Defective Brush-border Expression Intrinsic Factor ... · PDF fileinefficient BBM expression of IFCR due to a mutation ... TC-11, trans- cobalamin-11; IFCR, instrinsic factor-cobalamin

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 266, No. 7, Issue of March 5. pp. 4489-4494.1991 Printed in U. S. A.

Defective Brush-border Expression of Intrinsic Factor-Cobalamin Receptor in Canine Inherited Intestinal Cobalamin Malabsorption*

(Received for publication, August 13, 1990)

John C. Fyfe$$, Kalathur S . Ramanujamll, Krishnamurthy Ramaswamy 11, Donald F. Patterson$, and Bellur Seetharamll From the *Section of Medical Genetics, Veterinary School of the University of Pennsyluania, Philadelphia, Pennsylvania 19104- 6010, the VDiuision of Gastroenterology, Medical College of Wisconsin, and the IIVeterans Administration Medical Center, Milwaukee, Wisconsin 53226

Ligand binding activity of intrinsic factor-cobalamin receptor (IFCR) was determined in homogenates and isolated brush-border membranes (BBM) of ileum and kidney from dogs exhibiting simple autosomal reces- sive inheritance of selective cobalamin malabsorption (Fyfe, J. C., Giger, U., Hall, C. A., Jezyk, P. F., Klumpp, S. A., Levine, J. S., and Patterson, D. F. (1991) Pediatr. Res. 29, 24-31). IFCR activity of af- fected dog ileal homogenates was 3-4-fold higher than normal whereas IFCR activity in affected dog kidney homogenates was one-tenth of normal. The recovery of IFCR activity in the BBM of ileum and renal cortex of affected dogs was 30- and 20-fold less than normal, respectively. The dissociation constant (&) for intrin- sic factor-cobalamin was similar in BBM of both tissues and was the same in affected and normal dogs. In the affected dog ileal BBM, activities of alkaline phospha- tase and sucrase-isomaltase and vesicular transport of glucose and Na+-taurocholate were normal. Immuno- blots showed no IFCR cross-reactive material in the ileal or renal BBM of affected dogs. IFCR purified by affinity chromatography from kidney of both normal and affected dogs had an M, = 230,000. However, amino acid analysis revealed that the affected dog IFCR had more lysine than the normal, and protease cleavage of the purified IFCRs revealed different pep- tide maps. Asparagine-linked oligosaccharides of both proteins were sensitive to peptide N-glycosidase F cleavage, but only the affected dog IFCR was endogly- cosidase H sensitive. These results suggest that cobal- amin malabsorption in this canine family is caused by inefficient BBM expression of IFCR due to a mutation of IFCR and its retention in an early biosynthetic com- partment.

The gastrointestinal absorption of dietary cobalamin (vi- tamin BIZ) bound to gastric, and in dogs, pancreatic intrinsic factor (IF)’ (1) is mediated by an ileal receptor (IFCR) located

* This work was supported by National Institutes of Health Grants NIDDK 26638 and RR 02512, by an individual national research service award (to J. C. F.) from the Public Health Service, by the Lucille P. Markey Charitable Trust and the Kleberg Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

~ ~ ~~~

§To whom correspondence should be sent. ‘The abbreviations used are: IF, intrinsic factor; TC-11, trans-

cobalamin-11; IFCR, instrinsic factor-cobalamin receptor; BBM, brush-border membrane(s); PMSF, phenylmethylsulfonyl fluoride; Endo H, endoglycosidase H; PNGase F, peptide N-glycosidase F; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electropho- resis.

in the microvillus pits of the apical brush border of villus tip enterocytes (2). Available evidence suggests that after binding to IFCR, IF-cobalamin is internalized (3, 4) and cobalamin exits the enterocyte bound to the plasma transport protein, transcobalamin-I1 (TC-11) (5).

Reported inherited defects that cause selective cobalamin malabsorption in humans include lack of IF synthesis (6, 7), secretion of structurally abnormal IF (8, 9), TC-I1 deficiency (IO), and selective intestinal cobalamin malabsorption (11- 13). The last is a rare and poorly defined defect of enterocyte cobalamin transcytosis also known as Imerslund-Griisbeck syndrome (McKusick catalogue 261100) (14). This disease is inherited as a simple autosomal recessive trait and is charac- terized by normal production and function of IF and TC-11, normal gross and microscopic ileal morphology, and selective cobalamin malabsorption that is not corrected by oral IF administration (15). In most patients, proteinuria and/or amino aciduria is also present and persists despite correction of cobalamin deficiency (11). Because enterocyte transcytosis is a complex and not well understood process, it is likely that selective intestinal cobalamin malabsorption encompasses several molecular defects that may affect IFCR expression and function or subsequent steps in transfer of cobalamin to

IFCR has been localized to and purified from both ileal mucosa (16) and renal proximal tubular epithelium (17). Evidence of involvement of both of these tissues in Imerslund- Grasbeck syndrome suggests the hypothesis that the disease, in at least some patients, is caused by a defect of IFCR expression or function. Recently we described a family of dogs with the clinical, genetic, and laboratory features of selective intestinal cobalamin malabsorption seen in humans (18, 19). Electron microscopic examination of immunohistochemically treated ileal biopsies of the affected dogs showed IFCR to be present in intracellular structures but not in the apical brush- border membrane (BBM) (19). These earlier studies suggested that a possible cause of defective cobalamin transcytosis in the affected dogs was failure to express IFCR in the apical BBM. The current investigation was undertaken to test this hypothesis. Results of these studies show that both IFCR ligand binding activity and amounts of IFCR-immunoreactive protein are low in the ileal and renal BBM of affected dogs. Furthermore, based on the properties of renal IFCR purified from normal and affected dog kidneys, we propose that inef- ficient transport of IFCR to the apical membrane is due to synthesis of an altered IFCR that does not reach the medial Golgi compartment.

