J Am Soc Nephrol 10: 2297–2305, 1999 ATP Depletion Increases Tyrosine Phosphorylation of ␤-Catenin and in Renal Tubular Cells

JOHN H. SCHWARTZ, THEODORA SHIH, SARAH A. MENZA, and WILFRED LIEBERTHAL Evans Department of Clinical Research, Department of Medicine, Renal Division, Boston Medical Center, Boston, Massachusetts.

Abstract. This study examines the hypothesis that the loss of ameliorates the fall in transepithelial resistance induced by integrity of the junctional complex induced by ATP depletion ATP depletion. Also, using immunofluorescence microscopy it is related to alterations in tyrosine phosphorylation of the was demonstrated that ATP depletion results in a marked proteins ␤-catenin and plakoglobin. ATP diminution of E- staining in the basolateral membrane depletion of cultured mouse proximal tubular (MPT) cells of MPT cells. Vanadate mimics this effect of ATP depletion, induces a marked increase in tyrosine phosphorylation of both whereas genistein ameliorates the reduction in the intensity of ␤-catenin and plakoglobin. The tyrosine phosphatase inhibitor E-cadherin staining induced by ATP depletion. Because it is vanadate has the same effect in ATP-replete (control) mono- has been well established that hyperphosphorylation of the layers, whereas genistein, a tyrosine kinase inhibitor, reduces catenins leads to dissociation of the adherens junction and to phosphorylation of both proteins in ATP-replete monolayers dysfunction of the junctional complex, it is proposed that the and prevents the hyperphosphorylation of these proteins with increase in tyrosine phosphorylation of catenins observed in ATP depletion. This study also demonstrates that the fall in the MPT cells during ATP depletion contributes to the loss of transepithelial resistance of MPT monolayers induced by ATP function of the junctional complex associated with sublethal depletion can be reproduced by treatment of ATP-replete injury. monolayers with vanadate, whereas genistein substantially

Sublethal injury to renal tubular cells, induced by ATP deple- somes (11,12). In mature epithelia, the ZO and adherens junc- tion, leads to the rapid loss of functional integrity of the tight tion both completely circumscribe adjoining epithelial cells at junction (1,2), the loss of cell polarity (3,4), and severe im- the boundary between apical and basolateral membrane do- pairment of the epithelial permeability barrier (2,3,5). After mains and both play a role in maintaining the functional only a few minutes of ATP depletion, the transepithelial elec- integrity of the junctional complex (11,12). Although the ZO is trical resistance (TER) of renal epithelia falls to very low levels the component of the junctional complex that represents the and paracellular permeability increases (2,3,5). These func- barrier to the paracellular flux of molecules and ions across the tional changes are completely reversible if the ATP levels are (11,12), the adherens junction, which lies immedi- restored before lethal cell injury occurs (1,6). This loss of the ately basal to the ZO, is necessary not only for the formation of renal epithelial permeability barrier associated with sublethal the ZO but also for the maintenance of a functionally normal injury is believed to contribute, at least in part, to the “back- ZO (11,12). leak” of glomerular filtrate, which is believed to be an impor- The adherens junction of renal epithelial cells is comprised tant contributing factor to the profound loss of GFR associated of a number of proteins that include cadherin and the catenins. with acute ischemic renal injury (7–10). E-cadherin is a transmembrane protein that mediates adhesion The molecular events that lead to the rapid loss of function of adjacent cells to one another. The extracellular domain of of the junctional complex after ATP depletion remain uncer- the E-cadherin molecules of adjacent cells bind to one another tain. The junctional complex between epithelial cells is com- in a homophilic, calcium-dependent manner (11,13,14). The prised anatomically of at least three distinct structures: the intracellular domain of E-cadherin is connected to the actin zonula occludens (ZO), the adherens junction, and the desmo- by a complex of cytosolic proteins called the ␣ catenins (13,14). At least three different catenins ( -catenin, ␤-catenin and ␥-catenin [plakoglobin]) mediate the attachment Received January 22, 1999. Accepted May 11, 1999. of E-cadherin to actin by binding to an intermediary actin- Correspondence to Dr. John H. Schwartz, Boston University Medical Center, binding protein, ␣-actinin. There is now substantial evidence Evans Building Room 401, Boston Medical Center, One Boston Medical that tyrosine phosphorylation of the catenins plays an impor- Center Place, Boston, MA 02118. Phone: 617-638-7321; Fax: 617-638-8281; E-mail: [email protected] tant role in regulating the formation and functional integrity of 1046-6673/1011-2297 the adherens junction (14–16). The maintenance of an intact Journal of the American Society of Nephrology and stable adherens junction depends on maintaining the Copyright © 1999 by the American Society of Nephrology catenins in a dephosphorylated state. Increased tyrosine phos- 2298 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 2297–2305, 1999

