INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE MOST COMMONLY REPORTED gastrointestinal manifestation in cystic fibrosis (CF) is malabsorption due to pancreatic insufficiency, caused by occlusion of the pancreatic ducts by abnormally viscous mucus secretions. This alteration in intraluminal fluid and mucus content extends to other solid organs and hollow viscera of the gut, creating the clinical conditions of meconium ileus at birth and distal intestinal obstruction syndrome, chronic constipation with acquired megacolon, and rectal prolapse in older individuals (1).

Underlying these gastrointestinal complications is the basic defect in CF, which is caused by mutations in the gene that encodes the CF transmembrane conductance regulator (CFTR; Ref. 43). The most commonly reported mutation in CF, accounting for the genetic defect in 60-70% of the chromosomes of CF individuals, is a three-base-pair deletion that removes phenylalanine at position 508 (F508) of the CFTR protein (54). CF patients homozygous for this mutation exhibit the severe clinical phenotype of lung disease, pancreatic insufficiency, and predisposition to gastrointestinal obstruction (48). Electrophysiological analysis of the gut mucosa isolated from these individuals demonstrates a lack of cAMP-dependent fluid secretion (6, 37, 53). Reduced fluid transport, which is the primary cellular defect of CF epithelia, is believed to significantly enhance the pathology of mucus plugging in the gut.

Over the past six years, mouse models for CF that duplicate the pathophysiological phenotype of CF in the human intestine have been produced with the use of gene-targeting techniques. Transgenic CFTR-deficient animals fail to exhibit CFTR-mediated fluid secretion and present with gastrointestinal disease (15). In this study, we utilized the CFTR^m3Bay null mutation mouse as a model for CF to investigate whether infectious rotavirus or the novel rotaviral enterotoxin NSP4 (3) can cause diarrhea. These mice fail to express functional CFTR protein because of multiple stop codons engineered within exon 3 of the murine CFTR genome (23).

Rotaviruses are the leading cause of severe gastroenteritis in infants and young animals (25). We have recently shown that NSP4, a rotavirus nonstructural protein (16), can cause diarrhea in young mice (3). Furthermore, electrophysiological analyses of intact intestinal mucosa from mice revealed that NSP4 mobilizes Ca²⁺ to mimic the secretory effects of the cholinergic agonist carbachol (CCh) in potentiating cAMP-dependent fluid secretion (3). We therefore asked whether NSP4 could evoke a diarrheal response in CFTR-depleted mice that exhibit no functional cAMP-dependent secretory pathway. Here, we demonstrate that NSP4 injected intraperitoneally into CFTR-deficient (CFTR^/) mouse pups induces an age-dependent diarrhea. Hence, CFTR-mediated changes in intestinal fluid transport were not directly involved in NSP4-elicited diarrhea. However, NSP4 was found to mobilize Ca²⁺ in the crypt epithelia of both small and large intestine from wild-type (CFTR^+/+) as well as CFTR^/ pup and adult mice. These ex vivo data correlated with our in vitro findings in human gastrointestinal cell lines, in which NSP4 was shown to mobilize intracellular Ca²⁺ concentration ([Ca²⁺]_i) via phospholipase C (PLC) activation and inositol 1,4,5-trisphosphate (IP₃) production (13). To investigate in greater detail the relationship between NSP4-induced epithelial cell Ca²⁺ mobilization and our previously reported age-dependent effects of NSP4 on murine intestinal mucosa Cl secretory current generation (3), we measured plasma membrane halide permeability changes ex vivo within the epithelial cells of isolated distal colon crypts. We found that NSP4 elicited age-dependent I influx into mouse pup crypts, inhibited by either the removal of bath Ca²⁺ or the transport inhibitor DIDS. The fact that this Ca²⁺-dependent plasma membrane halide permeability pathway was not CFTR may provide a possible explanation as to why CFTR^/ mouse pups exposed to NSP4 develop age-dependent diarrhea.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reagents. Fura 2-AM and 6-methyoxy-N-(3-sulfopropyl) quinolinum (SPQ) were purchased from Molecular Probes (Eugene, OR). All other reagents, including DIDS, were purchased from Sigma (St. Louis, MO). NSP4 protein was fast-performance liquid chromatography- and affinity-purified from Sf9 insect cells infected with a recombinant baculovirus pAC461-G10 expressing simian rotavirus gene 10 (encoding NSP4), as described previously (13). The NSP4_114-135 peptide was synthesized at the University of Pittsburgh protein core laboratory and characterized as previously described (3).