TC-11.

EXPERIMENTAL PROCEDURES

Materials-The following reagents were purchased from commer- cial sources: [57Co]cyanocobalamin (15 wCi/pg), from Amersham

4489

by guest on May 12, 2018

http://ww

w.jbc.org/

Dow

nloaded from

4490 Defective Brush-border Expression of IF-Cobalamin Receptor COT.; '261-protein A (>30 pCi/pg), from ICN Radiochemicals (Imine, CAI; D-[U-'4C]glucose (340 mCi/mmol), from Du Pont-New England Nuclear; endo-P-N-acetylglucosaminidase H (Endo H) of Strepto- mycesplicatus, peptide N-glycosidase F (PNGase F) of Flavobacterium meningosepticum, and N-octyl glucoside, from Boehringer Mann- heim; trypsin and elastase, from Sigma. [22,23-3H]Taurocholate, so- dium salt was a gift from Dr. Alan Hoffman of the University of California at San Diego. Pure rat IF (20) and antiserum to canine ileal IFCR (16) were prepared as described earlier.

Normal dog kidneys were provided by the physiology department of the Medical College of Wisconsin. Tissues of affected dogs and age-matched normal dogs were obtained from the animal model breeding colony of the Veterinary School of the University of Penn- sylvania. Protocols for the collection of such tissues were approved by the Institutional Animal Care and Use Committee of the Univer- sity of Pennsylvania. Affected dogs were treated weekly by parenteral administration of 740 nmol (1 mg) of cyanocobalamin, and all dogs were healthy at the time of death. Dogs were fasted overnight and killed by intravenous pentabarbital injection. The intestine was re- moved, and ileal segments were opened longitudinally and rinsed in ice-cold 10 mM sodium-phosphate buffer, pH 7.4, containing 140 mM NaCl and 1 mM phenylmethylsulfonyl fluoride (PMSF). The mucosa was scraped from the muscularis on iced glass plates. The capsules of kidneys were removed, and renal cortex was separated from the medulla. These tissues were snap frozen in liquid Nz and stored at -70 "C for 10-30 weeks prior to further processing.

Brush-border Isolation, Activity, and Transport Measurements- Brush-border membranes were isolated from thawed ileal mucosa and renal cortex by the CaC1, aggregation method (21). IFCR ligand binding activity in tissue homogenates and isolated BBM was deter- mined using rat IF-[57Co]cobalamin by the rapid microcentrifuge centrifugation method (22) and, in purified IFCR fractions, by the (NH4),S04 precipitation method (16). IFCR activity and dissociation constants ( K d ) were calculated as described previously (23). Alkaline phosphatase activity was determined at pH 9.2 using p-nitrophenyl phosphate as substrate (24). Sucrase-isomaltase activity was deter- mined by the Tris-glucose oxidase-peroxidase method of Dahlqvist (25). Protein was determined by the method of Lowry et al. (26). Ileal apical brush-border vesicular uptake studies using 50 mM ["CJglucose or 10 mM [3H]taurocholate were carried out by the rapid filtration technique described previously (27). The statistical significance of the difference between group means (normal dogs versus affected dogs) was analyzed using Student's t test.

Zmmunoblotting of ZFCR in Ileal and Renal Membranes-Isolated renal cortex BBM were not processed further prior to electrophoresis and immunoblotting. However, ileal BBM IF-cobalamin-binding pro- teins were concentrated to detectable amounts prior to immunoblot- ting by the following method. Ileal BBM isolated from 2.5 g of mucosa was purified further by the MgS04 washing method (28) and was subsequently homogenized in 10 mM Tris-HC1, pH 7.4, containing 140 mM NaC1, 5 mM CaClz, and 1 mM PMSF. The BBM proteins were solubilized with 1% Triton X-100 and passed over IF-cobalamin- Sepharose columns (0.8 X 1 cm) that were then washed and eluted as described below for IFCR purification. The column eluants were dialyzed overnight against distilled water containing 1 mM PMSF and 2 mM benzamidine and were lyophilized. Denaturing polyacryl- amide gel electrophoresis in the presence of SDS (SDS-PAGE) was done as described by Laemmli (29) using 5 or 7.5% acrylamide in the separating gels. Immunoblotting was performed by the method of Burnette (30) with slight modifications. These included the addition of 0.4% bovine serum albumin and 0.1% Triton X-100 to the blocking and incubation buffers, 2% gelatin to the blocking buffer, and the passage of these buffers through 0.45-pm pore polycarbonate mem- brane filters just prior to use. Anti-dog ileal IFCR serum was diluted 1:10,000 in incubation buffer for immunoblotting. Preabsorbed serum was prepared by passage of the diluted antiserum over a column of purified dog kidney IFCR bound to rat IF-cobalamin-Sepharose.

