tRNAs, respectively (blue traces). In the presence of activated EF-1α, the rate constants are: 0.0015 min⁻¹ for Gln-tRNA^Gln, 0.013 min⁻¹ for Glu-tRNA^Gln, and 0.003 min⁻¹ for Glu-tRNA^Glu, respectively (black traces).

Fig. 7

Binding affinity of EF-1α to aminoacyl-tRNAs. Decreases in the fraction of Gln-tRNA^Gln (a) and Glu-tRNA^Gln (b) present over time in the presence of different concentrations of EF-1α at 0 μM (black inverted triangles), 0.044 μM (blue filled triangles), 0.17 μM (red filled circles), 0.7 μM (green filled diamonds), 2.2 μM (cyan circles), and 4.4 μM (pink squares). Each timecourse is fit to an exponential function to derive k_obs. (c) Rate constants for protection of spontaneous deacylation in the presence of EF-1α obtained from panels (a) and (b) are replotted and fit to a hyperbolic binding curve to derive the K_d for EF-1α binding to Gln-tRNA^Gln (0.29 ± 0.03 μM; red trace). No significant correlation of the rate constant for deacylation of Glu-tRNA^Gln, with increasing EF-1α concentration, could be made (blue inverted triangles indicate the data in this case).

Fig. 8

Model for two-step synthesis of Gln-tRNA^Gln without channeling in Archaea. In this model independent, distributive function of GluRS^ND and GatDE together with exclusion of Glu-tRNA^Gln binding by elongation factor suffices for selective glutamine incorporation in M. thermautotrophicus.