Escherichia coli enterotoxin receptors:
`localization in opossum kidney, intestine, and testis
`
`LEONARD R. FORTE, WILLIAM J. KRAUSE, AND RONALD H. FREEMAN
`Departments of Pharmacology, Anatomy, and Physiology, School of Medicine, University of Missouri,
`and Harry S. Truman Memorial Veterans Administration Medical Center, Columbia, Missouri 65212
`
`FORTE, LEONARD R., WILLIAM J. KRAUSE, AND RONALD H.
`FREEMAN. Escherichia coli enterotoxin receptors: localization in
`opossum kidney, intestine, and testis. Am. J. Physiol. 257 (Renal
`Fluid Electrolyte Physiol. 26): F874-F881, 1989.—The distri-
`bution of receptors for Escherichia coli enterotoxin were ex-
`amined in opossum kidney, intestine, and testis. E. coli enter-
`otoxin stimulated guanosine 3',5'-cyclic monophosphate
`(cGMP) production in renal cortex, testis, and small intestinal
`mucosa but had only a small effect in the colon. Atrial natri-
`uretic factor enhanced the cGMP content of renal cortex and
`small intestine but had no effect on testis or colon. The enter-
`otoxin receptors were observed to be localized in proximal
`tubules, to epithelial cells of crypts and villi of small intestine,
`to crypts of colon, and in seminiferous tubules. Both convoluted
`and straight portions of proximal tubules exhibited specific
`binding sites for 1"I-labeled enterotoxin. Glomeruli and distal
`tubules did not have receptors. Binding of '25I-enterotoxin to
`brush-border membranes of kidney cortex or intestinal mucosa
`and to testis membranes was markedly temperature dependent.
`The binding affinities of these receptors for E. coli enterotoxin
`were similar (i.e., IC50 = 0.4-0.5 nM). Daily administration of
`20 µg of enterotoxin intramuscularly to opossums increased
`urine cGMP excretion with no apparent changes in urine
`volume, Na+, or 1{± excretion. Thus receptors for heat-stable
`enterotoxins are localized to proximal tubules of kidney and to
`enterocytes and seminiferous tubules of intestine and testis,
`respectively. Apical membranes may be the site of enterotoxin
`receptors in these epithelia.
`
`guanosine 3',5'-cyclic monophosphate; receptor autoradiogra-
`phy; atrial natriuretic factor; brush-border membranes
`
`HEAT-STABLE PEPTIDES belonging to a class of diarrheal
`enterotoxins are produced by Escherichia coli, Yersinia
`enterocolitica, and other pathogenic enteric bacteria (13).
`Specific, high-affinity binding sites for these peptides are
`found on the apical membrane of intestinal epithelial
`cells (8, 9, 12, 14, 17, 23). These enterotoxins activate a
`membrane-bound form of guanylate cyclase, which leads
`to an increase in guanosine 3',5'-cyclic monophosphate
`(cGMP) content of enterocytes (8-10, 18-20, 28, 29).
`Analogues of cGMP cause changes in the transport of
`solute and water in the intestine similar to the effects of
`bacterial enterotoxins (10, 16, 20, 29). Thus the secretory
`diarrhea caused by this class of heat-stable (ST) enter-
`otoxins has been postulated to be mediated by the intra-
`cellular second messenger, cGMP (10). This cellular
`mechanism may be analogous to the form of transmem-
`brane signaling observed for hormones and neurotrans-
`
`secretion in the intestine and
`mitters, which promote
`in cultured intestinal cell lines via activation of adenylate
`cyclase (3, 6, 7, 26, 30, 32). E. coli enterotoxin also
`secretion in cultured T-84
`stimulated transepithelial
`cells, a human colon carcinoma cell line having receptors
`for the enterotoxin linked positively to the activation of
`guanylate cyclase (18, 21).
`The biological actions of ST enterotoxins were consid-
`ered to be restricted to the enterocytes of small or large
`intestine which expressed apical membrane receptors for
`these peptides (8, 19, 28). However, we reported recently
`that opossum kidney (OK), as well as cultured kidney
`cell lines (PtK-2) had specific, high-affinity binding sites
`for 'I -labeled enterotoxin (11). Moreover, E. coli enter-
`otoxin elicited large increases in kidney or intestinal
`cGMP production in vitro. Intravenous injection of E.
