NH3/NH4 + allosterically activates SLC4A11 by causing an acidic shift in the intracellular pK that governs H+(OH−) conductance

SLC4A11 is the most abundant membrane transport protein in corneal endothelial cells. Its functional presence is necessary to support the endothelial fluid pump that draws fluid from the corneal stroma, preventing corneal edema. Several molecular actions have been proposed for SLC4A11 including H2O transport and cell adhesion. One of the most reproduced actions that SLC4A11 mediates is a H+ (or OH−) conductance that is enhanced in the presence of NH4Cl. The mechanism by which this occurs is controversial with some providing evidence in favor of NH3-H+ cotransport and others providing evidence for uncoupled H+ transport that is indirectly stimulated by the effects of NH4Cl upon intracellular pH and membrane potential. In the present study we provide new evidence and revisit previous studies, to support a model in which NH4Cl causes direct allosteric activation of SLC4A11 by means of an acidic shift in the intracellular pK (pKi) that governs the relationship between intracellular pH (pHi) and SLC4A11 H+-conductance. These findings have important implications for the assignment of a physiological role for SLC4A11.


Introduction
The members of the SLC4 family of solute carrier proteins are mainly Na + -independent and Na + -dependent Cl − /HCO 3 − exchangers or Na + -coupled HCO 3 − (or CO 3 = ) transporters (Parker and Boron, 2013;Lee et al., 2022).SLC4A11 (originally named "BTR1": Bicarbonate Transporter Related Protein 1) was the last member of the SLC4 family to be cloned (Parker et al., 2001) and is the only member of the family that does not transport HCO 3 − /CO 3 = (Jalimarada et al., 2013;Ogando et al., 2013;Loganathan et al., 2016).Instead, SLC4A11 influences intracellular pH (pH i ) by conducting H + , or its thermodynamic equivalent OH − (Kao et al., 2015;Myers et al., 2016).For convenience hereafter, and in the absence of any definitive data in favor of one substrate over the other, we will assume that H + are the transported species.SLC4A11 is expressed at low levels in many tissues, but appears to have most functional impact in the cornea and inner ear as evidenced by the corneal dystrophies and progressive hearing loss that are caused by SLC4A11 mutations in both humans and mice (Vithana et al., 2006;2008;Desir et al., 2007;Lopez et al., 2009; Most studies of SLC4A11 have focused on its role in the cornea.Here, SLC4A11 is expressed in the basolateral membrane of the corneal endothelial cells that line the posterior (aqueous-humor facing side) of the cornea (Vilas et al., 2013).These cells are responsible for pumping fluid from the corneal stroma to prevent it from swelling and losing its ability to optimally transmit and refract light (Hodson, 1977).SLC4A11 is clearly valuable for endothelial pumping because autosomal recessive inheritance of SLC4A11 mutations cause congenital hereditary endothelial dystrophy (CHED), a non-progressive corneal thickening and opacification (Vithana et al., 2006).Furthermore, an apparently discrete set of SLC4A11 mutations result in autosomal dominant inheritance of a late-onset form of Fuchs endothelial corneal dystrophy (FECD4), which typically manifests from the 4th decade of life, preceded by the appearance of excrescences from the Descemet's membrane that underlie the endothelial cells, which are known as guttae (Weiss et al., 2024).However, the mechanism by which SLC4A11 loss compromises the endothelial pump remains elusive because there is little consensus about SLC4A11's molecular action.
The earliest proposal for SLC4A11 action was that it, like plantal Slc4-like proteins, performed borate transport: specifically electrogenic 2Na + -B(OH) 4 -cotransport (Frommer and von Wirén, 2002;Park et al., 2004).This hypothesis fell out of favor due to an inability of others to find evidence for this mode of action with mammalian SLC4A11 (Jalimarada et al., 2013;Ogando et al., 2013;Vilas et al., 2013;Kao et al., 2015;Loganathan et al., 2016).That original characterization study also attributed an EIPA-insensitive Na + /H + exchanger-like activity to SLC4A11, but neither the EIPA-sensitivity nor the Na + -dependence of this H + transport activity has proven to be universally repeatable, leaving open the possibility that such results were influenced by endogenous activities (Park et al., 2004;Ogando et al., 2013;Kao et al., 2015;2016;Zhang et al., 2015;Myers et al., 2016).Other reported SLC4A11 transport modes include a H 2 O permeability that is disturbed by diseasecausing mutations (Vilas et al., 2013), a role in extracellular matrix adhesion (Malhotra et al., 2020), and a conditional presence in mitochondria (Ogando et al., 2019).Any of these, alone or in combination, remain a feasible explanation for loss of endothelial pump function with SLC4A11 mutation.However, at present, none of these features has been investigated by more than one group.
In this study, we focus on the robust and repeatable Na +independent H + transport action of SLC4A11, an action that is stimulated by increases in pH i , extracellular pH (pH e ), and NH 4 Cl (Zhang et al., 2015;Myers et al., 2016;Kao et al., 2020;Quade et al., 2020) and which is also disrupted by disease-causing mutations (Kao et al., 2015;Quade et al., 2022).However, there remains a controversy over whether the NH 4 Cl-stimulated action of SLC4A11 represents SLC4A11-mediated NH 3 -H + cotransport (Zhang et al., 2015;Kao et al., 2020) or an indirect stimulation of SLC4A11mediated H + conductance caused by cellular depolarization and sub-membranous alkalinization of pH i due to NH 3 movement across the lipid bilayer (Myers et al., 2016).A recent attempt to distinguish NH 3 -H + cotransport from NH 4 + or H + transport using the Goldman-Hodgkin-Katz approach led to the conclusion that SLC4A11 operates in competing modes of NH 3 -H + cotransport or unaccompanied H + transport, with the inference that the presence of NH 3 /NH 4 + inhibits unaccompanied H + conduction (Kao et al., 2020).These are important distinctions as they inform the predicted direction of transport, the interpretation of SLC4A11 structure, and the ultimately the physiological role of SLC4A11.In our previous work, we demonstrated that SLC4A11-mediated H + transport is governed by an intracellular pK (pK i ), the value of which can be modulated by changes in pH e and by disease-causing mutations (Quade et al., 2020;Quade et al., 2022).At pH e = 7.50, pK i is too alkaline to be determined, but can be no more acidic than 7.6 (Quade et al., 2020).However, when we raise pH e , pK i is shifted into measurable range.For example, pK i for human SLC4A11 7.04 when pH e = 8.50 (Quade et al., 2022).This acidic-shift in pK i manifests as a rise in SLC4A11 current.We note that no study of SLC4A11 activity has directly measured transmembrane NH 3 / NH 4 + movement, relying instead on proxies such as pH and voltage changes.With that as context, here we revisit the phenomenon of NH 4 Cl-stimulation of SLC4A11-mediated H + currents to determine whether it could be explained by a direct allosteric effect of NH 3 /NH 4 + upon SLC4A11 pK i .