Experimentally induced rotaviral infection or administration of NSP4 protein and NSP4_114-135 peptide to mouse pups. The simian rotavirus strain SA11 clone 3 (18) was used to infect neonatal CD-1 or CFTR^/ (23) mouse pups between the ages of 6 and 8 days. The genotype of the CFTR^/ mice was determined by polyacrylamide gel analysis of PCR products generated from DNA extracted from the tails of 1- to 2-day-old mice with the use of primers, as described previously (23). Either 10 or 20 diarrheal dose 50 of SA11 was administered in 50 µl of medium 199 by intragastric gavage. Alternatively, purified NSP4 protein or NSP4_114-135 was administered to C57B CFTR^//CFTR^+/+ or CD-1 CFTR^+/+ mouse pups in a final volume of 50 µl of PBS by intraperitoneal injection. The doses used were 0.5 nmol of NSP4 protein or 100 nmol of NSP4_114-135 peptide because these doses were shown to be effective at inducing diarrhea in CD-1 CFTR^+/+ mouse pups (3).

Diarrheal activity measurements. To determine the presence of diarrhea following protein or peptide treatment, each mouse pup was examined every 1-2 h for the first 8 h and at 24 h after inoculation by gently pressing the abdomen. After virus inoculation, mouse pups were monitored twice a day for 4 days. Diarrhea was noted and scored from 1 to 4, with a score of 1 reflecting loose yellow stool and a score of 4 indicating completely liquid stool. A score of 2 (mucous with liquid stool, some loose but solid stool) and above was considered diarrhea. The scoring was performed on coded animals by a single person.

Crypt isolation and dye loading. Epithelial crypts from CFTR^/ mice (aged 8-12 days) were isolated from 3-cm segments of the mid to distal small intestine, 6 cm proximal to Bauhin's valve, and from the distal colon. After euthanasia by an overdose of ether and cervical dislocation, the entire small intestine and colon from the mouse were removed and flushed with ice-cold physiological saline. Individual intestinal segments were then mounted onto Perspex paddles and were immersed in Ca²⁺-free standard Krebs buffered saline (in mmol/l: 107 NaCl, 4.5 KCl, 0.2 NaH₂PO₄, 1.8 Na₂HPO₄, 10 glucose, and 10 EDTA), and continuously gassed with 5% CO₂-95% O₂ at 37°C for 10-20 min. The crypts were then separated from the overlying mucosa by mechanical vibration for 30 s into ice-cold KCl HEPES saline (in mmol/l: 100 potassium gluconate, 20 NaCl, 1.25 CaCl₂, 1 MgCl₂, 10 HEPES, 10 glucose, 5 sodium pyruvate, and 0.1% BSA, pH 7.4), resembling the intracellular medium. Suspended crypts were then deposited (1,200 rpm for 1 min) onto poly-L-lysine-coated microscope coverslips (0 oz) with the use of a Shandon Cytopsin cell preparation system (13). The coverslips with crypts were then attached with vacuum grease to the base of customized perfusion wells and loaded in the dark with either Ca²⁺-sensing or Cl-sensing dye.

Fura 2-AM loading. Isolated crypts were incubated with 10 µM fura 2-AM at room temperature for 15-20 min. The coverslips were then mounted on an inverted Nikon microscope and superfused with standard HEPES-buffered extracellular saline (in mmol/l: 140 NaCl, 4.7 KCl, 1.13 MgCl₂, 10 HEPES, 10 glucose, and 1 CaCl₂, pH 7.4) for 5 min before imaging.

SPQ loading. Isolated crypts were loaded with SPQ at room temperature by a 6-min exposure to hyposmotic (80 mosM Cl or NO₃) HEPES-buffered extracellular saline containing 5 mM SPQ. After loading, the crypts were allowed to recover in isosmotic (140 mosM Cl or NO₃) HEPES-buffered extracellular saline for 10-15 min. This period was used when performing anion channel inhibitor studies to preload the crypts so that the reported spectral interference of 0.5 mM DIDS with SPQ fluorescence was minimized (61). DIDS binds irreversibly to cellular membranes; therefore, crypts were loaded and the bathing saline was changed before experimentation. The solution temperature was maintained at 37°C throughout these studies by prewarming the extracellular solutions and by water-jacketing the oil-immersion lens of the inverted microscope.

NSP4/NSP4_114-135 addition to isolated crypts. Small (100 µl) volumes of either the NSP4 protein or NSP4_114-135 peptide were superfused onto the isolated crypts during regular bath flow by N₂ pressure injection with the use of a Picospritzer. Low-resistance glass pipettes were filled with either compound dissolved in HEPES-buffered extracellular saline, and the tip was maneuvered close to the isolated crypt. Peptide or protein was released directly into the vicinity of the crypt for 40 s before being washed away by the bath flow.

