Identification of a Novel Ankyrin Isoform (AnkG190) in Kidney and Lung That Associates with the Plasma Membrane and Binds α-Na,K-ATPase*

Ankyrins are a family of adapter molecules that mediate linkages between integral membrane and cytoskeletal proteins. Such interactions are crucial to the polarized distribution of membrane proteins in transporting epithelia. We have cloned and characterized a novel 190-kDa member of this family from a rat kidney cDNA library, which we term AnkG190 based on the predicted size and homology with the larger neuronal AnkG isoform. AnkG190 displays a unique 31-residue amino terminus, a repeats domain consisting of 24 repetitive 33-residue motifs, a spectrin binding domain, and a truncated regulatory domain. Probes derived from the unique amino terminus hybridize to an 8-kilobase message exclusively in kidney and lung and specifically to the kidney outer medullary collecting ducts by in situ hybridization. Transfections of Madin-Darby canine kidney and COS-7 epithelial cell lines with a full-length AnkG190 construct result in (a) expression at the lateral plasma membrane, (b) functional assembly with the cytoskeleton, and (c) interaction with at least one membrane protein, the Na,K-ATPase. Two independent Na,K-ATPase binding domains on AnkG190 are demonstrated as follows: one within the distal 12 ankyrin repeats, and a second site within the spectrin binding domain. Thus, ankyrins may interact with integral membrane proteins in a pleiotropic manner that may involve complex tertiary structural determinants.

Interactions between integral membrane proteins and the underlying spectrin-based cytoskeleton play key roles in several cellular activities, including motility, contact, and the maintenance of polarized membrane domains (1)(2)(3). Ankyrins are a family of conserved proteins that have emerged as crucial adapter molecules mediating such linkages, since they possess binding sites for several integral membrane proteins as well as for cytoskeletal elements (4 -6). Molecular cloning has identified three distinct genes, encoding for a variety of ankyrins that are expressed in vertebrates as tissue-specific and alternatively spliced isoforms. The ANK1 gene encodes for the 210and 220-kDa Ank1 (also called Ank R ) (see Refs. 6 and 12 for nomenclature) isoforms in red blood cells (7) and for multiple small isoforms in skeletal muscle (8). ANK2 encodes for a family of alternatively spliced brain isoforms referred to as Ank2 or Ank B (9,10). ANK3 is a newly described ankyrin gene that transcribes several Ank3 isoforms that display a generalized tissue distribution (hence also termed Ank G ). The largest Ank G protein described to date is a 480-kDa isoform localized at the axonal initial segment and Node of Ranvier (11). Alternatively spliced isoforms of this parent Ank G protein include a 215-kDa peptide encoded for by a 7-kb 1 mRNA in mouse epithelial tissues (12), a Golgi-associated 116-kDa species (Ank G 119) expressed predominantly in epithelial cells and muscle (13), and small 120/100-kDa Ank3 molecules associated with mouse macrophage lysosomes (14). Additional immunoreactive Ank3 isoforms at 200, 170, 120, and 105 kDa have been detected in mouse tissues, but incompletely characterized (12).
Such rich isoform diversity may be especially critical to the maintenance of a specific pattern of membrane protein distribution in polarized epithelial cells such as those lining various segments of the kidney tubules, each of which vectorially transports specific ions and nutrients. It is likely that such specialized tissues harbor additional isoforms of ankyrin. Indeed, antibodies raised against Ank R recognize a prominent 190-kDa ankyrin polypeptide in kidney tissue (13,(15)(16)(17)(18), the primary structure of which is unknown. However, Ank R -deficient NB/NB mice continue to express the 190-kDa kidney ankyrin, indicating that the latter is encoded by a gene distinct from ANK1, presumably an ANK3 gene (4,13,19). Based on data obtained from cross-reacting antibodies, this 190-kDa isoform is thought to be distributed in a polarized fashion along the basolateral membranes of both native and cultured renal epithelial cells, where it colocalizes with spectrin and Na,K-ATPase (3,4,(15)(16)(17)(18). Importantly, the polarized distribution of ankyrin (and Na,K-ATPase) is markedly disrupted following an ischemic kidney injury and is restored following recovery (20,21), suggesting an important role for the 190-kDa ankyrin in the generation and maintenance of renal cell polarity.
