- Email: [email protected]
H+ exchanger isoform NHE-3 is species specific
Glycosylation of the Na +/H + exchanger isoform NHE-3 is species specific GWEN L. BIZAL, RANDY L. HOWARD, CRESCENCE BOOKSTEIN, MRINALINI C. RAO, EUGENE B. CHANG, and MANOOCHER SOLEIMANI INDIANAPOLIS, INDIANA, and CHICAGO, ILLINOIS
The glycosylation of Na+/H + exchanger isoform NHE-3 was studied in brush border membrane (BBM) vesicles isolated from rabbit, dog, and rat kidney cortex. Western blot analyses were performed against BBM proteins by using polyclonal antibodies to an NHE-3 fusion protein. In rabbit kidney, NHE-3 antibody recognized a band with - 9 5 kd molecular mass. Treatment of rabbit cortical BBM with glycopeptidase F, at 16 U/ml, for 4 or 16 hours increased the apparent mobility of NHE-3 to 84 and 82 kd, respectively. Incubation of rabbit BBM proteins for 16 hours with endoglycosidase H, at 0.1 U/ml, did not alter the apparent mobility of NHE-3. Deglycosylation of NHE-3 with glycopeptidase F did not affect acid-stimulated, amiloride-sensitive sodium 22 influx in BBM vesicles as compared with that in controls (p > 0.05). Immunoblot analysis against BBM proteins from canine kidney cortex demonstrated the presence of an - 8 3 to 92 kd protein. Treatment of canine BBM with glycopeptidase F or endoglycosidase H or F for 16 hours did not alter the apparent mobility of NHE-3, suggesting that canine renal NHE-3 is not glycosylated. Treatment of canine kidney BBM with glycopepfidase F did not affect acid-stimulated 22Na+influx as compared with that in controls (p > 0.05). Immunoblot analysis against BBM proteins from rat kidney cortex demonstrated the presence of a sharp bond at 90 kd. Treatment of rat BBM with glycopeptidase F or endoglycosidase H or F for 16 hours did not alter the apparent mobility of NHE-3, suggesting that rot renal NHE-3 is not glycosylated. The above experiments suggest that NHE-3 glycosylation in mammalians is species specific and that glycosylation does not affect the exchanger activity. (J Lab Clin Med 1996;128:304-12)
Abbreviations: BBM = brush border membrane; HEPES/MES= N-2-hydroxyethylpiperazine-N2-ethanesulfonic acid/2-N-morpholino ethane sulfonic acid; IgG = immunoglobulin G; PBST= 0.1% Tween 20/phosphate-buffered saline solution; SDS-PAGE = sodium dodecyl sulfatepolyacrylamide gel electrophoresis
From the Department of Medicine, Indiana University School of Medicine and the Veterans Affairs Medical Center, Indianapolis; the Department of Medicine, University of Chicago; and the Department of Physiology and Biophysics, University of Illinois at Chicago. Supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK 46789 and a Merit Review Grant from the Department of Veterans Affairs (to M.S) and by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK 42086 (to E.B.C). Dr. Howard was the recipient of a Physician Scientist Award from the National Institutes of Health (DK-02116). Portions of this study have been published in abstract form (J Am Soc Nephrol 1994;5:262A). Submitted for publication Jan. 12, 1996; revision submitted March 27, 1996; accepted April 11, 1996. 304
T
h e majority o f H C O 3 - filtered at the glomerulus is r e a b s o r b e d in the kidney proximal tubule via the luminal N a + / H + exchanger. 1-3 R e c e n t molecular cloning experiments have identified the existence o f four structurally related N a + / H + exchanger isoforms, which have b e e n designated N H E - 1 , N H E - 2 , N H E - 3 , and N H E - 4 . 4-8 Functional, molecular, and immunologic studies have d e m o n s t r a t e d that N H E - 3 is Reprint requests: Manoocher Soleimani, MD, Division of Nephrology and Hypertension, Department of Internal Medicine, University of Cincinnati, 231 Bethesda Ave., MSB 5502, Cincinnati, OH 45267. 0022-2143/96 $5.00 + 0 5/1/74173
J Lab Clin Med Volume 128, Number 3
the Na+/H + exchanger isoform that is present in luminal membranes of proximal tubules. 9-13 A basolateral Na+/H + exchanger has also been identified in kidney proximal tubule cells and is presumed to be mostly involved with cell p H regulation. 2'3'14 Immunologic studies have demonstrated that NHE-1 is the basolateral Na+/H + exchanger isoform in kidney proximal tubules, as The amino acid sequences of Na+/I-I+ exchanger isoforms, deduced from their respective cDNAs, indicate that all N H E isoforms have consensus sequences for N-linked glycosylation in their extracytoplasmic loops. 4-s Several studies have examined N-linked glycosylation of NHE-1. The results indicate that NHE-1 is a glycoprotein, and they suggest the presence of at least two N-linked carbohydrate moieties. 16'17 In addition, recent studies illustrate that NHE-1 possesses O-linked glycosylation, as The presence of carbohydrate moieties on N H E - 3 is less known. The N H E - 3 c D N A in rabbit or rat encodes a protein with a molecular mass of - 9 2 kd. A recent investigation in which an antibody against an N H E - 3 - M B P fusion protein was used found that the exchanger had an apparent molecular mass of - 8 0 kd. 9 When a specific antiserum to an NHE-3G S T fusion protein was used, N H E - 3 was found to have a molecular mass of - 9 5 kd in rabbit kidney cortex. H-a3 Expression of the rat N H E - 3 c D N A in PS120 fibroblast cells deficient in endogenous exchanger shows the presence of an 82 kd protein, a8 NHE-3, however, was found to have a molecular mass of - 9 7 kd in experiments performed in the rat intestine, a° The reason for the difference in the mobility of the N H E - 3 exchanger in these several studies is not clear. One possible explanation is that the gel systems used in these studies are different. Another possibility is the presence of species difference with regard to processing of NHE-3. One plausible explanation, however, is that N H E - 3 might be present in more than one isoform. For example, differential glycosylation of N H E - 3 could result in the expression of several isoforms. As such, various N H E - 3 antibodies could detect, based on their affinities, different N H E - 3 isoforms. Indeed, preliminary reports show that the 80 kd N H E - 3 is not a glycoprotein. 19 The purpose of these experiments was to study glycosylation of N H E - 3 in m a m m a l i a n kidneys. Toward this end, brush border m e m b r a n e vesicle from rabbit, dog, and rat kidney cortex were isolated and studied for the presence of carbohydrate moieties on N H E - 3 with various N-glycanases. The results indicate that rabbit N H E - 3 is N-link glycosylated, whereas dog and rat N H E - 3 are not N-link glycosylated. The results further indicate that
Bizal et al.
305
N H E - 3 activity is not affected by the glycosylation state of the exchanger. METHODS Membrane vesicle preparation. Male New Zealand
white rabbits were killed by intravenous sodium pentobarbital administration. Male Sprague-Dawley rats were killed by intraperitoneal injection of sodium pentobarbital. Kidneys were removed and processed for cortical BBM vesicle preparation. Dog kidneys were obtained as described. 2° BBM vesicles were prepared from renal cortexes by a Ca ++ aggregation method zl as used previously.2z'23 The purification of BBM vesicles relative to the initial homogenate, assessed with alkaline phosphatase, was sevenfold to 10-fold for all groups. SDS-PAGE and immunoblot analysis. Membrane fractions (100 to 200 ixg/lane) were solubilized and subjected to vertical slab SDS-PAGE according to the Laemmli protocol. 24 Proteins were electrophoretically transferred to nitrocellulose at 200 mA for 15 hours. The nitrocellulose filters were blocked in PBST and then incubated with 10 Ixl of an NHE-3 immune serum diluted at 1:400 in PBST for 2 hours. The specificity of this antibody was recently demonstrated. 11 The excess antibody was removed with PBST, and the antigen-antibody complex was treated with alkaline phosphatase-conjugated goat antirabbit IgG diluted at 1:1000 in PBST. The nitrocellulose filters were developed with nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate dissolved in 70% or 100% N,N,-dimethylformamide. Carbohydrate analysis. BBM vesicles (10 mg) were washed in 5 ml of a glycolytic buffer consisting of 50 mmol/L KC1, 20 mmol/L NaH2PO4, pH 7.2. The membrane proteins were then resuspended to i mg protein per milliliter in the same buffer. Ethylenediaminetetraacetic acid (50 mmol/L) and a cocktail of proteinase inhibitors including pepstatin A, leupeptin, and phenylmethylsulfonyl fluoride were added to the membrane proteins as described, a5'26 Membrane proteins (100 txl) were incubated in the presence or absence of glycopeptidase F (at 8 and 16 U/ml), endoglycosidase H (0.1 U/ml), or endoglycosidase F (0.1 U/ml) for up to 16 hours at 37° C. The purpose of choosing the current enzymes (endoglycosidase F, endoglycosidase H, or glycopeptidase F) was to study the N-linked glycosylation of NHE-3, because these enzymes remove almost all N-linked carbohydrate moieties (high mannose, complex oligosaccharides, or both groups) of proteins. Endoglycosidase F removes carbohydrate moieties with the complex-type N-oligosaccharide biantennary group, whereas endoglycosidase H affects side chains with hybrid and high mannose-type groups and glycopeptidase F affects both types of carbohydrate side chains. ~2Na+ influx measurement. Intact BBM vesicles were treated with glycopeptidase F at 16 U/ml for 4 hours. Thereafter, vesicles were washed, preincubated in a HEPES/MES buffer, and assayed for Na+/H + exchanger activity with :2Na + influx and a rapid filtration method as
306
J Lab Clin Med September 1996
Bizal et al,
NHE-3 Immunoblot (rabbit kidney) KD
bit IgG and 5-bromo-4-chloro-3-indolyl phosphate were purchased from Boehringer Mannheim, Indianapolis, Ind. Nitrocellulose filters, nitro blue tetrazolium, and glycopeptidase F were purchased from Sigma Chemical Co., St. Louis, Mo. CNBr-activated Sepharose 4B was purchased from Pharmacia, Piscataway, N.J. SDS, acrylamide, and NN'-methylene-bis-acrylamide were purchased from BioRad Laboratories, Hercules, Calif. RESULTS
94-
71X
n m
o
~3
E O
Fig. 1. Immunoblotsof BBM from rabbit renal cortex (leftlane) and outer medulla (rightlane) when using NHE-3 antibody.BBM proteins (200 ixg/lane)were resolved on SDS-PAGE, transferred to nitrocellulose membranes, and probed with NHE-3 immune serum as described in Methods.