Results
2.1 Determining pK i for SLC4A11 at pH e = 8.50 In a previous study we determined that pK i for human SLC4A11 at pH e = 8.50 is 7.04 ± 0.01 (Quade et al., 2022).For this study we generated a contemporary set of control data to confirm the pK i of human SLC4A11 at pH e = 8.50 in new experimental hands.As we have previously reported, SLC4A11expressing oocytes slowly alkalinize upon exposure to pH e = 8.50 solution, and the rate of alkalinization can be enhanced by clamping the membrane potential (V m ) at a value more positive than the predicted reversal potential for H + (E H ).An example of this phenomenon can see seen in the first figure of Quade et al. (2022).As pH i rises, we gather a series of I-V plots such that each I-V plot can be assigned to a value of pH i .Figure 1A shows a selection of these plots gathered from a single SLC4A11-expressing oocyte as pH i is caused to rise under voltage-clamp.Note that the slope of the I-V relationship (i.e., membrane conductance, G m ) rises as pH i increases.The relationship between pH i and G m is shown for six SLC4A11-expressing cells in Figure 1B, in which the trace marked by black diamonds is the full data set from the cell shown in Figure 1A.The average G m.max was 90 ± 14 µS.Also shown in Figure 1B are resting pH i /G m relationships from six H 2 O-injected cells (crosses) for which the average G m was 2 ± 1 µS.In order to extract pK i values from the SLC4A11 data, we first normalize G m data from each cell to its own G m,max (Figure 1C) and fit those data to the Hill equation, generating the best-fit relationships described by a pK i and an apparent Hill coefficient (N app ) shown as gray lines in Figure 1D.The average of these relationships is represented as a black dottedline in Figure 1.We calculate an average pK i of 7.08 ± 0.04 for human SLC4A11 at pH e = 8.50, which is not different from the range that we had previously determined in Quade et al. (2022) (p = 0.32, two-tailed, unpaired t-test).

