The renal proximal tubule of vertebrates performs an essential role in controlling plasma SO42− concentration ([SO42−]). Although net tubular SO42− reabsorption is the predominate control process in terrestrial vertebrates, a facilitated secretory flux is also present. In contrast, marine teleosts obtain excess SO42− from drinking, and increased plasma [SO42−] is prevented predominately through net tubular secretion. Tubular SO42− secretion is accomplished by at least two electroneutral anion exchange processes in series. Movement of SO42− into the cell across the basolateral membrane is pH dependent, suggesting SO42−/OH− exchange. Luminal HCO3− and Cl− can facilitate SO42− movement out of the cell across the brush-border membrane. The molecular identities of the anion exchangers are unknown but are probably homologues of SO42− transporters in the mammalian SLC26 gene family. In all species tested, glucocorticoids increase renal SO42− excretion. Whereas glucocorticoids downregulate SO42− reabsorptive mechanisms in terrestrial vertebrates, they may also stimulate a mediated secretory flux. In the marine teleost, cortisol increases the level of SO42−/HCO3− exchange at the brush-border membrane, tubular carbonic anhydrase (CA) activity, CAII protein, and a proportion of tubular SO42− secretion that is CA dependent. CA activity is required for about one-half of this net SO42− secretion but is also required for about one-half of the net reabsorption in bird proximal epithelium. A CA-SO42−/anion exchanger metabolon arrangement is proposed that may speed both the secretory and reabsorptive processes.
- renal proximal tubule
- transport metabolon
inorganic so42− is the second most abundant anion in the sea and one of the most abundant anions in vertebrate plasma, with concentrations varying among species from 0.3 to 1.8 mM. This large tetrahedral divalent anion can be obtained from the diet (83) or oxidation of the sulfur-containing amino acids cysteine and methionine (77). Sulfoconjugation reactions are the initial detoxification process for many endogenous (e.g., steroid hormones) and xenobiotic compounds and are required for the activation of numerous biological molecules (e.g., heparin, cholecystokinin, and gastrin) (59). SO42− is a major component of glycosaminoglycans (dermatan sulfate, chondroitin sulfate, keratan sulfate, and heparan sulfate), which are structural components of numerous tissues, including cartilage, bone, myelin, and basal lamina (40). SO42− deficiency in chondrocytes causes phenotypes (chondrodysplasias) that are characterized by severe growth abnormalities (24). Elevated plasma SO42− can cause diuresis and acidosis (18), and increased SO42− concentration ([SO42−]) in the renal tubule lumen reduces Ca2+ and Mg2+ reabsorption (66).
RENAL SULFATE CLEARANCE
The kidneys are the major avenue for SO42− excretion in vertebrates. Nearly all of plasma SO42− is ultrafilterable in mammals (4), whereas 80% and 47% of plasma SO42− are ultrafilterable in birds (71) and fish (75), respectively. Variation in the estimates of the ultrafilterable fraction of plasma SO42− in vertebrates may arise from species differences or variations in method of determination. Urine content is determined by glomerular filtration, tubular reabsorption, and tubular secretion. This review will focus on the latter and will provide only a brief overview of tubular reabsorption because it has been described in detail elsewhere (2, 51, 52, 57).