previously described.27'28 In repeated experiments we found that incubation of vesicles at 37° C for 16 hours (in both the control group and the deglycosylated group) resulted in loss of vesicle integrity, making it impossible to study exchanger activity. The ice-cold medium used to dilute and wash the vesicles consisted of 170 mmol/L K gluconate, 10 mmol/L HEPES/tetramethyl ammonium hydroxide, pH 7.5. Each filter (0.45 Ixm; DAWP; Millipore Corp., Bedford, Mass.) was placed in 5 ml of scintillation fluid (Ready-Solv HP; Beckman Instruments Inc., Fullerton, Calif.), and radioactivity was determined by scintillation spectroscopy. The final composition of the experimental medium and other details of the protocols are given in the figure legends. All experiments were performed with vesicles treated with valinomycin (0.5 mg/ml) and pre-equilibrated in medium of appropriate composition to ensure that [K+]o = [K+]i during uptake measurements. Na uptake studies were performed in rabbit, rat, and dog kidney BBM vesicles. Materials. 22NaC1 (specific activity 906 mCi/mg) was obtained from DuPont NEN Research Products, Wilmington, Del.). Alkaline phosphatase-conjugated anti-rab-
We first studied the expression of NHE-3 in BBMs isolated from the rabbit kidney cortex and outer medulla. Fig. 1 demonstrates that the immune serum containing NHE-3 antibody detected a strong and diffuse band with an apparent molecular weight of - 9 5 kd in cortical (left lane) and outer medullary (right lane) luminal membranes. These results confirm recent reports on NHE-3 expression in the rabbit kidney, n'13 In the next series of experiments we determined whether the rabbit kidney NHE-3 is a glycoprotein. Rabbit cortical BBM proteins (100 ~g/100 txl) were incubated with glycopeptidase F at a low concentration (4 U/ml) for 16 hours at 37 ° C, resolved on SDS-PAGE, and electrotransferred, and the blot was probed with NHE-3 antiserum. As demonstrated in Fig. 2, the digestion of BBM with glycopeptidase F at a low concentration (lane 2) changed the apparent molecular mass of the exchanger (lane 1, control) from - 9 5 to - 9 1 kd. These results suggest that NHE-3 is a glycoprotein with a - 4 kd shift in molecular mass after treatment with 4 U/ml glycopeptidase F. To characterize the glycosylation of NHE-3 further, rabbit BBMs (100 ~g/100 rd) were completely digested with high concentrations of glycopeptidase F (16 U/ml) for 16 hours at 37 ° C. Thereafter, vesicles were subjected to SDS-PAGE, transferred to nitrocellulose membrane, and blotted against NHE-3 antiserum. As demonstrated in Fig. 3, digestion of BBMs with glycopeptidase F at high concentrations (lane 2) resulted in the appearance of three bands at 92, 84, and 82 kd. The 82 kd band was very strong when compared with the 84 and 92 kd proteins. Comparison of Fig. 2 and Fig. 3 indicates that NHE-3 is most likely present in more than one form. To study the issue of NHE-3 glycosylated isoforms further, BBM proteins were treated with endoglycosidase H (0.1 U/ml) for 16 hours or glycopeptidase F (16 U/ml) for 4 and 16 hours at 37 ° C. Membrane proteins were then resolved on SDS-PAGE and electrotransferred, and the blot was probed with NHE-3 antiserum. Fig. 4, lane 1, is an immunoblot
J Lab Clin Med Volume 128, Number 3
Bizal et al.
kD
307
kD
94
1
2
94-
Fig. 2. Immunoblot analysis of rabbit cortical BBM with NHE-3 immune serum after partial digestion with glycopeptidase F. BBM proteins (100 i~g/lane) were incubated for 16 hours at 37° C in the presence of glycopeptidase F at 4 U/ml (lane 2) or glycolytic buffer (lane 1). Proteins were resolved on SDS-PAGE.
71of NHE-3 treated with glycolytic buffer alone (no glycosidase). Lane 2 and lane 3 reflect digestion with glycopeptidase F for 4 and 16 hours, respectively. Fig. 4, lane 4, reflects 16 hours of digestion with endoglycosidase H at 0.1 U/ml. As shown, 4 hours of digestion with glycopeptidase F (lane 2) resulted in the appearance of two thin bands at 82 and 84 kd and another band at 92 kd. Complete digestion (16 hours) of NHE-3 with glycopeptidase F (lane 3) resulted in a stronger appearance of the 82 kd band and partial disappearance of the 92 kd band. Fig. 4, lane 4, indicates that digestion with endoglycosidase H had no effect on NHE-3 mobility. In additional experiments not shown we found that the effect of endoglycosidase F was very similar to that with glycopeptidase F. To examine the effect of NHE-3 deglycosylation on exchanger activity, BBM vesicles were treated with either glycolytic buffer (no glycosidase) or glycopeptidase F at 16 U/ml for 4 hours (as indicated in Methods, incubation of vesicles at 37° C for 16 hours resulted in the loss of vesicle integrity, making it impossible to study exchanger activity at that time point). Thereafter, vesicles were washed and preincubated in a HEPES/MES buffer (see the legend to Fig. 5). The time course of ZZNa+ influx in BBM vesicles was assayed in the presence of an inward pH gradient (pHo/pHi = 7.5/6.0) and --_ 1 mmol/L amiloride added to the external solution, Fig. 5 shows that partial deglycosylation of NHE-3 had no significant effect on the exchanger activity as compared with results in control vesicles. The 2-hour
1
2
Fig. 3. Immunoblot analysis of rabbit cortical BBM with NHE-3 immune serum after complete digestion with glycopeptidase F. BBM proteins (100 i~g/lane) were incubated for 16 hours at 37° C in the presence of glycopeptidase F at 16 U/ml (lane 2) or buffer (lane 1). Proteins were resolved on SDS-PAGE.
Z2Na+ uptakes (equilibrium values) were also comparable in both groups, indicating that enzymatic treatment did not affect vesicle integrity (Fig. 5). To determine whether deglycosylation affects the Michaelis-Menten constant of the exchanger, the initial rate of amiloride-sensitive 22Na influx in BBM vesicles was assayed in the presence of 3, 6, 9, and 12 mmol/L sodium. Kinetic analysis of the results by Hanes-Woolf plot did not show any significant change in the Michaelis-Menten constant of the exchanger in the deglycosylated state as compared with control values (KNa 13.4 mmol/L in glycopeptidase F-treated vesicles and 12.1 mmol/L in buffertreated vesicles, p < 0.05, n = 3). A limited doseresponse inhibition of the exchanger activity by dimethylamiloride at 1 and 10 t~mol/L in rabbit BBM vesicles showed no difference in the sensitivity of the exchanger to inhibition between glycosylated and nonglycosylated states (12% vs 15% inhibition at 1 ~zmol/L dimethylamiloride and 36% vs 41% inhibition at 10 p~mol/L dimethylamiloride for glycosylated and nonglycosylated NHE-3, respectively,
308
J Lab Olin Med September 1996
Bizal et al.
kD
94-
71-
1
2
3
4
Fig. 4. Time course of NHE-3 digestion with glycopeptidase F. BBM proteins (100 ixg/lane) were incubated for 16 hours with buffer (lane 1), 4 hours with glycopeptidase F (16 U/ml) (lane 2), 16 hours with glycopeptidase F (16 U/ml) (lane 3), or 16 hours with endoglycosidase H (0.1 U/ml) at 37° C.