Determining pK i for SLC4A11 at pH e = 8.50 in the presence of 1 mM NH 4 Cl
The response of SLC4A11-expressing oocytes in pH e = 8.50 solution was altered in two ways in the presence of 1 mM NH 4 Cl.First, as shown in the example in Figure 2A, the cell acidified rather than alkalinized (initial dpH i /dt = 5.1 ± 0.9 × 10 −4 pH units/s, n = 6).This was also a feature of H 2 O-injected cells (initial dpH i /dt = 5.5 ± 0.8 × 10 −4 pH units/s, n = 6, Figure 2A inset).Second, unique to SLC4A11-expressing cells, G m unexpectedly rose to a plateau (over period "a") and subsequently declined to a value close to its starting value (over period "b") when pH i was further acidified by clamping V m at a value more negative than E H . Let us first focus our attention on period "b," during which acidification causes a familiar decline.Figure 2B shows a selection of responses from the cell represented in Figure 2A.The pH i versus G m relationships gathered from seven SLC4A11-expressing cells during period "b" are plotted in Figure 2C.The average G m.max was 89 ± 5 µS, which is not different from that for the group of cells assayed in the absence of NH 4 Cl (p = 0.94, two-tailed unpaired t-test).Also shown in Figure 2C are resting pH i /G m relationships from six control cells (originally injected with H 2 O in place of SLC4A11 cRNA) that were acidified prior to assay by HCl injection (crosses).The average G m of these cells was 11 ± 2 µS.Best-fit normalized G m versus pH i data for SLC4A11-expressing cells is shown in Figure 2D.We calculate that the average pK i for SLC4A11-expressing cells at pH e = 8.50 in the presence of 1 mM NH 4 Cl is 6.28 ± 0.05, which is significantly more acidic than the pK i range determined in the absence of NH 4 Cl (p < 0.001, one-tailed, unpaired t-test).On the other hand, there was no significant difference in the value of N app compared to its value in the absence of NH 4 Cl (P = 0.17, two-tailed unpaired t-test).
Does the action of SLC4A11 during period "a" represent a second pK i that reports acid-activation?We hypothesized that period "a" represented the time during which pK i was transitioning between its ± NH 4 Cl values.If we assume that neither G m,max nor N app change in a single SLC4A11-expressing oocyte during the course of an experiment such as that in Figure 2A, we can solve the Hill equation to generate an apparent pK i (pK i,app ) for each point in the experiment at which we have paired values of G m and pH i .The results of this approach are shown in Figure 3A and support the hypothesis that period "a" is dynamic time of pK i adjustment, while period "b" represents a time over which SLC4A11 has assumed a new and relatively stable pK i .In three of our seven experiments, we extended the protocol to examine a period "c" (example shown in Figure 3B) during which we could examine the behavior of SLC4A11 in the pH-range of period "a", but after SLC4A11 has assumed its + NH 4 Cl pK i .As shown in Figure 3C, SLC4A11 exhibits a similar pK i during period "b" versus period "c" pK i (P = 0.19: two-tailed, paired t-test) and does not revisit the apparent acid-activated behavior exhibited during period "a."In summary for this section, we find that the presence of 1 mM NH 4 Cl results in a significant acidic shift in the value of pK i for human SLC4A11 at pH e = 8.50.

Determining the ion-selectivity of SLC4A11 in the presence of 1 mM NH 4 Cl
In an earlier study we demonstrated that for SLC4A11expressing oocytes, the relationship between V m and the transmembrane pH gradient (64 mV/pH-unit) was close to Nernstian with respect to H + (Myers et al., 2016).Figure 4A shows equivalent data gathered in the presence of 1 mM NH 4 Cl from experiments such as those shown in Figures 2A, 3B.The average slope of the relationship is 71 ± 5 mV/decade (Figure 4B) with an x-axis intercept of −1.25 ± 0.04 (Figure 4C).Data from one of the three Figure 3B-style experiments is highlighted in Figure 4A, with black data points taken from period "b" and white data points taken from period "c."In summary for this section, the slope of the relationship between V m and transmembrane pH gradient does not appear to be greatly disturbed by the presence of 1 mM NH 4 Cl, except for the unusual observation that the relationship does not intersect with the origin.The meaning of this observation is explored in Section 3.3.

Comparing the influence of pH e and [NH 3 ] on SLC4A11 pK i
Using a similar work-flow to that described above, we determined the pK i of SLC4A11 at pH e = 7.50 in the presence of 1 mM NH 4 Cl (Figures 5A-C).G m.max in this cohort of cells was 104 ± 7 µS, which is not different from the equivalent range reported from cells assayed at pH e = 8.50 + NH 4 Cl (p = 0.09: two-tailed, unpaired t-test).Because the ratio of NH 3 :NH 4 + is pH-sensitive, we also determined the pK i of SLC4A11 at pH e = 8.50 in the presence of 0.12 mM NH 4 Cl (Figures 6A-C).In both conditions, although [NH 4 Cl] is different, [NH 3 ] is the same (0.017 mM) Figure 6D summarizes the values of pK i that have been determined during this study.We find that pKi at pHe = 7.50 in the presence of 1 mM NH 4 Cl is 7.06 ± 0.05 (Figure 5C), and the pK i at pH e = 8.50 in the presence of 0.12 mM NH 4 Cl is 7.04 ± 0.05 (Figure 6C).These values are not significant different from each other (p = 0.68: two-tailed, unpaired t-test).
We can make two additional statistical comparisons: [1] at pH e = 8.50, pK i is not significantly different in the presence or absence of 0.12 mM NH 4 Cl (P = 0.50: two-tailed, unpaired t-test), and [2] in the presence of 1 mM NH 4 Cl, raising pH e from 7.50 to 8.50 has a significant acidifying effect on pK i (p< 0.01: two-tailed, unpaired t-test).The interpretation of these findings are discussed in Section 3.4.