phorylation of ␤-catenin and/or plakoglobin results in the onset of the experiment, 1 mg/ml lucifer yellow was added to the structural and functional disruption of the adherens junction apical solution. The entire volume of the basolateral compartment (1 (16,17). ml) was sampled at 30-min intervals and then replaced with fresh ␮ In these studies, we have examined the novel hypothesis that medium. Apical solution samples (1 l) were also obtained at the ATP depletion results in hyperphosphorylation of ␤-catenin same intervals as for the basolateral solution. The concentration of lucifer yellow in each compartment was determined by fluorescence and plakoglobin. We will also determine whether hyperphos- spectrofluorometry at an excitation wavelength of 425 nm and an phorylation of these cytoskeletal proteins contributes to the emission of 535 nm. The emission intensity of a standard curve was functional disturbance of the junctional complex associated linear over the range of samples tested and exhibited an r value of with ATP depletion. Ͼ0.95. Flux rates and membrane permeabilities were determined as described previously (2,6). Materials and Methods Cell Culture Preparation of MPT Lysates for Western Blotting The primary culture of mouse proximal tubule (MPT) cells was Confluent MPT cells were washed three times in cold phosphate- performed as described previously (18). In brief, collagenase digested buffered saline, scraped from the culture dish, and pelleted by cen- fragments were obtained from the renal cortices of mice (Harlan, trifugation at 1000 ϫ g for 10 min. The pellet was suspended in 4 vol Sprague Dawley, C57BL6) and placed in serum-free medium consist- of ice-cold homogenizing buffer containing 10 mM Tris-HCl, 150 ing of a 1:1 mixture of Dulbecco’s modified Eagle’s medium and mM NaCl, 50 mM NaF, 10 mM sodium pyrophosphate, 1 mM sodium Ham’s F-12 medium that contained 2 mM glutamine, 15 mM Hepes, vanadate, 5 mM ethylenediaminetetra-acetic acid, aprotinin (0.5 ␮g/ 5 ␮g/ml transferrin, 5 ␮g/ml insulin, 50 nM hydrocortisone, 500 ␮l), N-tosyl-L-phenylalanine chloromethyl ketone (2 ␮g/␮l), phenyl- ␮g/ml penicillin, and 50 ␮g/ml streptomycin. These cells have been methylsulfonyl fluoride (4 mM), DNAase (5 ␮g/ml), RNAase (5 identified previously as proximal tubular epithelial cells (18). ␮g/ml) and 1% Nonidet P-40. The suspended pellet was homogenized Cells were grown either on permeable filter supports (for epithelial by ten 1-s strokes in a Teflon homogenizer. The homogenate was permeability studies), in 6-well dishes (for Western blot analysis), or centrifuged for 10 min at 1000 ϫ g at 4°C to remove nuclei and on glass slides (for immunofluorescence studies). The filter supports remaining intact cells, and the supernatant was stored at Ϫ20°C for used for studies of epithelial permeability were 6.5-mm Transwell Western blotting and immunoprecipitation studies. filter inserts (Costar, Cambridge, MA) consisting of a polycarbonate membrane with 0.4-␮m pores coated with rat tail collagen. Immunoprecipitation ATP Depletion MPT lysates were immunoprecipitated using the anti-peptide ATP depletion was induced by chemical anoxia using sodium mouse monoclonal antibodies to ␤-or␥-catenin (Becton Dickinson, cyanide (CN) in the absence of glucose. MPT monolayers were rinsed Bedford, MA) according to the following protocol. The homogenate three times with Krebs-Hensleit buffer (KHB) that contained 1 mM was diluted to a protein concentration of 100 ␮g/ml with the homog- calcium and 1 mM magnesium, pH 7.40, at 37°C to remove residual enizing buffer that also contains 0.5% deoxycholate. Nonimmune substrates in the medium. Cells were then incubated for 1.5 h in serum (2 ␮l) and a 25% suspension of protein A-Sepharose 4B beads glucose-free KHB containing 5 mM CN. These conditions have been (30 ␮l) was added to a 900-␮l aliquot of the cell lysate. The mixture reported previously by our laboratory to lower ATP content to Ͻ5% was incubated at 4°C for 2 h and then centrifuged at 13,000 rpm in a of the control value (2,18). Control monolayers were incubated in microcentrifuge. The supernatant was incubated with 20 ␮l of anti-␤- CN-free KHB to which 10 mM glucose was added. or anti-␥-catenin antibody and 50 ␮l of a protein A-Sepharose 4B bead suspension that had been prereacted with goat anti-mouse IgG Transcellular Electrical Resistance (Sigma, St. Louis, MO) for 12 h at 4°C. In preliminary studies, we determined that the quantity of primary antibody used was more than TER, a sensitive marker of integrity, was measured adequate to precipitate all of the catenin present in the sample. The with an epithelial Voltohmeter (EVOM; World Precision Instruments, beads were pelleted by centrifugation and washed three times and New Haven, CT) in confluent cell monolayers grown on permeable suspended in sodium dodecyl sulfate sample buffer. filters. To control for intrinsic resistance of the filters, measurements of electrical resistance were obtained across cell-free inserts equili- brated in KHB at 37°C. The electrical resistance of the cell-free filter Immunoblot insert was then subtracted from all subsequent measurements. The Whole cell homogenates and immunoprecipitated samples, pre- electrodes were sterilized with 95% ethanol and were washed and pared as described above, were heated at 65 or 100°C for 5 min, equilibrated in sterile phosphate-buffered saline before use. To reduce respectively, before loading on a 7% polyacrylamide sodium dodecyl the variability of the determination, the electrodes were placed to the sulfate gel and run under reducing conditions as described previously same depth in the solutions bathing the cultured monolayer with a (6). Proteins were then electrophoretically transferred to nitrocellulose micromanipulator. Monolayers that did not have an initial resistance filters. After transfer, the filters were washed in 150 mM NaCl, 100 ⍀ ⅐ 2 of at least 175 cm were not used in this study. mM Tris-HCl, pH 7.5, and 0.05% Tween 20 (TBST), and blocked for 1 h in TBST containing 5% wt/vol nonfat powdered milk (TBSTM) Transcellular Permeability before incubation with the anti-phosphotyrosine antibody PY20 Another measure of the “gate” function of the tight junction is the (Transduction Laboratories, Lexington, KY) at a dilution of 1:1000 in barrier to the passage of molecules between epithelial cells. To ex- TBSTM at 4°C overnight. The filters were then washed three times amine permeability, we measured the unidirectional flux across con- with TBST and incubated in secondary horseradish peroxidase-la- fluent MPT monolayers grown on permeable supports of lucifer beled goat anti-mouse, 1:2000 in TBSTM for2hatroom temperature. yellow, a fluorescence, fluid-phase marker (MW ϭ 482) (3). At the After three additional washes with TBST, bound secondary antibody J Am Soc Nephrol 10: 2297–2305, 1999 Hyperphosphorylation of Catenins by ATP Depletion 2299 was detected using the enhanced chemiluminescence (ECL) system (Amersham).