p > 0.05 for both groups). Moreover, pH i dependence of the exchanger remained the same as shown by 60-second uptake studies performed at pH i 6.5 and 7.0 (410 _+ 36 picomoles sodium per milligram of protein per minute in controls vs 370 _+ 32 picomoles sodium per milligram of protein per minute in the deglycosylated state at pH i 6.5, and 160 _+ 16 picomoles sodium per milligram of protein per minute in controls vs 140 _+ 16 picomoles sodium per milligram of protein per minute in the deglycosylated state at pHi 7.0, p > 0.05, n = 3 for each group). In the next series of experiments, expression and possible glycosylation of NHE-3 in the canine kidney was examined. Fig. 6, A, demonstrates that the immune serum containing NHE-3 antibody detected a strong and diffuse band with an apparent molecular weight of 83 to 92 kd in canine kidney BBM. These results confirm recent reports on NHE-3 expression in the canine kidney,i~ In addition to the diffuse band, several smaller proteins in the 65 to 75 kd area were also occasionally detected. The immunodetection of these proteins was not consistent, suggesting the presence of proteolytic breakdown products of the 83 to 92 kd protein. We next examined N-linked glycosylation of NHE-3 in canine kidney. Canine cortical BBM proteins (100 ixg/100 txl) were incubated with glycopeptidase F, at 16 U/ml, for 16 hours at 37° C. Fig. 6, B (lefipanel), is an immunoblot analysis of NHE-3 and shows that treatment of canine BBM with glycopeptidase F did not alter the mobility of NHE-3. Treatment of BBM
vesicles with endoglycosidase H also did not affect the apparent mobility of NHE-3 (Fig. 6, B, right panel). Similarly, treatment of canine BBM vesicles with endoglycosidase F did not alter NHE-3 mobility (data not shown). Pretreatment of BBM vesicles with glycopeptidase F at 16 U/ml for 4 hours did not affect the exchanger activity as assessed by acidstimulated (pHo/pH i = 7.5/6.0), amiloride-sensitive Z2Na+ influx (104.7% ___ 6.3% in glycopeptidase F-treated vesicles vs 100% in controls, n = 3, p > 0.05). The above experiments demonstrate that NHE-3 is differentially glycosylated in rabbit and dog kidney. In the last series of experiments, possible Nlinked glycosylation of NHE-3 in the rat kidney was examined. Rat cortical BBM proteins (100 txg/100 ~1) were incubated with glycopeptidase F at 16 U/ml for 16 hours at 37° C. The proteins were then resolved on SDS-PAGE, electrotransferred, and blotted against NHE-3 antiserum. As demonstrated in Fig. 7 (lane 1), rat NHE-3 is present as a sharp band at - 9 0 kd. Pretreatment of rat BBM with glycopeptidase F (lane 2), endoglycosidase H (lane 3), or endoglycosidase F (lane 4) did not alter the mobility of NHE-3 as compared with controls (lane 1). Taken together, these results indicate that rat NHE-3 is not an N-linked glycoprotein. DISCUSSION
The presence or absence of carbohydrate moieties on NHE-3 and their functional role are poorly understood. Studies examining the role of glycosylation in exchanger activity have shown that treatment of rat renal cortical BBM with endoglycosidase F decreased the activity of the Na+/H + exchanger.29 These results suggested that NHE-3 might be a glycoprotein. However, a recent report demonstrated that rat NHE-3, when expressed in fibroblast cells, was not glycosylated,is A preliminary report regarding NHE-3 glycosylation showed that NHE-3 is not a glycoprotein.19 The results of the current experiments indicate that rabbit NHE-3, but not rat or dog NHE-3, is N-link glycosylated. There are several explanations with regard to these conflicting reports. It is possible that NHE-3 glycosylation may be cell specific. NHE-3 is an epithelium-specific isoform and is not expressed in non-epithelial cells. As such, the posttranslational modification of this exchanger may vary when it is expressed in a nonepithelial cell line. One should consider these facts when interpreting the results of recent experiments on NHE-3 expression in fibroblast cells, i8 Another possibility is that glycosylation of NHE-3 may be species or cell specific. Studies characterizing
J Lab Clin Med Volume 128, Number 3
Bizal et al.
10 sec
1 min
2min
control
570 + 40
1230 + 128
1480 _+ 145
790 + 89
deglycosylated
510 .+ 55
1080 _+ 141
1240 _+ 135
720 .+ 79
309
2hr
p>O.05 for all time points (n=3)
Fig. 5. Effect of partial deglycosylation on Na+/H + exchanger activity. BBM proteins were pretreated with buffer or glycopeptidase F (16 U/ml) for 4 hours at 37° C (similar to Fig. 3, lane 2). Thereafter, vesicles were pre-equih'brated for 120 minutes at 20°C in a medium consisting of 52 mmol/L tetramethyl ammonium gluconate, 60 mmol/L potassium gluconate, 52 mmol/L 2-N-morpholino ethane sulfonic acid, 42 mmol/L HEPES, and 21 mmol/L tetramethyl ammonium-hydroxide, pH 6.0. The uptake of 1 mmol/L Z2Na+ into BBM vesicles was assayed overtime in the presence of a medium consisting of 52 mmol/L tetramethyl ammonium gluconate, 60 mmol/L potassium gluconate, 31 mmol/L mannitol, 10 mmol/L/vIES, 42 mmol/L HEPES, 31 mmol/L tetramethyl ammonium hydroxide (pH 7.5), and 1 mmol/L amiloride. Values shown for uptake represent the mean +-_SEM for experiments performed in triplicate on three different membrane preparations.
kD
KD
.
~~!!i~iii!~!!!i;i,'i
94
kD
94
71
94
A
B
1
2
C
1
2
Fig. 6. A, Immunoblots of BBM from canine kidney cortex when using NHE-3 antibody. BBM proteins (200 ~g/lane) were resolved on SDS-PAGE, transferred to nitrocellulose membranes, and probed with NHE-3 immune serum as described in Methods. B, Immunoblot analysis of canine cortical BBM with NHE-3 immune serum after complete digestion with glycopeptidase F. BBM proteins (100 izg/lane) were incubated for 16 hours at 37° C in the presence of glycopeptidase F at 16 U/m1 (left panel, lane 2) or buffer (left panel, lane 1). Immunoblot analysis of canine cortical BBM with NHE-3 antibodies after complete digestion with endoglycosidase H. BBM proteins (100 tzg/lane) were incubated for 16 hours at 37° C in the presence of endoglycosidase H at 0.1 U/m1 (right panel, lane 2) or buffer (right panel, lane 1).
N H E - 3 in L L C - P K 1 cells f o u n d t h a t t r e a t m e n t o f t h e cells w i t h t u n i c a m y c i n for 24 h o u r s , w h i c h b l o c k s N-linked glycosylation in the endoplasmic reticul u m , significantly d e c r e a s e d a n t i p o r t e r activity, 3° in-
dicating that glycosylation of the Na+/H + exchanger i s o f o r m N H E - 3 is e s s e n t i a l for a n t i p o r t e r activity. Immunoblots of luminal membranes from tunicamyc i n - t r e a t e d L L C - P K 1 cells w i t h specific N H E - 3
310
J Lab Clin Med September 1996
Bizalet al.
kD
94-
1
2
3
4
Fig. 7. Immunoblot analysis of rat cortical BBM with NHE-3 immune serum after complete digestion with endoglycosidase H. BBM proteins (100 i~g/lane) were incubated for 16 hours at 37° C with glycolytic buffer alone (lane 1), glycopeptidase F at 16 U/ml (lane 2), endoglycosidase H at 0.1 U/ml (lane 3), or endoglycosidase F at 0.1 U/ml (lane 4).
antiserum showed that two NHE-3 bands at 94 and 90 kd did not reach the membrane. 3° These results suggest that the N-linked carbohydrates may have an important role in NHE-3 transport from the endoplasmic reticulum and insertion into the plasma membrane. In the present study, N-linked glycosylation of the Na+/H + exchanger isoform NHE-3 was examined in mammalian kidneys. Cortical BBM vesicles were isolated from rabbit, dog, or rat kidneys and were pretreated with glycopeptidase F, endoglysidase H, or endoglycosidase F for 16 hours. The membrane proteins were then blotted against an NHE-3-specific immune serum. The apparent molecular mass of NHE-3 was different in these three species. Rabbit NHE-3 was a - 9 5 kd protein, whereas rat and dog NHE-3 had apparent molecular masses of - 9 0 and 83 kd, respectively. Treatment of rabbit membrane proteins with glycopeptidase F significantly decreased the apparent molecular mass of NHE-3. However, similar treatment of dog or rat membrane proteins did not alter the mobility of NHE-3. These results indicate that NHE-3 is differentially glycosylated in mammalian species. Lack of deglycosylation of rat or dog kidney NHE-3 after pretreatment of membrane proteins with glycopeptidas F, endoglycosidase F, or endoglycosidase H does not exclude, with certainty, the possibility that NHE-3 in these species is nonglycosylated. These enzymes remove only N-linked carbohydrate moieties. As such, the results do not rule out the possibility of O-linked glycosylation of NHE-3 in rat or dog kidney.