Reabsorption of filtered SO42− occurs primarily in the proximal tubule (35) and is required for SO42− homeostasis in mammals and presumably all other terrestrial and freshwater vertebrates. SO42− clearance ratios [SO42− clearance/glomerular filtration rate (GFR)] >1 have not been measured in any terrestrial vertebrate, suggesting that net renal tubular SO42− secretion does not occur in these animals. SO42− loading apparently has no effect on the saturable reabsorptive transport maximum for SO42− (TmSO42−) in the renal tubule of the human, dog, and frog (3, 28, 48). In contrast, TmSO42− in rats is reduced ∼50% after plasma SO42− loading, indicating that adaptive mechanisms are in place to depress reabsorption, thus increasing excretion (27). Similarly, in domestic fowl, TmSO42− decreases with plasma SO42− loading, leading to the conclusion that a fraction of the excreted SO42− is the result of tubular secretion (30). Thiosulfate, an inhibitor of SO42− transporters, reduces an SO42− secretory flux in perfused renal tubules of the rabbit (13) and in renal proximal tubule epithelial monolayers (PTCs) prepared from the domestic chicken (Gallus gallus) (22) (Fig. 1). Brazy and Dennis (13) further demonstrated inhibition of the SO42− secretory flux in rabbit tubules by SITS, suggesting an anion exchange mechanism. Renal tubular thiosulfate secretion in dogs may share a common secretory mechanism with SO42− (6). The marine herring gull (Larus argentatus) can produce urine containing 77 mM SO42− after seawater ingestion (unpublished data by Knut Schmidt-Nielsen; cited in Ref. 4). Measurement of renal SO42− clearance in these animals may reveal net renal tubular SO42− secretion. Taken together, these data indicate that net SO42− secretion does not occur in renal tubules of terrestrial vertebrates (marine birds may be an exception). However, a facilitated unidirectional secretory flux appears to be operational and may contribute to renal SO42− excretion during periods of elevated plasma [SO42−].
Although mediated renal tubular SO42− secretion is present in terrestrial vertebrates, the highly expressed reabsorptive mechanism tends to obscure the process. Marine teleosts on the other hand have a highly expressed renal SO42− secretory system, and because they have no facilitated reabsorption, they are an ideal model system for investigating tubular secretion. The bony fishes maintain their plasma osmolality (∼300 mosmol/kgH2O) below that of the surrounding environment (∼1,000 mosmol/kgH2O) and thus undergo passive water loss. Dehydration is avoided through continuous drinking of seawater, which contains 25 mM SO42−. Water absorption by the intestine generates a high [SO42−] in the intestinal lumen (50–100 mM), and the steep concentration gradient promotes passive SO42− absorption across the intestinal mucosa (34, 60, 62). Although passive SO42− absorption dominates, an active secretory flux driven by an electroneutral brush-border SO42−/Cl− exchanger is also present (62). The major portion of the SO42− burden acquired from drinking is eliminated by renal excretion.
GFR can be low, intermittent, or even absent (species with aglomerular kidneys) in marine teleosts. Despite this, their urine contains very high concentrations (>50 mM) of SO42− (5, 33). In the marine summer flounder (Paralichthys lethostigma), renal SO42− clearance ratios average 12, and 93% of excreted SO42− is the result of tubular secretion (75). Furthermore, SO42− loading revealed a powerful secretory system with a transport maximum of 150 μmol·kg body wt−1·h−1 (75). In another marine species, the winter flounder (Pleuronectes americanus), perfused renal tubules were shown to concentrate SO42− in the lumen (7), and the transepithelial concentration and electrical potential (−1.9 mV, lumen negative) differences indicated that net SO42− secretion was active (10). Flounder PTCs in Ussing chambers were used to definitively demonstrate net active SO42− secretion (21). These epithelial sheets exhibit differentiated properties of proximal tubule cells, including ciliary activity, structural polarity, low transepithelial resistance, greater permeability to Na+ than Cl−, and phlorizin-sensitive, sodium-coupled glucose transport. In the presence of flounder saline (1 mM SO42−), typical measurements yield unidirectional secretory fluxes of 98.4 ± 6.67 nmol·cm−2·h−1, unidirectional reabsorptive fluxes of 4.8 ± 1.08 nmol·cm−2·h−1, and net fluxes of 93.6 ± 7.64 nmol·cm−2·h−1 under short-circuited conditions (69). Because freshwater generally contains very low levels of SO42−, and teleosts limit their drinking in freshwater, ingested food items should provide the only external source of SO42− available to freshwater teleosts. Although it has not been demonstrated, the kidney of freshwater teleosts should be a site of net SO42− reabsorption.
MECHANISMS OF TUBULAR REABSORPTION
Although the emphasis of this review is tubular secretion, the reabsorptive process is mentioned briefly to draw attention to a newly discovered role of CA in the process. Cellular phospholipid bilayers have a very low passive permeability to SO42−, and, therefore, transport proteins are required to efficiently move SO42− across the plasma membrane (38). Reabsorptive transport of SO42− across the brush-border membrane (BBM) is Na+ dependent in the intact proximal tubule (88) and in isolated BBM vesicles from mammalian (80, 87) and avian (71) species. The Na+-SO42− cotransporter (NaSi-1) has been identified from a rat renal cortex cDNA library by expression cloning (54). NaSi-1 protein localizes to the BBM of renal proximal tubule cells (20).