`coli enterotoxin caused 10- to 50-fold increases in urinary
`cGMP excretion in this species. Thus renal receptors for
`the E. coli enterotoxin and an associated guanylate cy-
`clase were present in the kidney and these receptors were
`activated by systemic administration of the enterotoxin.
`Therefore, ST enterotoxin receptors are more widely
`distributed in epithelial tissues than was previously con-
`sidered. In the present study we report that specific
`receptors for E. coli enterotoxin, which are positively
`coupled to guanylate cyclase, appear to be localized to
`the proximal tubule in the renal cortex of the North
`American opossum (Didelphis virginiana). Enterotoxin
`receptors were also localized to the enterocytes of both
`crypts and villi of small intestine, crypts of large intes-
`tine, and to seminiferous tubules of testis. Brush-border
`membranes prepared from renal cortex or small intes-
`tinal mucosa had high-affinity binding sites for 125I-
`enterotoxin.
`
`EXPERIMENTAL PROCEDURE
`
`Animals. Opossums were trapped locally using Hava-
`hart traps (Tomahawk Live Trap, Tomahawk, WI) under
`a permit from the Missouri Department of Conservation
`issued to W. J. Krause. Animals of both sexes were
`housed in the Laboratory Animal Medicine Facility of
`the School of Medicine. They were fed Purina dog chow
`(Ralston Purina, St. Louis, MO) and provided with water
`ad libitum. The animals appeared to be in good health
`when used in the experiments. Animals were killed by
`an overdose of ketamine administered intracardially fol-
`lowed by exsanguination.
`
`MYLAN EXHIBIT - 1034
`Mylan Pharmaceuticals, Inc. v. Bausch Health Ireland, Ltd. - IPR2022-00722
`
`

`

`E. COLI ENTEROTOXIN RECEPTORS
`
`F875
`
`Tissue cGMP measurements. Opossum tissues were
`dissected free and placed into ice-cold 0.9% NaCl. Tissue
`slices were prepared with a Stadie-Riggs microtome and
`placed
`into Dulbecco's modified Eagle's medium
`(DMEM) containing 20 mM N-2-hydroxyethylpipera-
`zine-N' -2-ethanesulfonic acid (HEPES) buffer, pH 7.4
`at 4°C until incubated with agonists (11). Tissue slices
`or intestinal mucosa (usually 125-150 mg wet wt) were
`incubated for 40 min at 37°C with E. coli enterotoxin-
`STa (Sigma Chemical, St. Louis, MO), atrial natriuretic
`factor (ANF, Peninsula Laboratories, Belmont, CA), or
`vehicle that was added to DMEM-HEPES (0.2 ml) plus
`1 mM methylisobutylxanthine (MIX). Perchloric acid
`was then added to a final concentration of 3.3%, slices
`homogenized, centrifuged, and the supernatants neutral-
`ized with 10 N KOH. After centrifugation the superna-
`tant solution was used to measure cGMP by radio-
`immunoassay (11). Mucosal cGMP content in small in-
`testine and colon was expressed per milligram protein
`(25).
`Iodination of the E. coli enterotoxin. Iodination of E.
`coli enterotoxin with 1251 was accomplished by means of
`lactoperoxidase as described by the manufacturer (En-
`zymobeads, Bio-Rad Laboratories, Richmond, CA). 125I-
`enterotoxin was purified by use of reverse-phase high-
`performance liquid chromatography (HPLC) with a C18
`column (3.9 mm x 30 cm µBondapak, Waters, Milford,
`MA). The peak of radioligand used in these experiments
`was that which corresponded to peak B of the HPLC
`purification method described by Thompson et al. (31).
`No detectable free 125I was observed in this peak from
`the linear gradient of acetonitrile. We routinely used the
`fraction of 1-25I-enterotoxin that provided the least
`amount of nonspecific binding in relation to total binding
`of the radioligand by intestinal brush-border membranes.
`Preparation of membranes from kidney, intestine, and
`testis. Testes from each opossum were homogenized for
`35 s in 3 ml/g tissue of a solution containing 0.25 M
`sucrose, 0.01 M
`tris(hydroxymethyl)aminomethane
`(Tris) • HC1, pH 7.4, 1 mM EDTA by means of a Tissum-
`izer (Tekmar, Cincinnati, OH). The homogenate was
`centrifuged at 43,500 g for 15 min and the pellet was
`resuspended in the same buffer. Membranes were stored
`frozen at —70°C. Brush-border membranes from opos-
`sum kidney cortex and small intestinal mucosa were
`prepared with the use of the magnesium precipitation
`technique described by Biber et al. (2). Kidney and
`intestinal brush-border membranes were stored frozen
`at —70°C in a solution containing 20% glycerol, 300 mM
`mannitol, 1 mM MgC12, 50 mM HEPES-Tris, pH 7.5.