Figs. 2 through 4 demonstrate that deglycosylation of rabbit NHE-3 with glycopeptidase F was both concentration and time dependent. Glycopeptidase F, at a low concentration (4 U/ml), decreased the apparent molecular weight of NHE-3 to -91 kd (Fig. 2). However, high concentrations of glycopeptidase F (16 U/ml) resulted in the appearance of three bands at 82, 84, and 91 kd (Figs. 3 and 4). Complete digestion of NHE-3 with glycopeptidase F (Fig. 4, lane 3) indicated that the 82 kd band is the most prominent and the 91 kd protein is the least prominent deglycosylated species. The 91 kd band intensity decreased while the abundance of the 82 kd protein increased significantly with an increase in the duration of the deglycosylation process (lanes 2 and 3, Fig. 4). This strongly suggests that the 91 kd isoform is a partially deglycosylated isoform, whereas the 82 kd band is a completely deglycosylated isoform. The abundance of the 84 kd band remained the same after complete digestion (lanes 2 and 3, Fig. 4), indicating that most likely it is a completely deglycosylated NHE-3 isoform. The resuits of the above experiments are compatible with the conclusion that the Na+/H + exchanger isoform NHE-3 exists in more than one form. The presence of two carbohydrate side chains with different molecular weights could result in the appearance of the 82 and 84 kd products. Alternatively, it is possible that the 95 kd protein is composed of two adjacent NHE-3 bands with identical carbohydrate side chains and the 82 and 84 kd products are the deglycosylated forms of these two bands.
J Lab Clin Med Volume 128, Number 3
Of interest are the molecular weights of the deglycosylated products in the present studies--82 and 84 kd. Two recent studies in fibroblast cells 18 and a mouse L-cell line 19 indicated that expression of NHE-3 yielded an 82 kd protein. Treatment of cells with N-glycosidase F is or tunicamycin 19 did not alter the mobility of NHE-3. These results indicate that the 82 kd protein expressed in fibroblasts or L-cells is not a glycoprotein. The results of these experiments are not in conflict with our studies (Figs. 2 through 4). The current experiments (Figs. 2 through 4) demonstrate that the 82 kd protein is a completely deglycosylated form of NHE-3. It is possible that posttranslational modification of NHE-3 in epithelial and nonepithelial cells is different. NHE-3 is an epithelium-specific isoform. 6 If glycosylation of NHE-3 is a posttranslational modification process unique to epithelial cells, then its expression in non-epithelial cells 18'19 should yield a nonglycosylated isoform--that is, an 82 kd protein. Alternatively, it is possible that the antibodies used in these experiments have different affinities for glycosylated or nonglycosylated isoforms and, as such, recognize different isoforms. Deglycosylation of rabbit NHE-3 did not alter the exchanger activity, indicating that the carbohydrate side chains have no role in exchanger function. This is very similar to NHE-1, where glycosylation was found to have no effect on the exchanger activityJ 7 With regard to the results of influx experiments, another plausible explanation is that a subpopulation of NHE-3 was deglycosylated and, as a result, the exchanger activity of that subpopulation decreased. The role of glycosylation in NHE-3 exchange activity needs further examination. Although one recent report indicated that it may play a role in the translocation of NHE-3 from the endoplasmic reticulum and insertion in the plasma membrane, 3° additional studies are needed to clarify this issue. In conclusion, the results of the above experiments demonstrate that the NHE-3 isoform is differentially glycosylated in mammalian kidneys. Rabbit NHE-3 is N-link glycosylated, whereas dog and rat NHE-3 are not N-link glycosylated. The results further demonstrate that glycosylation of NHE-3 is not required for the exchanger activity.
REFERENCES
1. Mahnensmith RL, Aronson PS. The plasma membrane sodium-hydrogen exchanger and its role in physiological and pathophysiological processes. Circ Res 1985;56:773-88. 2. Krapf R, Alpern RJ. Cell pH and transepitheial transport in the renal proximal tubule. J Membrane Biol 1993;131:1-10.
Bizal et al.
311
3. Soleimani M, Singh G. Physiologic and molecular aspects of the Na+/H ÷ exchangers in health and disease processes. J Invest Med 1995;43:419-30. 4. Sardet C, Franchi A, Pouyss6gur J. Molecular cloning, primary structure, and expression of the human growth factoractivatable Na+/H + antiporter. Cell 1989;56:271-80. 5. Orlowski J, Kandasamy RA, Shull GE. Molecular cloning of putative members of the Na/H exchanger gene family. J Biol Chem 1992;267:9331-9. 6. Tse C, Brant SR, Walker MS, Pouyssegur J, Donowitz M. Cloning and sequencing of a rabbit cDNA encoding an intestinal and kidney-specific Na+/H + exchanger isoform (NHE-3). J Biol Chem 1992;267:9340-6. 7. Wang Z, Orlowski J, Shull GE. Primary structure and functional expression of a novel gastrointestinal isoform of the rat Na/H exchanger. J Biol Chem 1993;268:11925-8. 8. Tse C, Levine SA, Yun CH, Montrose MH, Little PJ, Pouyssegur J, et al. Cloning and expression of a rabbit cDNA encoding a serum-activated ethylisopropylamiloride-resistant epithelial Na+/H + exchanger isoform (NHE-2). J Biol Chem 1993;267:9340-6. 9. Biemesderfer D, Pizzonia J, Abu-Alfa A, Markus E, Reilly RF, Igarashi P, et al. NHE-3: a Na+/H + exchanger isoform of renal brush border. Am J Physiol 1993;265:F736-42. 10. Bookstein C, Depaoli A, Musch MW, Rao MC, Chang EB. Na+/H + exchangers, NHE-1 and NHE-3, of rat intestine: expression and localization. J Clin Invest 1994;93:106-13. 11. Soleimani M, Bookstein C, Bizal GL, Musch MW, Hattabaugh YJ, Rao MC, et al. Localization of the Na+/H + exchanger isoform NHE-3 in rabbit and canine kidney. Biochim Biophys Acta 1994;1195:89-95. 12. Soleimani M, Bookstein C, McAteer J, Hattabaugh YJ, Bizal GL, Musch MW, et al. Effect of high osmolality on Na+/H ÷ exchange in renal proximal tubule cells. J Biol Chem 1994; 269:15613-8. 13. Soleimani M, Bookstein C, Sing G, Rao MC, Chang EB, Bastani B. Differential regulation of Na+/H + exchange and H+-ATPase by pH and HCO3- in kidney proximal tubules. J Membr Biol 1995;144:209-16. 14. Haggerty JG, Agarwal N, Reilly RF, Adelberg EA, Slayman CW. Pharmacologically different Na/I-I antiporters on the apical and basolateral surfaces of cultured porcine kidney cells (LLC-PK1). Proc Natl Acad Sci 1988;85:6797-801. 15. Biemesderfer D, Reilly RF, Exner M, Igarashi P, Aronson PS. Immunocytochemical characterization of Na÷-H + exchanger isoform NHE-1 in rabbit kidney. Am J Physiol 1992; 263:F833-40. 16. Sardet C, Councillon L, Franchi A, Pouyssegur J. Growth factors induce phosphorylation of the Na-/H ÷ antiporter, a glycoprotein of 110 kd. Science 1990;247:723-5. 17. Haworth RS, Frohlich O, Fliegel L. Multiple carbohydrate moeities on the Na+/H + exchanger. Biochem J 1993;289:63740. 18. Counillon L, Pouyssegur J, Reithmeier AF. The Na÷/H + exchanger NHE-1 possesses N- and O-linked glycosylation restricted to the first N-terminal extracellular domain. Biochemistry 1994;33:10463-9. 19. Abu-Alfa A, Biemesderfer D, Pizzonia J, Reilly RF, Igarashi P, Aronson PS. Glycosylation of Na+/H ÷ exchanger isoforms NHE1 and NHE3. Clin Res 1994;42:317A. 20. Soleimani M, Howard R. Presence of chloride/formate exchange in cardiac and vascular smooth muscle cells. Circ Res 1994;74:48-55. 21. Evers C, Hasse W, Murer H, Kinne R. Properties of brush
312
22.
23.
24.
25.
Bizal et al.
border membrane vesicles isolated from rat kidney cortex by calcium precipitation. Membrane Biochem 1978;1:202-21. Soleimani M, Bergman JA, Hosford MA, McKinney TD. Potassium depletion increases luminal Na+/H + exchange and basolateral Na+:CO3-:HCO3 - cotransport in rat renal cortex. J Clin Invest 1990;86:1076-83. Soleimani M, Bizal GL, McKinney TD, Hattabaugh YJ. 1992. Effect of in vitro metabolic acidosis on luminal Na+/I-I+ exchange and basolateral Na+:HCO3 - cotransport in rabbit kidney proximal tubules. J Clin Invest 1992;90:211-8. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227:680-5. Michalak M, Fliegel L, Wlasichuk KJ. Isolation and characterization of calcium binding glycoproteins of cardiac sarcolemmal vesicles. J Biol Chem 1990;265:5869-74.
J Lab Clin Med September 1996
26. Haworth RS, Frohlich O, Fliegel L. Biochem J 1993;289:637-40. 27. Soleimani M, Aronson PS. Effects of acetazolamide on Na/ HCO 3- cotransport in basolateral membrane vesicles isolated from rabbit renal cortex. J Clin Invest 1989;83:945-51. 28. Soleimani M, Aronson PS. Ionic mechanism of Na-HCO3 cotransport in rabbit renal basolateral membrane vesicles. J Biol Chem 1989;264:18302-8. 29. Yusufi AN, Szczeparska-Konkel M, Dousa TP. 1988. Role of N-linked oligosaccharides in the transport activity of the Na+/H + antiporter in rat renal brush border membrane. J Biol Chem 1988;263:13683-91. 30. Soleimani M, Hattabaugh Y, Singh G, Bookstein C, Rao MC, Chang EB. Inhibition of glycosylation decreases Na+/H + exchange activity, blocks NHE-3 transport to the membrane, and increases NHE-3 mRNA expression in LLC-PK1 cells. J Lab Clin Med 1996;127:565-73.