Electroneutral Na+-independent SO42−/HCO3− exchange has been demonstrated in renal basolateral membrane (BLM) vesicles isolated from the rat (65), chicken (71), and rabbit (44) (Table 1). In addition to SO42−/HCO3− exchange, Kuo and Aronson (44) found oxalate/HCO3− exchange in rabbit renal BLM vesicles and concluded that all three anions (SO42−/HCO3−/oxalate) are substrates for the same transporter. A transporter exhibiting SO42−/oxalate exchange similar to that found in renal BLM vesicles from mammals was cloned from rat liver and called sat-1 (11). Markovich et al. (53) injected rat kidney cortex mRNA into Xenopus oocytes and demonstrated Na+-independent SO42− uptake that was inhibited by oxalate and thiosulfate, which is similar to that found for liver sat-1. Size fractionation of the mRNA followed by expression in Xenopus oocytes and Northern hybridization provided further evidence for sat-1 in the renal cortex. Sat-1 was later localized to the BLM of proximal tubule cells with an anti-sat-1 antibody (41).
The involvement of CA in renal tubular SO42− reabsorption has recently been demonstrated in the chicken proximal tubule (22). CA catalyzes the reversible dehydroxylation of HCO3− and hydroxylation of CO2 (HCO3− ↔ OH− + CO2). Both CA biochemical activity and CA isoform II (CAII) protein are present in chick PTCs, and net SO42− reabsorption is inhibited ∼50% after CA inhibition. It is not known whether intracellular and/or extracellular CA activity is required for tubular SO42− reabsorption, because ethoxzolamide, a soluble CA inhibitor with a relatively high lipid-to-water partition coefficient, was used in the aforementioned study.
MECHANISMS OF TUBULAR SECRETION
Tubular SO42− secretion is the emphasis of this review, because the process is seldom discussed in other reviews of SO42− homeostasis. Although a facilitated SO42− secretory flux has been observed in birds and mammals, it is perhaps best expressed in marine teleosts. Anion exchange mechanisms for SO42− were identified in the renal tubular BBM and BLM of two species of marine flounder (75, 76). Concentrative SO42− uptake into renal BLM vesicles is pH dependent, suggesting SO42−/OH− exchange (Table 1) (75). SO42− transport by BBM vesicles is stimulated by HCO3−, Cl−, SCN− and S2O32−, with HCO3− serving as most effective counteranion (76). Anion exchange at the BBM is not pH dependent, distinguishing it from the basolateral SO42−/OH− exchanger (75). Unlike renal BBM vesicles from avian and mammalian species, an Na+ gradient has no effect on concentrative SO42− uptake into marine teleost renal BBM vesicles (75). Both basolateral SO42−/OH− exchange and brush-border SO42−/anion exchange are electroneutral processes that are inhibited by the anion exchange inhibitor DIDS. It is not clear for either pole of flounder proximal tubule epithelial cells whether a single anion exchanger or multiple exchangers facilitate SO42− transport.
Anion exchange mechanisms behaving similarly to those in the BBM and BLM of the marine teleost proximal tubule have also been found in the renal tubule of birds and mammals (Table 1). Although less effective than HCO3−, OH− was shown to trans-stimulate SO42− uptake into rat renal BLM vesicles (65). SO42− uptake into rat renal BBM vesicles is trans-stimulated by HCO3−, Cl−, OH−, SCN−, NO3−, and I−, with HCO3− being most effective (64). HCO3−-stimulated SO42− uptake into chicken BBM vesicles is inhibited by OH−, SCN−, and S2O32−, suggesting that these anions can act as substrates (71). The aforementioned anion exchange processes in the terrestrial vertebrate proximal tubule may perform a role similar to those in the marine teleost renal tubule.