Immunofluorescence Studies MPT cells grown on glass coverslips were fixed in 3.7% parafor- maldehyde and permeabilized with 0.2% Triton X-100. Detection of E-cadherin in these cells was performed by indirect immunofluores- cence using methods we have described previously (2). The anti-E- cadherin antibody (rat monoclonal antibody, clone DECMA-1; Sigma) was used at a dilution of 1:100. The secondary antibody was labeled with CY-3 (Molecular Probes, Eugene, OR) and used at a dilution of 1:500. The monolayers were photographed using a Nikon epifluorescence microscope, and the same exposure time, optimized for the control, was used for all coverslips to depict differences in staining intensity.

Phosphatase and Kinase Inhibitors Stock solution of vanadate (100 mM) and the sodium salt of okadaic acid (5 mM) were prepared in KHB. Genistein (1 mM), Figure 1. Effect of vanadate on transepithelial resistance of ATP- calyculin A (0.5 mM), H7 (1 mM), and H8 (10 mM) were dissolved replete (control) and ATP-depleted (cyanide [CN]-treated) mouse as stock solutions in DMSO. The final concentration of DMSO in the proximal tubular (MPT) monolayers grown on permeable supports. incubation media never exceeded 1.0%. All of these inhibitors were Ⅺ, control; छ, vanadate alone; ‚, CN and vanadate; E, CN. All obtained from Calbiochem (La Jolla, CA). values are means Ϯ SEM of six studies. *P Ͻ 0.05 comparing vanadate versus control; ϩP Ͻ 0.05 comparing vanadate versus Protein Assay vanadate ϩ CN. Protein concentrations were determined from a colorimetric dye binding assay (Bio-Rad) and expressed in milligrams per milliliter. obtained after 2 h was no different from that obtained with Statistical Analyses ATP depletion alone. Similar functional effects of vanadate Data are expressed as mean Ϯ SEM. Protocols that required com- were observed when the permeability of the epithelium was parisons between three or more groups were compared with ANOVA measured using a fluid phase marker (lucifer yellow) in parallel and Fisher post hoc test. When two groups were compared, analysis experiments. Epithelial permeability to lucifer yellow in- was performed with an unpaired, two-tailed t test. A P value of Ͻ0.05 creased 1 h after the addition of vanadate from a control value was defined as significant. of 3.9 Ϯ 0.3 ϫ 10Ϫ5 to 24.3 Ϯ 0.2 ϫ 10Ϫ5 cm/s (n ϭ 5; P Ͻ 0.05). This increase in permeability was less than the increase Results observed with CN-induced ATP depletion (54.7 Ϯ 0.4 ϫ 10Ϫ5 Transepithelial Resistance cm/s) (n ϭ 6; P Ͻ 0.05) or the combined effects of both We have documented in previous studies that ATP depletion vanadate and CN (65.3 Ϯ 0.5 ϫ 10 to 5 cm/s) (n ϭ 5; P Ͻ induced by incubating MPT cell monolayers in a glucose-free 0.05). medium containing CN results in a rapid fall in TER and a rise Because vanadate can potentially alter the activity of other in permeability to fluid phase markers (2,6). To evaluate the in addition to phosphotyrosine phosphatases, its func- possible role of changes in tyrosine phosphorylation in the tional effects cannot be definitively attributable to phosphoty- alteration in tight junction function, we first examined the rosine phosphatase inhibition alone. We therefore examined effects of phosphotyrosine protein kinase and phosphotyrosine the effect of other phosphatase inhibitors that have a different phosphatase inhibitors on transepithelial resistance. Treatment spectrum of action and specificity (Figure 2). Dephostatin (20 of control, dextrose-treated (ATP-replete) MPT monolayers ␮M), another phosphotyrosine phosphatase inhibitor, reduced with the phosphotyrosine phosphatase inhibitor vanadate at 1 TER in ATP-replete MPT monolayers by 61 Ϯ 9% (n ϭ 5; mM, a concentration shown in prior studies to increase tyrosine P Ͻ 0.05), an effect similar to that of vanadate. In contrast, two phosphorylation in intact cells (19), resulted in a decline in the different serine/threonine phosphatase inhibitors, okadaic acid transepithelial resistance from 281.2 Ϯ 25.2 to 170.8 Ϯ 18.2 (20 nM) and calyculin A (10 nM), had no effect on TER of ⍀ ⅐ cm2 (n ϭ 6; P Ͻ 0.05) in 20 min and a further decline to ATP-replete MPT monolayers (Figure 2). 143 Ϯ 12.2 ⍀ ⅐ cm2 in 60 min (Figure 1). Although the initial The phosphotyrosine protein kinase inhibitor genistein (19) rate of decline in TER induced by vanadate was similar to that at a concentration of 25 ␮M had no effect on TER in control, induced by ATP depletion, the reduction in TER after 60 min ATP-replete MPT monolayers (Figure 3). However, genistein was somewhat less in the vanadate-treated group than in the substantially ameliorated the reduction in TER induced by ATP-depleted monolayers (Figure 1). The combination of van- ATP depletion (Figures 3 and 4). As depicted in Figures 3 and adate and ATP depletion resulted in a more rapid decline in 4, the decline in TER after 60 min in genistein-treated ATP- TER than with either maneuver alone, but the minimum value depleted monolayers (32 Ϯ 7%) was significantly less than 2300 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 2297–2305, 1999

