Isolation and Characterization of CA XIV, a Novel Membrane-bound Carbonic Anhydrase from Mouse Kidney*

Carbonic anhydrase (CA) is involved in various physiological processes such as acid-base balance and transport of carbon dioxide and ions. In this study, we have succeeded in the isolation of a novel CA from the mouse kidney by use of the signal sequence trap method. It is a 337-amino acid polypeptide with a calculated molecular mass of 37.5 kDa, consisting of a putative amino-terminal signal sequence, a CA domain, a transmembrane domain, and a short hydrophilic carboxyl terminus, which we designated CA XIV.1 The CA domain of CA XIV is highly homologous with those of known CAs, especially extracellular CAs including CA XII, IX, VI, and IV. The expression study of an epitope-tagged protein has suggested that CA XIV is located on the plasma membrane. When expressed in COS-7 cells, CA XIV exhibits CA activity that is predominantly associated with the membrane fraction. By Northern blot analysis, the gene expression of CA XIV is most abundant in the kidney and heart, followed by the skeletal muscle, brain, lung, and liver. In situ hybridization has revealed that, in the kidney, the gene is expressed intensely in the proximal convoluted tubule, which is the major segment for bicarbonate reabsorption and also in the outer border of the inner stripe of the outer medulla. In conclusion, we have cloned a functional cDNA encoding a novel membrane-bound CA. This study will bring new insights into our understanding of carbon dioxide metabolism and acid-base balance.

of tissues where they participate in a broad range of physiological processes such as acid-base balance, carbon dioxide, and ion transport, respiration, body fluid generation, bone resorption, ureagenesis, gluconeogenesis, and lipogenesis (1,2). In mammals, CAs organize a family of 9 isoenzymes (CA I-VII, IX, and XII) with defined CA activity, 3 CA-related proteins (CA VIII, X, and XI), and 2 subtypes of receptor-type protein tyrosine phosphatase (RPTP ␤ and ␥) (3). These proteins share a well conserved CA or CA-like domain and differ in tissue distribution and subcellular localization (1,2). CA II is expressed in almost all tissues, whereas expression of others are restricted (CA VI and CA VII in the salivary gland; RPTP ␤ in the brain) (4,5). CA IX and XII are expressed more abundantly in various cancers than in corresponding normal tissues (6)(7)(8). Subcellularly, they are cytosolic (CA I, II, III, and VII), membrane-bound (CA IV, IX, and XII, RPTP ␤ and ␥), mitochondrial (CA V), or secreted (CA VI). The CA-like domain of RPTP ␤ does not have CA activity, but provides a binding site for a cell surface signal transducing molecule, contactin (9).
In the kidney, carbon dioxide metabolism and acid-base balance are regulated by two CA isoenzymes, CA II and CA IV (2,10). CA II catalyzes hydration of carbon dioxide in the cytoplasm, and CA IV catalyzes the reverse reaction on the plasma membrane. Thus, CA II and CA IV, together, enhance the net proton secretion and bicarbonate reabsorption (10). Functional importance of CAs has been investigated also in pathologic conditions. For instance, CA II and CA IV are up-regulated in the kidney in metabolic acidosis (11). Furthermore, CA II gene mutations cause a hereditary disease characterized by renal tubular acidosis, osteopetrosis, and cerebral calcification (12).
We have searched for novel soluble and membrane-bound proteins in the mouse kidney using the signal sequence trap method (13)(14)(15)(16). This method enables isolation of not only signal transducing molecules such as cytokines and receptors but also some sorts of enzymes and transporters (13)(14)(15)(16)(17)(18)(19). Here we describe the molecular cloning, sequence analysis, enzyme activity, and distribution of a novel membrane-bound CA, designated CA XIV.

EXPERIMENTAL PROCEDURES
Tissue Preparation and RNA Extraction-The whole kidney and other tissues were obtained from 8-week-old male BALB/c mice. Total RNA extraction was carried out as described by Chomczynski and Sacchi (20). Poly(A) ϩ RNA was purified using PolyATract (Promega, Madison, WI).