The glycosylation of Na+/H + exchanger isoform NHE-3 was studied in brush border membrane (BBM) vesicles isolated from rabbit, dog, and rat kidney cortex. Western blot analyses were performed against BBM proteins by using polyclonal antibodies to an NHE-3 fusion protein. In rabbit kidney, NHE-3 antibody recognized a band with - 9 5 kd molecular mass. Treatment of rabbit cortical BBM with glycopeptidase F, at 16 U/ml, for 4 or 16 hours increased the apparent mobility of NHE-3 to 84 and 82 kd, respectively. Incubation of rabbit BBM proteins for 16 hours with endoglycosidase H, at 0.1 U/ml, did not alter the apparent mobility of NHE-3. Deglycosylation of NHE-3 with glycopeptidase F did not affect acid-stimulated, amiloride-sensitive sodium 22 influx in BBM vesicles as compared with that in controls (p > 0.05). Immunoblot analysis against BBM proteins from canine kidney cortex demonstrated the presence of an - 8 3 to 92 kd protein. Treatment of canine BBM with glycopeptidase F or endoglycosidase H or F for 16 hours did not alter the apparent mobility of NHE-3, suggesting that canine renal NHE-3 is not glycosylated. Treatment of canine kidney BBM with glycopepfidase F did not affect acid-stimulated 22Na+influx as compared with that in controls (p > 0.05). Immunoblot analysis against BBM proteins from rat kidney cortex demonstrated the presence of a sharp bond at 90 kd. Treatment of rat BBM with glycopeptidase F or endoglycosidase H or F for 16 hours did not alter the apparent mobility of NHE-3, suggesting that rot renal NHE-3 is not glycosylated. The above experiments suggest that NHE-3 glycosylation in mammalians is species specific and that glycosylation does not affect the exchanger activity. (J Lab Clin Med 1996;128:304-12)
Abbreviations: BBM = brush border membrane; HEPES/MES= N-2-hydroxyethylpiperazine-N2-ethanesulfonic acid/2-N-morpholino ethane sulfonic acid; IgG = immunoglobulin G; PBST= 0.1% Tween 20/phosphate-buffered saline solution; SDS-PAGE = sodium dodecyl sulfatepolyacrylamide gel electrophoresis
From the Department of Medicine, Indiana University School of Medicine and the Veterans Affairs Medical Center, Indianapolis; the Department of Medicine, University of Chicago; and the Department of Physiology and Biophysics, University of Illinois at Chicago. Supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK 46789 and a Merit Review Grant from the Department of Veterans Affairs (to M.S) and by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK 42086 (to E.B.C). Dr. Howard was the recipient of a Physician Scientist Award from the National Institutes of Health (DK-02116). Portions of this study have been published in abstract form (J Am Soc Nephrol 1994;5:262A). Submitted for publication Jan. 12, 1996; revision submitted March 27, 1996; accepted April 11, 1996. 304
T
h e majority o f H C O 3 - filtered at the glomerulus is r e a b s o r b e d in the kidney proximal tubule via the luminal N a + / H + exchanger. 1-3 R e c e n t molecular cloning experiments have identified the existence o f four structurally related N a + / H + exchanger isoforms, which have b e e n designated N H E - 1 , N H E - 2 , N H E - 3 , and N H E - 4 . 4-8 Functional, molecular, and immunologic studies have d e m o n s t r a t e d that N H E - 3 is Reprint requests: Manoocher Soleimani, MD, Division of Nephrology and Hypertension, Department of Internal Medicine, University of Cincinnati, 231 Bethesda Ave., MSB 5502, Cincinnati, OH 45267. 0022-2143/96 $5.00 + 0 5/1/74173
J Lab Clin Med Volume 128, Number 3
the Na+/H + exchanger isoform that is present in luminal membranes of proximal tubules. 9-13 A basolateral Na+/H + exchanger has also been identified in kidney proximal tubule cells and is presumed to be mostly involved with cell p H regulation. 2'3'14 Immunologic studies have demonstrated that NHE-1 is the basolateral Na+/H + exchanger isoform in kidney proximal tubules, as The amino acid sequences of Na+/I-I+ exchanger isoforms, deduced from their respective cDNAs, indicate that all N H E isoforms have consensus sequences for N-linked glycosylation in their extracytoplasmic loops. 4-s Several studies have examined N-linked glycosylation of NHE-1. The results indicate that NHE-1 is a glycoprotein, and they suggest the presence of at least two N-linked carbohydrate moieties. 16'17 In addition, recent studies illustrate that NHE-1 possesses O-linked glycosylation, as The presence of carbohydrate moieties on N H E - 3 is less known. The N H E - 3 c D N A in rabbit or rat encodes a protein with a molecular mass of - 9 2 kd. A recent investigation in which an antibody against an N H E - 3 - M B P fusion protein was used found that the exchanger had an apparent molecular mass of - 8 0 kd. 9 When a specific antiserum to an NHE-3G S T fusion protein was used, N H E - 3 was found to have a molecular mass of - 9 5 kd in rabbit kidney cortex. H-a3 Expression of the rat N H E - 3 c D N A in PS120 fibroblast cells deficient in endogenous exchanger shows the presence of an 82 kd protein, a8 NHE-3, however, was found to have a molecular mass of - 9 7 kd in experiments performed in the rat intestine, a° The reason for the difference in the mobility of the N H E - 3 exchanger in these several studies is not clear. One possible explanation is that the gel systems used in these studies are different. Another possibility is the presence of species difference with regard to processing of NHE-3. One plausible explanation, however, is that N H E - 3 might be present in more than one isoform. For example, differential glycosylation of N H E - 3 could result in the expression of several isoforms. As such, various N H E - 3 antibodies could detect, based on their affinities, different N H E - 3 isoforms. Indeed, preliminary reports show that the 80 kd N H E - 3 is not a glycoprotein. 19 The purpose of these experiments was to study glycosylation of N H E - 3 in m a m m a l i a n kidneys. Toward this end, brush border m e m b r a n e vesicle from rabbit, dog, and rat kidney cortex were isolated and studied for the presence of carbohydrate moieties on N H E - 3 with various N-glycanases. The results indicate that rabbit N H E - 3 is N-link glycosylated, whereas dog and rat N H E - 3 are not N-link glycosylated. The results further indicate that
Bizal et al.
305
N H E - 3 activity is not affected by the glycosylation state of the exchanger. METHODS Membrane vesicle preparation. Male New Zealand
white rabbits were killed by intravenous sodium pentobarbital administration. Male Sprague-Dawley rats were killed by intraperitoneal injection of sodium pentobarbital. Kidneys were removed and processed for cortical BBM vesicle preparation. Dog kidneys were obtained as described. 2° BBM vesicles were prepared from renal cortexes by a Ca ++ aggregation method zl as used previously.2z'23 The purification of BBM vesicles relative to the initial homogenate, assessed with alkaline phosphatase, was sevenfold to 10-fold for all groups. SDS-PAGE and immunoblot analysis. Membrane fractions (100 to 200 ixg/lane) were solubilized and subjected to vertical slab SDS-PAGE according to the Laemmli protocol. 24 Proteins were electrophoretically transferred to nitrocellulose at 200 mA for 15 hours. The nitrocellulose filters were blocked in PBST and then incubated with 10 Ixl of an NHE-3 immune serum diluted at 1:400 in PBST for 2 hours. The specificity of this antibody was recently demonstrated. 11 The excess antibody was removed with PBST, and the antigen-antibody complex was treated with alkaline phosphatase-conjugated goat antirabbit IgG diluted at 1:1000 in PBST. The nitrocellulose filters were developed with nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate dissolved in 70% or 100% N,N,-dimethylformamide. Carbohydrate analysis. BBM vesicles (10 mg) were washed in 5 ml of a glycolytic buffer consisting of 50 mmol/L KC1, 20 mmol/L NaH2PO4, pH 7.2. The membrane proteins were then resuspended to i mg protein per milliliter in the same buffer. Ethylenediaminetetraacetic acid (50 mmol/L) and a cocktail of proteinase inhibitors including pepstatin A, leupeptin, and phenylmethylsulfonyl fluoride were added to the membrane proteins as described, a5'26 Membrane proteins (100 txl) were incubated in the presence or absence of glycopeptidase F (at 8 and 16 U/ml), endoglycosidase H (0.1 U/ml), or endoglycosidase F (0.1 U/ml) for up to 16 hours at 37° C. The purpose of choosing the current enzymes (endoglycosidase F, endoglycosidase H, or glycopeptidase F) was to study the N-linked glycosylation of NHE-3, because these enzymes remove almost all N-linked carbohydrate moieties (high mannose, complex oligosaccharides, or both groups) of proteins. Endoglycosidase F removes carbohydrate moieties with the complex-type N-oligosaccharide biantennary group, whereas endoglycosidase H affects side chains with hybrid and high mannose-type groups and glycopeptidase F affects both types of carbohydrate side chains. ~2Na+ influx measurement. Intact BBM vesicles were treated with glycopeptidase F at 16 U/ml for 4 hours. Thereafter, vesicles were washed, preincubated in a HEPES/MES buffer, and assayed for Na+/H + exchanger activity with :2Na + influx and a rapid filtration method as
306
J Lab Clin Med September 1996
Bizal et al,
NHE-3 Immunoblot (rabbit kidney) KD
bit IgG and 5-bromo-4-chloro-3-indolyl phosphate were purchased from Boehringer Mannheim, Indianapolis, Ind. Nitrocellulose filters, nitro blue tetrazolium, and glycopeptidase F were purchased from Sigma Chemical Co., St. Louis, Mo. CNBr-activated Sepharose 4B was purchased from Pharmacia, Piscataway, N.J. SDS, acrylamide, and NN'-methylene-bis-acrylamide were purchased from BioRad Laboratories, Hercules, Calif. RESULTS
94-
71X
n m
o
~3
E O
Fig. 1. Immunoblotsof BBM from rabbit renal cortex (leftlane) and outer medulla (rightlane) when using NHE-3 antibody.BBM proteins (200 ixg/lane)were resolved on SDS-PAGE, transferred to nitrocellulose membranes, and probed with NHE-3 immune serum as described in Methods.