SO42−/anion exchangers important for SO42− secretion most likely represent one or more of the SO42− transporters within the mammalian SLC26 gene family. There have been 11 SLC26 genes (SLC26A1-SLC26A11) identified and seven of these have been shown to transport SO42− when expressed in heterologous systems (58). Of the seven that transport SO42−, five (SLC26A1, SLC26A2, SLC26A6, SLC26A7, and SLC26A11) are expressed in the mammalian kidney. SLC26A1 encodes the SO42−/HCO3−/oxalate exchanger (sat-1) identified in the BLM of mammalian renal proximal tubule cells (41) (Table 2). In addition, SLC26A1 exhibits Cl− transport activity when expressed in Xenopus oocytes (45). SLC26A2 (DTDST), which appears to be expressed ubiquitously (32), has affinities for many anions, including SO42−, HCO3−, Cl−, oxalate, and S2O32− (79). The putative Cl−/formate exchanger SLC26A6 (PAT-1, CFEX) has been localized to the BBM of mammalian renal proximal tubule cells (43). In addition to Cl− and formate, SLC26A6 has affinities for oxalate, HCO3−, and SO42− (39, 93). SLC26A7 has been localized to the BLM of rat outer medullary collecting duct cells where it has been hypothesized to function in HCO3− reabsorption (63). When expressed in Xenopus oocytes, SLC26A7 exhibits HCO3−, Cl−, oxalate, and SO42− transport activity (47, 63). SLC26A11 mRNA transcript is present in the mammalian kidney and displays DIDS-sensitive Na+-independent SO42− transport when expressed in insect Sf9 cells (92).
Whereas anion exchangers seem certain to play a role in tubular SO42− secretion, exactly how they are involved in vivo is less certain. Luminal [SO42−] in perfused renal tubules of the winter flounder is ∼10 mM (7). With the use of estimates of luminal volume (42) and measurements of [SO42−] in the luminal (10) and tubular fluid spaces (73), Renfro (70) estimated total intracellular [SO42−] to be 2 mM. Hence, the driving force for SO42− exit across the BBM has to be questioned because the anion exchange mechanism is electroneutral and SO42− appears to move against a chemical gradient. Although HCO3− can trans-stimulate SO42− uptake into BBM vesicles (Table 1), it is not apparent whether luminal concentrations in vivo are high enough to facilitate SO42− exit across the BBM. BBM vesicle experiments have shown that Cl− can trans-stimulate SO42− transport (Table 1). Cl− appears to be a plausible substrate to drive SO42− exit from the cell because luminal [Cl−] is the same as or slightly higher than in plasma (∼160 mM) (10) and the Nernstian concentration of intracellular Cl− would be ∼15 mM at membrane potential difference of −60 mV. The data presented in Fig. 2 show that in the absence of luminal HCO3−, Cl− is able to maintain transepithelial SO42− secretion. The movement of Cl− down its chemical gradient into the cell likely energizes the uphill movement of SO42− out of the cell across the BBM.
Our unpublished data also show that in the absence of luminal Cl−, 5 mM HCO3− is able to maintain transepithelial SO42− secretion by flounder PTCs (Fig. 2), albeit at a reduced level. In contrast, complete removal of luminal Cl− and HCO3− together and reduction of luminal SO42− to 0.1 mM to prevent SO42−/SO42− exchange completely abolishes net SO42− secretion (Fig. 2). With the use of a conservative estimate of 7.3 for intracellular pH and a dissociation constant (pK) for OH− + CO2 ↔ HCO3− of 6.3 at 20°C, and assuming that intracellular PCO2 is identical to extracellular PCO2 (4 mmHg), the calculated intracellular HCO3− concentration would be ∼2 mM. The tubular fluid in the early proximal tubule of marine teleosts would have a [SO42−], pH (∼7.7), and [HCO3−] (∼5 mM) identical to plasma, and thus [HCO3−] in the initial filtrate should be high enough to energize SO42− exit across the BBM. The final urine of marine teleosts is acidic (∼6.6) and, therefore, contains low levels of HCO3− (<1 mM) (33, 50, 74). Assuming that the proximal tubule is responsible for most HCO3− reabsorption and proton secretion, the late proximal tubule luminal [HCO3−] should be greatly reduced compared with the calculated intracellular [HCO3−]. Luminal [HCO3−] in the Japanese bullfrog (Rana catesbeiana) proximal tubule is slightly lower than intracellular [HCO3−] (29). It can be concluded, therefore, that whereas HCO3− may drive SO42− secretion in early proximal tubule segments, it cannot energize SO42− exit across the BBM in the late segments where the Cl− gradient may be essential.