NH 4 Cl increases SLC4A11 G m by shifting pK i in the acidic direction
Our data show that, at pH e = 8.50, the extracellular presence of 1 mM NH 4 Cl results in a significant acidic shift in pK i such that, at almost any value of pH i between 6.0 and 7.3 (i.e., the range bounded by the two ± NH 4 Cl traces: solid and dashed black-lines in Figure 7A), G m would be increased.A formal determination of pK i at physiological pH e in the absence of NH 4 Cl has been precluded by a practical limitation on how high we can raise oocyte pH i , but we have previously estimated that the value must be more alkaline than 7.6 (Quade et al., 2020).The ability of NH 4 Cl to cause an acidic shift in pK i , enables us to determine a pK i range of 7.06 ± 0.05 at pH e = 7.50 in the present study.Thus, we can imagine that, even with the conservative estimate of the NH 4 Cl-free pK i , the implication is similar: the presence of NH 4 Cl would cause an increase in G m at typical physiological values of pH i (e.g., 7.0-7.3,which is included in the range bounded by the solid and dashed gray-lines in Figure 7A).Critically, this stimulatory effect of NH 4 Cl represents only an increase in G m caused by a redefinition of the pH i versus G m relationship.Neither G m,max nor N app are significantly altered by NH 4 Cl (at least as determined at pH e = 8.50) so the redefinition appears to represent a direct acidic-translation of the relationship.Although we are wary of assigning any meaning to the numerical value of N app (which is a function of the number of titratable moieties within SLC4A11), the observation of an unchanging N app at least implies that mechanism by which SLC4A11 responds to pH i is similarly complex in the absence and presence of NH 4 Cl.
We had once before investigated the role of NH 4 Cl in stimulating SLC4A11 (Myers et al., 2016).We found, as others had before us (Zhang et al., 2015), that the presence of 5 mM NH 4 Cl causes an increase in SLC4A11 G m (or I m at fixed V m in the case of those other studies).However, although we appreciated at that time that SLC4A11 was pH i -dependent, we did not consider the possibility that the relationship between pH i and G m could be modulated.The addition of 5 mM NH 4 Cl to (even H 2 O-injected) oocytes causes a rapid and robust depolarization (Musa-Aziz et al., 2009).This action in itself is sufficient to drive SLC4A11-mediated H + efflux, resulting in a cellular alkalinization that further activates SLC4A11 (Myers et al., 2016).Thus, in that study, when we increased G m to G m.max at pH e = 8.50 by adding 5 mM NH 4 Cl and saw no change in that value upon NH 4 Cl removal (maintaining a voltage clamp at 0 mV to mimic the depolarizing effect of NH 4 Cl presence), we assumed that the presence of NH 4 Cl was merely causing G m to rise according to the prescribed pH i vs. G m relationship.In light of our new data, we reinterpret those data as likely to have been gathered at a pH i value that was greater than the NH 4 Cl-free pK i , where G m values for both ± NH 4 Cl relationships are close to G m,max .That is to say that we reinterpret our data in favor of a model in which NH 4 Cl causes a direct rather than indirect allosteric activation of SLC4A11.

Data do not conclusively support a more important role for NH 3 versus NH 4 + for SLC4A11 action
Models in which SLC4A11 transports NH 3 /NH 4 + favor a NH 3 : nH + cotransport mechanism over NH 4 + transport because NH 3 increases in concentration with rising pH, thereby providing an explanation for the greater stimulation of SLC4A11 currents/ conductance at pH e = 8.50 than pH e = 7.50.We might use the same logic to conclude that NH 3 is more likely than NH 4 + to be the allosterically activating species.However, our data provide an alternative explanation: as shown in Figures 7A,B, at the typical resting pH i range for SLC4A11-expressing oocytes [6.9-7.0:(Quade et al., 2020)], the shift in pK i makes less difference to G m at pH e =  7A,B/upper panel) than at pH e = 8.50 (dark gray arrow in Figures 7A,B/lower panel).In this case it is not necessary to invoke a model in which the abundance of the substrate/activator is pH dependent; NH 4 + (whose fractional abundance is relatively unchanged between pH e 7.50 and 8.50) could be considered equally likely to be the species responsible for the observed stimulation.