previously described.27'28 In repeated experiments we found that incubation of vesicles at 37° C for 16 hours (in both the control group and the deglycosylated group) resulted in loss of vesicle integrity, making it impossible to study exchanger activity. The ice-cold medium used to dilute and wash the vesicles consisted of 170 mmol/L K gluconate, 10 mmol/L HEPES/tetramethyl ammonium hydroxide, pH 7.5. Each filter (0.45 Ixm; DAWP; Millipore Corp., Bedford, Mass.) was placed in 5 ml of scintillation fluid (Ready-Solv HP; Beckman Instruments Inc., Fullerton, Calif.), and radioactivity was determined by scintillation spectroscopy. The final composition of the experimental medium and other details of the protocols are given in the figure legends. All experiments were performed with vesicles treated with valinomycin (0.5 mg/ml) and pre-equilibrated in medium of appropriate composition to ensure that [K+]o = [K+]i during uptake measurements. Na uptake studies were performed in rabbit, rat, and dog kidney BBM vesicles. Materials. 22NaC1 (specific activity 906 mCi/mg) was obtained from DuPont NEN Research Products, Wilmington, Del.). Alkaline phosphatase-conjugated anti-rab-
We first studied the expression of NHE-3 in BBMs isolated from the rabbit kidney cortex and outer medulla. Fig. 1 demonstrates that the immune serum containing NHE-3 antibody detected a strong and diffuse band with an apparent molecular weight of - 9 5 kd in cortical (left lane) and outer medullary (right lane) luminal membranes. These results confirm recent reports on NHE-3 expression in the rabbit kidney, n'13 In the next series of experiments we determined whether the rabbit kidney NHE-3 is a glycoprotein. Rabbit cortical BBM proteins (100 ~g/100 txl) were incubated with glycopeptidase F at a low concentration (4 U/ml) for 16 hours at 37 ° C, resolved on SDS-PAGE, and electrotransferred, and the blot was probed with NHE-3 antiserum. As demonstrated in Fig. 2, the digestion of BBM with glycopeptidase F at a low concentration (lane 2) changed the apparent molecular mass of the exchanger (lane 1, control) from - 9 5 to - 9 1 kd. These results suggest that NHE-3 is a glycoprotein with a - 4 kd shift in molecular mass after treatment with 4 U/ml glycopeptidase F. To characterize the glycosylation of NHE-3 further, rabbit BBMs (100 ~g/100 rd) were completely digested with high concentrations of glycopeptidase F (16 U/ml) for 16 hours at 37 ° C. Thereafter, vesicles were subjected to SDS-PAGE, transferred to nitrocellulose membrane, and blotted against NHE-3 antiserum. As demonstrated in Fig. 3, digestion of BBMs with glycopeptidase F at high concentrations (lane 2) resulted in the appearance of three bands at 92, 84, and 82 kd. The 82 kd band was very strong when compared with the 84 and 92 kd proteins. Comparison of Fig. 2 and Fig. 3 indicates that NHE-3 is most likely present in more than one form. To study the issue of NHE-3 glycosylated isoforms further, BBM proteins were treated with endoglycosidase H (0.1 U/ml) for 16 hours or glycopeptidase F (16 U/ml) for 4 and 16 hours at 37 ° C. Membrane proteins were then resolved on SDS-PAGE and electrotransferred, and the blot was probed with NHE-3 antiserum. Fig. 4, lane 1, is an immunoblot
J Lab Clin Med Volume 128, Number 3
Bizal et al.
kD
307
kD
94
1
2
94-
Fig. 2. Immunoblot analysis of rabbit cortical BBM with NHE-3 immune serum after partial digestion with glycopeptidase F. BBM proteins (100 i~g/lane) were incubated for 16 hours at 37° C in the presence of glycopeptidase F at 4 U/ml (lane 2) or glycolytic buffer (lane 1). Proteins were resolved on SDS-PAGE.
71of NHE-3 treated with glycolytic buffer alone (no glycosidase). Lane 2 and lane 3 reflect digestion with glycopeptidase F for 4 and 16 hours, respectively. Fig. 4, lane 4, reflects 16 hours of digestion with endoglycosidase H at 0.1 U/ml. As shown, 4 hours of digestion with glycopeptidase F (lane 2) resulted in the appearance of two thin bands at 82 and 84 kd and another band at 92 kd. Complete digestion (16 hours) of NHE-3 with glycopeptidase F (lane 3) resulted in a stronger appearance of the 82 kd band and partial disappearance of the 92 kd band. Fig. 4, lane 4, indicates that digestion with endoglycosidase H had no effect on NHE-3 mobility. In additional experiments not shown we found that the effect of endoglycosidase F was very similar to that with glycopeptidase F. To examine the effect of NHE-3 deglycosylation on exchanger activity, BBM vesicles were treated with either glycolytic buffer (no glycosidase) or glycopeptidase F at 16 U/ml for 4 hours (as indicated in Methods, incubation of vesicles at 37° C for 16 hours resulted in the loss of vesicle integrity, making it impossible to study exchanger activity at that time point). Thereafter, vesicles were washed and preincubated in a HEPES/MES buffer (see the legend to Fig. 5). The time course of ZZNa+ influx in BBM vesicles was assayed in the presence of an inward pH gradient (pHo/pHi = 7.5/6.0) and --_ 1 mmol/L amiloride added to the external solution, Fig. 5 shows that partial deglycosylation of NHE-3 had no significant effect on the exchanger activity as compared with results in control vesicles. The 2-hour
1
2
Fig. 3. Immunoblot analysis of rabbit cortical BBM with NHE-3 immune serum after complete digestion with glycopeptidase F. BBM proteins (100 i~g/lane) were incubated for 16 hours at 37° C in the presence of glycopeptidase F at 16 U/ml (lane 2) or buffer (lane 1). Proteins were resolved on SDS-PAGE.
Z2Na+ uptakes (equilibrium values) were also comparable in both groups, indicating that enzymatic treatment did not affect vesicle integrity (Fig. 5). To determine whether deglycosylation affects the Michaelis-Menten constant of the exchanger, the initial rate of amiloride-sensitive 22Na influx in BBM vesicles was assayed in the presence of 3, 6, 9, and 12 mmol/L sodium. Kinetic analysis of the results by Hanes-Woolf plot did not show any significant change in the Michaelis-Menten constant of the exchanger in the deglycosylated state as compared with control values (KNa 13.4 mmol/L in glycopeptidase F-treated vesicles and 12.1 mmol/L in buffertreated vesicles, p < 0.05, n = 3). A limited doseresponse inhibition of the exchanger activity by dimethylamiloride at 1 and 10 t~mol/L in rabbit BBM vesicles showed no difference in the sensitivity of the exchanger to inhibition between glycosylated and nonglycosylated states (12% vs 15% inhibition at 1 ~zmol/L dimethylamiloride and 36% vs 41% inhibition at 10 p~mol/L dimethylamiloride for glycosylated and nonglycosylated NHE-3, respectively,
308
J Lab Olin Med September 1996
Bizal et al.
kD
94-
71-
1
2
3
4
Fig. 4. Time course of NHE-3 digestion with glycopeptidase F. BBM proteins (100 ixg/lane) were incubated for 16 hours with buffer (lane 1), 4 hours with glycopeptidase F (16 U/ml) (lane 2), 16 hours with glycopeptidase F (16 U/ml) (lane 3), or 16 hours with endoglycosidase H (0.1 U/ml) at 37° C.