Signal Sequence Trap-Signal sequence trap was performed as described (13)(14)(15)(16). 2 g of poly(A) ϩ RNA from the mouse kidney was reverse transcribed by random hexamer priming using SuperScript II reverse transcriptase (Life Technologies, Inc.), and deoxyadenosine (dA) tails were added at the 3Ј end of the first strand cDNA. The second strand was synthesized with a specific primer containing polydeoxythymidine (dT) and an EcoRI restriction site and ligated with a SacI * This work was supported in part by research grants from Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists, the Japanese Ministry of Education, Science, Sports, and Culture, and the Japanese Ministry of Health and Welfare. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB005450.
adaptor. The cDNA fragments of 300 -700 base pairs in size were isolated by an agarose gel electrophoresis and subcloned into the EcoRI and SacI sites of pcDL-SR␣-Tac(3Ј) vector (21), to generate a fusion protein with interleukin-2 receptor ␣ chain (or Tac antigen) lacking its own signal sequence (22). The expression plasmid library thus obtained was transfected into COS-7 cells by the lipofection method using Transfectam (Sepracor, Marlborough, MA). The fusion proteins containing artificial amino-terminal signal sequences were sorted to the cell surface, retained on the plasma membrane by a transmembrane domain of Tac antigen, and detected by cell-surface immunostaining with anti-Tac antibody (23). Otherwise, they remained intracellularly and were not recognized by the antibody.
Rapid Amplification of 3Ј-cDNA Ends (3Ј-RACE)-The 3Ј-RACE experiment was performed using Marathon cDNA amplification kit (CLONTECH Inc., Mountain View, CA) (24). 1 g of poly(A)ϩ RNA from the mouse kidney was reverse transcribed by oligo(dT) 15 priming using SuperScript II reverse transcriptase. After synthesis of the second-strand cDNA by the method of Gubler and Hoffman (25) and ligation with the marathon cDNA adaptor, polymerase chain reaction was carried out using a 5Ј-gene-specific primer (5Ј-TGGAAAATC-AAGTCCCTGGAAGTC-3Ј) (nucleotides Ϫ147 to Ϫ124, Fig. 1) and an adaptor primer (5Ј-CCATCCTAATACGACTCACTATAGGGC-3Ј). The 3Ј-RACE product was subcloned into the pCR II vector (Invitrogen Corp., San Diego, CA) for sequencing.
DNA Sequencing-Nucleotide sequences were determined on both strands by the dideoxy chain-termination method using Dye Terminator cycle sequencing kit FS and 373B DNA sequencer (Applied Biosystems Inc., Foster City, CA).
Phylogenetic Tree Construction-Phylogenetic tree of CAs and CArelated proteins was constructed by comparing the amino acid sequences of their CA or CA-like domains using the alignment algorithm of GeneWorks software (Oxford Molecular Group Inc., Campbell, CA), which is based on the unweighted pair group method with arithmetic mean (26). Transient Expression and Measurement of CA Activity-A full-length mouse CA XIV cDNA (nucleotides Ϫ12 to 1150) was synthesized by reverse transcription-polymerase chain reaction using the following primers: sense, 5Ј-TGTGGGGATAATATGTTGTTCTTCG-3Ј; antisense, 5Ј-GGGGTCCCTGGTGTATAGAGAGGG-3Ј. The nucleotide sequences were confirmed by sequencing. The cDNA was ligated into the EcoRI restriction site of an expression vector pCXN2, a derivative of pCAGGS (27). The plasmid was transfected into COS-7 cells using Transfectam. Cells were harvested 72 h later and homogenized in a buffer containing 50 mM Tris-SO 4 (pH 7.5) and 1 mM benzamidine (Sigma). The cell homogenate was centrifuged at 100,000 ϫ g for 30 min and separated into the membrane fraction and cytosolic fraction. CA activity was measured essentially as described by Sundaram et al. (28). SDS-resistant CA activity was determined by measuring samples preincubated with 0.2% SDS before assay (29). Protein concentrations were determined by the Lowry method (30) using bovine serum albumin as a standard. CA activity was calculated by the following formula: CA activity (units/mg) ϭ (T 0 /T Ϫ 1)/C; where T 0 represents the reaction time for buffer alone (s); T, sample reaction time (s); C, protein content (mg). The reaction times were measured in triplicate, and the mean values were used to calculate CA activities.
Northern Blot Analysis-Northern blot analysis was performed as described with [ 32 P]dCTP-labeled cDNA insert of clone G31C5 (nucleotides Ϫ265 to 282) (15). The blot was used to expose BAS-III imaging plate (Fuji, Kanagawa, Japan) for 40 h.