`Radioligand binding experiments. Kidney or intestinal
`brush-border membranes were thawed, sedimented by
`centrifugation at 30,900 g for 30 mM, and then resus-
`pended in a buffer containing 50 mM Tris • HC1, pH 7.5,
`0.1 mM EDTA, 150 mM NaCl. Kidney (20 µg), intestine
`(10 µg), or testis (150 µg) membranes were incubated in
`this buffer containing either 30,000 counts per minute
`(cpm; for kidney and intestine membranes) or 100,000
`cpm (for testis membranes) of 125I-enterotoxin in the
`presence of a range of concentrations of unlabeled enter-
`otoxin from 10 pM to 1µM. Incubation times and tem-
`
`peratures are given in the appropriate figure legends. At
`the end of the incubation, 2 ml of ice-cold calcium,
`magnesium-free phosphate-buffered saline (PBS) was
`added to each tube. This suspension was filtered by 25
`mm diameter glass microfiber filters (GF/F, Whatman
`International, Maidstone, England). The filters had been
`pretreated with 0.1% polyethylenimine (Sigma Chemi-
`cal) before use. The filters were washed three times with
`4 ml each of cold wash solution. Radioactivity was meas-
`ured by gamma scintillation spectrometry.
`In vivo autoradiography of ' 25I-enterotoxin distribution
`in kidney. A 150-g male opossum was anesthetized with
`ketamine and then injected with 30 x 106 cpm of 1251-ST
`intravenously. Ten minutes later, a time when peak urine
`cGMP responses to E. coli ST were previously observed
`(9), the animal was exsanguinated and perfused via the
`ventricle with phosphate-buffered saline (PBS) until all
`blood was removed from the animal. Kidneys were then
`frozen in liquid N2. Frozen sections (14 µm) were cut at
`—20°C and thaw-mounted on cover slips. After air-drying
`the sections were placed on X-ray film and exposed for
`3 mo.
`In vitro autoradiography of 1251-enterotoxin binding to
`kidney, small intestine, colon, and testis. Opossum tissues
`were frozen in N2, and 14-µm sections were cut in a
`cryostat maintained at —20°C. Sections were thaw-
`mounted onto gelatin-coated slides and air-dried.
`Mounted sections were stored at —70°C until used. Each
`section was incubated with 50 µI of DMEM-2-(N-mor-
`pholino)ethanesulfonic acid (MES), pH 5.5, containing
`0.5% bovine serum albumin (BSA) for 15 min at 37°C.
`Then 50 µl of DMEM-MES containing 500-2,000 cpm/
`µl of 19-enterotoxin ± 1µM unlabeled enterotoxin and
`0.5% BSA were added to measure total and nonspecific
`binding of this peptide to the tissue sections. The sec-
`tions were incubated for 15 min at 37°C, washed with a
`gentle stream of ice-cold PBS, and then washed 3 times
`in 50 ml of ice-cold PBS for 5 mM at each wash. Sections
`were air-dried and placed into contact with Kodak X-
`Omat AR X-ray film (Eastman Kodak, Rochester, NY)
`for detection of radioactivity. The slides were then coated
`with Kodak NTB-2 emulsion, sealed in light-tight boxes,
`and stored at 4°C for 3-5 wk until developed.
`Chronic treatment of opossums with E. coli enterotoxin.
`Five opossums raised in our colony (7 mo of age, 1.5-2.0
`kg body wt) were placed into metabolism cages for at
`least 1 wk to allow the animals to become accustomed to
`the new housing conditions. Then 24-h urine samples
`were collected for 7-10 days comprising a control period.
`Each animal received 20 µg of E. coli enterotoxin by
`intramuscular injection per day for three consecutive
`days. Urine was also collected for several days after
`treatment was discontinued. Urine cGMP was measured
`by radioimmunoassay and Na+ and IC+ by flame photom-
`etry.
`Statistical analysis. The data were analyzed for differ-
`ences in tissue cGMP content between control and en-
`terotoxin- or ANF-stimulated values by Student's paired
`t test.
`Materials. Tissue culture medium (DMEM) was pur-
`chased from GIBCO Laboratories (Grand Island, NY).