In this study, we describe the cDNA cloning and characterization of the 190-kDa ankyrin isoform from rat kidney, which we term Ank G 190 based on its predicted size and homology to the 480-kDa neuronal Ank G isoform. Ank G 190 displays a unique 31-residue amino terminus, a repeats domain consisting of 24 repetitive 33-residue motifs characteristic of ankyrins, a highly conserved spectrin binding domain, and a truncated regulatory domain. Complementary DNA probes derived from the unique amino terminus hybridize to an 8-kb message exclusively in kidney and lung, and cRNA probes derived from the same unique sequences hybridize specifically to kidney outer medullary collecting duct cells by in situ hybridization. Transfections of MDCK (Madin-Darby canine kidney) and COS-7 epithelial cell lines with full-length Ank G 190 constructs result in (a) expression at the lateral membrane, (b) functional assembly with the cytoskeleton, and (c) interaction with at least one integral membrane protein, namely Na,K-ATPase. These findings significantly extend the known diversity of ankyrin isoforms, directly confirm the interaction of Ank G 190 with Na,K-ATPase, and suggest that ankyrins may interact with integral membrane proteins in a complex, pleiotropic manner that requires complex tertiary structural determinants not easily deducible from the primary structure.

EXPERIMENTAL PROCEDURES
Isolation of the Ank G 190 cDNA-All molecular biologic procedures were carried out using standard methods (22). Oligonucleotides bracketing a highly conserved 525-bp region extending from repeat 11 to 16 of human erythrocyte and brain ankyrins (7,9,10) were used in a high stringency PCR reaction (23), with rat kidney RNA reverse-transcribed with random hexamer priming and avian myeloblastosis virus reverse transcriptase (Boehringer Mannheim) as template. The sense primer was GGCTTTACCCCCTTACACATCGCCTGCAAAAAGAAC and the antisense primer was CTTGGCCGCCACGTGCAGAGGGGTAAATC-CTTTCTTGGT. This yielded a single 525-bp product which was subcloned into the TA vector (Invitrogen) and sequenced by the dideoxynucleotide chain termination method (U. S. Biochemical Corp.). Since its amino acid sequence was only about 55% homologous to previously described ankyrins (7,9,10), the PCR product was used to generate a random primed 32 P-labeled probe to screen a rat kidney cDNA library at high stringency (Stratagene). Three full-length clones were isolated, which were identical in sequence and differed only in the length of the 5Ј-untranslated region; the largest clone was completely sequenced in both directions.
Northern Blot Analysis-Oligonucleotides targeted to the unique amino terminus of Ank G 190 ( Fig. 1) were used in standard PCR reactions, with the full-length clone as template. The sense primer was GGAGACGCCGACCCGGGCGCAGAGCC (from within the 5Ј-untranslated region) and the antisense primer was TCGTGAATAACAGGATC-CAAAGTCACCCTG (corresponding to residues 19 -28). The 146-bp PCR product was sequenced to confirm its identity, labeled with 32 P and random primers, and hybridized to a rat multiple tissue Northern blot (CLONTECH) at high stringency. Following detection of hybridized messages by autoradiography, the blot was stripped and probed with a previously described PCR product corresponding to residues 1254 -1340 of Ank G 190; these residues represent a very highly conserved region within the spectrin binding domain of ankyrins (13). Finally, the blot was probed with actin as a control for RNA loading.