Uptake of SO42− into flounder renal BLM vesicles is electroneutral (75). The OH− gradient should be unfavorable for SO42− uptake via SO42−/OH− exchange given that intracellular pH in marine teleosts is ∼0.6 pH units lower than plasma (15). Similarly, the SO42− gradient across the BLM should also be unfavorable because plasma [SO42−] in the flounder is ∼0.6 mM, whereas total intracellular [SO42−] is ∼2 mM. Even so, there exists the possibility that the level of free intracellular SO42− available for exchange is much less than the estimated 2 mM. Intracellular Cl− can bind to proteins and phospholipid surfaces and can be sequestered in organelles making the exchangeable [Cl−] much less than the total (38). If SO42− behaves similarly, then there exists the possibility that the exchangeable intracellular [SO42−] is less than the interstitial [SO42−]. Another possibility, however, is that the pH of an unstirred layer of fluid in the peritubular interstitices may be lower than intracellular pH thus providing an OH− gradient to drive the process.
In mammals, H+ secretion indirectly drives proximal tubular HCO3− reabsorption, and 40% of tubular HCO3− reabsorption is CA dependent (67). Proximal tubular CA is also required for a fraction (50%) of tubular SO42− reabsorption in chickens (22). Although the role of CA in the marine fish kidney has long been questioned because CA inhibition has no effect on urine pH (see Table 3) or [HCO3−] (37), it is now recognized that CA activity subserves proximal tubular SO42− secretion in this tissue (74). CA inhibition in vivo reduces renal tubular SO42− secretion ∼40% (Table 3). CA activity is localized to the proximal tubule, and treatment of flounder PTCs with the cell soluble CA inhibitors, ethoxzolamide or methazolamide, reduces net active SO42− secretion 40–60% (Fig. 3). Only the active secretory flux is sensitive to CA inhibition. The polymer-linked CA inhibitor, polyoxyethylene-aminobenzolamide, which is restricted to the extracellular space, has no effect on SO42− secretion, indicating that extracellular CA activity has no role in tubular SO42− secretion (Fig. 3).
CAII, which is a soluble (cytosolic) isoform, has been shown to physically associate with numerous mammalian ion transporters including AE1 (band 3) (85), Na+/H+ exchanger isoform 1 (NHE1) (46), and Na+-HCO3− cotransporter isoform 1 (NBC-1) (1, 31), and, in all cases, this association enhances transport activity. An allosteric effect can be ruled out because coexpression of AE1 (85) and NBC-1 (1) with an inactive CAII mutant (V143Y), which retains binding affinity, inhibits transport activity. The complex that is formed from the association of CA with a transport protein has been termed a “transport metabolon” (85). A metabolon is defined as a series of physically associated enzymes involved in a coupled metabolic pathway that allows metabolites to move efficiently between active sites by limiting the loss of intermediates to diffusion (84, 85). Even for enzymes that approach “kinetic perfection,” such as CA, the rate-limiting step in enzymatic catalysis is diffusion (19). Copeland (19) uses the electron transport system as an example of an enzyme pathway that has overcome the diffusion limitation by stating, “Because of the proximity of the enzymes in the membrane, the product leaves the active site of one enzyme and is presented to the active site of the next enzyme without the need for diffusion through solution.” A CA transport metabolon can be envisioned in the same manner, with presentation of the product of enzymatic catalysis (HCO3− or OH−) in close proximity to the anion-binding site on the transporter or, alternatively, enzymatic catalysis may remove substrate from the vicinity of the anion-binding site. For anion exchange, this association may effectively create a functional asymmetry, which favors the transport of a particular substrate (HCO3− or OH−) in one direction.