Figure 4. Comparison of the effect of protein tyrosine kinase inhibitor Figure 2. Comparison of the effect of protein tyrosine phosphatase genistein with protein kinase A and protein kinase C inhibitors on inhibitors with protein phosphatase 1 and 2 inhibitors on transepithe- ATP depletion-induced reduction of transepithelial resistance MPT lial resistance of ATP-replete monolayers grown on permeable sup- monolayers grown on permeable supports. Values are means Ϯ SEM ports. Values are means Ϯ SEM of six studies obtained 60 min after of six studies obtained 60 min after the addition of either diluent the addition of either diluent (Control) or 1 mM vanadate, 20 ␮M (Control) or 5 mM cyanide or a combination of cyanide plus 25 ␮M dephostatin, 20 nM okadaic acid, or 10 nM calyculin A. *P Ͻ 0.05 genistein or 25 ␮MH7or25␮M H8. *P Ͻ 0.05 comparing CN with compared with control. control; †P Ͻ 0.05 comparing vanadate versus vanadate plus genistein.

Figure 5. Reversal in the vanadate-induced reduction of transepithe- lial resistance by genistein. Ⅺ, control; छ, vanadate alone; E, Figure 3. Effect of genistein on transepithelial resistance of ATP- genistein alone; ‚, vanadate and genistein. All values are means Ϯ replete and ATP-depleted MPT monolayers grown on permeable SEM of six studies. *P Ͻ 0.05 comparing vanadate with control; supports. Ⅺ, control; E, genistein alone; ‚, CN and genistein; छ,CN †P Ͻ 0.05 CN versus CN plus genistein. alone. All values are means Ϯ SEM of six studies. *P Ͻ 0.05 comparing CN and genistein or CN alone with control; †P Ͻ 0.05 comparing CN and genistein versus CN alone. assertion, we investigated whether the reduction in TER in- duced by vanadate could be ameliorated by the tyrosine kinase inhibitor genistein. In Figure 5, we demonstrate that 1 mM after ATP depletion alone (69 Ϯ 11%, P Ͻ 0.05, n ϭ 6). This vanadate reduces TER after1hby149Ϯ 12 ⍀ ⅐ cm2. protective effect of genistein on the reduction in TER induced However, with the simultaneous addition of both 1 mM van- by ATP depletion could not be reproduced by other types of adate and 25 ␮M genistein, the reduction in TER was reduced kinase inhibitors (20). Neither the protein kinase C inhibitor by one-third, to only 49 Ϯ 5 ⍀ ⅐ cm2. Thus, the effect of H7 (25 ␮M) nor the protein kinase A inhibitor H8 (25 ␮M) vanadate must be dependent on enhanced tyrosine phosphory- affected the reduction in TER induced by ATP depletion (Fig- lation. ure 4). If the vanadate-induced reduction in TER is primarily de- Phosphorylation of Proteins in Whole Cell pendent on tyrosine hyperphosphorylation and not by other Homogenates of MPT Cells mechanisms, then one would predict that its action should be Based on the above observations, we would predict that ATP antagonized by a tyrosine kinase inhibitor. To verify this depletion induces changes in the function of via J Am Soc Nephrol 10: 2297–2305, 1999 Hyperphosphorylation of Catenins by ATP Depletion 2301 tyrosine phosphorylation of proteins. To evaluate this possibil- regulate the stability of the ZA (E-cadherin complexes), we ity, we first analyzed whole cell lysates of MPT cells by examined the effect of ATP depletion on catenin phosphory- immunoblot analysis to assess the degree of protein tyrosine lation. In these studies, ␤-catenin and plakoglobin were immu- phosphorylation and to determine the general effects of ATP noprecipitated from whole cell homogenates of MPT cells. The depletion, vanadate, and genistein on tyrosine phosphorylation immunoprecipitates were then subjected to immunoblot anal- of cell proteins (Figure 6). As expected, lysates of MPT cells ysis and probed first with an anti-tyrosine phosphate antibody incubated with 1 mM vanadate for 1 h demonstrated a gener- and then with an anti-catenin antibody. The amount of ␤-cate- alized increase in the immunodetectable degree of tyrosine nin and plakoglobin expressed was not measurably changed by phosphorylation (Figure 3, lane 2) compared to control cells ATP depletion or exposure to either vanadate or genistein (lane 1), whereas 25 ␮M genistein reduced phosphorylation of (Figure 7B). The treatment of ATP-replete (control) monolay- all protein bands. (Figure 3, lane 3). Most protein bands from ers with vanadate (Figure 7, lane 2) enhanced tyrosine phos- MPT cells treated with CN for 1 h (Figure 3, lane 4) demon- phorylation of ␤-catenin (upper band) by 398 Ϯ 42% and strated a decrease in tyrosine phosphorylation. However, sev- plakoglobin (lower band) by 283 Ϯ 23% above control (lane eral protein bands (one at 50 kD and one at 160 kD) were 1). Genistein treatment of ATP-replete monolayers reduced the hyperphosphorylated compared with control MPT cells. There tyrosine phosphorylation of ␤-catenin by 30 Ϯ 9% and plako- also was a marked increase in the tyrosine phosphorylation of globin by 35 Ϯ 6% below control (Figure 7, lane 3). ATP proteins