Broad phylogenetic analysis of cation/proton antiporters reveals transport determinants

Abstract

Cation/proton antiporters (CPAs) play a major role in maintaining living cells' homeostasis. CPAs are commonly divided into two main groups, CPA1 and CPA2, and are further characterized by two main phenotypes: ion selectivity and electrogenicity. However, tracing the evolutionary relationships of these transporters is challenging because of the high diversity within CPAs. Here, we conduct comprehensive evolutionary analysis of 6537 representative CPAs, describing the full complexity of their phylogeny, and revealing a sequence motif that appears to determine central phenotypic characteristics. In contrast to previous suggestions, we show that the CPA1/CPA2 division only partially correlates with electrogenicity. Our analysis further indicates two acidic residues in the binding site that carry the protons in electrogenic CPAs, and a polar residue in the unwound transmembrane helix 4 that determines ion selectivity. A rationally designed triple mutant successfully converted the electrogenic CPA, EcNhaA, to be electroneutral.

Fig. 1

The CPA1 subtree. Unrooted tree with the main clades highlighted in different colors. The names of K⁺-selective clades are underlined. The prevalent motifs that correlate with each clade are listed. The invariable positions 5 and 8, a fingerprint of all CPA1s, are highlighted in bold. Bootstrap values are marked for the branches that separate the main clades. Names of representative CPAs are presented at the leaves

Fig. 2

The CPA2 subtree. Unrooted tree with the main clades highlighted in different colors and the prevalent motifs that correlate with each clade listed as in Fig. 1. The names of K⁺-selective clades are underlined; whereas. the names of known electrogenic clades and sub-clades are marked with an asterisk. Bootstrap values are marked for the branches that separate the main clades. Names of representative CPAs are presented at the leaves

Fig. 3

The CPA motif is located in the core domain, near the ion binding site. a Periplasmic view of the electrogenic EcNhaA, showing the dimerization domain to the left and core domain to the right. The frame to the right displays a close-up view of the motif. The eight residues of the motif are shown using sticks representation, each with its sequence number and corresponding position in the motif (subscript). Their respective helices are designated in italics, and the conserved polar interaction between TM-5 and TM-10 is marked with a solid line. b The electroneutral MjNhaP1, shown as in a. The numbered residues correspond to their equivalents in EcNhaA. For example, D132 and E156 in MjNhaP1 correspond to D133 and L159 in EcNhaA, respectively. c Schematic two-dimensional representation of EcNhaA’s structure. The membrane boundaries are shown as dashed lines, the helices are numbered TM-1-through-TM-12, and the motif residues are indicated with yellow spheres

Fig. 4

Model structure of HsNHA2. Left: periplasmic view, colored by the ConSurf evolutionary conservation bar at the bottom. The dimerization domain is on the left and core domain on the right. Right: close-up view of the unique polar interaction between TM-3, TM-5, and TM-10 characteristic of HsNHA2-like PCAs. The three residues involved are highly conserved. E215 could potentially salt bridge with R432₈ on TM-10. In one plausible scenario D278₆ would be protonated and hydrogen bonded to R432₈, leaving only D279₇ to alternate its protonation state upon transport, resulting in electroneutral transport. Modeling is based on the structure of TtNapA (PDB id 5BZ2) as template

Fig. 5

The NhaA triple mutant supports cell growth under neutral but not alkaline pH. EP432 E. coli, carrying deletions of both nhaA and nhaB genes, was transformed with EcNhaA, either WT or the P108E_A160S_D163N triple mutant, or an empty pBR322 vector, as negative control. Growth resistance to Na⁺/Li⁺ was performed as described in Methods section. a Cells were grown in modified L broth (LBK, with NaCl replaced by KCl) and then b–d tested for growth resistance to Li⁺ and Na⁺. e A summary of the number of cells grown under each condition. The standard deviation was 3–5%