`
`

`

`F876
`
`E. COLT ENTEROTOXIN
`
`RECEPTORS
`
`Na "'I was supplied by Amersham, 14-17 peCi/µg (Ar-
`lington Heights, IL). Other chemicals and reagents were
`purchased from various suppliers.
`
`RESULTS
`Opossum tissues were examined for the presence of E.
`coli enterotoxin-stimulated guanylate cyclase activity by
`measuring cGMP levels of either tissue slices or intes-
`tinal mucosa suspensions exposed to vehicle, entero-
`
`TABLE 1. Effect of E. coli enterotoxin and ANF
`on cGMP levels of opossum kidney, testis,
`and intestine in vitro
`
`cGMP, fmolimg
`
`Tissue
`
`Basal
`
`Enterotoxin
`
`ANF
`56+9*
`Kidney cortex (11)
`446±128*
`34±11
`17±4
`Testis (6)
`30±9
`79±20*
`398±77
`Small intestine (4)
`3,842±1,177*
`887±213*
`Colon (5)
`2,183±1,657
`1,764±914 2,436±1,654
`Values are means ± SE; number of experiments in parentheses. E.
`colt enterotoxin concentration; kidney and small intestine 0.5 µM;
`other tissues 1 µM; atrial natriuretic factor (ANF) concentration was
`0.1 AM. Guanosine 3',5'-cyclic monophosphate (cGMP) content is
`expressed as fmol/mg wet wt for kidney and testis, and fmol/mg protein
`for small intestine and colon mucosa. * P < 0.05 compared with basal,
`Student's paired t test.
`
`FIG. 1. Localization of 19-enterotoxin in kidney after in vivo ad-
`ministration of peptide to an opossum. See EXPERIMENTAL PROCEDURE
`for details of experiment.
`
`toxin, or ANF. Under these conditions E. coli enterotoxin
`elicited a 4.6-fold increase in cGMP content of testis
`(Table 1). Colon did not respond to the enterotoxin with
`a significant increase in cGMP, although cGMP levels
`were generally higher in tissues exposed to that peptide.
`The effect of E. coli enterotoxin on renal cortical and
`small intestinal mucosa cGMP content (i.e., 13-fold and
`10-fold, respectively) was substantially greater than the
`cGMP response of testis. ANF increased the cGMP
`content of kidney cortex and intestine, but this peptide
`had no apparent effect on cGMP content of testis or
`colon under these conditions.
`In vivo labeling of opossum kidney enterotoxin recep-
`tors was accomplished by injecting the ' 2'1-labeled pep-
`tide into an opossum. The distribution of radioactivity
`was highest in renal cortex, much lower in medulla, and
`intermediate in the renal papilla (Fig. 1). When kidney
`sections were labeled in vitro with 125I-enterotoxin, we
`found that specific binding of this radioligand occurred
`only in the renal cortex (Fig. 2). Radioactivity associated
`with the medulla or papilla was equivalent to nonspecific
`binding of the radioligand (denoted by arrows in Fig. 2).
`In contrast, the testis had a lower apparent binding of
`125I-enterotoxin to receptors, which appeared to be dis-
`tributed evenly throughout the tissue. Small intestine
`and colon were both labeled by 'I-enterotoxin over the
`mucosa. We consistently found that brain (cerebral cor-
`tex) and skeletal muscle had no specific binding of 125I-
`enterotoxin (data not shown).
`Examination of the ' 25I-enterotoxin labeled sections of
`OK by light microscopy revealed that enterotoxin recep-
`tors were localized to the proximal tubule (Fig. 3). The
`pars recta appeared to have a somewhat greater level of
`radioactivity than did the convoluted portion of the
`proximal tubule. Glomeruli and distal tubules did not
`have receptors for this peptide. Also, it appears unlikely
`that enterotoxin receptors occur in either the loop of
`Henle or collecting tubules because there was no specific
`binding of "I-enterotoxin in the inner medulla or pa-
`pilla. Enterotoxin receptors in small and large intestine
`were restricted to the enterocytes of both crypts and villi.
`No receptors were observed in association with submu-
`cosal connective tissue or smooth muscle. The 125I-enter-
`
`FIG. 2. Specific binding of ' 26I-enter-
`°toxin to kidney cortex (A), small intes-
`tine (s), and colonic mucosa (C) and
`testes (D) revealed by in vitro autoradi-
`ography. Data are representative exper-
`iments of at least 6 different experiments
`with frozen sections from these tissues
`(see EXPERIMENTAL PROCEDURES for
`details). Exposure of X-ray film varied
`from 4 to 16 h and data are intended to
`demonstrate tissue localization and not
`receptor density. Sections of brain (cor-
`tex) and skeletal muscle have consist-
`ently shown no specific binding of 'aI-
`enterotoxin (data not shown).