The association of mammalian CAII and AE1 occurs between a histidine-rich region in the NH2 terminal of CAII and a hydrophobic residue followed by 2–3 acidic residues in the COOH terminal of AE1 (85, 90, 91). In addition to CAII acceleration of transport activity, the COOH-terminal motif of AE1 (LDADD) increases CAII activity (81). Numerous other ion transporters, including Pendrin (SLC26A4), a member of the mammalian SLC26 gene family, possess potential binding sites for CAII (85). Figure 4 shows that the SO42− transporters SLC26A1, SLC26A2, SLC26A7, and SLC26A11, which are expressed in the mammalian kidney, also possess potential CAII binding sites at their COOH terminals. SLC26A6, the putative Cl−/formate exchanger in the BBM of mammalian proximal tubule cells does not appear to contain a CAII binding motif, at least in the last 40 amino acids of the COOH terminal. Because the cytoplasmic portion of the COOH-terminal tail could not be confidently identified it cannot be ruled out that SLC26A6 contains binding sites for CAII.
CAII has been cloned from numerous fish species and the amino terminus contains histidine residues that may form an electrostatic interaction with either glutamate or aspartate residues. A portion of CAII (329 bp) was recently cloned from the winter flounder proximal tubule, and CAII protein was shown to be both cytosolic and membrane associated (61). Furthermore, CAII protein migrated differently than purified bovine CAII on nondenaturing gels, suggesting that it associates with another membrane component. Unpublished data from our laboratory show that CAII immunolocalizes to the cytosol, as well as in or near the BBM and BLM of intact flounder proximal tubule cells (Fig. 5).
IS THERE A REGULATED RELATIONSHIP BETWEEN SO42− TRANSPORT AND CA?
Dietary SO42− and K+ availability, acid/base status, heavy metals, and membrane fluidity affect renal SO42− handling (2, 51, 52, 57). Hormones that stimulate SO42− reabsorption include growth hormone, insulin-like growth factor I, progesterone, estradiol, vitamin D3, and triodothyronine (2, 51, 52, 57).
In all species tested, glucocorticoids stimulate renal SO42− excretion. Renal SO42− clearance is elevated in rats treated (in vivo) with the glucocorticoid methylprednisolone (78). Furthermore, a reduced Vmax value for Na+-dependent SO42− transport in renal BBM vesicles and reductions in NaSi-1 mRNA and protein levels occur concurrently with the increase in renal SO42− excretion. The Vmax for Na+-dependent SO42− transport in renal BBM vesicles is similarly reduced in domestic chickens (G. gallus) treated with the long-lived cortisol agonist dexamethasone (72). Treatment of chick PTCs with supraphysiological concentrations of cortisol for 24 h selectively reduces net transepthelial SO42− reabsorption ∼50% (22). The decrease in net SO42− reabsorption is the result of an increase in the unidirectional secretory flux and decrease in the unidirectional reabsorptive flux (Fig. 6). This study provided evidence that both the facilitated SO42− reabsorptive and secretory fluxes in the avian proximal tubule are regulated by glucocorticoids.
In teleosts, cortisol is essential for osmoregulation in both freshwater and seawater (23, 25, 55, 56). Net renal tubular SO42− secretion ceases (SO42− clearance ratio <1) after acclimation of winter flounder to 10% seawater (69). The cessation in net SO42− secretion is associated with the complete inhibition of SO42−/HCO3− exchange in the BBM. Treatment of 10% seawater acclimated winter flounder with dexamethasone in vivo restores SO42−/HCO3− exchange in the subsequently isolated BBM vesicles to levels comparable to seawater-acclimated controls. Interestingly, intravenous SO42− infusion into flounder acclimated to 10% seawater stimulated net renal SO42− secretion after 2.5 h, indicating that the secretory mechanism can be elicited rapidly. The mechanism by which SO42− secretion is acutely stimulated has yet to be determined.
Reduction of flounder PTC culture medium cortisol to 10% of normal serum concentration for 9–12 days reduces SO42− secretion to 33% of the control value (69). Nonspecific effects, however, could not be ruled out because transepithelial resistance and phlorizin-sensitive glucose current, measures of proximal tubule-like function, were also significantly altered by cortisol removal. Cortisol is a known differentiating factor (26), and altered electrical properties in this study imply partial dedifferentiation of the epithelium.
