Who’s on first …Na⁺, HCO₃⁻ or CO₃²⁻ ?

In 1983 Boron and Boulpaep first described an electrogenic Na⁺ and HCO₃⁻ cotransport activity in the kidney proximal tubule of the salamander Ambystoma tigrinum (Boron and Boulpaep, 1983). While the red blood cell Band 3, i.e., Cl⁻-HCO₃⁻ exchange protein, had been studied for several decades and a Na⁺ dependent Cl⁻-HCO₃⁻ exchange activity was known from snail neuron and barnacle muscle, this salamander kidney study was the first description that membrane potential could directly influence HCO₃⁻ transport. The authors postulated that if this was a single protein’s activity, then a stoichiometry of at least 1Na⁺ : 2HCO₃⁻ ions was required, but that 1:3 was more energetically favorable for the voltage and ion gradient conditions of the kidney proximal tubule. Soon thereafter, several investigators identified similar activities in mammalian proximal tubules, corneal endothelium, liver, intestine and retinal Müller cells (Brady et al., 2021). These studies initiated a stoichiometry debate that ultimately argued that Na⁺ and HCO₃⁻ efflux at the basolateral membrane of the kidney proximal tubule should be 1:3. Conversely, for all other tissues mediating Na⁺ and HCO₃⁻ influx by this putative protein, should have a stoichiometry of 1:2. Using basolateral membrane vesicles, Soleimani and Aronson supported an equivalent 1:3 stoichiometry by showing that this activity was supported by CO₃²⁻ and SO₃²⁻ (Soleimani and Aronso, 1989) leading to an iterative model of 1Na⁺: 1HCO₃⁻; 1CO₃²⁻ cotransport.

In 1997 the cloning of an Ambystoma DNA sequence revealed that a single protein functioned as an electrogenic Na⁺ and HCO₃⁻ cotransporter (Romero et al., 1997). NBCe1, as the protein was named, turned out to be a new and fourth member of the SLC4 gene family (SLC4A4) accompanying the three cloned anion exchangers. The mammalian kidney sequences soon followed (NBCe1-A), but an N-terminal variant from the mammalian pancreas and heart (NBCe1-B) was soon identified (Brady et al., 2021). This “molecular sequence switch” seemed to conveniently segregate with transport direction and apparent ion-coupling; i.e., Na⁺/nHCO₃⁻ efflux by the kidney as the role of NBCe1-A and the Na⁺/nHCO₃⁻ influx by other tissues as the role of NBCe1-B. Biology is rarely so simple. The rat NBCe1-A clone, i.e., “kidney isoform,” was found to have a 1 Na⁺ :2 HCO₃⁻ stoichiometry when expressed in Xenopus oocytes (Sciortino and Romero, 1999). In recent years, data with human NBCe1-A have also revisited the hypothesis that CO₃²⁻ is an ionic substrate (Moss and Boron, 2020).

The current paper (Wu et al., 2022) tests the human NBCe1-A and -B isoforms using varying [Na⁺], [HCO₃⁻], [CO₃²⁻] and membrane voltage (V_m or V_clamp) while probing the cryo-EM structure predicted binding sites (Huynh et al., 2018). This is the first study to probe the ionic coupling of NBCe1-B as well as NBCe1-A. The authors delimit the Na⁺ binding site as well as two anion binding sites. Together these experiments reveal that directionality of ion transport rather than isoform primary sequence seems to dictate which ions bind and are transported. Specifically, when Na⁺ and HCO₃⁻ move from the outside world to the intracellular world, the anion sites accommodate the trigonal planar HCO₃⁻. Nevertheless, when ions move out of the cell, one of these same anion sites accommodates the also trigonal planar CO₃²⁻ (Figure 9,10, 13 from Wu et al., 2022).