Purification of Renal IFCR from Normal ana' Affected Dogs- Purification of normal and affected dog kidney IFCR was carried out essentially by the method described for the purification of rat kidney IFCR (17). All steps were carried out a t 5 "C. Briefly, 200 g of renal cortex containing 84 nmol (normal dog) or 10 nmol (affected dog) of IF-cobalamin binding activity was homogenized in 800 ml of 10 mM Tris-HCl, pH 7.4, containing 140 mM NaCl, 1 mM PMSF, and 2 mM benzamidine. The homogenate was extracted in 1% Triton X-100. CaClz was added (5 mM final concentration) to the supernatant obtained by centrifugation (100,000 X g, 60 min). The solubilized fraction was then passed over an affinity column of rat IF-cobalamin-

Sepharose at a flow rate of 20-30 ml/h. After extensive washing with 2 liters of 10 mM Tris-HC1, pH 7.4, containing 140 mM NaCl and 0.1% Triton X-100, IFCR was eluted from the column with 10 mM Tris-HC1, pH 5.0, containing 140 mM NaCI, 0.1% Triton X-100, and 5 mM Na,EDTA. The eluted fractions were neutralized immediately to pH 7.4 with 1 M Tris, and 5 mM CaCI,, 1 mM PMSF, and 2 mM benzamidine were added (final concentrations).

Properties and Enzymatic Digestion of Purified ZFCR-Amino acid composition of the purified IFCR proteins was determined by auto- mated chromatography (Beckman 6300 amino acid analyzer, Imine, CA) after acid hydrolysis in the Protein-Nucleic Acid Shared Facility of the Medical College of Wisconsin. For Endo H digestion, purified lFCR (5-15 pl) was evaporated to dryness and denatured at 95 "C for 3 min in 25 pl of 50 mM sodium citrate buffer, pH 5.5, containing 2 mM PMSF and 0.02% SDS (final concentrations). The denatured protein was then incubated overnight at 37 "C with 15 milliunits of Endo H (1 milliunit/pl). For PNGase F digestion, the dried protein was denatured at 95 "C for 3 min in 30 p1 of 100 mM potassium phosphate buffer, pH 7.4, containing 5 mM Na2EDTA, 2 mM PMSF, and 0.05% SDS (final concentrations). N-Octyl glucoside (0.6%) was added to the solution, which was then incubated overnight a t 37 "C with 0.6 unit of PNGase F (0.2 unit/pl).

For trypsin and elastase digestions, purified IFCR fractions were dialyzed overnight against 10 mM Tris-HC1, pH 7.4, to remove pro- tease inhibitors. 20 pg of IFCR protein was incubated in 500 p1 of 10 mM Tris-HC1, pH 7.4, containing 10 mM CaCl, and 5 pg of trypsin (11,000 units/mg) or elastase (90 units/mg) for 8 h at 37 "C. Protease activity was terminated by the addition of 2 mM benzamidine. Control digestions of each protein were carried out similarly but without added protease. Proteins in gels were visualized by silver staining (31).

RESULTS

IFCR, Alkaline Phosphatase, and Sucrase-isomaltase Actiui- ties in Normal and Affected Dogs-The Ca2+-dependent ligand binding activity of IFCR using rat IF-[67Co]cobalamin and activities of other apical brush-border enzymes were studied in homogenates (Table I) and isolated BBM (Table 11) of ileal mucosa and renal cortex. Tissues were derived from six nor- mal and six affected dogs, all of which were 13 weeks old at the time of death. IFCR activity in the ileal homogenates of affected dogs was 3-4 times higher than in normal dogs (Table I). However, in the kidney homogenates, the reverse was true. The homogenates of renal cortex of affected dogs had one- tenth the IFCR activity present in normal dog renal cortex. The noted differences in the receptor activity between tissues and between normal and affected dogs were statistically sig- nificant ( p < 0.001) and were not due to differences in the affinity (&) of IFCR for IF-cobalamin. Normal dogs had about 140-fold more activity in the kidney than in ileal mucosa. Despite the reduction of IFCR activity in the kidney of affected dogs, it was still 3-5-fold more than in their ileal mucosa. Alkaline phosphatase activity in homogenates ofboth tissues and sucrase-isomaltase activity in ileal homogenates were similar in normal and affected dogs. There was no significant difference of total protein between normal and affected dogs in either tissue.

These results suggest that IFCR activity of affected dogs is modulated, albeit differently, in these two tissues. However, for normal cobalamin transcytosis, IFCR must be expressed in the BBM. Thus, IFCR activity was measured in isolated BBM from ileal mucosa and renal cortex of the same affected and normal dogs (Table 11). The specific activity of IFCR in the affected dogs' BBM was 12-fold less in ileum and 180- fold less in renal cortex than in BBM of the corresponding tissue of normal dogs. Again, this difference was highly sig- nificant ( p < 0.001) and was not caused by large differences in K d . The fractional yield of IFCR activity recovered in BBM was similar in the ileum and kidney of affected dogs (0.01 f 0.004 and 0.007 +. 0.002 of total homogenate IFCR activity, respectively). This was 30- and 20-fold less than in the re-

by guest on May 12, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Defective Brush-border Expression of IF-Cobalamin Receptor 4491 TABLE I