In Situ Hybridization-In situ hybridizations were performed as described previously (34). The 146-bp PCR product containing sequences unique to Ank G 190 was subcloned into the TA vector (Invitrogen) in both directions, to yield sense and antisense constructs. 35 S-Labeled antisense and sense cRNA probes were generated by linearizing the constructs with the appropriate restriction endonuclease, followed by in vitro transcription with T7 RNA polymerase (Promega, Madison, WI) in the presence of [ 35 S]␣-thio-UTP (NEN Life Science Products, 1100 Ci/mmol).
Transfection of Ank G 190 Constructs-PCR techniques were utilized to create a series of Ank G 190 expression constructs, each containing the eight-residue FLAG tag (IBI Biosystems) added to the carboxyl terminus. Primers were targeted to Ank G 190 residues as shown in Fig. 7A; the FLAG sequence and a stop codon were added to the 3Ј end of each antisense primer. PCR products were verified by sequencing, subcloned into the pcDNA3 eukaryotic expression vector (Invitrogen), and transfected into MDCK and COS-7 cells either transiently or stably with G418 (400 g/ml) selection (Life Technologies, Inc.). Subconfluent cells were exposed to the DNA-LipofectAMINE complex in Opti-MEM (Life Technologies, Inc.) for 6 h, and the transiently transfected cells were analyzed upon reaching confluence, after an additional 48 h incubation in normal growth medium.
Extractions, Immunofluorescence, and Immunoprecipitations-Both MDCK and COS-7 cells, transiently transfected with FLAG-tagged full-length Ank G 190 in pcDNA3, were grown to confluence in 150-cm 2 culture flasks in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and sequentially extracted in situ as described (13,19). Briefly, cells were first extracted for 10 min at 4°C in buffer 1 (yielding fraction 1) containing 10 mM Pipes, pH 6.8, 100 mM NaCl, 300 mM sucrose, 3 mM MgCl 2 , 0.5% Triton X-100, and protease inhibitors. The second extraction (yielding fraction II) was for 20 min at 4°C in buffer 2 containing 250 mM ammonium sulfate instead of NaCl and 1% Triton X-100. Equal amounts of total proteins in each extract were analyzed by SDS-PAGE and Western blotting with various antibodies, to determine their distribution in the soluble (fraction 1) or cytoskeletal (fraction 2) fractions.
MDCK cells stably transfected with the full-length Ank G 190-FLAG fusion construct were grown to confluence on coverslips and processed for indirect immunofluorescence microscopy as described (19,24). All incubations were at room temperature. Cells were washed with PBS, fixed, and permeabilized with 100% acetone for 20 min, blocked with normal goat serum for 60 min, and incubated first with the M2 monoclonal antibody against FLAG (IBI Biosystems) for 1 h and then with rhodamine-conjugated goat anti-mouse secondary antibody for 30 min (Amersham Pharmacia Biotech). The coverslips were mounted on slides with Crystalmount (Biomeda) and viewed with a microscope equipped for epilumination (IX70, Olympus). COS-7 cells transiently transfected with each of the five Ank G 190 constructs were grown to confluence in six-well plates and processed for immunoprecipitations as described (24). Cells were lysed for 20 min at 4°C in 2 ml of IP buffer (10 mM Tris-HCl, pH 7, 150 mM NaCl, 5 mM EDTA, 1 mM EGTA, 0.5% deoxycholate, 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 0.5 mM Pefablock), the lysate was centrifuged for 1 min at 10,000 ϫ g, and precleared by a 60-min incubation at 4°C with 25 l of nonimmune serum and 200 l of a 50% protein A-Sepharose solution (Amersham Pharmacia Biotech). The cleared supernatant was incubated for 4 h at 4°C with 10 g of the M2 monoclonal antibody against FLAG and for an additional 2 h at 4°C with 200 l of a 50% protein A-Sepharose solution. The lysate was centrifuged for 1 min at 10,000 ϫ g, washed three times with IP buffer, and the pellet subjected to SDS-PAGE and Western analysis with a monoclonal antibody against ␣-Na,K-ATPase (Upstate Biotechnology).