Second, Wu and coworkers find that HCO₃⁻ binding in NBCe1A/B precedes Na⁺ binding. This is unusual in that other Na⁺ coupled transporters, for which ordered binding has been determined, bind the Na⁺ cation first. This initial Na⁺ binding provides favorable protein energetics in the form of a favorable chemical gradient and often an accompanying protein conformational change to accommodate or bind the other ions or solutes. Figure 4F & 4G (Wu et al., 2022) show that decreasing [HCO₃⁻] results in an increased apparent K_m for Na⁺ while changing [Na⁺] has no obvious effect of the K_m for HCO₃⁻ for either NBCe1-A or NBCe1-B. Additionally, the modeling and mutation experiments reveal that the more distal spaces allow the more selective coordination by the proximal immediate side-chains, e.g., F840A on I803A and I803A on T436S (Figure 8, Wu et al., 2022).

Investigators focused on NBCe1-A stoichiometry reasoning that tissue experiments implied that “electrogenic Na⁺ bicarbonate cotransporter activity” in kidney proximal tubule epithelia had the outwardly directed transport and therefore a 1:3 stoichiometry. Importantly, Wu and colleagues extend their study to include the NBCe1-B isoform in documenting [Na⁺] dependence, [HCO₃⁻] dependence and stoichiometry (Wu et al., 2022). While differing in absolute magnitude, the NBCe1-A and NBCe1-B isoforms are remarkably similar. Through modeling, molecular dynamics and some experimentation, the authors provide additional evidence that the CO₃²⁻ anion as well as the HCO₃⁻ anion may be coordinated and transported by NBCe1. On some level this could be attributed to biophysical or thermodynamic “nuance,” but the authors work provides some of the first mechanistic explanation accounting for apparent stoichiometry differences between tissues and preparations. Moreover, this work indicates that the nature of stoichiometry differences is associated with directionality rather than sequence or binding domain differences, specifically IRBIT-binding to the autoinhibitory domain of the N-terminus of NBCe1-B. That is, Na⁺ and HCO₃⁻ moving into a cell (influx) has a 1:2 ion coupling. Conversely, Na⁺ and HCO₃⁻ moving OUT of a cell (efflux) has a 1Na⁺:1HCO₃⁻:1CO₃²⁻ ion coupling that is simplified to 1Na⁺:3HCO₃⁻ coupling. Of course when ions actually move, it is the electrochemical gradient (i.e., ionic chemical gradients and V_clamp) that determine directional transport. That said, why apparently different coordination occurs in an efflux mode and why this “change” does not seem to occur at both anion sites, is not known.

The paper by Wu and coworkers (2022) emphasizes the importance to repeat experiments especially when new tools, e.g. molecular structures, are available. Interestingly, the modeling experiments indicate that the NBCe1 structure can also accommodate two CO₃²⁻ ions indicating that 3-net negative charges could be moved through NBCe1. Why this coupling directionality exists is not immediately clear but may be a reflection that membrane transporters are a type of “enzyme.” Enzymes create internal protein spaces where chemistry different from solution chemistry may occur for some biological advantage. Though not highlighted, Wu and coworkers finding that HCO₃⁻ binding precedes Na⁺ binding does seem to eliminate the possibility that NBCe1 could function as a NaCO₃⁻ channel. While there will likely be additional evaluations of NBCe1-mediated ion binding and transport, we now know where discrete ions bind and that NBCe1-stoichiometry varies according to transport direction and not isoform sequence differences. Finally, Figure 13 (Wu et al., 2022) while providing a mechanistic, conceptual framework for ordered binding and unbinding of ions, does not incorporate the known dimer structure of NBCe1(Brady et al., 2021, Huynh et al., 2018). That is, the current work does not address if monomer-binding sites in the know dimer structure might be cooperative or interfering or irrelevant. It is attractive to speculate that monomer asymmetries in the NBCe1-dimer might allow discrete confirmational states to be extracted from the more infrequent cryoEM states (Huynh et al., 2018). How these differences occur, will mostly be the subject of future investigations of the authors and others in the field.