Cortisol has recently been shown to alter CA expression in flounder PTCs (61). Control flounder PTCs were supplemented with high levels of cortisol (7.3 × 10−6 M), and removal of cortisol for 5 days resulted in cortisol concentrations (6.3 × 10−9 M) at the lower end of the physiological range (8.4 × 10−9-3.5 × 10−7 M). The 5-day exposure was chosen knowing that longer treatments cause partial dedifferentiation of this epithelium (69). Both CA biochemical activity (30%) and CAII protein abundance (65%) are reduced in flounder PTCs maintained in reduced cortisol for 5 days. In addition, net active SO42− secretion is 28% lower after the reduced cortisol treatment (Fig. 7). Methazolamide treatment sufficient to block 100% of the CA biochemical activity and methazolamide in combination with reduced cortisol inhibit net active SO42− secretion to the same extent (∼60%). The reduced cortisol level does not alter transepithelial resistance or phlorizin-sensitive glucose current, indicating that the epithelium remains in a differentiated state. The observations that reduction of cortisol to very low levels 1) lowers CA expression (biochemical and protein), 2) inhibits net active SO42− secretion, but 3) has no effect on the CA-independent fraction of SO42− secretion led the authors to conclude that cortisol influences a component of SO42− transport that is CA dependent.
In his classical review of the physiology of CA, Maren (49) pointed out that inhibitor concentrations sufficient to inhibit 99.2% of CA enzymatic activity have no physiological effect on functional parameters of kidney, eye, pancreas, and stomach. However, renal HCO3− excretion increases and intraocular pressure decreases at inhibitor concentrations that inhibit 99.7% of enzymatic activity, indicating that 1% of CA activity is required for normal physiological function in the majority of tissues. Despite this, a few investigations demonstrated that relatively small changes in CA activity can be physiologically relevant. Renal tubular CA activity is elevated 78% in rabbits undergoing metabolic acidosis, and increased renal acid excretion accompanies the induction of CA activity (14). As noted above, small changes in CA activity induced by cortisol alter the CA-dependent fraction of tubular SO42− secretion (61). If CA is in such excess (∼500-fold), why should relatively small changes in enzymatic activity cause physiologically important changes in tubular anion transport? This may occur if 1) CA physically associates with the anion transporter, 2) the association stimulates transport activity, and 3) the treatments (e.g., metabolic acidosis and cortisol) directly affect the number of functional associations.
HYPOTHETICAL MODELS OF TUBULAR SULFATE SECRETION
Consideration of the known properties of transepithelial SO42− secretion gives rise to several hypothetical models (Fig. 8). In Fig. 8A, SO42− enters the cell across the BLM in exchange for cellular OH−. At the opposite pole of the cell, inward HCO3− movement facilitates SO42− exit across the BBM. Intracellular CA activity accelerates the formation of OH− ions, the substrate for basolateral SO42−/OH− exchange. CO2 formed from the dehydroxylation of HCO3− diffuses across the BLM and reacts with OH− to reform HCO3− in the interstitium.
It is well documented that mammalian CAII binds to and enhances the activity of the Cl−/HCO3− exchanger, AE-1 (68, 86). Thus it is entirely feasible that CAII binds to and enhances brush-border SO42−/HCO3− exchange activity in marine teleosts as well. At normal intracellular pH, the ratio of HCO3− to CO2 is 20:1. Therefore, it should be questioned whether an association of CAII with the brush-border anion exchanger (SO42−/HCO3−) would be stimulatory. On the other hand, an association of CAII with the basolateral SO42−/OH− exchanger would most likely enhance transport because CAII could cause rapid regeneration of transported OH− ions from HCO3− near the intracellular OH− binding site.