[Ca²⁺]_i imaging. All experiments were carried out with the aid of a high-resolution camera imaging system as described previously (13). In brief, light emitted from fura 2-loaded cells at 510 nm was captured by an intensified video camera after exposure to both 340- and 380-nm excitation light. The camera signal was then digitally encoded and processed with image analysis software (IMAGE1/FL, Universal Imaging, Media, PA). The background-subtracted images were ratioed on a pixel-by-pixel basis to yield a bitmap field. Calibration of the fura 2 dye fluorescence was carried out with the use of the ionophore ionomycin under Ca²⁺-free and Ca²⁺-saturating conditions as described previously, and the [Ca²⁺]_i was calculated according to the Grynkiewicz equation (21). The [Ca²⁺]_i values of individual field pixels obtained by this procedure were color coded and displayed on an RGB monitor before being stored on the hard disk. Six collection areas were chosen along the longitudinal axis (base to neck) of the crypt for spatial and time-dependent analysis of [Ca²⁺]_i. The averaged ratio signal obtained from each of the six areas was digitally saved as a log file. The collected values from the six areas imaged within a single experiment were averaged together to give an experimental observation of one (n = 1). Values obtained from similarly placed collection areas along the longitudinal axis of different crypts were also averaged to see if any spatial localization to the Ca²⁺ signal existed. All protocols were completed within 1 h after crypt isolation.

Intracellular I imaging. The same high-resolution camera imaging system described above was utilized for these studies. Excitation for SPQ fluorescence was provided by a barrier filter centered at 365 ± 10 nm, reflected into the microscope objective by a 400-nm dichroic mirror. A neutral density filter (1.5 OD) was included in the light path to minimize photobleaching. The fluorescence emission from the SPQ-loaded cells passed through a 450 ± 25-nm barrier filter before being detected by the intensifier/camera. Fluorescent images were acquired at 0.8-s intervals and averaged over eight frames to yield a 540 × 480 field of mean pixel intensities. Six collection areas were then chosen within the longitudinal axis (base to midregions) of the crypt for spatial and time-dependent analysis of cellular SPQ fluorescence. SPQ does not exhibit excitation or emission shifts on anion interaction. Thus ratiometric imaging is not possible, and dye fluorescence is affected by volume changes within the cell (58). Both cAMP- and Ca²⁺-mobilizing agonists have been reported to induce shrinkage without compensatory volume recovery in the lower to midregion of rat distal colonic crypts over a 5- to 30-min time course (11). To determine the extent of this effect over the time period of SPQ quenching (1-3 min), colonic crypts isolated from the distal colon of the mouse were loaded with the AM form of calcein, a dye inert to Ca²⁺ and pH changes in cells within the physiological range (8, 9). Cell volume measured over the first 30 s of agonist addition to the bath did not change appreciably (<2%) within the lower/midcrypt region (data not shown). These regions were utilized for all studies. Furthermore, to minimize possible volume effects on dye fluorescence, we chose an experimental design in which SPQ fluorescence quench by halide influx was employed rather than dequenching following halide efflux. This ensured that agonist-induced cell shrinkage, if present, would act to decrease rather than to potentiate changes in plasma membrane halide permeability. Thus our reported values of SPQ quench may modestly underestimate agonist-evoked halide influx. The collected values from the six areas imaged within a single experiment were averaged together to give an experimental observation of one (n = 1).

Calculation of Stern-Volmer constants for halide quenching of SPQ. SPQ fluorescence is quenched collisionally by halide anions with different potency. Nonphysiological anions such as thiocyanate (SCN), Br, F, and I quench strongly in free solution with an efficacy greater than Cl (Cl < Br < SCN < I), whereas NO₃ does not quench at all (27). The Stern-Volmer equation that describes this interaction is F_o/F_anion = 1 + K_q[anion], where F_o is the fluorescence in the absence of anion, F_anion is the fluorescence in the presence of a given anion concentration, K_q (slope) is the Stern-Volmer quenching constant (in M¹), and [anion] is the concentration of halide utilized (27). Stern-Volmer quenching constants for Cl, Br, and I anions were calculated from linear curves obtained over a concentration range of 0-150 mM halide (see RESULTS). The near-linear Stern-Volmer quench curve for I (Fig. 1) and the fact that plasma membrane I permeability changes occur largely through conductive pathways (see DISCUSSION and Refs. 36 and 63) led us to choose this anion to quantify agonist effects in isolated crypts.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Stern-Volmer plot for quenching of 6-methoxy-N-(3-sulfopropyl)quinolinum (SPQ) fluorescence by Cl, Br, and I. Isolated crypts from distal colon of wild-type cystic fibrosis transmembrane conductance regulator (CFTR+/+) and heterogeneous (CFTR+/) mice were loaded with SPQ. Intracellular SPQ quench by different intracellular halide concentrations was induced with dual ionophore technique. Fo and F represent background and dye leak subtracted values of fluorescence in absence and in presence of quenching halide, respectively. Values are means ± SD.