compared to either control or only vanadate treatment depletion with CN increased phosphorylation of ␤-catenin by when MPT cells were treated with both CN and vanadate 348 Ϯ 37% and plakoglobin by 355 Ϯ 23% (Figure 7, lane 4). (Figure 3, lane 5). This latter change indicates that neither ATP The effect of ATP depletion on phosphorylation of both pro- depletion nor kinase activity is rate-limiting for protein ty- teins was comparable to that of vanadate alone. The combina- rosine phosphorylation. In addition, the combination of tion of genistein and CN (Figure 7, lane 5) reduced phosphor- genistein and ATP depletion (Figure 3, lane 6) reduced phos- ylation of ␤-catenin and plakoglobin by 50 Ϯ 11% and 65% Ϯ phorylation more than any one of these maneuvers alone. 10% below control, respectively. Thus, genistein prevented the These observations are consistent with the hypothesis that ATP hyperphosphorylation of both ␤-catenin and plakoglobin in- depletion is associated with alterations in the activity of both duced by ATP depletion (Figure 7, lane 4). The combination of tyrosine kinases and phosphatase with resultant dephosphory- vanadate and CN (Figure 7, lane 6) resulted in hyperphospho- lation of some proteins and hyperphosphorylation of others. rylation of both ␤-catenin (428 Ϯ 33%) and plakoglobin (393 Ϯ 53%), an effect somewhat greater than that induced by Phosphorylation of ␤-Catenin and Plakoglobin vanadate (Figure 7, lane 2) or ATP depletion alone (Figure 4, Since an intact zonula adherens (ZA) is required to maintain lane 4). a functional ZO and phosphorylation of catenins is known to In addition to determining the effect of ATP depletion on the phosphorylation of ZA proteins, we also examined the effect of ATP depletion on two ZO proteins: occludin and ZO-1. These proteins were immunoprecipitated from whole cell homoge- nates with either an antibody to ZO-1 or to occludin. The immunoprecipitates obtained (Figure 8) were then subjected to immunoblot analysis, and the blots were probed first with an antibody to either ZO-1 or to occludin and then with an anti-tyrosine phosphate antibody (PY20). Neither the amount of ZO-1 (Figure 8, lanes 1 to 2) and occludin (Figure 8, lanes 3, 4) expressed nor the degree of tyrosine phosphorylation of these proteins was affected by ATP depletion (Figure 8, lanes 2, 4, 6, and 8).

Immunofluorescence of E-Cadherin Tyrosine phosphorylation of ␤-catenin and/or plakoglobin has been associated with the withdrawal of E-cadherin com- plexes from the lateral membrane. Immunohistochemical stud- ies were performed to determine whether the maneuvers that induce tyrosine hyperphosphorylation also induce withdrawal Figure 6. Effect of tyrosine kinase inhibitors, tyrosine phosphatase of E-cadherin from the lateral membrane. Immunofluorescence inhibitors, and ATP depletion on tyrosine phosphorylation of proteins microscopy demonstrated that the distribution of E-cadherin in in MPT cells. An immunoblot of whole cell homogenates (50 ␮g protein/lane) from monolayers treated with vehicle (control) (lane 1), MPT cells is comparable to that described in other epithelial vanadate (lane 2), genistein (lane 3), CN (lane 4), CN and vanadate monolayers (11,16). E-cadherin is present primarily at cell-cell (lane 5), and CN and genistein (lane 6) and probed with an anti- borders and in the basolateral membrane (Figure 9a). One hour tyrosine phosphate antibody. The blot shown is representative of four after treatment of MPT monolayers with 5 mM CN, there is a experiments. marked decline in the intensity of basolateral E-cadherin stain- 2302 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 2297–2305, 1999

Figure 7. Phosphotyrosine phosphorylation of ␤-catenin (92 kD) and plakoglobin (82 kD). Whole cell homogenates from monolayers treated with vehicle (control) (lane 1), vanadate (lane 2), genistein (lane 3), CN (lane 4), CN and genistein (lane 5), and CN and vanadate (lane 6) were immunoprecipitated with an antibody directed against both ␤-catenin and plakoglobin and then analyzed by Western blot for tyrosine phosphorylation (A) or ␤-catenin and plakoglobin (B). The blot shown is representative of four experiments. ing (Figure 9b). Treatment of MPT monolayers with 1 mM to many of the functional abnormalities of the renal epithelium vanadate resulted in a generalized diminution in E-cadherin following sublethal injury (1,2). In this study, we have exam- staining in a pattern similar to that seen after CN (Figure 9c). ined the role played by alterations of tyrosine phosphorylation E-cadherin staining in MPT monolayers treated with the com- of adherens junction proteins in the loss of functional integrity bination of CN and genistein was similar to that in control of the junctional complex that is associated with chemical monolayers (Figure 9d), suggesting that genistein ameliorates anoxia (1,2,23). the effect of CN on the redistribution of E-cadherin. Genistein It is now well established that the barrier