Fig. 6

Valinomycin effect on Na⁺-induced proton efflux confirms electroneutrality. For measuring the Li⁺/H⁺ antiporter activity in pH 8.5, E. coli EP432 cells expressing WT or the P108E-A160S-D163N mutant were grown in LBK (pH 7.0) and everted membrane vesicles were isolated. The ΔpH across the membranes was determined using acridine orange, a fluorescence probe of ΔpH. The reaction mixture (2.5 mL) was modified according to Schuldiner and Fishkes to get the maximal effect of valinomycin (2 µM). It contained, protein (100 µg), Tris-Cl (10 mM, pH 7.8), MgSO₄ (10 mM), and KCI (Na⁺ free) 0.14 M. After a 1-min incubation, 9-amino- acridine (0.5 µM) was added. The ΔpH was generated by the addition of ATP (2 mM). Where indicated, 2 µM of an ethanolic solution of valinomycin was added 1 min prior to the ATP. At the onset of the reaction, d-lactate (2 mM) was added (downward facing arrow) and the fluorescence quenching was recorded until a steady-state level of ΔpH (100% quenching) was reached. Then, 10 mM NaCl was added (upward facing arrow), and the new steady state of fluorescence obtained (dequenching) was monitored. Whereas valinomycin/K⁺ had no effect on the Na⁺-induced dequenching by the mutant, it increased it dramatically in the WT. The experiments were repeated at least three times with practically identical results

Fig. 7

A proposed model of electroneutrality versus electrogenicity. The four main variations of the putative interaction between the conserved polar residues on TM-5 and TM-10 are shown. Helices are numbered according to EcNhaA’s topology. The relevant residues are shown as circles, with acidic in red, basic in blue, and polar in yellow. Interactions that are expected to remain intact throughout the transport cycle according to the proposed model are marked with solid lines in a, c, d, and the interaction that would alternate is marked with a dashed line in b. a The electroneutral CPA1 subtree would feature a salt bridge between arginine in position 8 and glutamate in position 5. The protonation state of these residues would be fixed, while that of D₇ would change upon proton transport. b Electrogenic CPA2s, characterized by acidic residues in positions 6 and 7, presumably interacting with the two protons, and a basic residue in position 8. Upon interacting with a proton, D₆ would alternate between salt bridging and hydrogen bonding with K₈. c Putative electroneutral transport by CPA2s that lack the acidic residue in position 6. Positions 6 and 8 would hydrogen bond with each other. Similar to electroneutral CPA1s, the protonation state of these residues would be fixed, while that of D₇ would alter upon interaction with the proton. d Electroneutral mammalian NHA-like CPA2s, surprisingly featuring two acidic residues at positions 6 and 7, similar to the electrogenic CPA2s shown in b. This group also includes an arginine at position 8 that is prevalent in electroneutral CPAs, and a uniquely conserved glutamate on TM-3 that potentially could salt bridge with each other. This slight change in the dielectric environment of D₆ could, theoretically, prevent its protonation and deprotonation, resulting in electroneutral transport

Fig. 8

A proposed model for potassium-selectivity in CPA1s. Side view (a, b), top view (c, d), and schematic two-dimensional projection (e, f) of the binding site of the Na⁺-selective PaNhaP (PDB id 4CZA). a, c, and e are the wild-type structure and b, d, and f are a model structure of a proposed PaNhaP mutant, which we suggest to be K⁺-selective. Residues that participate in ion coordination are shown as sticks. The positions where we suggest mutations are in bold. A thallium ion (bronze) and a water molecule (red) are shown in the binding site. e vs. f The suggested additional coordinating interaction between the protein and the ion and the slight increase in the pore diameter resulting from the A128S (light-blue) and P131A (light-yellow) mutations, respectively, are shown. The A128S substitution is expected to provide another polar interaction with the ion, consistent with the expected larger coordination number of potassium ions compared to sodium. The P131A substitution is expected to slightly increase the cavity to accommodate the larger potassium ions a vs. b