All cells are faced with a continuous acid load due to factors such as metabolic production of acid, passive HCO3− efflux, and passive proton influx (12). In the case of the marine teleost proximal tubule, the basolateral SO42−/OH− exchanger also contributes to the acid loading processes. The models in Fig. 8 show that Na+/H+ exchange activity, as in almost all cells, is an important defense against the tendency for acid loading processes to cause intracellular acidification. Maintenance of normal pH assures normal intracellular [HCO3−]. Na+,K+-ATPase activity is necessary to generate the Na+-gradient needed to drive Na+/H+ exchange. SO42− uptake across the peritubular surface of flounder renal tubules is dependent on the Na+ gradient (73, 75), and there is precedent for Na+/H+ exchange activity in teleost renal BBM vesicles (Anguilla anguilla) (89, 94). NHE-3 was recently cloned from Osorezan dace (Tribolodon hakonensis), and NHE-3 mRNA was found in its kidney (36). Although we hypothesized that an association between CAII and an SO42− anion exchanger aids SO42− secretion, a physical association between CAII and an Na+/H+ exchanger may be important as well. NHE-1 is a ubiquitous Na+/H+ exchanger involved in intracellular pH regulation (82). Like hNHE-1, the COOH terminal of winter flounder NHE-1 contains numerous acidic residues that could serve as binding sites for CAII (Fig. 4). It remains to be seen whether an Na+/H+ exchanger is required for maximum tubular SO42− secretion. Na+-HCO3− cotransport activity may also be important but has not yet been implicated in renal tubular SO42− transport.
A typical value for tubular SO42− secretion rate in winter flounder is 20 μmol·kg−1·h−1, which exceeds the estimated maximum possible rate of tubular HCO3− reabsorption (9 μmol·kg−1·h−1) (74). As luminal [HCO3−] decreases, continued SO42− secretion requires that another anion facilitate SO42− exit from the cell. In Fig. 8B, luminal Cl− energizes transfer of SO42− from cell to lumen. Cl− secretion has been demonstrated in proximal tubules of the dogfish shark (Squalus acanthias) and seawater and freshwater acclimated killifish (Fundulus heteroclitus) (9, 16, 17). The proposed mechanism includes the movement of Cl− into the cell on an Na+-K+-2Cl− cotransporter, accumulation of intracellular Cl− above electrochemical equilibrium, and exit of Cl− across the BBM on a cAMP-activated conductance pathway (possibly CFTR) (8). Thus Cl− gained through electroneutral SO42−/Cl− exchange at the BBM could be recycled to the lumen via this Cl− secretory process (Fig. 8, B and D).
SO42− secretion is reduced just 50% after CA inhibition, indicating that CA-independent mechanisms are responsible for a portion of SO42− secretion. Although orders of magnitude slower than the CA catalyzed reaction, the uncatalyzed dehydroxylation of HCO3− may contribute OH− ions for SO42−/OH− exchange (Fig. 8C). At present, the possibility cannot be ruled out that an anion other than OH−, such as a recyclable metabolic intermediate (e.g., oxalate), also facilitates SO42− translocation across the BLM (Fig. 8D).
CONCLUDING REMARKS AND PERSPECTIVES
The renal proximal tubule is an important point of control for the maintenance of plasma [SO42−] in vertebrates. Whereas a facilitated secretory flux exists in the terrestrial vertebrate proximal tubule the highly expressed reabsorptive flux dominates control of net excretion. In contrast, the marine teleost proximal tubule has highly expressed mechanisms that promote net tubular SO42− secretion, with no evidence of facilitated reabsorption. Unique SO42−/anion exchangers at the BBM and BLM facilitate secretion. On the basis of their similarity in substrate specificity the SO42−/anion exchangers in the marine teleost proximal tubule are probably homologues of SO42− transporters (SLC26A1, SLC26A2, SLC26A6, SLC26A7, and SLC26A11) in the mammalian SLC26 gene family. At least two of these transporters, SLC26A1 and SLC26A6, are present in the mammalian proximal tubule (41, 43).
CA has been shown to physically associate through an electrostatic interaction with numerous mammalian ion transporters, thus increasing their transport activities. CA activity is associated with a large fraction of SO42− secretion in the marine teleost proximal tubule and SO42− reabsorption in the avian proximal tubule, and a physical association of CA with transporters involved in SO42− transport may be necessary for maximum transport activity. Many of the SO42− transporters in the mammalian SLC26 gene family contain potential CAII binding motifs at their COOH termini, raising the possibility that these transporters may physically interact with CAII. Future work should be directed at identifying the SO42−/anion exchangers in the marine teleost proximal tubule and assessing the functional interaction of CA with these SO42− transporters.
The unpublished work reported here was supported by the National Science Foundation.
The authors thank Dr. D. Miller (NIH/NIEHS) for assistance with fluorescence confocal microscopy.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society