IFCR, alkaline phosphatase, and sucrase-isomaltase activities in normal and affected dog ileal and kidney homogenates IFCR

Alkaline IF-[67Co]cobalamin

bound

Sucrase- K d

phosphatase isomaltase Protein

prnollg tissue nM unitslg tissue mg/g tissue Ileum ( n = 6)

Normal 7.7 f 2.1 0.49 f 0.13 6.0 f 2.4 0.87 f 0.37 79 f 9.3 Affected 31 f 7.0" 0.56 f 0.11 8.0 f 3.3 1.4 f 0.50 94 f 14

Normal 1100 f 450 0.40 f 0.09 7.9 f 2.7 140 f 11 Affected 110 f 20" 0.32 f 0.07 5.6 f 1.8 150 f 16

Kidney ( n = 6)

Difference from normal dog tissue is statistically significant, p < 0.001.

TABLE I1 IFCR, alkaline phosphatase, and sucrase-komaltase activities in n o r m 1 and affected dog ileal

and kidney apical brush-border membranes IFCR

IF-["Cojcobalamin bound

Alkaline Sucrase- phosphatase isomaltase

K d

Protein

prnolfmgprotein nM unitslrng protein mg/g tissue

Normal 0.93 f 0.43 0.23 f 0.08 1.0 f 0.1 0.20 f 0.07 5.6 f 1.0 Affected 0.08 f 0.03" 0.36 f 0.13 1.5 f 0.6 0.24 f 0.03 5.3 f 2.5

Normal 18 f 3.4 0.15 f 0.01 0.51 f 0.19 7.2 f 0.5 Affected 0.10 f 0.02" 0.09 f 0.01 0.26 f 0.08 7.3 f 0.9

Ileum ( n = 6)

Kidney ( n = 6)

Difference from normal dog tissue is statistically significant, p < 0.001.

spective tissues of the normal dogs (0.32 f 0.07, ileum; 0.13 & 0.05, kidney). In contrast, alkaline phosphatase, sucrase- isomaltase, and total protein recovered in BBM isolated from both tissues were not significantly different in normal and affected dogs. These data suggest that ileum and kidney of affected dogs share a similar defect of IFCR expression which is selective for the receptor. Moreover, these results also suggest that the previously observed failure of intestinal co- balamin transcytosis in the affected dogs (19) is due to lack of IFCR expression in the BBM rather than a general mem- brane transport defect.

Na+-dependent Transport of D-Glucose and Taurocholate- T o confirm that other transport functions of the BBM were normal in the affected dogs, BBM transport of D-glucose and taurocholate, an ileal specific function, were studied. Na+- dependent transport of ['4C]glucose and [3H]taurocholate was measured in ileal BBM vesicles of normal and affected dogs. Vesicles of age-matched affected and normal dogs (n = 6 in each group) exhibited concentrative uptake energized by an electrochemical gradient of Na+ as evidenced by 4-5-fold overshoots for taurocholate and 6-10-fold overshoots for glu- cose uptake. The vesicle sizes based on equilibrium uptake values for glucose were comparable for normal and affected dogs and ranged from 1 to 2 pl/mg protein. Comparison of initial rates (0.1 min) revealed no difference between normal and affected dogs in the intestinal BBM vesicle uptake of glucose (normal = 183 k 22 pmol/mg protein; affected = 178 f 25 pmol/mg protein) or taurocholate (normal = 223 f 62 pmol/mg protein; affected 238 f 53 pmol/mg protein).

Immunoblots of Ileal and Renal Apical Membranes-To examine whether the reduced IFCR activity in BBM of af- fected dogs was associated with reduced amounts of IFCR protein in these membranes, immunoblotting was carried out. Immunoblots of IF-cobalamin-binding proteins from ileal BBM of affected and normal dogs were probed with polyclonal antiserum raised against purified dog ileal IFCR (Fig. 1). These revealed the absence of a 190-kDa IFCR band from the

FIG. 1. Immunoblot of IF-cobalamin binding proteins from ileal BBM of normal and affected dogs. BBM were isolated from 2.5 g of ileal mucosas of three normal (lanes 1-3) and three affected dogs (lanes 4-6). IF-cobalamin binding proteins were concentrated by affinity chromatography from Triton X-100-solubilized BBM. The recovered proteins were separated by SDS-PAGE and electroblotted to a nitrocellulose membrane. The blot was incubated sequentially with anti-dog ileal IFCR serum diluted 1:10,000 and '"1-protein A and autoradiographed.

BBM of affected dogs (lanes 4-6) which was present in the normal dog BBM (lanes 1-3). It was noted in earlier experi- ments, however, that a 230-kDa IFCR band was present on immunoblots of both normal and affected dog ileal BBM when those preparations were contaminated with intracellular material (not shown). A 148-kDa band that was present in both normal and affected dogs was judged to be nonspecific cross-reactive material because it remained on blots probed with anti-IFCR serum that had been preabsorbed with puri- fied IFCR bound to IF-cobalamin-Sepharose and on blots probed with preimmune serum. Immunoblots of renal cortex BBM proteins (Fig. 2) revealed the absence of a 230-kDa IFCR band from the BBM of affected dogs (lanes 5-7) which was present in normal dog renal BBM (lanes 2-4). These data confirm that IFCR is expressed poorly in the BBM of both ileum and renal cortex of affected dogs.