I

Standard experimental procedure for measuring agonist-stimulated I influx. SPQ-loaded crypts were perfused with HEPES-buffered NaNO₃ containing extracellular saline, and, after establishment of a stable fluorescence signal, bath NO₃ was replaced with I (140 mosM). The quenched SPQ fluorescent signal (NO₃ does not quench SPQ nor compete with I) was recorded over predefined measurement windows along the base to midcrypt axis. Basal rates of I influx (J_I) were then calculated with the use of the Stern-Volmer quench constants for I obtained for either mouse pup or adult crypts (see Calculation of Stern-Volmer constants for halide quenching of SPQ and RESULTS). Cl was then reintroduced, and I was simultaneously removed from the bath. A single agonist was administered during this recovery phase (forskolin, FSK) or after establishment of a new stable fluorescence signal (CCh or NSP4) when bath I was reintroduced in Cl-free saline. The resulting agonist-stimulated SPQ quench curve was recorded, and J_I agonist was calculated. Agonist-induced changes in I influx rate were then calculated from pooled values of J_I agonist and J_I basal as J_I = (J_I agonist  J_I basal). In addition to this approach, mean ± SD rates of fluorescence change at 1.6-s intervals over the first 40 s of fluorescence quench were graphically displayed in histogram format as change in agonist-induced quench rate, R_I = (F/t_agonist  F/t_basal). This later approach allowed us to determine latency and to record changes in dye quench unrelated to initial (time 0) changes in quench rate. All SPQ-loaded crypts were superfused with saline at the same bath perfusion rate (4 ml/min); solution exchange in the experimental chamber (200 µl) occurred within 3 s.

Statistical analysis. To determine statistical significance of differences between observations within an experiment, the paired Student's t-test was used. For statistical differences between experiments, the unpaired Student's t-test was used. Between 6 and 12 separate experimental observations pooled from different dye loadings were routinely collected for each experimental condition. The Fisher's exact test was used to estimate the probability of difference for diarrheal activity measurements within CF and non-CF groups in which the tabulated frequency of occurrence was too small for ² analysis.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of orally administered rotavirus and injected viral NSP4 peptide(s) in mouse pups. Inoculation of CFTR^/ mice with infectious SA11cl3 rotavirus caused diarrhea in four of five mice within 24-48 h (Table 1). When CFTR^+/+ or CFTR^+/ mouse pups from the same C57B1/J6 background were infected with the SA11 virus, ~78% of the mice exhibited diarrhea (Table 1). Intraperitoneal injection of intact NSP4 protein or NSP4_114-135 was similarly found to elicit diarrhea in CFTR^+/+ or CFTR^+/ mouse pups (66 and 70% responding, respectively). CFTR^/ mouse pups inoculated with these reagents exhibited diarrhea in 50% and 43% of these mice, respectively (Table 1). Diarrheal content was graded as equally severe as that encountered in CFTR^+/+ mice (3). Although the incidence of diarrhea caused by NSP4 protein or NSP4_114-135 appeared lower in the CF mouse studies, there was no significant difference between CFTR^+/+, CFTR^+/, and CFTR^/ littermates (Fisher's exact test for 2 × 2 contingency tables; Ref. 32). All three reagents (SA11 virus, NSP4 protein, or NSP4_114-135) were effective diarrheal agents. In contrast, when infectious SA11cl3 virus was given to either adult CFTR^+/+, CFTR^+/, or CFTR^/ mice, none exhibited diarrhea (Table 1).

View this table: [in this window] [in a new window]   Table 1.   Induction of diarrhea by rotavirus infection or NSP4 in CFTR/ mice

NSP4 peptide mobilizes [Ca²⁺]_i levels in epithelial cells isolated from the small and large bowel of both CFTR^+/+ and CFTR^/ mice. Rotaviral protein production has been correlated with changes in [Ca²⁺]_i homeostasis (39) and with Ca²⁺-dependent cytotoxicity during viral infection in cultured cells (33, 34). With the use of the human colonic epithelial cell line HT-29 clone 19A, which expresses plasmalemmal proteins found in both small and large intestinal cells, we recently demonstrated that exogenously added NSP4 and NSP4_114-135 peptide causes [Ca²⁺]_i mobilization by PLC-mediated IP₃ production (13). We therefore investigated whether the diarrheal effects of NSP4 and NSP4_114-135 peptide in vivo were also associated with native mucosal intestinal cell [Ca²⁺]_i mobilization. Fluorescence video microscopy was used to study the effects of exogenous NSP4 addition on [Ca²⁺]_i in intestinal crypts isolated from either small or large bowel of 8- to 13-day-old mouse pups.

