The Envelope Glycoprotein of Simian Immunodeficiency Virus Contains an Enterotoxin Domain

Virology 277, 250–261 (2000) doi:10.1006/viro.2000.0626, available online at http://www.idealibrary.com on

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector The Envelope Glycoprotein of Simian Immunodeficiency Virus Contains an Enterotoxin Domain

C. L. Swaggerty,* A. A. Frolov,*,1,2 M. J. McArthur,*,2 V. W. Cox,* S. Tong,† R. W. Compans,† and J. M. Ball*,3

*Department of Pathobiology, College of Veterinary Medicine, Texas A&M University, College Station, Texas 77843-4467; and †Department of Microbiology and Immunology, Emory University School of Medicine, Atlanta, Georgia 30322 Received February 3, 2000; returned to author for revision April 7, 2000; accepted September 6, 2000

By the use of a mouse model, the enteropathic effects of the simian immunodeficiency virus (SIV) surface unit (SU) envelope glycoprotein were explored. Purified SU (0.01–0.45 nmol) was administered intraperitoneally to 6- to 8-day-old mouse pups and induced a dose-dependent diarrheal response. Surgical introduction of SU into adult mouse intestinal loops revealed fluid accumulation without histological alterations and SU-treated unstripped intestinal mucosa induced chloride (ClϪ) secretory currents in Ussing chambers. Similarly to rotavirus NSP4, the first described viral enterotoxin, SU induced a

transient increase in intracellular calcium levels and increased inositol 1,4,5-triphosphate (IP3) levels in HT-29 cells. These

data indicate the calcium response is mediated by IP3. The presence of diarrhea and fluid accumulation within intestinal loops in the absence of histological alterations and induction of ClϪ secretory currents demonstrate that SIV contains an enterotoxic domain localized within SU and is the second viral enterotoxin described. © 2000 Academic Press Key Words: calcium mobilization; chloride secretion; diarrhea; envelope glycoprotein; SIV; viral enterotoxin.

INTRODUCTION al., 1998; Kuhn et al., 1997; Coue¨del-Courteille et al., 1997;

4 Smit-McBride et al., 1998; Heise et al., 1993b). The GI Simian immunodeficiency virus (SIV ) is a good model tract contains a majority of the lymphoid tissue in the for studying human immunodeficiency virus (HIV) infection, pathogenesis, and progression to disease (Veazey body, including a large population of activated-memory T et al., 1998; Desrosiers and Ritter, 1989; Daniel et al., lymphocytes, the cell subset in which HIV replicates 1985). SIV disease in macaques is comparable to human most efficiently (Veazey et al., 1998; Kraehenbuhl and acquired immunodeficiency syndrome (AIDS) and in- Neutra, 1992; Zeitz et al., 1988; Cerf-Bensussan and Guy- duces similar pathological and morphological changes Grand, 1991). In addition, galactosylceramide (GalCer) is in the intestinal mucosa (Heise et al., 1991, 1993a; Kotler a glycosphingolipid that is expressed in intestinal tis- et al., 1984; Ullrich et al., 1989; Letvin et al., 1985; Kurtz et sues and mediates HIV-1 gp120 binding via a domain in al., 1985; Nelson et al., 1988; Kaup et al., 1998). the V3 loop (Hammache et al., 1998; Bhat et al., 1991; SIV infects mononuclear cells throughout the gastro- Fantini et al., 1993; Cook et al., 1994; Yahi et al., 1995). intestinal (GI) tract of rhesus macaques (Heise et al., HIV-1 surface unit (SU) glycoprotein reacts preferentially ␣ 1993a,b) and SIV-infected macaques have infected mac- with hydroxylated GalCer, the predominant GalCer rophage and T lymphocytes in gut-associated lymphoid species expressed by HT-29 cells, a colonic epithelial tissue as early as 1 week postinfection (Heise et al., cell line (Hammache et al., 1998). The presence of the 1994). Further, CD4ϩ T lymphocytes localized in the GI GalCer receptor and a large number of activated T lym- tract are depleted as early as 7 days post-SIV infection phocytes in the intestine may facilitate infection and (Veazey et al., 1998; Kewenig et al., 1999). Others report replication of HIV and/or SIV. the GI tract is a primary target for SIV infection (Veazey et Alterations in GI function, including chronic diarrhea, malabsorption, and significant weight loss are devastating characteristics of infection with HIV or SIV and ac- 1 Present address: Washington University School of Medicine, Cen- count for the frequently observed wasting syndrome in ter for Cardiovascular Research, Cardiovascular Division, St. Louis, MO AIDS or simian AIDS (SAIDS) (Ehrenpreis et al., 1992; 63110. Heise et al., 1993a). There is a wide range of variability in 2 These authors contributed equally to this work. 3 To whom correspondence and reprint requests should be ad- the number of cases reported in which no identifiable dressed. Fax: (979) 845-9231. E-mail: [email protected]. enteric pathogen can be attributed to the GI dysfunction 4 2ϩ Abbreviations used: [Ca ]i, intracellular calcium concentration; associated with SIV/HIV infections (Heise et al., 1991, DD50, diarrhea dose 50; GalCer, galactosylceramide; GI, gastrointesti- 1993b; Ullrich et al., 1989; Lambl et al., 1996). It has been nal; HIV, human immunodeficiency virus; IP, intraperitoneally; NSP4, nonstructural protein 4; PI, phosphatidylinositol; Sf9, Spodoptera frugi- postulated that HIV infection and/or replication, the perda; SIV, simian immunodeficiency virus; SU, surface unit; wt, wild- virion, or a viral protein could be responsible for the HIV type baculovirus. enteropathy (Kotler et al., 1984). To determine whether