`
`

`

`E. CM ENTEROTOXIN RECEPTORS
`
`F877
`
`5.
`
`4
`
`I
`
`alitt=" 4.
`
`It;
`
`is, to
`
`,
`
`•
`
`• %V'
`v.% fi5 11 s.
`
`ill. X
`
`C
`
`FIG. 3. A: radioautograph of a nonstained section from proximal small intestine labeled with 'I-enterotoxin
`demonstrates receptor localization on intestinal epithelial cells covering villi (v) and lining intestinal glands (arrows).
`Lumen (L) of intestine is to extreme right. Muscularis externa (E) and connective tissue elements forming submucosa
`and lamina propria are unlabeled. Magnification x20. B: portion of colon labeled with "I-enterotoxin and viewed
`with dark-field microscopy shows intense labeling of intestinal lining epithelium. Magnification x100. C: radioauto-
`graph of a nonstained section taken through full thickness of renal cortex and an adjacent portion of medulla shows
`"I-enterotoxin labeling of tubules in cortex. Medulla (M) is unlabeled. Tubules concentrated near corticomedullary
`junction show more intense labeling (small arrows) and these often follow a straight course perpendicular to surface
`of cortex. Glomeruli (large arrows) are unlabeled. Magnification x20. D: region of testis labeled with "I-enterotoxin
`demonstrates receptor localization primarily in seminiferous tubules. Dark field. Magnification x250.
`
`otoxin receptors were located in the seminiferous tubules
`of the testis. Thus enterotoxin receptors in these tissues
`appeared to be expressed only in the epithelial cells of
`the proximal renal or seminiferous tubules and the in-
`testinal mucosa.
`The receptors for ST enterotoxins have been shown to
`be localized to apical membranes of the intestine (8, 9,
`12, 14, 17, 23). Therefore, we prepared brush-border
`membranes from OK cortex and small intestinal mucosa
`by a magnesium-precipitation technique (2). Total cel-
`lular membranes were prepared from testis. The char-
`acteristics of 125I-enterotoxin binding to these prepara-
`tions were investigated to compare properties of the
`receptors in these tissues. We found that little specific
`
`binding of 125I-enterotoxin to receptors in brush-border
`membranes from kidney (Fig. 4) or intestine (not shown)
`occurred at 4°C relative to the binding observed at 37°C.
`Similar data were obtained with the testis membrane
`preparation (Fig. 4). The binding of 125I-enterotoxin to
`kidney, intestinal, and testicular membranes was inhib-
`ited by unlabeled peptide between 0.01 and 10 nM of
`enterotoxin (Fig. 5). The relative affinities of these recep-
`tor sites for the E. coti enterotoxin were similar under
`these conditions. The IC5c, values for E. coil enterotoxin
`inhibition of 'I-enterotoxin binding to kidney, intes-
`tine, and testis membranes were 0.5 ± 0.09 (n = 4), 0.4
`± 0.2 (n = 3), and 0.4 ± 0.04 nM (n — 3), respectively.
`We previously reported that intravenous administra-
`
`

`

`F878
`
`K COLI ENTEROTOXIN RECEPTORS
`
`•
`
`•
`
`• --___ •
`
`-•
`
`(17
`
`aJ
`
`0
`
`100
`
`90
`
`BO
`
`1200
`
`1000
`
`BOO
`
`600-
`
`400-
`
`POO
`
`cGMP nmol/day
`
`12
`
`10
`
`8
`
`6
`
`4
`
`2
`
`1:1
`
`LLJ
`
`2
`
`4
`
`6
`Days
`
`8
`
`10
`
`FIG. 6. Chronic administration of E. coli enterotoxin to opossum
`increases urine cGMP excretion. Data are means of 5 animals for each
`point. Each animal received 20 µg of E. coli enterotoxin intramuscularly
`as a single injection on days 4, 5, and 6. Peak changes in urine cGMP
`occurred 24 h later at days 5, 6, and 7 of urine collection. u
`• Na+
`excretion, •—• lc+ excretion.
`
`of this peptide (via cGMP) on ion transport in that
`nephron segment could be substantially affected by com-
`pensatory responses of the nephron segments distal to
`the proximal tubule.