Small intestinal crypts. Crypts from the small intestine were chosen in preference to villi, which dissociated during isolation. Addition of 100 nM NSP4 protein, a dose shown to promote near-maximal [Ca²⁺]_i rises in vitro (13), elicited rapid rises in [Ca²⁺]_i that lasted 2-5 min in CFTR^+/+, CFTR^+/, and CFTR^/ small intestinal crypts. Resting [Ca²⁺]_i values of 126 ± 28 nM (CFTR^/, n = 13) and 110 ± 63 nM (CFTR^+/+ and CFTR^+/, n = 14) increased to peak values of 205 ± 54 nM and 260 ± 116 nM, respectively (Table 2). A representative example (Fig. 2, A-C) shows consecutive images taken before, during, and after NSP4 addition to an epithelial crypt isolated from a CFTR^/ mouse. No significant difference was seen between mouse CFTR genotypes with respect to basal, peak, or net Ca²⁺ values (P > 0.01, Student's t-test; Table 2). These results indicated that NSP4 was an effective Ca²⁺-mobilizing agonist in the small intestine and that Ca²⁺ homeostasis, both basal and agonist-stimulated, is unaffected by the CFTR gene knockout. Not all experimental crypts responded; 27 of 108 (25%) gave [Ca²⁺]_i rises. Increasing the dose of NSP4 from 100 to 500 nM and above did not increase the number of NSP4-responsive crypts (3 of 18 responded = 17%; data not shown).

View this table: [in this window] [in a new window]   Table 2.   Summary of the effects of 100 nM NSP4 and 100 µM CCh on [Ca2+]i levels in either CFTR/ or CFTR+/+,+/ small and large intestine mouse crypts

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Consecutive pseudocolor images of retroviral enterotoxin protein NSP4-induced intracellular Ca2+ concentration ([Ca2+]i) rise in an isolated CFTR-deficient (CFTR/) small intestinal mouse crypt, loaded with fura 2. A: resting Ca2+ level of ~120 nM before agonist stimulation. B: peak Ca2+ level approaching 220 nM immediately after addition of 100 nM NSP4. C: gradual return of crypt to resting Ca2+ levels ~1.5 min after addition of NSP4.

Colonic crypts. When isolated crypts from the distal colon of CFTR^+/+, CFTR^+/, and CFTR^/ mice were superfused with 100 nM NSP4 protein, [Ca²⁺]_i rises similar to those obtained from small intestinal crypts were recorded (P > 0.1, Student's t-test; Table 2). Peak NSP4-induced [Ca²⁺]_i values were similar in all three genetic backgrounds, although basal values were slightly, but not significantly, higher in CFTR^/ mice when all samples were pooled (P > 0.01, Student's t-test; Table 2). In our in vitro studies, we found that the NSP4-induced [Ca²⁺]_i mobilization was susceptible to predepletion by other agonists (remaining unresponsive for tens of minutes; Ref. 13). In contrast, NSP4-induced [Ca²⁺]_i release in both ex vivo colonic and ileal mucosal cell preparations was not desensitized by repetitive agonist administration (n = 15). In the CFTR^/ mouse crypt preparation shown in Fig. 3, identical [Ca²⁺]_i rises were seen for consecutive NSP4 or CCh followed by NSP4 challenge. Ca²⁺ homeostasis in the isolated crypt therefore appears to be more tightly regulated than in colonic cell lines.

View larger version (11K): [in this window] [in a new window]   Fig. 3.   A representative example of a single crypt. Average [Ca2+]i values from 6 measurement areas along length of an isolated CFTR/ colonic mouse crypt, showing effect of 100 nM NSP4 (A), 500 µM carbachol (CCh, B), and 100 nM NSP4 (C). SD values are shown as error bars.

View larger version (9K): [in this window] [in a new window]   Fig. 4.   [Ca2+]i measurements taken from 3 different regions separated by ~15 µm along longitudinal axis of a CFTR/ small intestinal crypt, showing magnitude of NSP4 [Ca2+]i rise at crypt base (A), midregion of crypt (B), and near neck of crypt (C). Agonist (100 nM NSP4) was washed out of bath ~10 min after addition.