FIG. 1. SU induced a dose-dependent diarrhea in 6- to 8-day-old Balb/C mouse pups. Each dose was administered in 50 ␮l of sterile PBS and was determined to be free of endotoxin. The x-axis shows the dose of SU administered and the y-axis shows the percentage of pups that developed diarrhea. The number of animals that developed diarrhea/total number of animals given a particular dose is shown above each dose administered. Control pups were inoculated with purified PBS, wt protein, or HPIV2 F protein (1 nmol), purified exactly like SIV SU, and did not develop diarrhea.

The calculated DD50 of SU was 0.13 nmol based on the calculations of Reed and Muench (1938).

HIV/SIV has a direct role in GI dysfunction and diarrhea, neurotensin and NSP4. Thus, SU alone may be sufficient infected subjects who had no secondary infections and to alter intestinal ion secretion and contribute to the suffered with chronic HIV-related diarrhea were admin- pathogenesis of HIV and SIV infections by acting as a istered antiretroviral protease inhibitors. These patients viral enterotoxin. We propose SIV SU encodes a toxic reported a significant decrease in stool frequency, in- domain that contributes to the associated GI dysfunction crease in body weight, decrease in recurrence of diar- and subsequent disease, in particular in the absence of rhea, and a longer mean survival time (Bini and Cohen, other enteric infections. 1999; Foudraine et al., 1998). The fact that the GI dysfunction improves upon treatment with protease inhibi- RESULTS tors strongly suggests a direct involvement for HIV-in- SU induced a dose-dependent diarrhea in a mouse duced GI dysfunction. pup model The first viral enterotoxin, rotavirus NSP4, was reported in 1996 and provides a basis for the isolation and SU was inoculated intraperitoneally (IP) into 6- to identification of other viral toxins (Ball et al., 1996). Ente- 8-day-old Balb/C mouse pups and examined for diar- rotoxins stimulate net secretion in ligated intestinal seg- rheagenic activity. Six- to 8-day-old pups were chosen ments without inducing histological alterations, or stim- because their intestines have not yet undergone closure ulate secretion as measured in Ussing chambers (Sears (the formation of tight junctions). Closure of the intestines 2ϩ et al., 1995; Ball et al., 1996). NSP4 mobilizes [Ca ]i via reduces the uptake of materials administered IP into the phospholipase C activation and inositol 1,4,5-triphos- abdominal cavity (Henning and Leeper, 1984; Ball et al., Ϫ phate (IP3) production and induces Cl secretory cur- 1996). SU induced diarrhea in our mouse model when rents (Dong et al., 1997; Ball et al., 1996). Likewise, HIV-1 pups were administered 0.30 nmol of protein. The diar- 2ϩ SU increases [Ca ]i in cultured epithelial intestinal cells rhea response to SU was specific, as administration of (Dayanithi et al., 1995) and induces diarrhea in mouse an equimolar concentration of purified baculovirus wild- pups (J. M. Ball and M. K. Estes, unpublished data). type (wt) protein or 1 nmol purified human parainflu- Addition of neutralizing antibodies directed against the enza-2 virus fusion protein (HPIV2 F) expressed in a 2ϩ V3 loop ablates the increase in [Ca ]i. The response is baculovirus system and purified as SU, or the same also diminished by pretreating the cells with neuroten- volume of buffer, had no effect (Fig. 1). To determine Ϫ 2ϩ sin, a stimulator of Cl secretion via IP3-mediated Ca whether the response was dose-dependent, varying mobilization, or by administering SU immediately fol- doses of SU were administered (Fig. 1). The incidence of lowed by neurotensin (Dayanithi et al., 1995). These data diarrhea ranged from 0% to 100%, following administra- 2ϩ indicate SU mobilizes [Ca ]i by a pathway common to tion of 0.01 and 0.45 nmol of purified SU, respectively. 252 SWAGGERTY ET AL.