p > 0.05 for both groups). Moreover, pH i dependence of the exchanger remained the same as shown by 60-second uptake studies performed at pH i 6.5 and 7.0 (410 _+ 36 picomoles sodium per milligram of protein per minute in controls vs 370 _+ 32 picomoles sodium per milligram of protein per minute in the deglycosylated state at pH i 6.5, and 160 _+ 16 picomoles sodium per milligram of protein per minute in controls vs 140 _+ 16 picomoles sodium per milligram of protein per minute in the deglycosylated state at pHi 7.0, p > 0.05, n = 3 for each group). In the next series of experiments, expression and possible glycosylation of NHE-3 in the canine kidney was examined. Fig. 6, A, demonstrates that the immune serum containing NHE-3 antibody detected a strong and diffuse band with an apparent molecular weight of 83 to 92 kd in canine kidney BBM. These results confirm recent reports on NHE-3 expression in the canine kidney,i~ In addition to the diffuse band, several smaller proteins in the 65 to 75 kd area were also occasionally detected. The immunodetection of these proteins was not consistent, suggesting the presence of proteolytic breakdown products of the 83 to 92 kd protein. We next examined N-linked glycosylation of NHE-3 in canine kidney. Canine cortical BBM proteins (100 ixg/100 txl) were incubated with glycopeptidase F, at 16 U/ml, for 16 hours at 37° C. Fig. 6, B (lefipanel), is an immunoblot analysis of NHE-3 and shows that treatment of canine BBM with glycopeptidase F did not alter the mobility of NHE-3. Treatment of BBM
vesicles with endoglycosidase H also did not affect the apparent mobility of NHE-3 (Fig. 6, B, right panel). Similarly, treatment of canine BBM vesicles with endoglycosidase F did not alter NHE-3 mobility (data not shown). Pretreatment of BBM vesicles with glycopeptidase F at 16 U/ml for 4 hours did not affect the exchanger activity as assessed by acidstimulated (pHo/pH i = 7.5/6.0), amiloride-sensitive Z2Na+ influx (104.7% ___ 6.3% in glycopeptidase F-treated vesicles vs 100% in controls, n = 3, p > 0.05). The above experiments demonstrate that NHE-3 is differentially glycosylated in rabbit and dog kidney. In the last series of experiments, possible Nlinked glycosylation of NHE-3 in the rat kidney was examined. Rat cortical BBM proteins (100 txg/100 ~1) were incubated with glycopeptidase F at 16 U/ml for 16 hours at 37° C. The proteins were then resolved on SDS-PAGE, electrotransferred, and blotted against NHE-3 antiserum. As demonstrated in Fig. 7 (lane 1), rat NHE-3 is present as a sharp band at - 9 0 kd. Pretreatment of rat BBM with glycopeptidase F (lane 2), endoglycosidase H (lane 3), or endoglycosidase F (lane 4) did not alter the mobility of NHE-3 as compared with controls (lane 1). Taken together, these results indicate that rat NHE-3 is not an N-linked glycoprotein. DISCUSSION
The presence or absence of carbohydrate moieties on NHE-3 and their functional role are poorly understood. Studies examining the role of glycosylation in exchanger activity have shown that treatment of rat renal cortical BBM with endoglycosidase F decreased the activity of the Na+/H + exchanger.29 These results suggested that NHE-3 might be a glycoprotein. However, a recent report demonstrated that rat NHE-3, when expressed in fibroblast cells, was not glycosylated,is A preliminary report regarding NHE-3 glycosylation showed that NHE-3 is not a glycoprotein.19 The results of the current experiments indicate that rabbit NHE-3, but not rat or dog NHE-3, is N-link glycosylated. There are several explanations with regard to these conflicting reports. It is possible that NHE-3 glycosylation may be cell specific. NHE-3 is an epithelium-specific isoform and is not expressed in non-epithelial cells. As such, the posttranslational modification of this exchanger may vary when it is expressed in a nonepithelial cell line. One should consider these facts when interpreting the results of recent experiments on NHE-3 expression in fibroblast cells, i8 Another possibility is that glycosylation of NHE-3 may be species or cell specific. Studies characterizing
J Lab Clin Med Volume 128, Number 3
Bizal et al.
10 sec
1 min
2min
control
570 + 40
1230 + 128
1480 _+ 145
790 + 89
deglycosylated
510 .+ 55
1080 _+ 141
1240 _+ 135
720 .+ 79
309
2hr
p>O.05 for all time points (n=3)
Fig. 5. Effect of partial deglycosylation on Na+/H + exchanger activity. BBM proteins were pretreated with buffer or glycopeptidase F (16 U/ml) for 4 hours at 37° C (similar to Fig. 3, lane 2). Thereafter, vesicles were pre-equih'brated for 120 minutes at 20°C in a medium consisting of 52 mmol/L tetramethyl ammonium gluconate, 60 mmol/L potassium gluconate, 52 mmol/L 2-N-morpholino ethane sulfonic acid, 42 mmol/L HEPES, and 21 mmol/L tetramethyl ammonium-hydroxide, pH 6.0. The uptake of 1 mmol/L Z2Na+ into BBM vesicles was assayed overtime in the presence of a medium consisting of 52 mmol/L tetramethyl ammonium gluconate, 60 mmol/L potassium gluconate, 31 mmol/L mannitol, 10 mmol/L/vIES, 42 mmol/L HEPES, 31 mmol/L tetramethyl ammonium hydroxide (pH 7.5), and 1 mmol/L amiloride. Values shown for uptake represent the mean +-_SEM for experiments performed in triplicate on three different membrane preparations.
kD
KD
.
~~!!i~iii!~!!!i;i,'i
94
kD
94
71
94
A
B
1
2
C
1
2
Fig. 6. A, Immunoblots of BBM from canine kidney cortex when using NHE-3 antibody. BBM proteins (200 ~g/lane) were resolved on SDS-PAGE, transferred to nitrocellulose membranes, and probed with NHE-3 immune serum as described in Methods. B, Immunoblot analysis of canine cortical BBM with NHE-3 immune serum after complete digestion with glycopeptidase F. BBM proteins (100 izg/lane) were incubated for 16 hours at 37° C in the presence of glycopeptidase F at 16 U/m1 (left panel, lane 2) or buffer (left panel, lane 1). Immunoblot analysis of canine cortical BBM with NHE-3 antibodies after complete digestion with endoglycosidase H. BBM proteins (100 tzg/lane) were incubated for 16 hours at 37° C in the presence of endoglycosidase H at 0.1 U/m1 (right panel, lane 2) or buffer (right panel, lane 1).
N H E - 3 in L L C - P K 1 cells f o u n d t h a t t r e a t m e n t o f t h e cells w i t h t u n i c a m y c i n for 24 h o u r s , w h i c h b l o c k s N-linked glycosylation in the endoplasmic reticul u m , significantly d e c r e a s e d a n t i p o r t e r activity, 3° in-
dicating that glycosylation of the Na+/H + exchanger i s o f o r m N H E - 3 is e s s e n t i a l for a n t i p o r t e r activity. Immunoblots of luminal membranes from tunicamyc i n - t r e a t e d L L C - P K 1 cells w i t h specific N H E - 3
310
J Lab Clin Med September 1996
Bizalet al.
kD
94-
1
2
3
4
Fig. 7. Immunoblot analysis of rat cortical BBM with NHE-3 immune serum after complete digestion with endoglycosidase H. BBM proteins (100 i~g/lane) were incubated for 16 hours at 37° C with glycolytic buffer alone (lane 1), glycopeptidase F at 16 U/ml (lane 2), endoglycosidase H at 0.1 U/ml (lane 3), or endoglycosidase F at 0.1 U/ml (lane 4).
antiserum showed that two NHE-3 bands at 94 and 90 kd did not reach the membrane. 3° These results suggest that the N-linked carbohydrates may have an important role in NHE-3 transport from the endoplasmic reticulum and insertion into the plasma membrane. In the present study, N-linked glycosylation of the Na+/H + exchanger isoform NHE-3 was examined in mammalian kidneys. Cortical BBM vesicles were isolated from rabbit, dog, or rat kidneys and were pretreated with glycopeptidase F, endoglysidase H, or endoglycosidase F for 16 hours. The membrane proteins were then blotted against an NHE-3-specific immune serum. The apparent molecular mass of NHE-3 was different in these three species. Rabbit NHE-3 was a - 9 5 kd protein, whereas rat and dog NHE-3 had apparent molecular masses of - 9 0 and 83 kd, respectively. Treatment of rabbit membrane proteins with glycopeptidase F significantly decreased the apparent molecular mass of NHE-3. However, similar treatment of dog or rat membrane proteins did not alter the mobility of NHE-3. These results indicate that NHE-3 is differentially glycosylated in mammalian species. Lack of deglycosylation of rat or dog kidney NHE-3 after pretreatment of membrane proteins with glycopeptidas F, endoglycosidase F, or endoglycosidase H does not exclude, with certainty, the possibility that NHE-3 in these species is nonglycosylated. These enzymes remove only N-linked carbohydrate moieties. As such, the results do not rule out the possibility of O-linked glycosylation of NHE-3 in rat or dog kidney.
Figs. 2 through 4 demonstrate that deglycosylation of rabbit NHE-3 with glycopeptidase F was both concentration and time dependent. Glycopeptidase F, at a low concentration (4 U/ml), decreased the apparent molecular weight of NHE-3 to -91 kd (Fig. 2). However, high concentrations of glycopeptidase F (16 U/ml) resulted in the appearance of three bands at 82, 84, and 91 kd (Figs. 3 and 4). Complete digestion of NHE-3 with glycopeptidase F (Fig. 4, lane 3) indicated that the 82 kd band is the most prominent and the 91 kd protein is the least prominent deglycosylated species. The 91 kd band intensity decreased while the abundance of the 82 kd protein increased significantly with an increase in the duration of the deglycosylation process (lanes 2 and 3, Fig. 4). This strongly suggests that the 91 kd isoform is a partially deglycosylated isoform, whereas the 82 kd band is a completely deglycosylated isoform. The abundance of the 84 kd band remained the same after complete digestion (lanes 2 and 3, Fig. 4), indicating that most likely it is a completely deglycosylated NHE-3 isoform. The resuits of the above experiments are compatible with the conclusion that the Na+/H + exchanger isoform NHE-3 exists in more than one form. The presence of two carbohydrate side chains with different molecular weights could result in the appearance of the 82 and 84 kd products. Alternatively, it is possible that the 95 kd protein is composed of two adjacent NHE-3 bands with identical carbohydrate side chains and the 82 and 84 kd products are the deglycosylated forms of these two bands.