Other Methods-The isolation of total RNA from rat kidney used standard methods (25). SDS-PAGE and Western blotting were as described (26). Immunodetection of transferred proteins was by enhanced chemiluminescence (Amersham Pharmacia Biotech). Antibody to nonerythroid spectrin (␤II) was a kind gift from Dr. Jon Morrow (Yale University). Antibody to Ank G 119 (Golgi ankyrin) has been previously characterized (13,24).

RESULTS
Isolation and Characterization of a Novel Ankyrin (Ank G 190) from Rat Kidney-A comparison of cDNA sequences encoding human erythrocyte (7, 27) and brain (10)  region of high conservation extending from repeat 11-16. Oligonucleotides bracketing this region were used in a PCR reaction with reverse-transcribed rat kidney RNA as template, yielding a 525-bp PCR product whose amino acid sequence was only 55% homologous to red cell or brain ankyrin. Using the PCR product as a probe, three full-length clones were isolated from a rat kidney cDNA library, which were identical in sequence and differed only in the length of the 5Ј-untranslated region. The largest clone provided a 6134-bp contiguous cDNA sequence for Ank G 190 (Fig. 1, GenBank TM accession number AF069525), with a putative polyadenylation signal (AATAAA) upstream of the poly(A) ϩ tail, and a single open reading frame encoding for a protein of 1763 amino acids with a predicted mass of 190 kDa. We term this protein Ank G 190, based on its predicted size and homology to the 480-kDa neuronal Ank G isoform (10). Compared with Ank G 480, Ank G 190 possesses a unique 31-residue amino terminus, and is over 95% homologous within the repeats domain and spectrin binding domain; however, it is devoid of the large serine-rich region characteristic of the regulatory domain of Ank G 480 (Fig. 2). Sequence searches with GenBank TM reveal that rat Ank G 190 is only 80% homologous to mouse epithelial ankyrin (Ank3), with significant sequence diverge not only in the unique amino-terminal residues but also within the regulatory domain, where rat Ank G 190 is completely devoid of a 197-residue motif present in mouse Ank3 (12). Rat Ank G 190 is only 55-60% homologous to human red cell or human brain ankyrins. Nevertheless, Ank G 190 retains the domain structure of ankyrins, with a characteristic repeats domain of 24 repetitive 33-residue motifs, a highly conserved spectrin binding domain, and a truncated regulatory domain (compared with Ank G 480 and mouse Ank3, Fig. 2).
Ank G 190 Is Expressed Exclusively in Kidney and Lung-High stringency rat multiple tissue Northern blot analysis with cDNA probes derived from the unique amino terminus of Ank G 190 identified a single Ϸ8-kb message exclusively in kidney and lung (Fig. 3A). Hybridization of the same blot with a cDNA probe targeted to a highly conserved region within the spectrin binding domain yielded the same 8-kb message in kidney and lung, with additional smaller transcripts in the 4.5-6-kb range, as described previously (11)(12)(13). Somewhat larger isoforms were detected in skeletal muscle, heart, and brain (multiple transcripts) only with the probe from the spec-trin binding domain. Hybridization with actin revealed approximately equal loading of message in all lanes (not shown). Collectively, these data suggest that Ank G 190 is an alternatively spliced, tissue-specific isoform of the Ank3 family.
Ank G 190 Is Expressed by the Outer Medullary Collecting Duct Cells of the Rat Kidney-In situ hybridization of rat kidney tissue sections with an antisense cRNA probe generated from the unique amino-terminal sequences of Ank G 190 showed that it is distributed primarily in the outer medullary region, indicating that it is expressed by the outer medullary collecting duct cells (Fig. 4). The sense cRNA probes did not hybridize at all to the same region (not shown). These results are comparable to those recently obtained for mouse epithelial Ank3, the expression of which was found to be predominantly in the outer medullary collecting ducts (12).