(light gray arrow in Figures
Our data do not speak definitively to the nature of the activating species.We believe, because the time course of pK i shift (>10 min: Figure 3A) is much slower than the rate of solution turnover in the bath (<1 min) that the allosteric activation likely requires the activating species to accumulate intracellularly in order to exert its effect on SLC4A11.Because the handling of NH 3 /NH 4 + by oocytes is unusually complex (Musa-Aziz et al., 2009), it is difficult to specifically relate this time course of activation to the accumulation of either species.On the one hand NH 3 is presumed to be the more membrane permeable of the two species.On the other hand, in contrast to the expected alkalinization observed in mammalian cells exposed to NH 4 Cl, (even H 2 O-injected) oocyte pH i paradoxically acidifies as if NH 4 + is accumulating faster, perhaps entering via non-selective cation channels as NH 3 is sequestered in sub-membranous granules (Burckhardt and Frömter, 1992).For this reason, it is not clear whether there is a diagnostically useful differential in their rate of accumulation that could point to one species over the other.

SLC4A11 retains H + selectivity in the presence of NH 4 Cl
Another critical parameter that is unchanged in the presence of NH 4 Cl is its H + selectivity, because the V m of SLC4A11-expressing cells exhibits a close-to-Nernstian response to changes in the transmembrane pH gradient (Figure 4).This suggests that SLC4A11 remains a selective H + conductor rather than assuming a novel NH 3-coupling mode.Although previous studies have claimed to provide evidence of NH 3 -coupled H + transport using a similar approach (Zhang et al., 2015;Kao et al., 2020), we believe that these finding should be interpreted with caution as the approach violates necessary assumptions of reversal potential calculations: chiefly that substrates can only cross the membrane via SLC4A11 (not true for NH 3 ) and that V m is dominated by the action of SLC4A11 (not true due to the depolarizing action of NH 3 /NH 4 + ).In the absence of a specific inhibitor for SLC4A11, to distinguish the behavior of the protein from that of the system, the best that such calculations may achieve is a description of the permeability of the system (i.e., SLC4A11 and its membrane environment).One such study concluded that SLC4A11 was capable of both NH 3 :H + cotransport and H + transport (Kao et al., 2020).If so, this is not the typical action of an obligatorily coupled cotransport protein and could equally describe the behavior of a H + conductor in an NH 3permeable membrane.
There is one unusual aspect to our reversal potential data that requires further explanation.In previous studies we have found that the relationship between the transmembrane gradient and V m for wild-type SLC4A11 crosses the x-axis at a value close to zero as expected for a H + conductor.In the present study performed in the presence of NH 4 Cl, we find that the relationship is substantially offset from the origin and intersects the x-axis at −1.25 pH-units as if we had underestimated pH i by 1.25 units.This is not impossible, as seven of the ten data sets were gathered while SLC4A11 was mediating H + influx and thus pH immediately below the membrane may have been more acidic than bulk pH i being measured at the tip of our microelectrode, which is impaled deeper into the cell.However, three of these data sets were gathered while SLC4A11 was mediating H + efflux, and exhibited the same offset, so we do not believe that this can be the correct explanation.An alternate explanation is illustrated in Figure 8.Here we show that, if we consider these data as being representative of the system and interpret them using a modified Goldman-Hodgkin-Katz equation, we can reproduce the offset by implementing a small permeability to a depolarizing cation.Two examples are provided: in Model 1, we add a sodium permeability to the system.Because the abundance of H + is so small compared to the abundance of Na + , the relative permeability of Na + to H + must be very small not to completely overwhelm V m .In this instance, P Na /P H = 6 × 10 −10 provides a good fit to our observations.In Model 2 we add an NH 4 + permeability to the system.We assume that [NH 4 Cl] is 1 mM on both sides of the membrane and calculate [NH 4 + ] i for each value of pH i .In this instance P NH4 / P H = 6 × 10 −8 provides a good fit for our data.Although the relationship curves off to an asymptote, the initial slope is Nernstian with respect to H + , and its projection (dashed gray line) crosses the x-axis at −1.25 pH units.We note that the x-intercept of our data is also a projection and we do not know whether our data would also curve in a similar way if extended towards the x-axis.In any case this is not a critical issue as the relationship can be made to conform to the dashed line if we lower our estimate of [NH 4 Cl] i to 0.1 mM.As either permeability is trivial compared to that of H + , it does not appear to be a major confounding factor to our hypothesis.In the absence of a specific blocker, we cannot know whether the additional permeability is intrinsic to SLC4A11 or to the system in general.However, as we have only observed this in the presence of NH 4 + , and because NH 4 + is a depolarizing influence in even H 2 O-injected cells, we tentatively suggest that the x-axis offset represents the previously described endogenous NH 4 + conductance.