`
`DISCUSSION
`
`These experiments revealed that renal receptors for E.
`coli enterotoxin are localized to the proximal tubule,
`presumably in the brush-border membranes of proximal
`tubular epithelial cells. Moreover, receptor autoradiog-
`raphy revealed that ' 25I-enterotoxin labeled the receptor
`sites in epithelial cells of the seminiferous tubules of
`opossum testis. Enterotoxin receptors were localized to
`enterocytes of both crypt and villus regions in the small
`intestine and crypts of the large intestine. Therefore, the
`
`16
`
`12
`
`B
`
`4
`
`0
`
`Testis
`
`Kidney
`
`,.o
`
`---- A
`
`-A -----
`
`-
`
`.
`60
`
`-
`
`-
`
`-
`90
`
`-
`-
`1203
`
`30
`
`----- . , 1
`
`30
`
`60
`
`90
`
`120
`
`Time - minutes
`
`FIG. 4. Time and temperature dependence for binding of 1"I-enter-
`°toxin to testis and kidney membranes. See EXPERIMENTAL PROCE-
`DURE for methods of membrane preparation. Incubations were carried
`out at 37°C •—•, A
`A, or 4°C o
`o, o-
`L. Total binding of
`125I-enterotoxin is illustrated by •
`•, o- - - -o and nonspecific
`binding (with 1µM enterotoxin) by A-A,
`- - -L. Data are means
`± SE of triplicates in a representative experiment. Scale on left axis
`also applies to data in right-hand panel.
`
`•
`I - - -
`
`•
`
`H
`4
`
`4
`
`* Kidney
`0 Intestine
`
`A Testes
`
`9
`
`6
`
`3
`
`a
`
`O
`
`I-
`
`CU
`
`5.2
`
`M
`
`4.7
`
`4.2 c` Cu
`
`3.7 "
`
`3.2
`
`-
`-
`-
`-
`-
`0 —H
`0 12 11 10 9 8 7
`0 12 11 10 9 8 7
`—Log
`rE Coll ST] M
`FIG. 5. Comparison of radioligand binding curves for 125I-entero-
`toxin binding to kidney and intestine brush-border membranes with
`testis membranes. Data are a representative experiment of a minimum
`of 3 binding isotherms carried out with each membrane preparation.
`ST, heat stable.
`
`tion of K coli enterotoxin to the anesthetized opossum
`resulted in a large increase in urine cGMP excretion
`without any change in urinary excretion of H20, Nat,
`Kt, Mg', Ca', and phosphate (11). In contrast,
`ANF elicited a smaller increase in cGMP excretion but
`caused a marked diuresis, natriuresis, and calciuria. In
`the present study we investigated the effects of chronic
`administration of E. coli enterotoxin on renal function.
`Intramuscular injection of 20 µg enterotoxin daily for 3
`days resulted in a marked increase in urine cGMP excre-
`tion in conscious, unrestrained opossums that persisted
`for 1 day after discontinuation of treatment (Fig. 6). The
`peptide did not appear to influence urine Nat, K±, or
`H2O excretion when administered on a daily basis and
`did not produce diarrhea by this route of administration.
`Lack of an effect of the enterotoxin on Na± and H2O
`excretion in this and earlier experiments (11) even
`though a marked stimulation of cGMP excretion oc-
`curred is consistent with the localization of E. coli enter-
`otoxin receptors to the proximal tubule. Putative effects
`
`

`

`E. COLI ENTEROTOXIN RECEPTORS
`
`F879
`
`conclusion can be made that the enterotoxin receptor
`protein may be broadly expressed in several different
`epithelial cells of the opossum. The enterotoxin stimu-
`lates chloride secretion and also inhibits salt absorption
`in the intestine (6, 10, 13, 16, 29). These effects, which
`are the apparent cause of "traveler's diarrhea" can be
`reproduced by cGMP analogues in experimental systems
`(20). If the apical membrane receptor for ST enterotoxins
`is broadly expressed in epithelia, perhaps this membrane
`protein (via cGMP) influences salt transport in kidney,
`testis, and other epithelia by a mechanism similar to that
`which regulates salt transport in the small and large
`intestine.