function of the alone had no effect on E-cadherin staining (data not shown). junctional complex is due to the unique anatomical character- istics of the ZO (11,12). Although the molecular mechanisms Discussion involved in the regulation and maintenance of the permeability This study is the first to demonstrate that when MPT mono- barrier are still not well understood, it has been demonstrated layers are subjected to ATP depletion, the tyrosine residues of that occludin, a protein that forms an integral part of the some proteins within the cell are dephosphorylated, while extracellular domain of the ZO (24,25), contributes directly to others become hyperphosphorylated. Other studies that have the permeability characteristics of the tight junction (22). The examined the effects of ATP depletion on serine-threonine function of other proteins associated with the ZO, including the phosphorylation of renal epithelial cell proteins have reported cytoplasmic proteins ZO-1 and ZO-2, is less clear but probably that chemical anoxia induces dephosphorylation of the cy- relates to assembly and localization of occludin to the ZO (25). toskeletal protein ezrin (21), as well as other renal tubular cell However, it is also clear that the formation and maintenance proteins (22). of a functionally intact ZO also depends indirectly on the Our study, which focuses entirely on tyrosine phosphoryla- presence of an intact adherens junction. The importance of the tion, demonstrates that ATP depletion causes complex and adherens junction to tight junction integrity and function has variable effects on the multiplicity of kinases and phosphatases been demonstrated in two ways. First, maintenance of normal that modulate protein tyrosine phosphorylation. Because ty- TER and epithelial permeability characteristics has been dem- rosine phosphorylation plays an integral role in control of onstrated to be dependent on the presence of extracellular cell-cell and cell matrix adhesion, we examined the hypothesis calcium (11) even though the ZO itself is not structurally that the dysregulation of tyrosine phosphorylation contributes dependent on calcium (11,26). Thus, the well-known deleteri- J Am Soc Nephrol 10: 2297–2305, 1999 Hyperphosphorylation of Catenins by ATP Depletion 2303

Figure 8. Effect of ATP depletion on expression and tyrosine phosphorylation of occludin (68 kD) and ZO-1 (225 kD). ZO-1 and occludin were immunoprecipitated from whole cell homogenate of control (lanes 1 and 3) and ATP-depleted (lanes 2 and 4) MPT monolayers. The immunoprecipitate was analyzed by Western blot for the amount of proteins immunoprecipitated (A) and for the degree of tyrosine phosphorylation of these proteins (B). ous effects of low extracellular calcium concentrations on the phorylation was important in the process by which ATP de- barrier function of the junctional complex is due to the calcium pletion induced changes in tight junction function. The dependency of E-cadherin, the protein that mediates normal pronounced reduction in TER and increase in permeability of cell- at the adherens junction (11,12,25). Second, MPT monolayers subjected to the tyrosine phosphatase inhib- the importance of an intact adherens junction in maintaining itors (vanadate and dephostatin) support the idea that the ty- the functional integrity of the tight junction is also suggested rosine phosphatases are involved in regulation of the integrity by studies in which neutralizing antibodies to cadherin inhibit of the junctional complex (17). The absence of any effect of normal ZO assembly and function (12). Thus, substantial evi- genistein on TER or permeability in ATP-replete monolayers is dence indicates that the normal function of the tight junction not surprising since inhibition of kinases involved in adherens depends indirectly on the presence of an intact adherens junc- junction integrity would be expected to have either no effect or, tion. if anything, might increase the “tightness” of adherens junction Although the mechanisms governing regulation of the adhe- mediated cell-cell adhesion. However, the observation that rens junction remain incompletely defined, recent studies have genistein, but not H7 or H8, ameliorates the fall in TER and demonstrated that the binding of catenins to cadherin, and the monolayer permeability associated with either addition of van- subsequent anchorage of E-cadherin to the cytoskeleton, is adate or ATP depletion is entirely novel. This finding is in regulated by tyrosine phosphorylation of the catenins (14–17). keeping with the role of tyrosine kinases and or phosphatases Tyrosine phosphorylation of ␤-catenin and/or plakoglobin re- in the regulation of tight junction function and with our hy- sults in the structural and functional disruption of the adherens pothesis that selective tyrosine hyperphosphorylation is in- junction (17). On the other hand, the dephosphorylation of duced by ATP depletion. these proteins is necessary for the establishment