Properties of IFCR from Normal and Affected Dog Kidney- The low levels of IFCR in the BBM of affected dogs suggested selective cobalamin malabsorption in these animals to be the result of an IFCR defect causing inefficient transport of the receptor to the BBM. To investigate possible differences in

by guest on May 12, 2018

http://ww

w.jbc.org/

Dow

nloaded from

4492 Defective Brush-border Expression of IF-Cobalumin Receptor

A B

FIG. 2. IFCR from renal cortex of normal and affected dogs. A, for immunoblotting, 50 pg of renal cortex BBM protein from three normal (lanes 2-4) and three affected dogs (lanes 5-7) were separated by SDS-PAGE and electroblotted to a nitrocellulose membrane. The blot was incubated sequentially with anti-dog ileal IFCR serum di- luted 1:10,000 and '"I-protein A and autoradiographed. Lane 1 con- tained 1 pg of IFCR purified by affinity chromatography from normal dog kidney, and lane 8 contained 0.3 pg of similarly purified IFCR from affected dog kidney. B, IFCR was purified from normal (lane 1 ) and affected dog (lane 2 ) renal cortex by affinity chromatography. 1 pg of normal and 0.3 pg of affected dog IFCR protein were electro- phoresed on a 5% gel, and protein bands were visualized by silver staining.

TABLE I11 Amino acid composition of canine renal ZFCR Residue Normal

IFCR Affected

IFCR

Asx Thr Ser Glx Pro GlY Ala Val Met Ile

Leu TYr Phe His LY s Arg Trp

Total

mol % 13.4 6.9 9.7

10.0 9.3 9.7 4.6 5.3 0.7 5.3 8.1 3.3 5.6 3.0 1.7 3.5

ND"

97.5%

13.4 6.7 9.0

10.0 8.5 9.1 4.3 5.5 0.7 5.6 8.3 3.4 5.7 3.1 2.9

ND 3.7

100% ND, not determined.

the IFCR from normal and affected dogs, the receptor protein was purified from normal and affected dog kidney. Although the physiologically significant tissue in cobalamin malabsorp- tion is the ileum, insufficient ileal mucosa from affected dogs was available. Kidney was chosen as a source of IFCR because it manifested the BBM expression defect of IFCR in the affected dogs and contained high levels of IFCR activity. Each protein obtained by affinity chromatography migrated as a single band on SDS-PAGE (5% gels) visualized by either immunoblotting (Fig. 2 4 , lanes 1 and 8 ) or silver staining (Fig. 2B, lanes 1 and 2). Affected and normal dog IFCR thus purified from total homogenates of renal cortex comigrated with the renal IFCR detected in normal dog renal BBM, each having an apparent M, of 230,000. However, the two proteins had slightly different amino acid compositions (Table 111). The mol % of lysine in the normal dog kidney IFCR was 1.70% and was 2.93% in the affected dog IFCR. Real differ- ences of serine, proline, and glycine may also exist, but these data differed by less than the 10% interassay variability of the amino acid composition determination.

Both purified kidney IFCR proteins exhibited single bands on SDS-PAGE when initially prepared. These were stored for several months at 5 "C in the neutralized column elution buffer containing protease inhibitors. After 2-3 weeks a 222-

IFCR (Fig. 3, lane 2 and Fig. 4, lane 3) . This did not occur in the stored normal dog IFCR even after 6 months, suggesting that the affected dog IFCR was intrinsically less stable than normal.

When the purified kidney IFCR proteins were subjected to Endo H digestion (Fig. 3), the gel mobility of normal dog IFCR did not shift detectably (lane 3 ) . In contrast, the mo- bility of the 230-kDa band of the affected dog IFCR shifted by about 5 kDa, and the 222-kDa band shifted by about 14 kDa (lane 4 ) . A minor portion of the affected dog IFCR 230- kDa band was resistant to Endo H digestion. After PNGase F digestion, the gel mobilities of the normal dog IFCR (lane 5 ) and of the 230-kDa band of affected dog IFCR (lane 6 ) were both shifted by about 3 kDa. The 222-kDa band of the affected dog IFCR was shifted by about 10 kDa. These data indicate that IFCR purified from normal dog kidney has asparagine-linked complex or hybrid-type oligosaccharides whereas the asparagine-linked oligosaccharides of affected dog kidney IFCR are predominantly high mannose type. Although estimating the mobility shifts or apparent size of proteins >200 kDa by SDS-PAGE is inaccurate at best the shift of the Endo H-digested protein bands of the affected dog IFCR suggests that dog renal IFCR has three to seven asparagine-linked glycosylation sites, each high mannose oligosaccharide representing about 1.7 kDa.