In Situ Hybridization Analysis-Sense and antisense [ 35 S]CTP-labeled cRNAs were generated from a partial cDNA fragment of the 3Ј-RACE product (nucleotides Ϫ147 to 698) using T7 and SP6 RNA polymerases (Promega Corp.). In situ hybridization analysis was performed as described previously (15,31). In brief, 10-m cryosections of the mouse kidney were mounted on poly-L-lysine-coated slides, fixed with paraformaldehyde, acetylated, and hybridized with the cRNA probes. Slides were washed, dehydrated, and apposed to Hyperfilm ␤-max films (Amersham Int., Buckinghamshire, UK) for 10 days, or dipped into autoradiographic emulsion (NTB-2, Eastman Kodak) and exposed for 10 weeks and counterstained with hematoxylin and eosin.

RESULTS AND DISCUSSION
Isolation and Sequence Analysis of Mouse CA XIV-5,000 clones from the mouse kidney cDNA library were screened by the signal sequence trap method, and 25 positives were isolated. A positive clone G31C5 with a 555-base pair insert encoded a novel 94-amino acid polypeptide. To obtain the fulllength cDNA, 3Ј-RACE was performed, and a 1.6-kilobase fragment was obtained and sequenced. The full-length cDNA encoded a novel CA-like protein with 337 amino acid residues, which had a calculated molecular mass of 37.5 kDa (Fig. 1). Hydropathy analysis (32) predicted that the protein consists of a signal sequence, a CA domain, a transmembrane domain, and a short hydrophilic carboxyl terminus, which we designated CA XIV. CA XIV possessed an N-glycosylation motif in its CA domain and several potential phosphorylation sites by protein kinases A and C in the carboxyl terminus (33). When clone G31C5 was expressed in COS-7 cells as an epitope-tagged protein, it was sorted to the cell surface and detected by cell surface immunostaining (see "Experimental Procedures"). These findings suggest that CA XIV is located on the plasma membrane with its CA domain facing extracellular space.
The amino acid sequence of the CA domain of CA XIV was highly homologous with those of other members of the CA family (Fig. 2). The amino acid identities were 43% for human CA IX and XII, 38% for CA VI and VII, 35% for CA I, II, III, and IV, 32% for CA V and VIII, and 31% for RPTP ␤ and ␥ (3). The homologies were especially high in the putative CA active site residues including 3 conserved zinc-binding histidine residues ( Fig. 1) (3, 34). When CA XIV was compared with CA VII, which is considered the common ancestor of the CA family (3), as many as 30 of 36 active site residues were identical. These findings imply that CA XIV possesses CA activity (3).
Homology search using TBLASTN algorithm (35) revealed the presence of partial cDNA fragments of the putative human counterpart for CA XIV in the data bases, which are Ͼ80% identical to mouse CA XIV at the amino acid level, e.g. expressed sequence tag clones with GenBank TM accession numbers H82563, R87427, and AA401879. The structural conservation beyond species may suggest the physiological importance of CA XIV.
Phylogenetic Tree Construction-To examine the evolutional history of CA XIV among the CA family, a phylogenetic tree was constructed (Fig. 2). The tree consisted of three clusters (3). The first cluster included intracellular CAs (CA I-III, V, and VII), and the third cluster included RPTP ␤ and ␥. CA XIV was most closely related to CA XII, followed by CA IX, VI, and IV, forming the second cluster.
As shown in this study, CA XIV likely localizes on the plasma membrane, whereas CA IX is a membrane-bound protein that exists on the plasma membrane and in the nucleus (6), CA VI is a secreted type of CA (4), and CA IV is a CA attached to the plasma membrane by glycosylphosphatidylinositol (GPI) linkage (36,37). Thus, the second cluster may be categorized as extracellular CAs. Consistent with their subcellular localization, CA IV, VI, IX, XII, and XIV possessed putative signal sequences in their amino termini (Fig. 2).
Measurement of CA Activity-To examine whether CA XIV has CA activity, the full-length cDNA was expressed in COS-7 cells. The homogenate of mock-transfected cells exhibited low CA activity, which may be attributed to endogenous CAs (Table  I). Transfection with the CA XIV cDNA resulted in an increase in CA activity by 5.2-fold. The membrane fraction of the homogenate possessed 2.1-fold higher activity than the whole homogenate. After the SDS treatment, which inactivates soluble CAs (29), the CA activity decreased only by 27%. These findings indicate that CA XIV is located predominantly in the membrane fraction.