`Data presented in this manuscript suggest that enter-
`otoxin receptors do not occur in glomeruli, distal tubules,
`thick ascending limb of the loop of Henle, or collecting
`tubules. The occurrence of high-affinity receptors for
`1-25I-enterotoxin in the proximal tubule (convoluted and
`pars recta) detected by autoradiography and in brush-
`border membranes does not rule out the possibility that
`other nephron segments and basolateral membranes may
`have this protein. A population of receptors at substan-
`tially lower concentration (i.e., density) may not be re-
`vealed by the in vitro autoradiography technique used in
`these experiments. Moreover, we have not measured 125I-
`enterotoxin binding to putative receptors of kidney baso-
`lateral membranes. However, other studies using such
`approaches with membranes isolated from intestinal mu-
`cosa revealed that enterotoxin receptors were localized
`to the brush-border membranes (4, 8, 9, 12, 17, 23, 28).
`Thus it may be postulated that this receptor protein,
`which in the OK cell line was 120 kDa in mass (11), may
`be an apical membrane protein in proximal tubular cells
`of OK. A similar size of the enterotoxin receptor of about
`100 kDa was observed in the intestine of rats (8).
`ST enterotoxins have three physiological actions. The
`E. coli enterotoxin stimulates cGMP production in the
`intestine (8-10, 18-20, 29), kidney (11), and testis (pres-
`ent study), stimulates chloride secretion in the intestine
`and in crypt cell-like T-84 cells (13, 21), and inhibits
`sodium chloride absorption in the intestine (10, 13, 16,
`29). Thus "traveler's diarrhea" elicited by heat-stable
`enterotoxins is a secretory form of diarrhea that is similar
`to the disorder caused by heat-labile enterotoxins such
`as cholera toxin (13). Binding of the E. coli enterotoxin
`to apical membranes of T-84 cells caused a marked
`increase in apical membrane chloride permeability, prob-
`ably by opening chloride channels in this membrane (21).
`Because analogues of cGMP mimic the effects of E. coli
`enterotoxin on intestinal ion transport, it has been pos-
`tulated that cGMP serves as a second messenger for the
`ST enterotoxins, similar to the adenosine 3' ,5' -cyclic
`monophosphate-mediated effects of cholera toxin to
`cause secretory diarrhea (10, 13, 16, 20). However, ad-
`ministration of E. coli enterotoxin to the opossum acutely
`(11) or chronically (daily) did not appear to influence
`urinary Nat, Cr, or water excretion even though large
`increases in urine cGMP excretion were observed in
`these experiments. In view of the localization of the
`enterotoxin receptors to the proximal renal tubule, it
`may be postulated that the enterotoxin could influence
`
`salt transport in this segment in a fashion analogous to
`the peptides' actions on intestinal salt transport (10, 13,
`16, 21, 29). However, nephron segments distal to the
`proximal tubule may compensate for putative changes in
`filtrate composition elicited by the enterotoxin's ac-
`tion(s) on proximal epithelial cells. It is of interest that
`the intestine, unlike the nephron, has receptors for E.
`coli enterotoxin throughout the small and large intestine.
`Moreover, enterotoxin receptors occurred both in the
`enterocytes of crypts and villi, which agrees with the
`cGMP responses to the enterotoxin that were shown in
`the epithelial cells of crypts or villi (8). Such a distribu-
`tion of receptors throughout the length of the intestine
`may contribute substantially to the massive diarrhea
`produced by ST enterotoxins in mammals (13). The
`action(s) of E. coli enterotoxin on proximal tubular func-
`tion, if any exist, may require micropuncture and/or
`perfusion of proximal tubules to assess the effects of this
`peptide on transport in this nephron segment. Because
`OK and PtK-2 kidney cell lines express enterotoxin
`receptors (11), it may be possible to gain new information
`regarding the physiological actions of ST enterotoxins
`on renal function by use of these proximal tubulelike cell
`lines. Because E. coli enterotoxin regulates apical chlo-
`ride permeability in T-84 intestinal cells via control of
`chloride channels (21), it is tempting to speculate that
`this peptide could exert a similar influence on apical
`membrane chloride channels in proximal tubular cells.
`Such a mechanism implies that the 120-kDa enterotoxin
`receptor protein (11) may serve as a regulator of apical
`chloride channels in kidney, intestine, and perhaps other
`epithelia.