Partial peptide mapping of the purified IFCR proteins by trypsin and elastase treatment revealed different patterns for the normal and affected renal IFCR (Fig. 4). When receptor fractions were subjected to tryptic cleavage, normal dog IFCR produced four major fragments (210, 183, 155, and 107 kDa; lane 4 ) . Affected dog IFCR produced six major fragments (210, 178, 147, 107, 72, and 60 kDa; lane 5 ) . Tryptic digestion resulted in the complete disappearance of the 230-kDa IFCR of the normal dogs and both the 230- and 222-kDa bands of

kDa

i 1 :: & 225 - 222

200- s

FIG. 3. Endo H and PNGase F digestion of purified renal IFCR from normal and affected dogs. 1.5 pg of affinity-purified renal IFCR from normal (lanes 1,3, and 5) and affected dogs (lanes 2,4, and 6 ) were digested exhaustively with Endo H (lanes 3 and 4 ) or PNGase F (lanes 5 and 6). The proteins in lanes 1 and 2 were incubated as for PNGase F digestion but without enzyme. The di- gested proteins were subjected to SDS-PAGE on a 5% gel and visualized by silver staining.

kDa

200-

116-

97-

66-

43-

FIG. 4. Protease digestion of purified renal IFCR from nor- mal and affected dogs. 20 pg of purified renal IFCR from normal (lanes 2,4, and 6 ) and affected dogs (lanes 3 ,5 , and 7) were incubated for 8 h at 37 "C with no enzyme (lanes 2 and 3 ) , 5 pg of trypsin (lanes 4 and 5), or 5 pg of elastase (lanes 6 and 7). Peptide fragments in aliquots containing 1.5 pg of protein were separated by SDS-PAGE on a 7.5% gel and were visualized by silver staining. Lune 1 contained

kDa band appeared in fractions of the affected dog kidney molecular mass standards.

by guest on May 12, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Defective Brush-border Expression of IF-Cobalamin Receptor 4493

the affected dog IFCR. The major fragment resulting from tryptic cleavage of the normal dog IFCR had an Mr of 210,000 whereas that from affected dog IFCR had an M, of 107,000. The apparent size of the 178- and the 147-kDa fragments of the trypsin-treated affected dog IFCR differed from the cor- responding fragments of the normal dog IFCR by about 5 and 8 kDa, respectively. The 72- and the 60-kDa fragments of the affected dog IFCR were not detected in the tryptic digestion of the normal dog IFCR. When subjected to elastase cleavage, the normal dog IFCR was not degraded completely, but frag- ments varying from 165 to 222 kDa were formed (lane 6). Elastase treatment of the affected dog IFCR degraded both bands completely and produced eight identifiable fragments varying from 74 to 210 kDa ( l a n e 7).

DISCUSSION

Ileal enterocyte transcytosis of IF-cobalamin is an efficient process (32), the details of which are not understood com- pletely. Nevertheless, the initial events involve binding of IF- cobalamin to IFCR that is expressed in the microvillus pits of apical BBM (2). After the binding of IF-cobalamin, the complex is internalized, and cobalamin is transferred slowly to TC-I1 (3). Recently (18, 19) we described a family of dogs that exhibit classical signs of cobalamin deficiency early in life including growth failure, low serum cobalamin concentra- tions, methylmalonic aciduria, homocystinemia, and megalo- blastic dyshematopoeisis. These signs were completely revers- ible by parenteral administration of physiological doses of cobalamin (3.7 nmol/day). Orally administered radiolabeled cobalamin was not absorbed by affected dogs although all other gastrointestinal functions and histologic morphology were normal. Breeding experiments showed that selective cobalamin malabsorption in this canine family was inherited as a simple autosomal recessive trait.

As shown previously (19), IF and TC-I1 were normal in the affected dogs. However, immunoelectron microscopy of ileal biopsies from affected dogs showed that immunocross- reactive IFCR, although present in many intracellular tubu- lovesicular structures, was absent from the microvillus pits and terminal web area of the apical BBM. These observations suggested that cobalamin deficiency in these dogs was caused by a defect involving targeting of IFCR which, in turn, inter- rupted cobalamin transcytosis by ileal enterocytes. The cur- rent investigation was undertaken to test this hypothesis.

The results presented here suggest that the disease in these dogs is caused by failure to express IFCR in the ileal BBM. This conclusion is based on observations that IFCR ligand binding activity in isolated ileal BBM was low (Table 11) despite increased total IFCR activity in ileal homogenates (Table I) and that a 190-kDa immunocross-reactive IF-cobal- amin binding protein was absent from the ileal BBM of affected dogs (Fig. 1). These results are consistent with the immunohistochemical findings reported previously (19). When IFCR was examined in renal BBM of affected dogs, results similar to those of affected dog ileum were obtained. BBM recovery of IFCR activity was low, and immunocross- reactive IFCR protein was absent (Fig. 2). The defect in both tissues appeared specific to IFCR because recoveries of BBM alkaline phosphatase and sucrase-isomaltase activities as well as vesicular transport of glucose and taurocholate were normal in the affected dogs.

On immunoblots of normal dog BBM, the apparent size of the ileal BBM IFCR was about 40 kDa less than that of the renal BBM IFCR. This difference is very similar to the size difference noted for IFCR in ileal and renal BBM in rats (17). This suggests that in both species ileal IFCR may be synthe-

sized as a 230-kDa precursor that undergoes extracellular proteolytic modification following insertion into the apical membrane and exposure to serine proteases of the intestinal lumen. Consistent with this idea was the observation of a 230- kDa IFCR band on immunoblots of affected dog ileal BBM contaminated with intracellular material. Another line of evidence supports this hypothesis: the mature IFCR immu- noprecipitated from the apical BBM of [35S]methionine-la- beled human colon adenocarcinoma (Caco-2) cells in culture is 230 kDa (33). Proteolytic modification of renal IFCR would not occur after insertion into the renal proximal tubular BBM because appropriate proteases are lacking in the tubular lu- men. Further studies are needed to validate this hypothesis, but membrane protein modification by pancreatic proteases in the intestinal lumen is known to occur to other intestinal BBM proteins such as sucrase-isomaltase (34) and aminopep- tidase (35).