Ank G 190 Assembles with the Cytoskeletal Fraction, at the Lateral Membrane of Transfected Cells-Since antibodies that specifically recognize Ank G 190 were not available, we used transfected MDCK (dog kidney collecting duct cell line) and COS-7 (monkey kidney) cells to examine the biochemical and cellular properties of an Ank G 190-FLAG fusion construct. In situ extraction (13,19) followed by Western blot analysis with FLAG antibody revealed that in both cell types, the transiently or stably transfected Ank G 190 assembled predominantly in fraction II, or the Triton-inextractable, cytoskeletal fraction (Fig. 5). In the same cells, a similar segregation into fraction II was also detected for native nonerythroid spectrin and ␣-Na,K-ATPase. In contrast, the Golgi-associated ankyrin (Ank G 119) in these cells was exclusively found in the soluble fraction I, as previously reported (13).
The association of Ank G 190 with the plasma membraneassociated cytoskeleton was confirmed by immunofluorescence microscopy. In confluent, polarized MDCK cells stably transfected with the Ank G 190-FLAG fusion construct, the antibody against FLAG produced a predominantly lateral membrane staining (Fig. 6). The nonerythroid spectrin and ␣-Na,K-ATPase were also predominantly localized to the lateral membrane in these cells, as expected (not shown) (13,16).
Ank G 190 Interacts with ␣-Na,K-ATPase via Its Distal 12 Repeats and Spectrin Binding Domains-Given the association of Ank G 190 with the plasma membrane by extractions and immunofluorescence, and the putative role of ankyrins as molecules linking integral membrane proteins (such as Na,K- ATPase) with the underlying cytoskeleton, it was of interest to demonstrate a functional interaction between Ank G 190 and Na,K-ATPase. Indeed, immunoprecipitations with FLAG antibody of COS-7 cells transiently transfected with the full-length Ank G 190-FLAG fusion construct (construct I, Fig. 7A), followed by Western analysis with antibody against ␣-Na,K-ATPase, revealed a functional complex between these proteins. Furthermore, a similar analysis using a series of Ank G 190 constructs showed that both the distal 12 repeats (construct III) and the spectrin binding domain (construct IV) were capable of interacting with ␣-Na,K-ATPase, whereas the proximal 12 repeats and the regulatory domain were devoid of Na,K-ATPase binding activity.

DISCUSSION
Ankyrins are critical adapter molecules that are thought to maintain the polarized distribution of integral membrane proteins by mediating their linkage with the underlying spectrinbased cytoskeleton (4 -6). Renal epithelial cells express a prominent 190-kDa ankyrin polypeptide (13,(15)(16)(17)(18); data with cross-reacting antibodies have suggested that it co-distributes with spectrin and Na,K-ATPase along the basolateral membrane (3, 4, 15-18, 20, 21). The present study identifies the primary structure of this novel ankyrin which is expressed exclusively in kidney and lung, termed Ank G 190 based on its predicted size and homology to the larger neuronal Ank G isoform (10). Ank G 190 displays a unique amino terminus, a highly conserved repeats domain and spectrin binding domain, and a truncated regulatory domain. Ank G 190 is expressed primarily by the outer medullary collecting duct cells in rat kidney. In cultured epithelial cells, Ank G 190 assembles predominantly with the Triton-inextractable cytoskeletal fraction at the lateral plasma membrane and forms a functional complex with ␣-Na,K-ATPase.
Interactions between ankyrin and several integral membrane proteins have been documented, and the structural basis of some such linkages have been identified. The initial cloning and characterization of red cell ankyrin, and the elucidation of its characteristic repeats domain of 24 tandemly arrayed 33residue motifs (7,27), lead to the postulate that different repeats, singly or in combination, constitute the binding sites for integral membrane proteins. Recent evidence suggests that this membrane binding domain consists of four independently folded subdomains of six repeats each (28). Thus, in in vitro assays, both the voltage-dependent sodium channel (29) and the anion exchanger (30 -32) bind red cell ankyrin predominantly via the distal 12 repeats, whereas neurofascin appears The PCR product used to screen the rat kidney cDNA library for Ank G 190 is also shown (angled stripes).