and mainte- The immunoblots shown in Figures 7 and 8 demonstrate that nance of a stable adherens junction (16). A family of phospha- interventions like vanadate, genistein, and ATP depletion mod- tases, the receptor tyrosine phosphatases, is believed to partic- ulate the degree of tyrosine phosphorylation of the ZA proteins ipate in the regulation of ␤-catenin phosphorylation and in ␤-catenin and plakoglobin, but have no effect on the tyrosine adherens junction assembly and disassembly (27–29). We phosphorylation of ZO proteins. ATP depletion as well as therefore examined the possibility that the alterations in tight vanadate increased phosphorylation of both ␤-catenin and pla- junction function associated with ATP depletion are the con- koglobin. Genistein reduced the degree of tyrosine phosphor- sequence of changes in the degree of tyrosine phosphorylation ylation of both proteins in control monolayers and ameliorated of the catenins. the ATP depletion-induced hyperphosphorylation of these pro- Initial studies using tyrosine kinase and phosphatase inhib- teins. In contrast, the ZO proteins occludin and ZO-1 had no itors were consistent with our hypothesis that tyrosine phos- change in tyrosine phosphorylation induced by ATP depletion. 2304 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 2297–2305, 1999

Figure 9. Immunohistochemical distribution of E-cadherin in MPT cells subjected to ATP depletion. Photomicrographs of MPT cells grown on glass coverslips stained for E-cadherin with a rat monoclonal antibody for mouse E-cadherin. (a) Control monolayer. ATP-replete MPT cells demonstrate the typical basolateral distribution of E-cadherin with predominant staining at cell-cell borders. (b) CN-treated monolayers. After 1 h exposure to 5 mM CN, there is a marked reduction in the intensity of the E-cadherin staining. (c) Vanadate-treated monolayer. MPT cells treated with 1 mM vanadate for 1 h demonstrate a reduction in the intensity of E-cadherin staining, an effect similar to that induced by CN. (d) Genistein and CN. The CN-induced reduction in E-cadherin staining (b) can be ameliorated by the simultaneous incubation with 1 mM genistein. To facilitate an assessment of the relative degree of E-cadherin staining, the exposure time for each photomicrograph was identical. Magnification: ϫ400.

These observations indicate that the target that is tyrosine- multiple proteins that regulate actin stress fibers (tensin, vin- phosphorylated during ATP depletion is not within the ZO but culin, cortactin, tallin, and annexin II) (31). Thus, it is also the ZA. Therefore, it is reasonable to suggest that the change in possible that tyrosine phosphorylation of one or more of these ZA functions is likely to be secondary, at least in part, to proteins in response to ATP depletion contributes not only to redisruption of E-cadherin complexes by phosphorylation of loss of epithelial permeability, but also to the changes in the catenin. actin cytoskeleton and loss of cell-substrate adhesion associ- There are numerous tyrosine kinases that can phosphorylate ated with sublethal injury (2). catenins. These include epidermal growth factor, human We also demonstrate that CN and vanadate result in loss of growth factor, and the Src family of tyrosine kinase (14,30,31). E-cadherin staining from the cell-cell junctions and basolateral Although there is some precedent for activation of Src kinase membrane of MPT monolayers (Figure 9, a through c). Fur- activity by hypoxia of endothelial cells (32) and myocytes (33), thermore, genistein ameliorates the effects of CN on E-cad- we have not identified the tyrosine kinases activated during herin staining. These observations suggest that alterations in ATP depletion in renal tubular cells. However, the potential tyrosine phosphorylation of the catenins ultimately disrupt the protein targets for this kinase are also likely to be numerous localization of E-cadherin to the adherens junction. Since and include other proteins in addition to the catenins evaluated changes in phosphorylation of catenins are known to regulate in this study. For example, the substrates for Src include the assembly and disassembly of the adherens junction J Am Soc Nephrol 10: 2297–2305, 1999 Hyperphosphorylation of Catenins by ATP Depletion 2305