Finding that both ileum and kidney of the affected dogs had similar abnormalities suggests that these defects are the result of a single mutation of a gene that is expressed in both tissues. This is supported by the fact that all of the affected dogs studied were members of the same family in which this rare disease is inherited via an apparently simple autosomal recessive allele (19). Hence, it is improbable that two muta- tions of similar but distinct genes would occur together in the affected lineage. It is also improbable that the primary defect is in a gene that modifies the expression of IFCR in various tissues because the results obtained in studies of purified renal IFCR indicate that there is a subtle alteration of the affected dog IFCR protein. Although the apparent M, of renal IFCR was the same when isolated from normal or affected dogs, the amino acid compositions, stability during storage, and peptide maps of the two proteins were different. Most likely this is due to a mutation of the coding region of the structural IFCR gene.

Normal dog IFCR purified from total renal cortex homog- enates was resistant to Endo H but sensitive to PNGase F digestion. The same results have been obtained with purified rat renal IFCR' (17). This suggests that normal maturation and targeting of the receptor to the BBM are associated with the conversion of high mannose asparagine-linked oligosac- charides to either complex or hybrid type, a processing event that is a function of the Golgi apparatus (36). It further suggests that in both species, at steady state, most of the renal IFCR is in the mature, Endo H-resistant form. In contrast, the affected dog renal IFCR was largely Endo H sensitive, indicating that it retained high mannose-type oligosaccha- rides. A minor portion of the affected dog renal IFCR was Endo H resistant and PNGase F sensitive, suggesting that, at steady state, a small fraction of IFCR had reached the medial Golgi compartment. Clearly, however, only a small amount of IFCR activity and no immunocross-reactive material were detectable at the BBM. An interpretation of these results is that in the affected dogs, IFCR does not move efficiently through the endoplasmic reticulum and Golgi compartments during vectorial transport to the BBM. Retention of a protein in an early biosynthetic compartment has been noted in several other inherited diseases. These include certain var- iants of intestinal sucrase-isomaltase deficiency (37), low density lipoprotein receptor deficiency of humans and Watan- abe heritable hyperlipidemic rabbits (38), al-antitrypsin de- ficiency (39), P-hexosaminidase deficiency (40), and others.

An emerging paradigm is that certain mutations produce an altered protein that folds improperly co- or post-transla-

'K. S. Ramanujam, J. C . Fyfe, and B. Seetharam, unpublished results.

by guest on May 12, 2018

http://ww

w.jbc.org/

Dow

nloaded from

4494 Defective Brush-border Expres

tionally in the endoplasmic reticulum and that further move- ment and processing of such a protein are impeded (41). The renal IFCR of affected dogs was more sensitive to in vitro proteolysis, and treatment with trypsin and elastase resulted in formation of smaller fragments (<lo0 kDa) than those obtained by similar treatment of the normal dog renal IFCR. This may have been the result of improper folding of the affected dog protein, making more sites available to proteo- lytic attack. Alternatively, the affected dog protein may have new cleavage sites in an altered amino acid sequence. Con- ceivably, either change could be sufficient to alter the progress of the affected dog IFCR through the endoplasmic reticulum or cis-Golgi, resulting in turn in the observed lack of oligosac- charide processing. Further studies are currently under way to establish the site of retention of IFCR and the nature of the mutation that causes this rare canine disease. Further work is also needed to ascertain whether a similar type of IFCR defect causes Imerslund-Griisbeck syndrome in some patients.

Acknowledgments-We are indebted to Dr. Nancy Dahms, Medical College of Wisconsin, Milwaukee, for many helpful discussions and critical review of this manucript.

REFERENCES

1. Batt, R. M., and Horadgoda, N. U. (1989) Am. J. Physiol. 257,

2. Levine, J. S., Allen, R. H., Alpers, D. H., and Seetharam, B.

3. Kapadia, C. R., Serfilippi, D., Voloshin, K., and Donaldson, R.

4. Seetharam, B., Presti, M., Frank, B., Tiruppathi, C., and Alpers,

5. Rothenberg, S. P., Weiss, J. P., and Cotter, R. (1978) Br. J.

6. Lillibridge, C. B., Brandborg, L. L., and Rubin, C. E. (1967)

7. Levine, J. S., and Allen, R. H. (1985) Gastroenterology 88, 1132-

8. Katz, M., Mehlman, C. S., and Allen, R. H. (1974) J. Clin. Inuest.

9. Yang, Y., Ducos, R., Rosenberg, A. J., Catrou, P. G., Levine, J. S., Podell, E. R., and Allen, R. H. (1985) J. Clin. Inuest. 76,