FIG. 3. Ank G 190 mRNA expression is restricted to kidney and lung.
A rat multiple tissue Northern blot (CLONTECH) was hybridized to probes derived from Ank G 190-specific sequences (A) and from a region within the spectrin binding domain that is highly conserved among ankyrins (B). The Ank G 190-specific probe yielded an 8-kb message exclusively in kidney and lung. The spectrin binding domain probe recognized several additional transcripts, presumably representing other alternatively spliced AnkG/Ank3 isoforms (9 -13). Hybridization with actin showed equal loading of all lanes (not shown).  (13,19), followed by Western blotting with antibodies as indicated. Note that Ank G 190 codistributes with spectrin and ␣-Na,K-ATPase predominantly in fraction II, whereas the Golgi-associated Ank G 119 is found exclusively in fraction I (13). Molecular masses in kDa are shown. Similar results were obtained in cells stably transfected with Ank G 190 (not shown).
to interact independently with both repeats 7-12 as well as repeats 13-24 (32). However, previous in vitro competition assays have shown that the spectrin binding domain of red cell ankyrin can also associate with liposomes containing Na,K-ATPase (31). Our studies demonstrate for the first time that Ank G 190, the isoform of ankyrin associated with the plasma membrane of polarized kidney epithelial cells, forms a functional immunoprecipitable complex with ␣-Na,K-ATPase in vivo. Furthermore, we show that two distinct ankyrin domains, namely the distal 12 ankyrin repeats as well as the spectrin binding domain, are capable of this interaction in vivo, whereas the proximal 12 repeats and the regulatory domains are devoid of Na,K-ATPase binding ability. These results are compatible with previous findings that both red cell and kidney ankyrins bind to two distinct cytoplasmic domains of ␣-Na,K-ATPase (19,33). Furthermore, our results provide a structural basis for the ability of even significantly truncated ankyrin isoforms, such as the Golgi-associated Ank G 119 (which contains only the distal 12 ankyrin repeats and a diminutive regulatory domain), to not only bind Na,K-ATPase but indeed to mediate its trafficking from endoplasmic reticulum to Golgi apparatus (13,24).
In summary, ankyrins interact with integral membrane proteins such as Na,K-ATPase in a complex, pleiotropic manner that requires recognition of multiple binding sites on both ligands. Such combinatorial complexity presumably allows not only for the acquisition of the high binding affinities that characterize ankyrin-Na,K-ATPase interactions (16,19,31) but also for the ability of ankyrins to accommodate an increasing number of integral membrane proteins and sequester them into specialized domains of polarized epithelial cells. With the kidney Ank G 190 clone in hand, it will be important in future studies to determine the minimal Na,K-ATPase-binding residues on ankyrin, the relative contribution of the two independent binding sites, and the consequences of interference with this binding on the generation and maintenance of polarized distribution of Na,K-ATPase in epithelial cells.  FIG. 7. Ank G 190 interacts with ␣-Na,K-ATPase via its distal 12 repeats and spectrin binding domains. A, overlapping constructs I-V were generated by PCR, with primers targeted to the residues shown. B, untransfected (UnTx) or transiently transfected COS-7 cells (with constructs I-V as shown) were lysed and subjected to Western blotting with FLAG antibody. Note that each construct was expressed as a peptide of the appropriate size. The lane marked LYS contains wild type COS-7 cell lysate, which does not cross-react with the FLAG antibody. Also, no signal was detected in the untransfected lane. C, Western blot with ␣-Na,K-ATPase antibody of transfected cells following immunoprecipitation with FLAG antibody. Note that the 97-kDa ␣-Na,K-ATPase was detected in the wild type lysate (LYS) and only in the immunoprecipitates of cells transfected with constructs I (fulllength Ank G 190), III (distal 12 repeats), and IV (spectrin binding domain).