The relationship between pK i shifts caused by extracellular alkalinization versus NH 4 Cl addition
As previously observed, in the absence of NH 4 Cl, a shift of pH e from 7.50 to 8.50, is itself sufficient to acid-shift SLC4A11 pK i by more than 0.5 pH-units (Quade et al., 2020).If we compare SLC4A11 pK i determined at pH e = 7.50 + 1 mM NH 4 Cl to pK i determined at pHe = 8.50 + 0.12 mM NH 4 Cl (conditions in which [NH 3 ] is the same and pK i is not different) we may conclude that, in the presence of NH 4 Cl, pK i has become pH e -independent and is determined by [NH 3 ] alone.In that case we could consider [NH 3 ] and pH e as activators that share a common mechanism.If SLC4A11 is less pH e dependent in the presence of NH 4 Cl, that would imply that the stimulatory effect of pH e on SLC4A11 may have limited relevance to its in vivo action.
On the other hand, if we compare SLC4A11 pK i determined at pH e = 7.50 + 1 mM NH 4 Cl to pK i determined at pH e = 8.50 + 1 mM NH 4 Cl, values which are ~0.8 units apart, we may conclude that the phenomenon of pH e -dependence is preserved in the presence of 1 mM NH 4 Cl and that the actions of NH 4 Cl and pH e are, at least in part, mechanistically independent and additive.The common ground between these two ways of looking at the data are that the presence of NH 4 Cl alters the relationship between pH e and pK i such that a more acidic pK i can be achieved at a given pH e .Interestingly, this is the opposite to a phenomenon that we have observed in relation to certain pathological SLC4A11 mutants such as R125H, in which the mutation causes pK i to become more alkaline at a given pH e (Quade et al., 2022).Unfortunately, the technical limitations on how far we can extend pH i and the low resolution of our pK i assay (due to the large standard error intrinsic to the data sets) currently preclude us from more detailed exploration of the relationship between these activating parameters.

Implications for the physiological role of SLC4A11
The fluid pumping action of corneal endothelial cells requires the action of a basolateral Na/K-ATPase to generate the transmembrane sodium gradient that drives the Na + /CO 3 = cotransporter, drawing osmolytes from the stromal fluid to discourage fluid accumulation.The pump is energized by glutaminolysis that feeds α-ketoglutarate into the TCA cycle to generate ATP; a side product of this reaction is NH 3 (Zhang et al., 2017).We have previously hypothesized that acid-loading by SLC4A11 could be useful to stabilize pH i during robust bicarbonate pump function, responding to Na + /CO 3 = cotransporter action by sensing a local rise in pH i (Myers et al., 2016;Nehrke, 2016).Our new data indicate that SLC4A11 is a H + conductor both in the presence and absence of NH 4 Cl and thus that NH 3 /NH 4 + is an allosteric activator rather than a cotransported substrate of SLC4A11.This action is incompatible with a role of Models of additional, non-H + permeabilities in SLC4A11-expressing cells in the presence of NH 4 Cl Two models based on a modified Goldman-Hodgkin-Katz equation in which the implementation of a small cationic permeability (either Na + in the case of Model 1 or NH 4 + in the case of Model 2) causes an offset in the projected x-axis intercept, as observed in our experiments conducted at pH e = 8.50 + 1 mM NH 4 Cl (per Figure 4).corneal endothelial SLC4A11 in mediating the export of excess NH 3 from glutaminolysis, and incompatible with SLC4A11 being able to harness an outwardly-directed NH 3 gradient to mediate H + efflux.Because the electrochemical gradient for H + is typically inwardly directed, we predict that a rise in intracellular NH 3 /NH 4 + promotes H + influx independent of any rise in pH i .The ability of NH 3 /NH 4 + to acid shift pK i implies that the ability of SLC4A11 to support pump function is potentiated by NH 3 /NH 4 + , the generation of which could be considered to be a proxy for the energetic requirements of the pump.

Summary
The presence of NH 4 Cl causes an acidic shift in the pK i of human SLC4A11, which translates to an increase in G m at physiological values of pH i .The presence of NH 4 Cl does not affect the H + selectivity of SLC4A11, thus we conclude that NH 4 Cl is an allosteric activator of SLC4A11-mediated H + conductance.The influence of increasing [NH 4 Cl] upon SLC4A11 activity is reminiscent of the influence of increasing pH e , but further work will be required to determine whether these share a common mechanism.

Oocyte preparation and culture
Ovaries were harvested from female Xenopus laevis (Xenopus Express, Brooksville, FL) in accordance with the protocol approved by the University at Buffalo Institutional Animal Care and Use Committee.Frogs were anesthetized in 0.2% tricaine solution, ovariectomized, and euthanized by exsanguination.Extracted tissue was cut into ~1 cm 2 pieces and washed in a Ca 2+ -free solution (82 mM NaCl, 2 mM KCl, 20 mM MgCl 2 , 5 mM HEPES, pH 7.50).Oocytes were liberated by digestion in 2 mg/mL type 1A collagenase solution, and the isolated cells were washed further in the Ca 2+ -free solution to remove the collagenase prior to resuspension in a physiological buffer (ND96: 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 5 mM HEPES, pH 7.50, 200 mOsmol/kg H 2 O).Until experimental use, oocytes were cultured at 18 °C in OR3 medium (14 g/L Leiboviz's L-15 medium powder, 5 mM HEPES, 20 mL/L 100x penicillin-streptomycin, pH 7.50, 200 mOsmol/kg H 2 O).