TABLE 1 cific, 0.03- to 0.45-nmol samples of purified SU were SIV SU-Promoted Fluid Accumulation in Adult Balb/C Mouse administered IP to 7-day-old CD-1 outbred mouse Intestinal Loops pups (n ϭ 25). Diarrhea in the CD-1 pups was observed, although the instance of diarrhea was less Treatment PBS HPIV2 F SIV SU than that seen in the Balb/C pups (0% for the 0.03-nmol n 12 4 11 dose, 10% for the 0.3-nmol dose, 20% for the 0.45-nmol Dose (nmol) 1 0.3 dose). The rate of diarrhea was similar to that ob- Average weight (g) served in the Balb/C animals in that onset was within per length (cm) 0.0467 0.0411 0.0550 2 h and persisted for up to 8 h post injection. The Range (g/cm) 0.0376–0.0702 0.0352–0.0442 0.0387–0.1081 Standard deviation severity of the diarrhea was also comparable to the (SD) of g/cm values 0.0089 0.0042 0.0193 Balb/C pups. CD-1 pups administered purified wt pro- Number of animals tein or PBS did not develop diarrhea. These data that differ from the indicate that the enterotoxic effect was not specific to HPIV2 F control by 2 SD/total number Balb/C pups; however, genetic make-up of the animals of animals may contribute to the number of animals which expe- examined (%) 1/12 (8.3%)a 7/11 (63.6%)b rience diarrhea as seen with NSP4 (J. M. Ball and M. K. Estes, unpublished data). a One of 12 animals statistically different from the HPIV2 F control based on Student’s t test (P Ͻ 0.05). b Of animals examined, 63.9% are statistically different from the SU induced fluid accumulation in mouse intestinal HPIV2 F control (P Ͻ 0.05, Student’s t test). loops SU was surgically introduced into 4- to 6-week-old The diarrhea dose 50 (DD ) was determined to be 0.13 50 Balb/C mouse intestinal loops to monitor protein-specific nmol, based on the calculations of Reed and Muench fluid accumulation. The average SU-treated loops were (1938). Six to 10 pups were tested at each dose of SU, 21 0.0550 g/cm compared to an average of 0.0411 and pups were examined with wt control or phosphate-buffered saline (PBS), and an additional nine pups were 0.0467 g/cm for the HPIV2 F-treated and PBS control- administered 1 nmol of HPIV2 F protein (greater than treated intestinal loops, respectively (Table 1). Seven of twice the highest SU dose administered) and did not 11 (63.6%) of the SU-treated loops were statistically develop diarrhea. The onset of diarrhea induced by SU heavier/cm as the result of an accumulation of fluid when was typically within 2 h post exposure and continued for compared to HPIV2 F control-treated loops (Student’s t up to 8 h, occasionally for up to 12 h. When SU-treated test, P Ͻ 0.05). The SU-treated mice often had diarrhea animals had severe diarrhea (3 to 4ϩ), they often exhib- during recovery. This observation was surprising since ited lower body temperatures and/or were lethargic com- the SU inoculum was administered directly into the in- pared to those animals administered the control inocu- testinal segments. Fluid accumulation in adult mice in- lum. testinal loops supports the conclusion that the SU-in- To ensure the response was not mouse strain-spe- duced diarrhea is not age-dependent, as would be ex-

FIG. 2. Hematoxylin and eosin staining of the mucosa from the mouse intestinal loops. (A) PBS-treated loop with a normal population of lymphocytes (white, closed arrow) and goblet cells (white, open arrow). The columnar epithelium and the microvilli brush-border are intact (black, closed arrow). (B) SU (0.3 nmol)-treated loop with a normal number of lymphocytes (white, closed arrow) and goblet cells (white, open arrow). The columnar epithelium and microvilli brush-border (black, closed arrow) are unaltered, indicating there are no histological alterations induced by the addition of SU to the intestinal loop. (C) HPIV2 F (1 nmol) protein-treated loop with an equal number of lymphocytes (white, closed arrow) and goblet cells (white, open arrow). The columnar epithelium and the microvilli brush-border remained intact (black, closed arrow). SIV SU IS A VIRAL ENTEROTOXIN 253

2ϩ 2ϩ FIG. 3. Effects of neurotensin on [Ca ]i in Fluo-4-loaded HT-29 cells. Images taken from a representative experiment demonstrating [Ca ]i is 2ϩ mobilized upon the addition of 5 nM neurotensin. (A) Basal [Ca ]i level prior to the addition of neurotensin (0 s). (B) At 28 s after the addition of 2ϩ 2ϩ neurotensin, maximum [Ca ]i was mobilized (600 nM). (C) The response was transient as [Ca ]i had returned to resting levels by 147 s after the 2ϩ addition of the neurotensin. Basal [Ca ]i was subtracted from values shown. 2ϩ 2ϩ FIG. 4. SU induced a dose-dependent increase in [Ca ]i in Fluo-4-loaded HT-29 cells: (A) 125 nM SU increased the [Ca ]i level to 48 nM; (B) 150 2ϩ nM SU, to 651 nM; and (C) 300 nM SU, to 449 nM. Basal [Ca ]i was subtracted from values shown.

pected, since SIV infection and enteropathy are not age- the mucosa of any specimens examined from the dependent. treatment or control groups.