NSP4 peptide mobilizes [Ca²⁺]_i levels in crypt epithelial cells from both pup and adult mice. Isolated colonic crypts from either normal (CFTR^+/+) C57B or BALB/c mice were loaded with fura 2 and exposed to 100 nM NSP4. In the example shown (BALB/c, Table 3), NSP4 mobilized Ca²⁺ in crypts from both 7- to 12-day- and >25-day-old mice. Although the mean number of BALB/c crypt responses was higher than that observed in the C57B background (data not shown), there was no age dependency to the magnitude of the NSP4-induced [Ca²⁺]_i rise (values were not statistically significant, unpaired t-test, P > 0.05, n = 13; Table 3). When CCh (100 µM) was applied after a NSP4-induced Ca²⁺ mobilization, 75% of pup crypts and 80% of adult crypts responded with a second [Ca²⁺]_i rise (n = 13). In this instance, the net agonist-induced [Ca²⁺]_i rise was significantly smaller in adult crypts (P < 0.001, unpaired t-test; Table 3). When CCh alone was applied to crypts, a similar age-dependent trend in Ca²⁺ mobilization was observed, which failed to be significant because of the great variability in the magnitude of the peak [Ca²⁺]_i values (P > 0.05; Table 3). These results indicated that events distal to NSP4-elicited [Ca²⁺]_i mobilization must regulate the age- and Ca²⁺-dependent Cl secretory activity of NSP4 observed in the ex vivo mouse intestine (3).

Calculation of Stern-Volmer constants for Cl, Br, and I in adult and pup mouse crypts. To separate carrier-mediated anion transport events from conductive anion transport in SPQ-loaded crypts, the halide I was used as a surrogate anion for Cl. I is not a substrate for either the Na⁺-K⁺-2Cl cotransporter or Cl/HCO exchanger (36, 63). This choice was not, however, without technical drawbacks, because the Stern-Volmer relationships for Br, I, and SCN in free solution are reported to become positively nonlinear at >15 mM halide (24). Nonlinear increases in K_q above this concentration suggest that, in addition to dynamic collision interactions, static interactions occur (i.e., those leading to nonfluorescent halide-SPQ complex formation; Ref. 28). Irrespective of the exact nature of this quenching (assumed to be static buffering), the dye would become noncalibrated at high halide concentrations, compromising I usefulness as a surrogate anion for conductive Cl influx. The Stern-Volmer relationship for Cl on the other hand has been demonstrated to remain linear when measured both in free solution and cells and shown to be the product of the dynamic collision-quench rate constant K_q (24). Stern-Volmer relationships for Cl, I, and Br were constructed in SPQ-loaded crypts with the use of the dual ionophore technique (Ref. 7; see MATERIALS AND METHODS). We found that apparent Stern-Volmer quench constants for both Br and I, calculated from halide concentrations of either <15 mM or <150 mM, exhibited modest <15% and slight <5% nonlinear increases in slope, respectively. This deviation was not large enough to accurately determine an additional buffering coefficient (28). Thus K_q values for Br and I were estimated by linear curve fitting and found to be 38.5 M¹ and 51.4 M¹, respectively.¹ The K_q value for Cl was found to be 21.5 M¹ (measurements made in both pup and adult crypts, n = 8 per halide; Fig. 1), reflecting very well the ratio of SPQ collisional quench (1:1.7:1.9) reported for colonocytes (45). These values were slightly lower than those calculated in the rabbit crypts (45) and higher than those estimated in a variety of cancer cell lines (which range between 12 M² and 18 M² depending on cell type; Ref. 58). Because we failed to measure changes in tubular geometry (shrinkage) associated with agonist stimulation when confining our measurements to basal crypt areas, we did not resort to more theoretical measurements of the Stern-Volmer constant (22). With the use of these values, 50% of cellular SPQ quench would be achieved at 26, 19.6, and 47 mM Br, I, and Cl, respectively.

Calculation of basal and agonist-stimulated halide permeability. The dynamic K_q values determined above were used to calculate influx rates for the respective anions that were dependent on plasma membrane permeability and not on alterations in SPQ quench rate. Basal influx was in the order I > Cl > Br (Table 4) for CFTR^+/+, CFTR^+/, and CFTR^/ pup crypts. An age-dependent but not significant (P > 0.05) trend in basal influx was seen, with adult values being somewhat higher than those in pups. The lack of CFTR was associated with a statistically significant (P < 0.001) decrease in basal halide influx rate (J_halide) in CFTR^/ mice (Table 4).

View this table: [in this window] [in a new window]   Table 4.   Basal halide influx in CFTR-expressing (CFTR+/+,+/) and nonexpressing (CFTR/) mouse pups

I agonist

I

View this table: [in this window] [in a new window]   Table 5.   Summary of the effects of forskolin, CCh, and NSP4 on I  influx rates in CFTR+/+,+/ and CFTR/ mice

FSK-stimulated I influx into crypt cells was age independent and required CFTR. FSK, as predicted from previous studies (10, 45), induced significant I influx in both CFTR^+/+ and CFTR^+/ mouse pup and adult crypts; J_I agonist increased 3.7- to 4-fold over J_I basal and were statistically significant (P < 0.001, n = 5; Table 5). Age-dependent differences in the magnitude of J_I were absent (P > 0.05, n = 5; Table 5). Time-dependent analysis revealed that the peak FSK-induced quenching rate (R_I agonist) occurred 8-10 s after bath I exchange (Fig. 5A).