J Lab Clin Med Volume 128, Number 3
Of interest are the molecular weights of the deglycosylated products in the present studies--82 and 84 kd. Two recent studies in fibroblast cells 18 and a mouse L-cell line 19 indicated that expression of NHE-3 yielded an 82 kd protein. Treatment of cells with N-glycosidase F is or tunicamycin 19 did not alter the mobility of NHE-3. These results indicate that the 82 kd protein expressed in fibroblasts or L-cells is not a glycoprotein. The results of these experiments are not in conflict with our studies (Figs. 2 through 4). The current experiments (Figs. 2 through 4) demonstrate that the 82 kd protein is a completely deglycosylated form of NHE-3. It is possible that posttranslational modification of NHE-3 in epithelial and nonepithelial cells is different. NHE-3 is an epithelium-specific isoform. 6 If glycosylation of NHE-3 is a posttranslational modification process unique to epithelial cells, then its expression in non-epithelial cells 18'19 should yield a nonglycosylated isoform--that is, an 82 kd protein. Alternatively, it is possible that the antibodies used in these experiments have different affinities for glycosylated or nonglycosylated isoforms and, as such, recognize different isoforms. Deglycosylation of rabbit NHE-3 did not alter the exchanger activity, indicating that the carbohydrate side chains have no role in exchanger function. This is very similar to NHE-1, where glycosylation was found to have no effect on the exchanger activityJ 7 With regard to the results of influx experiments, another plausible explanation is that a subpopulation of NHE-3 was deglycosylated and, as a result, the exchanger activity of that subpopulation decreased. The role of glycosylation in NHE-3 exchange activity needs further examination. Although one recent report indicated that it may play a role in the translocation of NHE-3 from the endoplasmic reticulum and insertion in the plasma membrane, 3° additional studies are needed to clarify this issue. In conclusion, the results of the above experiments demonstrate that the NHE-3 isoform is differentially glycosylated in mammalian kidneys. Rabbit NHE-3 is N-link glycosylated, whereas dog and rat NHE-3 are not N-link glycosylated. The results further demonstrate that glycosylation of NHE-3 is not required for the exchanger activity.
REFERENCES
1. Mahnensmith RL, Aronson PS. The plasma membrane sodium-hydrogen exchanger and its role in physiological and pathophysiological processes. Circ Res 1985;56:773-88. 2. Krapf R, Alpern RJ. Cell pH and transepitheial transport in the renal proximal tubule. J Membrane Biol 1993;131:1-10.
Bizal et al.
311
3. Soleimani M, Singh G. Physiologic and molecular aspects of the Na+/H ÷ exchangers in health and disease processes. J Invest Med 1995;43:419-30. 4. Sardet C, Franchi A, Pouyss6gur J. Molecular cloning, primary structure, and expression of the human growth factoractivatable Na+/H + antiporter. Cell 1989;56:271-80. 5. Orlowski J, Kandasamy RA, Shull GE. Molecular cloning of putative members of the Na/H exchanger gene family. J Biol Chem 1992;267:9331-9. 6. Tse C, Brant SR, Walker MS, Pouyssegur J, Donowitz M. Cloning and sequencing of a rabbit cDNA encoding an intestinal and kidney-specific Na+/H + exchanger isoform (NHE-3). J Biol Chem 1992;267:9340-6. 7. Wang Z, Orlowski J, Shull GE. Primary structure and functional expression of a novel gastrointestinal isoform of the rat Na/H exchanger. J Biol Chem 1993;268:11925-8. 8. Tse C, Levine SA, Yun CH, Montrose MH, Little PJ, Pouyssegur J, et al. Cloning and expression of a rabbit cDNA encoding a serum-activated ethylisopropylamiloride-resistant epithelial Na+/H + exchanger isoform (NHE-2). J Biol Chem 1993;267:9340-6. 9. Biemesderfer D, Pizzonia J, Abu-Alfa A, Markus E, Reilly RF, Igarashi P, et al. NHE-3: a Na+/H + exchanger isoform of renal brush border. Am J Physiol 1993;265:F736-42. 10. Bookstein C, Depaoli A, Musch MW, Rao MC, Chang EB. Na+/H + exchangers, NHE-1 and NHE-3, of rat intestine: expression and localization. J Clin Invest 1994;93:106-13. 11. Soleimani M, Bookstein C, Bizal GL, Musch MW, Hattabaugh YJ, Rao MC, et al. Localization of the Na+/H + exchanger isoform NHE-3 in rabbit and canine kidney. Biochim Biophys Acta 1994;1195:89-95. 12. Soleimani M, Bookstein C, McAteer J, Hattabaugh YJ, Bizal GL, Musch MW, et al. Effect of high osmolality on Na+/H ÷ exchange in renal proximal tubule cells. J Biol Chem 1994; 269:15613-8. 13. Soleimani M, Bookstein C, Sing G, Rao MC, Chang EB, Bastani B. Differential regulation of Na+/H + exchange and H+-ATPase by pH and HCO3- in kidney proximal tubules. J Membr Biol 1995;144:209-16. 14. Haggerty JG, Agarwal N, Reilly RF, Adelberg EA, Slayman CW. Pharmacologically different Na/I-I antiporters on the apical and basolateral surfaces of cultured porcine kidney cells (LLC-PK1). Proc Natl Acad Sci 1988;85:6797-801. 15. Biemesderfer D, Reilly RF, Exner M, Igarashi P, Aronson PS. Immunocytochemical characterization of Na÷-H + exchanger isoform NHE-1 in rabbit kidney. Am J Physiol 1992; 263:F833-40. 16. Sardet C, Councillon L, Franchi A, Pouyssegur J. Growth factors induce phosphorylation of the Na-/H ÷ antiporter, a glycoprotein of 110 kd. Science 1990;247:723-5. 17. Haworth RS, Frohlich O, Fliegel L. Multiple carbohydrate moeities on the Na+/H + exchanger. Biochem J 1993;289:63740. 18. Counillon L, Pouyssegur J, Reithmeier AF. The Na÷/H + exchanger NHE-1 possesses N- and O-linked glycosylation restricted to the first N-terminal extracellular domain. Biochemistry 1994;33:10463-9. 19. Abu-Alfa A, Biemesderfer D, Pizzonia J, Reilly RF, Igarashi P, Aronson PS. Glycosylation of Na+/H ÷ exchanger isoforms NHE1 and NHE3. Clin Res 1994;42:317A. 20. Soleimani M, Howard R. Presence of chloride/formate exchange in cardiac and vascular smooth muscle cells. Circ Res 1994;74:48-55. 21. Evers C, Hasse W, Murer H, Kinne R. Properties of brush
312
22.
23.
24.
25.
Bizal et al.
border membrane vesicles isolated from rat kidney cortex by calcium precipitation. Membrane Biochem 1978;1:202-21. Soleimani M, Bergman JA, Hosford MA, McKinney TD. Potassium depletion increases luminal Na+/H + exchange and basolateral Na+:CO3-:HCO3 - cotransport in rat renal cortex. J Clin Invest 1990;86:1076-83. Soleimani M, Bizal GL, McKinney TD, Hattabaugh YJ. 1992. Effect of in vitro metabolic acidosis on luminal Na+/I-I+ exchange and basolateral Na+:HCO3 - cotransport in rabbit kidney proximal tubules. J Clin Invest 1992;90:211-8. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227:680-5. Michalak M, Fliegel L, Wlasichuk KJ. Isolation and characterization of calcium binding glycoproteins of cardiac sarcolemmal vesicles. J Biol Chem 1990;265:5869-74.
J Lab Clin Med September 1996
26. Haworth RS, Frohlich O, Fliegel L. Biochem J 1993;289:637-40. 27. Soleimani M, Aronson PS. Effects of acetazolamide on Na/ HCO 3- cotransport in basolateral membrane vesicles isolated from rabbit renal cortex. J Clin Invest 1989;83:945-51. 28. Soleimani M, Aronson PS. Ionic mechanism of Na-HCO3 cotransport in rabbit renal basolateral membrane vesicles. J Biol Chem 1989;264:18302-8. 29. Yusufi AN, Szczeparska-Konkel M, Dousa TP. 1988. Role of N-linked oligosaccharides in the transport activity of the Na+/H + antiporter in rat renal brush border membrane. J Biol Chem 1988;263:13683-91. 30. Soleimani M, Hattabaugh Y, Singh G, Bookstein C, Rao MC, Chang EB. Inhibition of glycosylation decreases Na+/H + exchange activity, blocks NHE-3 transport to the membrane, and increases NHE-3 mRNA expression in LLC-PK1 cells. J Lab Clin Med 1996;127:565-73.