(15,16,27,28), and since integrity of the adherens junction is 16. Kypta RM, Su H, Reichardt LF: Association between a trans- necessary for normal tight junction function, it is reasonable to tyrosine phosphatase and the cadherin-catenin propose that ATP depletion alters the function of the junctional complex. J Cell Biol 134: 1519–1529, 1996 complex, at least in part, via its effect on the phosphorylation 17. Staddon JM, Herrenknecht K, Rubin LL: Evidence that tyrosine of ␤-catenin and plakoglobin. phosphorylation may increase tight junction permeability. J Cell Sci 108: 609–619, 1995 Acknowledgment 18. Sheridan AM, Schwartz JH, Kroshian VM, Tercyak AM, LaRaia L, Masino S, Lieberthal W: Mouse proximal cells are more This work was supported by National Institutes of Health Grant susceptible to injury than MDCK cells: Role of differences in DK385101. lipid metabolism. Am J Physiol 265: F342–F350, 1993 References 19. Feick P, Gilhaus S, Schultz I: Pervanadate stimulates amylase release and protein tyrosine phosphorylation of paxillin and 1. Canfield P, Geerdes A, Molitoris B: Effect of reversible ATP p125FAK in differentiated AR4–2J pancreatic acinar cells. J Biol depletion on tight junction integrity. Am J Physiol 261: F1038– Chem 273: 16366–16373, 1998 F1045, 1991 2. Kroshian VM, Sheridan A, Lieberthal W: Functional and cy- 20. Hagiwara M, Inagaki M, Watanabe M, Ito M, Onoda K, Tanaka toskeletal changes induced by sublethal injury in proximal tubu- T, Hidaka H: Selective modulation of calcium-dependent myosin lar epithelial cells. Am J Physiol 266: F21–F30, 1994 phosphorylation by novel protein kinase inhibitors, isoquinoline- 3. Molitoris BA, Dahl RH, Falk SA: Ischemia-induced loss of sulfonamide derivatives. Mol Pharm 32: 7–12, 1987 epithelial polarity: Role of the tight junction. J Clin Invest 84: 21. Chen JJ, Cohn A, Mandel LJ: Dephosphorylation of ezrin as an 1334–1339, 1989 early event in renal microvillar breakdown and anoxic injury. 4. Molitoris B, Geerdes A, McIntosh J: Dissociation and redistri- Proc Natl Acad Sci USA 92: 7495–7499, 1995 bution of Naϩ,Kϩ-ATPase from its surface membrane actin 22. Kobryn C, Mandel L: Decreased protein phosphorylation in- cytoskeletal complex during cellular ATP depletion. J Clin Invest duced by anoxia in proximal renal tubules. Am J Physiol 267: 88: 462–469, 1991 C1073–C1079, 1994 5. Molitoris BA: Putting the actin cytoskeleton into perspective: 23. Mandel L, Bacallao R, Zampighi G: Uncoupling of the molecular Pathophysiology of ischemic alterations. Am J Physiol 272: “fence” and paracellular “gate” functions in epithelial tight junc- F430–F433, 1997 tions. Nature 361: 552–555, 1993 6. Borkan SC, Wang Y-H, Lieberthal W, Burke PR, Schwartz JH: 24. Furuse M, Hirase T, Itoh M, Nagafuchi A, Yonemura S, Tsukita Heat stress ameliorates ATP depletion-induced sublethal injury in S, Tsukita S: Occludin: A novel integral membrane protein mouse proximal tubule cells. Am J Physiol 272: F347–F355, 1997 localizing at tight junctions. J Cell Biol 123: 1777–1788, 1993 7. Donohoe JF, Venkatachalam MA, Bernard DB, Levinsky NG: 25. Gumbiner BM: Cell adhesion: The molecular basis of tissue Tubular leakage and obstruction in acute ischemic renal failure. architecture and morphogenesis. Cell 84: 345–357, 1996 Kidney Int 13: 208–222, 1978 26. Stevenson B, Siciliano J, Mooseker M, Goodenough D: Zonulae 8. Venkatachalam MA, Bernard DB, Donohoe J, Levinsky NG: Isch- occludentes in junctional complex-enriched fractions from emic damage and repair in the rat proximal tubule: Differences mouse liver: Preliminary morphological and biochemical char- among S1, S2 and S3 segments. Kidney Int 14: 31–49, 1978 acterization. J Cell Biol 98: 1209–1221, 1984 9. Lieberthal W: Biology of ischemic and toxic renal tubular cell 27. Brady-Kalnay S, Tonks N: Protein tyrosine phosphatases as injury: Role of nitric oxide and the inflammatory response. Curr adhesion receptors. Curr Opin Cell Biol 7: 650–657, 1995 Opin Nephrol Hypertens 7: 289–295, 1998 28. Tonks N, Neel B: From form to function: Signalling by protein 10. Brady HR, Brenner BM, Lieberthal W: Acute renal failure. In: tyrosine phosphatases. Cell 87: 365–368, 1996 The Kidney, edited by Brenner B, Philadelphia, Saunders, 1996, 29. Denu J, Stuckey J, Saper M, Dixon J: Form and function in pp 1200–1252 protein phosphorylation. Cell 87: 361–364, 1996 11. Gumbiner B: Structure, biochemistry and assembly of epithelial 30. Balkovetz DF, Pollack AL, Mostov KE: Hepatocyte growth tight junctions. Am J Physiol 253: C749–C758, 1987 12. Citi S: The molecular organization of tight junctions. J Cell Biol factor alters the polarity of MDCK cell monolayers. J Biol Chem 121: 485–489, 1993 272: 3471–3477, 1997 13. Geiger B, Ayalon O: . Annu Rev Cell Biol 8: 307–332, 31. Brown MT, Cooper JC: Regulation, substrates and function of 1992 Src. Biochim Biophys Acta 1287: 121–149, 1996 14. Aberle H, Schwartz H, Kemler R: Cadherin-catenin complex: 32. Tsokias L, Zhou XM, Foster D, Brugge JS, Sukhatme VP: Protein interactions and their implications for cadherin function. Hypoxic induction of human vascular endothelial growth factor J Cell Biochem 61: 514–523, 1996 expression through c-Src activation. Nature 375: 577–581, 1995 15. Balsamo J, Leung T, Ernst H, Zanin M, Hoffman S, Lilien J: 33. Seko Y, Tobe K, Takahashi N, Kaburagi Y, Kadowaki T, Yazaki Regulated binding of a PTP1B-like phosphatase to N-cadherin: Y: Hypoxia and hypoxia/regeneration activate Src family ty- Control of cadherin mediated adhesion by dephosphorylation of rosine kinases and p21ras in cultured rat cardiac myocytes. ␤-catenin. J Cell Biol 134: 801–812, 1996 Biochem Biophys Res Commun 226: 520–535, 1996