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Change in SPQ fluorescence quench rate in mouse distal colon crypts elicited by cAMP-mobilizing agonist forskolin (FSK, 10 µM). A: values obtained from adult and pup CFTR+/+ and CFTR+/ mouse crypts, showing a lack of any age dependency. B: values obtained from pooled CFTR+/+ and CFTR+/ and CFTR/ pup crypts, demonstrating that effects of FSK on SPQ quench rate were dependent on CFTR expression. Values are means ± SD. FU, fluorescence unit.

/

I agonist

I

I agonist

I

I

I

+/

I

/

3

I

+/

3

CCh-stimulated I influx into crypt cells was age independent and required CFTR expression. CCh (100 µM) J_I agonist values were twofold greater than J_I basal values in wild-type and heterozygous CF genotypes and were significantly different (P < 0.001, n = 5; Table 5). The resulting J_I values from both groups displayed no age dependency (J_I pup = 0.62 ± 0.30 M³/s vs. J_I adult = 0.69 ± 0.24 M³/s, n = 5, P > 0.05; Table 5). These changes approximated to 51 and 40% of the values elicited by FSK in pup and adult crypts, respectively (Table 5). CCh-elicited R_I changes were greatest 8-10 s after bath I exchange, with a latency similar to that recorded for FSK (Fig. 6A). CCh was therefore a potent age-independent stimulator of plasma membrane I influx in CFTR-containing crypts, causing identical but smaller changes in I influx rate compared with FSK.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Change in SPQ fluorescence quench rate in mouse distal colon crypts elicited by Ca2+-mobilizing agonist CCh (100 µM). A: values obtained from pooled adult and pup CFTR+/+ and CFTR+/ mouse crypts, showing a lack of any age dependency. B: values obtained from pooled CFTR+/+ and CFTR+/ and CFTR/ pup crypts, demonstrating that effects of CCh on SPQ quench rate were delayed and significantly smaller in magnitude when CFTR was absent from crypt. Values are means ± SD.

/

I agonist

I basal

I

+/

/

I agonist

3

+/

I

3

I

/

I

I

I

/

/

NSP4-stimulated I influx into crypt cells was age dependent and largely independent of CFTR expression. NSP4 protein (100 nM) was found to elicit changes in plasma membrane I influx in crypts isolated from wild-type and heterozygous mouse CFTR genotypes. However, unlike either FSK or CCh, the effects of NSP4 were clearly age dependent (Table 5). In the pup crypt, NSP4 stimulated plasma membrane I influx by 2.7-fold, whereas, in the adult crypt, NSP4-stimulated I influx was <1.1-fold. Corresponding P values for J_I agonist vs. J_I basal were significantly different for pup (P < 0.001) but not adult (P > 0.05) crypts. Net changes in NSP4-elicited influx rate were as follows: J_I pup = 0.67 ± 0.42 M³/s vs. J_I adult = 0.06 ± 0.02 M³/s, and were significantly different (P < 0.001, n = 5; Table 5). When the NSP4-induced change in I influx rate was displayed in histogram format, peak changes in influx rate had a lag of 10 s after introduction of I into the bath, similar to that recorded for both FSK and CCh (Fig. 7A).

View larger version (26K): [in this window] [in a new window]   Fig. 7.   Change in SPQ fluorescence quench rate in mouse distal colon crypts elicited by NSP4 (100 nM). A: values obtained form pooled adult and pup CFTR+/+ and CFTR+/ mouse crypts, demonstrating that NSP4 effects were age dependent. B: values obtained from pooled CFTR+/+ and CFTR+/ and CFTR/ pup crypts, demonstrating that effects of NSP4 on SPQ quench rate were independent of CFTR expression. Values are means ± SD.