10. Hakami, N., Neiman, P. E., Canellos, G. P., and Lazerson, J.

11. Imerslund, 0. (1960) Acta Puediatr. Scand. 49, (Suppl. 119) 1-

12. Grasbeck, R., Gordin, R., Kantero, I., and Kuhlback, €3. (1960)

G344-G349

(1984) J. Cell Biol. 98, 1111-1118

M., Jr. (1983) J. Clin. Inuest. 71, 440-448

D. H. (1985) Am. J. Physiol. 248, G326-G331

Haematol. 40, 401-414

Gastroenterology 52, 792-809

1136

53,1274-1283

2057-2065

(1971) N. Engl. J. Med. 285, 1163-1170

115

Acta Med. Scand. 167,289-296

'sion of IF-Cobalamin Receptor

13. Ben-Bassat, I., Feinstein, A., and Ramot, B. (1969) Israel J. Med. Sci. 5,62-68

14. McKusick, V. A. (1990) Mendelian Inheritance in Man, 9th ed, pp. 1411-1412, Johns Hopkins University Press, Baltimore

15. MacKenzie, I. L., Donaldson, R. M., Trier, J. S., and Mathan, V. I. (1972) N . Engl. J. Med. 286, 1021-1025

16. Seetharam, B., Alpers, D. H., and Allen, R. H. (1981) J. Biol. Chem. 256,3785-3790

17. Seetharam, B., Levine, J. S., Ramasamy, M., and Alpers, D. H. (1988) J. Biol. Chem. 263,4443-4449

18. Fyfe, J. C., Jezyk, P. F., Giger, U., and Patterson, D. F. (1989) J. Am. Anim. Hosp. Assoc. 25, 533-539

19. Fyfe, J. C., Giger, U., Hall, C. A., Jezyk, P. F., Klumpp, S. A., Levine, J. S., and Patterson, D. F. (1991) Pediutr. Res. 29, 24- 31

20. Seetharam, B., Bakke, J. E., and Alpers, D. H. (1983) Biochem. Biophys. Res. Commun. 115,238-244

21. Schmitz, J., Preiser, H., Maestracci, D., Ghosh, B. K., Cerda, J. J., and Crane, R. K. (1973) Biochem. Biophys. Acta 323, 98- 112

22. Hagedorn, C. H., and Alpers, D. H. (1977) Gastroenterology 73, 1019-1022

23. Hooper, D. C., Alpers, D. H., Burger, R. L., Mehlman, C. S., and

24. Garen, A., and Levinthal, C. (1960) Biochim. Biophys. Acta 38,

25. Dahlqvist, A. (1964) Anal. Biochem 7, 18-25 26. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J.

(1951) J. Biol. Chem. 193, 265-275 27. Harig, J. M., Barry, J. A., Rajendran, V. M., Soergel, K. H., and

Ramaswamy, K. (1989) Am. J. Physiol. 256, G618-G623 28. Hopfer, U., Nelson, K., Perrotto, J., and Isselbacher, K. J. (1973)

J. Biol. Chem. 248, 25-32 29. Laemmli, U. K. (1970) Nature 227, 680-685 30. Burnette, W. N. (1981) Anal. Biochem. 112, 195-203 31. Merrill, C. R., Goldman, D., Sedman, S. A., and Ebert, M. H.

32. Seetharam, B., and Alpers, D. H. (1985) Nutr. Reu. 43,97-102 33. Ramanujam, K. S., Seetharam, S., Ramasamy, M., and See-

34. Hauri, H. P., Quaroni, A,, and Isselbacher, K. J. (1979) Proc.

35. Sjostrom, H., Norh , O., Jeppesen, L., Staun, M., Svensson, B.,

36. Kornfeld, R., and Kornfeld, S. (1985) Annu. Rev. Biochem. 54,

37. Hauri, H. P., Roth, J., Sterchi, E. E., and Lentze, M. J. (1985) Proc. Natl, Acad. Sci. U. S. A. 8 2 , 4423-4427

38. Schneider, W. J., Brown, M. S., and Goldstein, J. L. (1983) Mol. Biol. Med. 1, 353-367

39. Sifers, R. N., Brashears-Macatee, S., Kidd, V. J., Muensch, H., and Woo, S. L. C. (1988) J. Biol. Chem. 263, 7330-7335

40. Zokaeem, G., Bayleran, J., Kaplan, P., Hechtman, P., and Neu- feld, E. F. (1987) Am. J . Hum. Genet. 40, 537-547

41. Lodish, H. F. (1988) J. Biol. Chem. 263,2107-2110

Allen, R. H. (1973) J. Clin. Inuest. 52, 3074-3083

470-483

(1981) Science 211, 1437-1438

tharam, B. (1991) Am. J. Physiol., in press

Natl. Acad. Sci. U. S. A. 76, 5183-5186

and Christiansen, L. (1978) Eur. J. Biochem. 88, 503-511

631-664

by guest on May 12, 2018

http://ww

w.jbc.org/

Dow

nloaded from

J C Fyfe, K S Ramanujam, K Ramaswamy, D F Patterson and B Seetharaminherited intestinal cobalamin malabsorption.

Defective brush-border expression of intrinsic factor-cobalamin receptor in canine

1991, 266:4489-4494.J. Biol. Chem. 

  http://www.jbc.org/content/266/7/4489Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/266/7/4489.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on May 12, 2018

http://ww

w.jbc.org/

Dow

nloaded from