Northern Blot Analysis-Tissue distribution of CA XIV gene expression was investigated by Northern blot analysis (Fig. 3). The CA XIV mRNA was 1.8 kilobases in size and was expressed most abundantly in the kidney and heart, followed by the skeletal muscle, liver, brain, and lung. No CA XIV transcript was detected in the intestine and spleen. The tissue distribution of CA XIV mRNA is very similar to that of CA IV mRNA, which is expressed in the colon, lung, brain, kidney, and heart but not in the intestine and spleen (37).
In Situ Hybridization Analysis-The intrarenal localization of CA XIV was determined by in situ hybridization analysis with the antisense and sense cRNA probes (Fig. 4). At autoradiograph with the antisense probe, strong hybridizing signals were observed in the cortex. At photomicrograph, these signals were confined to the proximal convoluted tubule, which is known as the most important segment for bicarbonate reabsorption in the nephron (10). At autoradiograph, moderate signals were also seen in the outer border of the inner stripe of the outer medulla. In view of shape, position, number, and comparison with previous reports using histochemical methods (38,39), these signals most likely correspond to the initial portion of long loop of the thin descending limb of Henle. Physiological roles of CAs in the thin descending limb of Henle still remain to be elucidated (10,40). Identification of specific inhibitors for CA XIV may facilitate our understanding concerning such issues (2,10). No specific signals were seen in sections hybridized with the sense probe (not shown).
In the kidney, two CA isoenzymes, CA II and CA IV, have been well characterized (10). CA II is a cytosolic isoenzyme localized in the intercalated cells of the distal tubule and collecting duct, and the thin descending limb of Henle (41,42). On the other hand, CA IV is a membrane-bound isoenzyme localized in the proximal convoluted tubule and the thick ascending limb of Henle (43). Thus, the intrarenal localization of CA XIV is distinct from those of CA II and CA IV, suggesting its unique function in the kidney.
Comparison of CA XIV with CA IV-CA XIV resembles CA IV in that it has a putative signal sequence in its amino terminus, is membrane-bound, and likely locates on the plasma  (26). Values indicate estimated genetic distances, which are proportional to the lengths of horizontal lines. Thick lines at branching points represent error bars whose lengths indicate standard errors of the branch positions. Note that all proteins are from human, except CA XIV from mouse. CA VIII is also called CA-related protein; CA IX, MN protein; RPTP ␤, receptor-like protein tyrosine phosphatase . ‚, putative signal sequence; CA, CA or CA-like domain; TM, transmembrane domain; GPI, anchoring site to glycosylphosphatidylinositol; PTPase, protein tyrosine phosphatase domain. membrane. Furthermore, the tissue distribution of CA XIV is also very similar to that of CA IV. However, CA XIV and CA IV differ in two points. First, the intrarenal localizations of CA XIV and CA IV differ. Although they both localize in the proximal convoluted tubule, CA XIV likely localizes also in the thin descending limb of Henle, whereas CA IV localizes in the thick ascending limb of Henle. Second, CA XIV is presumed to locate on the plasma membrane by a transmembrane domain, whereas, in case of CA IV, the transmembrane domain in the extreme carboxyl terminus of the precursor protein is cleaved off, and the mature protein is bound to the plasma membrane by a GPI anchor (36,37). The GPI-anchored form of CA IV is released from the membrane by phosphatidylinositol-specific phospholipase C (PI-PLC) treatment (36). Although CA IV was defined as a membrane-bound isoenzyme of CA (2, 10), approximately half of the membrane-bound CA activity in the kidney still remains after PI-PLC treatment (29). One interpretation for this finding is that some part of CA IV exists as a membrane-spanning form, which is PI-PLC insensitive (29), yet it is also possible that other membrane-bound CAs are present in the kidney. A novel CA described here, CA XIV, is one of such candidates.
Conclusion-In this study, we describe the isolation and characterization of a novel CA designated CA XIV. The primary structure and the expression studies suggest that CA XIV is a type I membrane protein localized on the plasma membrane. Because CA XIV mRNA is expressed abundantly in various tissues and CA XIV has enzyme activity, CA XIV may play important roles in carbon dioxide metabolism and acid-base balance in the kidney and other tissues.  (magnification, ϫ10). B, brightfield photomicrograph of cortex (magnification, ϫ150). Cx, cortex; OM, outer medulla; IM, inner medulla; OS, outer stripe; IS, inner stripe; G, glomerulus.