cRNA preparation and injection
Our starting material was a clone of human SLC4A11-B in pBSXG4 vector.The construct was linearized with HindIII, which cuts at a site downstream of the open reading frame providing a termination point for transcription.The linearized DNA was purified using a MinElute PCR Purification Kit (QIAgen, Germantown, MD) used as a template for cRNA synthesis using the T7 mMESSAGE mMACHINE kit (Invitrogen, Carlsbad, CA).cRNA was purified using an RNeasy MinElute Cleanup KIt (QIAgen).25 ng of cRNA or H 2 O was injected into each oocyte using a Nanoject programmable injector (Drummond Scientific, Broomall, PA).

Electrophysiology
Oocytes were placed into chamber (RC-3Z: Warner Instruments, Hamden, CT) on an anti-vibration table (Vision IsoStation; Newport Corp., Irvine, CA) and were superfused at 2 mL/min with solutions fed from syringe pumps (Harvard Apparatus, Holliston, MA).Borosilicate glass capillaries (BF200-156-10: Sutter Instrument, Novato, CA) were pulled into microelectrodes (such that they exhibited a tip resistance of 0.1-2 MΩ when filled with saturated KCl solution) using a micropipette puller (P-1000: Sutter Instrument).Oocytes were impaled with two such KCl-filled microelectrodes (one currentpassing and one voltage-sensing) connected to an oocyte clamp (OC275: Warner Instruments, Hamden, CT).A bath clamp (725I: Warner Instruments) was used to hold the potential of the chamber fluid at 0 mV.Current-voltage (I-V) plots were gathered in 20 mV, 100 ms steps, returning to the spontaneous membrane potential for 100 ms between each step.H + -selective microelectrodes were pulled in the same manner as voltage electrodes but the tips were filled with hydrogen ionophore I/cocktail B (Sigma Aldrich) and backfilled with a solution composed of 40 mM KH 2 PO 4 , 15 mM NaCl, pH 7.0.These electrodes were connected to a dual-channel electrometer (HiZ-223: Warner Instruments).Complete technical details can be found in the 2013 review by Lee, Boron, and Parker (Lee et al., 2013).Signals were digitized via a Digidata 1550 unit and captured using Clampex 10.4 software (Molecular Devices LLC, San Jose) and custom continuous acquisition software (written by Mr. Dale Huffman for Walter Boron's laboratory at Case Western Reserve University, Cleveland, OH).

Electrophysiology solutions
pH 7.50 solutions contained 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 5 mM HEPES, 200 mOsmol/kg H 2 O. pH 8.50 solutions had the same composition but were buffered with 5 mM Bicine in place of HEPES.NH 4 Cl was added to NH 4 Clcontaining solutions as a powder, and pH was readjusted as necessary.

Data analysis
Data are presented as means ± standard error of the mean.Slope conductance (G m ) was determined from the slope of a linear trendline fit to I-V data in Microsoft Excel.Normalized G m data was plotted against pH i (expressed as [OH − ]) and fit to the Hill equation using the solver add-in of Excel to determine the values of EC 50 and N app that would result in the minimum root square difference between the observed data and the outcome of the Hill equation: EC 50 was converted into a value of pK i using the following equation: For Figure 3A, we calculated pK i,app by solving the Hill equation for EC 50 at each pair of G m and [OH − ] (i.e., pH i ) data points.We assumed that G m,max and N app were constants: 79 µS and 7 respectively, corresponding to the data gathered from this cell during period "b."Calculations of [NH 3 ] at a given pH value and [NH 4 Cl] assume a pK a for the NH 3 /NH 4 + equilibrium of 9.25.Statistical analysis was performed in Excel using unpaired t-tests, one-or two-tailed as necessary.For multiple comparison, ANOVA was performed using MiniTab software.

FIGURE 1 SLC4A11
FIGURE 1 SLC4A11 behavior in the absence of NH 4 Cl (extracellular pH = 8.50) (A) A representative selection of current-voltage (I-V) relationships gathered from a single SLC4A11-expressing oocyte as intracellular pH (pH i ) rises.(B) The relationship between pH i and slope conductance (G m ) from the full-set of I-V relationships gathered from the SLC4A11-expressing cell shown in panel (A) (black diamonds) and from five other SLC4A11-expressing cells (each represented by its own symbol).The pH i versus G m relationship at resting pH i for five H 2 O-injected oocytes is represented by crosses.(C) SLC4A11 data from panel (B), normalized to its respective maximum G m (G m,max ).(D) Best-fit lines for each cell to the Hill equation are shown in gray.The dashed black line represents the Hill equation generated using the average pK i and apparent Hill coefficients (N app ) of the n = 6 replicates, as shown in the inset table.