/

I agonist

I basal

I agonist

+/

I

/

I

+/

I

/

NSP4-induced I influx into pup crypt cells was dependent on [Ca²⁺]_i mobilization. To investigate the relationship between NSP4-induced Ca²⁺ mobilization and plasma membrane I influx, J_I basal, J_I agonist, and J_I were measured in CFTR^+/+ and CFTR^+/ mouse pup crypts loaded and bathed in Ca²⁺-free HEPES-buffered saline (nominally Ca²⁺-free plus 1 mM EGTA). Basal I influx rate was not appreciably lower in this instance than in crypts bathed with standard (1 mM CaCl₂) extracellular saline (J_I basal Ca²⁺-free = 0.32 ± 0.08 M³/s vs. J_I basal Ca²⁺-containing = 0.4 ± 0.14 M³/s, values were not significantly different, P > 0.05, n = 4 and 15, respectively; Table 5). However, in all crypts the absence of bath Ca²⁺ completely inhibited NSP4-mediated I influx (J_I Ca²⁺-free = 0.02 ± 0.06 M³/s vs. J_I Ca²⁺-containing = 0.67 ± 0.42 M³/s, P < 0.001, n = 4 and 5, respectively; Table 5). These findings demonstrated that NSP4-stimulated plasma membrane I influx was dependent on [Ca²⁺]_i mobilization.

NSP4 protein-induced I influx in mouse pup crypt cells was inhibited by DIDS. The effectiveness of NSP4, but not CCh, at eliciting significant I influx into SPQ-loaded CFTR^/ mouse pups raised the question as to the nature of this novel conductive pathway. To begin to characterize this pathway, crypts were preexposed to 0.5 mM DIDS and then washed before experimentation. DIDS inhibits a variety of Cl channels identified in epithelial cells, including the outward-rectifying Cl channel, the swelling-activated Cl conductance channel, and the Ca²⁺-sensitive Cl channel, but does not affect cAMP-stimulated CFTR Cl channels (reviewed in Ref. 19). DIDS also inhibits a variety of organic osmoltye exchange mechanisms (52) and the Cl/HCO₃ exchanger (63). Basal I influx rate in the presence of DIDS was lower than controls but was not significantly different (J_I basal DIDS = 0.29 ± 0.13 M³/s vs. J_I basal control = 0.4 ± 0.14 M³/s, P > 0.05, n = 5; Table 5). This general halide transport inhibitor completely abolished the NSP4-elicited increase in plasma membrane I influx in all crypts (J_I agonist DIDS = 0.34 ± 0.14 M³/s vs. J_I agonist control = 01.07 ± 0.32 M³/s, P < 0.001, n = 4), and the net change in I influx rate reflected this difference (J_I DIDS = 0.05 ± 0.06 vs. J_I control = 0.67 ± 0.24, P < 0.001, n = 4; Table 5). Thus the ontogenically regulated NSP4-activated plasma membrane I influx pathway present in mouse pup crypts was DIDS sensitive.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Relationship between the diarrhea effects of infectious rotavirus and NSP4. The induction of diarrhea by NSP4 protein and NSP4_114-135 peptide compared with active rotavirus indicates that the enterotoxic effects of NSP4 may account for the initial diarrheal phase of rotaviral infection (3). During this period, diarrhea occurs without significant ultrastructural damage to the small intestinal mucosa (38, 57). This contrasts with diarrhea seen during the later stages of rotaviral infection in some animal species, in which villus blunting has been suggested to contribute to diarrhea by reducing small intestinal absorptive area (5). This later histopathological feature of rotaviral infection is not seen in all animal models; in the mouse, significant villus blunting is rare (38, 49). Therefore, loss of absorptive area is likely to be auxiliary to the underlying pathophysiological basis of diarrhea in this model. Supporting this hypothesis, we have shown that NSP4 and its active NSP4_114-135 peptide fail to cause significant histological damage to the mucosa when injected intraluminally or interperitoneally into wild-type mice (unpublished observations and Ref. 3).

Correlation between [Ca²⁺]_i mobilization and fluid transport in the gut. Addition of Ca²⁺-mobilizing agonists such as CCh to intestinal mucosal sheets from either CFTR^/ mice or CF patients fails to evoke a Cl secretory current (6, 37, 53). This lack of response suggests that the cellular effects of Ca²⁺ mobilization on the Cl secretory current in the gastrointestinal tract occur secondary to cAMP-dependent activation of CFTR channels and are limited to the upregulation of basolateral membrane ion transporters and K⁺ channels (51, 60). Results from NSP4-induced [Ca²⁺]_i mobilization in both native small and large intestinal crypts (Fig. 2, A-C, Fig. 3, and Table 2) support the hypothesis that the enterotoxic activity of NSP4 in normal mucosa may be partly due to the secondary effects of increased cellular Cl uptake at the basolateral plasma membrane and cellular hyperpolarization. These facilitating but not controlling mechanisms for intracellular Cl conduction across the luminal membrane do not explain all of our observations. Two important questions relating to the role of NSP4-induced [Ca²⁺]_i mobilization, namely the age-dependent effects of NSP4 on Cl secretion and diarrhea and the diarrheal effects of NSP4 in CF mice, remained