Evidence for an allosteric mechanism of substrate release from membrane-transporter accessory binding proteins

Abstract

Numerous membrane importers rely on accessory water-soluble proteins to capture their substrates. These substrate-binding proteins (SBP) have a strong affinity for their ligands; yet, substrate release onto the low-affinity membrane transporter must occur for uptake to proceed. It is generally accepted that release is facilitated by the association of SBP and transporter, upon which the SBP adopts a conformation similar to the unliganded state, whose affinity is sufficiently reduced. Despite the appeal of this mechanism, however, direct supporting evidence is lacking. Here, we use experimental and theoretical methods to demonstrate that an allosteric mechanism of enhanced substrate release is indeed plausible. First, we report the atomic-resolution structure of apo TeaA, the SBP of the Na(+)-coupled ectoine TRAP transporter TeaBC from Halomonas elongata DSM2581(T), and compare it with the substrate-bound structure previously reported. Conformational free-energy landscape calculations based upon molecular dynamics simulations are then used to dissect the mechanism that couples ectoine binding to structural change in TeaA. These insights allow us to design a triple mutation that biases TeaA toward apo-like conformations without directly perturbing the binding cleft, thus mimicking the influence of the membrane transporter. Calorimetric measurements demonstrate that the ectoine affinity of the conformationally biased triple mutant is 100-fold weaker than that of the wild type. By contrast, a control mutant predicted to be conformationally unbiased displays wild-type affinity. This work thus demonstrates that substrate release from SBPs onto their membrane transporters can be facilitated by the latter through a mechanism of allosteric modulation of the former.

Fig. 1.

(A) X-ray structure of the open apo-state of TeaA. The structure was solved to 2.2-Å resolution by molecular replacement, using the structure of ectoine-bound TeaA (PDB ID code 2VPN) as a template for phasing. The asymmetric unit comprises four TeaA monomers, about 250 solvent molecules and 41 Zn²⁺ ions. Residues 44 to 47 and 311 to 316 in the C-terminal domain are disordered. The long helix α9 and strand β4, depicted in orange, extend across TeaA, flanking both N- (blue) and C- (red) terminal domains. (B) In the open state, the flanking helix α9 (orange) is fully formed, while in the bound state it unravels around Gly248/Leu249. As the N- and C-terminal domains close around the substrate, helix α9 becomes noticeably kinked.

Fig. 2.

Comparison of the apo (A) and ectoine-bound (B) structures of TeaA. Protein and substrate are represented as cartoons and sticks, respectively. Water molecules are shown as dotted spheres, and Mg²⁺ and Zn²⁺ as solid spheres in cyan and green, respectively. The apo-protein adopts an open conformation, in which the N-terminal (blue) and C-terminal (red) lobes are further away than in the ectoine-bound structure. (Insets) Architecture of the ligand-binding sites in the apo (A) and ectoine-bound (B) structure of TeaA. Ectoine-binding residues in the N domain are shown in blue, and those in the C domain in red, with ectoine in black. The dotted line indicates the distance between the Cβ atom of Trp167 and the Cε₁ of Phe209 in the two structures. These atoms are about 4 Å further apart in the apo-structure compared to the ectoine-bound structure.

Fig. 3.

(A) All-atom molecular simulation model of TeaA. The system comprises the protein (and ectoine), approximately 12,000 TIP3P water molecules, 54 Na⁺, and 20 Cl^- ions (approximately 100 mM plus counterions), enclosed in a truncated-octahedron periodic box. Glu121 is set in the protonated state. Wild-type and mutant simulation systems, in the apo or bound states, are based on the X-ray structures; these were constructed through a series of energy minimizations and constrained dynamics of the protein and its environment. (B) Conformational free-energy landscapes for the apo and ectoine-bound states, obtained from bias-exchange molecular dynamics simulations. Contours are plotted in increments of 1 kcal/mol. The statistical error in these landscapes is discussed in the SI Appendix, Supplementary Materials.

Fig. 4.

Microscopic mechanism of ectoine recognition by TeaA. Representative snapshots of configurational clusters along the binding reaction are depicted, highlighting the side chains involved in direct ectoine contacts as well as the fold of helix α9 and the overall protein conformation (water, ions, and most side chains are omitted for clarity). The calculated free energies of the corresponding clusters are also plotted. The connectivity between this clusters is established based on their characteristic free energy and observed transitions during the bias-exchange simulations (see SI Appendix, Supplementary Methods).

Fig. 5.

Conformational free-energy landscape for apo TeaA after triple Ala mutation of G248, L249, and S250 in helix α9, computed as those in Fig. 3. Contours are plotted in increments of 1 kcal/mol.

Fig. 6.

Shift in the open-to-closed equilibrium of TeaA upon mutation of helix α9. The change upon mutation in the free-energy difference between the closed and open conformations () is equivalent to the difference in the “free energy of mutation” of each of these states (); these differences may be computed using alchemical-perturbation free-energy calculations (see SI Appendix, Supplementary Material). The calculated ΔΔG values for the G248A/L249A/S250A (AAA) and G248P mutants can be interpreted as an estimate of the conformational bias induced by the mutation, relative to the wild type.

Fig. 7.

Calorimetric titrations at 25 °C of (A) 117 μM TeaA mutant AAA by adding 28 × 5 μl 5 mM ectoine in 50 mM sodium phosphate at pH 7.3; and (B) 42 μM TeaA mutant G248P by adding 23 × 4 μl 1 mM ectoine, same medium.

Fig. 8.

Structural dynamics of the binding cleft in unbiased 1-microsecond simulations of wild-type TeaA and its AAA mutant. The protein conformations obtained in each simulation (snapshots taken at 20-ps intervals) were combined in a single ensemble and clustered based on pairwise similarity according to a global descriptor, namely the rmsd of the complete C^α trace of the protein (the cutoff difference within each cluster is 1 Å). Only clusters with a significant population (> 10%) were analyzed further. The variability of the structure of the binding cleft across all the snapshots in each cluster was then quantified, now using a local descriptor. This descriptor is the pairwise rmsd of a substructure that includes only the protein residues involved in ectoine coordination (N, C, C^α, C^β, and C^γ atoms). Note that this local descriptor is not correlated with the global one used for clustering a priori, because the binding-cleft structure is defined by < 5% of the total number of residues in the protein. The curve in the plot labeled WT + AAA is a histogram of these rmsd values, combining all the pairs in each of the clusters. The characteristic variability in the binding-cleft structure among simulation snapshots is 0.97 ± 0.27 Å. To assess whether the AAA mutation has a direct influence on the structure of the binding cleft, the distribution was recomputed after removing from the analysis all the snapshots from the AAA simulation (labeled WT), or from the wild-type simulation (labeled AAA). As the figure demonstrates, these distributions are not significantly different from the one above. On average, the characteristic variability in the binding-cleft structure becomes 0.99 ± 0.28 Å and 0.91 ± 0.26 Å, respectively. In conclusion, this analysis shows that the AAA mutation does not directly perturb the structure of the binding cleft, even though it influences the global conformational dynamics of the protein (Fig. 5).

Fig. P1.

Schematic of the cycle of conformational changes in SBPs involved in bacterial membrane import. SBPs sequester their substrate by using a high-affinity mechanism akin to a Venus flytrap. The substrate, however, must be released upon interaction of the SBP with suitable membrane importers, regardless of their much weaker affinity.