FIGURE 2 SLC4A11
FIGURE 2 SLC4A11 behavior in the presence of 1 mM NH 4 Cl (extracellular pH = 8.50).(A) A representative example of the pH i (top) and G m (bottom) response of an SLC4A11-expressing oocyte to the addition of 1 mM NH 4 Cl.The first G m point in the data series (white cross) was gathered prior to the presence of NH 4 Cl.The inset shows a representative pH i response to the addition of 1 mM NH 4 Cl (point of solution change indicated by gray triangle) of a H 2 O-injected cell at pH e = 8.50.(B) A representative selection of I-V relationships gathered from a single SLC4A11-expressing oocyte as pH i falls during period "b".(C) The relationship between pH i and G m from the full-set of I-V relationships gathered from the SLC4A11-expressing cell shown in panel (B) (black diamonds) and from six other SLC4A11-expressing cells (each represented by its own symbol).The pH i versus G m relationship for six H 2 O-injected oocytes after acidification by HCl injection is represented by crosses.(D) Best-fit lines for each cell to the Hill equation are shown in gray.The dashed black line represents the Hill equation generated using the average pK i and N app of the n = 7 replicates, as shown in the inset table.

FIGURE 3
FIGURE 3Validation of change in SLC4A11 pK i upon addition of NH 4 Cl.(A) G m data from Figure2Areplotted with the apparent pK i (pK i,app ) calculated from the pH i at each value of G m .(B) A representative experiment similar to that shown in Figure2A, extended into a third experimental period in which G m is monitored during a return to starting pH i .(C) G m values from panel (B) plotted for each of the three experimental periods, normalized to G m,max for each period.The inset table shows pK i and N app for periods "b" and "c" calculated from best-fit data to the Hill equation for three such experiments.

FIGURE 4
FIGURE 4 The ion-selectivity of SLC4A11 in the presence of NH 4 Cl.(A) The relationship between V m and the transmembrane pH gradient (pH i -pH e ), determined from the experiments represented in Figures 2, 3, are shown by solid gray lines.One example data set is shown in full from period "b" (filled circles) and period "c" (open circles).The Nernstian slope of ideal H + selectivity is shown as a gray dashed line.(B) The average slopes calculated from the data in panel (A) are shown as filled (data gathered during period "b") and open (data gathered during period "c") circles.The cross represents the average of these data.(C) The average x-intercepts calculated from the data in panel (A) are shown as filled (data gathered during period "b") and open (data gathered during period "c") circles.The cross represents the average of these data.

FIGURE 5 SLC4A11
FIGURE 5 SLC4A11 behavior in the presence of 1 mM NH 4 Cl (extracellular pH = 7.50).(A) A representative selection of current-voltage (I-V) relationships gathered from a single SLC4A11-expressing oocyte as intracellular pH (pH i ) rises.(B) The relationship between pH i and slope conductance (G m ) from the full-set of I-V relationships gathered from the SLC4A11-expressing cell shown in panel (A) (black diamonds) and from five other SLC4A11-expressing cells (each represented by its own symbol).(C) Best-fit lines for each cell to the Hill equation are shown in gray.The dashed black line represents the Hill equation generated using the average pK i and apparent Hill coefficients (N app ) of the n = 6 replicates, as shown in the inset table.

FIGURE 6 SLC4A11
FIGURE 6 SLC4A11 behavior in the presence of 0.12 mM NH 4 Cl (extracellular pH = 8.50).(A) A representative selection of current-voltage (I-V) relationships gathered from a single SLC4A11-expressing oocyte as intracellular pH (pH i ) rises.(B) The relationship between pH i and slope conductance (G m ) from the full-set of I-V relationships gathered from the SLC4A11-expressing cell shown in panel (A) (black diamonds) and from five other SLC4A11-expressing cells (each represented by its own symbol).(C) Best-fit lines for each cell to the Hill equation are shown in gray.The dashed black line represents the Hill equation generated using the average pK i and apparent Hill coefficients (N app ) of the n = 6 replicates, as shown in the inset table.(D) A summary of the pK i of SLC4A11 detemined under the various conditions tested in this study.

FIGURE 7
FIGURE 7 Models of SLC4A11 action in the presence and absence of NH 4 Cl.(A) Cartoon representing the pH i versus G m relationships for SLC4A11 at pHe 7.50 (gray lines) and 8.50 (black lines) in the presence (solid lines) and absence (dashed lines) of 1 mM NH 4 Cl.(B) Cartoon showing how the scheme in panel (A) could explain how NH 4 Cl can achieve greater potency with respect to enhancing SLC4A11-mediated conductance at pH e = 8.5 (lower panel) versus 7.5 (upper panel), as observed in prior studies such as Zhang et al., 2015 or Myers et al., 2016.I-V plots are cartoons, gray circles represent cells expressing SLC4A11 (gray boxes) with black arrows indicating relative magnitudes of conductance.