`These experiments revealed that a third tissue, testis,
`expresses receptors for the E. coli enterotoxin. These
`receptors appear to be, like those of OK and intestine,
`coupled positively to the activation of an apical mem-
`brane guanylate cyclase. Membrane-bound guanylate cy-
`clase (22) is also postulated to be an effector mechanism
`for ANF, because atrial peptides stimulate cGMP pro-
`duction in a number of tissues, including kidney, intes-
`tine, and testis (1, 22). We found that ANF stimulated
`cGMP production in the small intestine and kidney
`cortex, but ANF was ineffective under these conditions
`in the testis and colon mucosa of opossums. Thus the
`relative distribution of the ANF and enterotoxin recep-
`tors that are coupled to guanylate cyclase appears to be
`different in these opossum tissues. Receptors for ANF
`that may be responsible for at least part of the natriuretic
`action of this peptide are found in cells of the inner
`medullary collecting ducts (24, 33). If ANF receptors
`exist in the terminal portion of the collecting tubule in
`the opossum, a nephron segment with no apparent en-
`terotoxin receptors, then these receptors may contribute
`substantially to the natriuresis and diuresis elicited by
`ANF in vivo (11). Regulation of Na+ channel permeabil-
`ity of collecting tubules by a cGMP-mediated process
`has been invoked as a cellular mechanism of ANF action
`(15, 24, 33). Both OK proximal tubules and the proximal
`tubular-like (5) OK cell line have enterotoxin receptors
`coupled positively to guanylate cyclase (11). OK cells
`also have ANF receptors (27). Thus a useful approach
`
`

`

`F880
`
`E. COLI ENTEROTOXIN RECEPTORS
`
`for future research may be exploring the effects of E. coli
`enterotoxin and ANF on Na+ permeability by use of the
`OK cell line as an experimental model for proximal
`tubular cells of the opossum.
`In conclusion, receptors for E. coli enterotoxin are
`localized to epithelial cells of the proximal renal tubule,
`seminiferous tubules of testis, and enterocytes in both
`crypts and villi of the small intestine and crypts of the
`colon. These receptors appeared to be associated with
`apical membranes of both kidney cortex and small intes-
`tine of the opossum. The kidney and testis receptors,
`similar to the intestine, are coupled to activation of
`guanylate cyclase. Occurrence of receptors for this family
`of peptides secreted by pathogenic enteric bacteria in
`tissues other than the intestine suggests that this 120-
`kDa membrane protein (11) may have a more general
`role in epithelial cell function than was previously con-
`sidered.
`
`The technical assistance of Sammy Eber, Pam Thorne, Jennifer
`Casati, M. Corlija, and Richard Poelling in carrying out this study is
`greatly appreciated. The help provided by Dr. Wynn Volkert in estab-
`lishing the tissue autoradiography methods was instrumental in com-
`pleting those experiments and we very much appreciate this assistance.
`We also thank Kris Nelson for preparing the manuscript for publica-
`tion.
`This research was supported by funds from the Medical Research
`Service Department of the Veterans Affairs, by the National Institutes
`of Health Grants DK-32848 and HL-10612, and by a Weldon Spring
`Research Award from the University of Missouri.
`Address for reprint requests: L. R. Forte, Dept. of Pharmacology,
`School of Medicine, University of Missouri, Columbia, MO 65212.
`
`Received 9 November 1988; accepted in final form 3 July 1989.
`
`REFERENCES
`
`1. ATLAS, S. A., AND T. MAACK. Effects of atrial natriuretic factor
`on the kidney and the renin-angiotensin-aldosterone system. En-
`docrinol. Metab. Clin. N. Am. 16: 107-143, 1987.
`2. BIBER, J., B. STIEGER, AND H. MURER. A high yield preparation
`for rat kidney brush border membranes: different behavior of
`lysosomal markers. Biochem. Biophys. Acta 647: 169-176, 1981.
`3. BRIDGES, R. J., W. RUMMEL, AND B. SIMON. Forskolin induced
`chloride secretion across the isolated mucosa of rat colon descen-
`dens. Arch. Pharmacol. 323: 355-360, 1983.
`4. COHEN, M. B., M. R. THOMPSON, G. J. OVERMANN, AND R.
`GIANNELLA. Association and dissociation of Escherichia coli heat-
`stable enterotoxin from rat brush border membrane receptors.
`Infect. Immunol. 55: 329-334, 1987.
`5. COLE, J. A., S. L. EBER, R. E. POELLING, P. K. THORNE, AND L.
`R. FORTE. A dual mechanism for regulation of kidney phosphate
`transport by parathyroid hormone. Am. J. Physiol. 253 (Endocrinol.
`Metab. 16): E221-E227, 1987.
`6. DIIARMSA