SU does not induce histological alterations in SU induced intracellular calcium mobilization in intestinal loops cultured HT-29 colonic cells

2ϩ Upon removal of the intestinal loops in the preced- A calibrated assay that measures [Ca ]i concentra- ing study, the tissues were analyzed for the presence tion in response to a stimulus was utilized to determine of histological alterations by hematoxylin and eosin whether SU elicited a Ca2ϩ-specific response in cultured, staining (Fig. 2). The mucosa of the SU-treated intes- Fluo-4-loaded HT-29 cells. A dose response of 1–500 nM tinal loops appeared identical (Fig. 2B) to the PBS- neurotensin was tested with the Fluo-4-loaded HT-29 treated and HPIV2 F control-treated loops (Figs. 2A cells. The optimal concentration for our system was 5 nM 2ϩ and 2C, respectively). Minimal accumulation of neu- neurotensin, which was used in subsequent [Ca ]i mo- trophils was observed within and around vessels of bilization studies as the positive control (Dong et al., the mesentery in both treated and nontreated intesti- 1997; Bozou et al., 1989; Morris et al., 1990). HT-29 cells nal tissues. The accumulation of neutrophils was con- express the neurotensin receptor (Dong et al., 1997). sidered to be a secondary result due to the manipu- Neurotensin (5 nM) added to Fluo-4-loaded HT-29 cells 2ϩ lation involved in tying off the loops and subsequent induced a rapid and transient increase in [Ca ]i (Fig. 3). 2ϩ handling during injection of SU or controls. No signif- The onset of [Ca ]i mobilization was seen 20 s after the 2ϩ icant lesions or other alterations were observed within addition of agonist with a maximum [Ca ]i response 254 SWAGGERTY ET AL.

FIG. 5. Average kinetics and dose response of the calcium response FIG. 6. SU induced a dose-dependent increase in IP3 levels in HT-29 of five representative cells treated with 5 nM neurotensin (E), 125 nM cells. The x-axis is the treatment administered (all were 45 s with n ϭ ϫ छ F SU ( ),150nMSU( ), 300 nM SU ( ), and HPIV2 F (—). Fluorescence 3 or 4). The y-axis is the fold increase in IP3 levels above background intensity values are within 2 SD of the mean for all cells analyzed. levels in untreated cells. Addition of 75 and 150 nM SU increased IP3

levels 1.3- and 1.6-fold, respectively. Neurotensin (10 nM) increased IP3

levels 1.7-fold. Wt baculovirus protein and HPIV2 F did not increase IP3 2ϩ levels above the basal levels. (600 nM) after 28 s. [Ca ]i returned to baseline fluorescence levels by 147 s. The basal fluorescence (63 nM) was subtracted from all values shown. To determine ϩ nM SU increased IP levels an average of 1.3- and 1.6- whether SU mobilized [Ca2 ] , varying doses of purified 3 i fold, respectively, above basal levels in control cells (n ϭ protein (10–300 nM) were added to Fluo-4-loaded HT-29 2ϩ 3). Neurotensin (10 nM) served as the positive control cells and [Ca ]i levels were monitored. As shown in Fig. ϩ and increased IP levels 1.7-fold above basal levels. 4, SU induced a dose-dependent increase in [Ca2 ] with 3 i Cells treated with wt protein or 150 nM HPIV2 F protein 125–300 nM. Addition of 125, 150, and 300 nM SU in- ϩ (n ϭ 4) yielded IP levels comparable to those of mock- creased [Ca2 ] by 48, 651, and 449 nM, respectively 3 i treated cells (Fig. 6). (basal fluorescence subtracted). These data indicate the ϩ 2 Ϫ maximal [Ca ]i was mobilized at a dose of 150 nM SU, SU induced Cl secretory currents since an additional increase was not seen with 300 nM Ϫ SU. The differences in the effective dose between SU To fulfill the definition of an enterotoxin, Cl secretory and neurotensin may be a result of the number of spe- currents were measured across unstripped mouse intes- cific receptors on HT-29 cells, but further studies are tinal mucosa. To ensure tissue reactivity to a known required to determine the level of SU-specific receptors. agonist, forskolin (Fsk), a cAMP agonist (Frizzell and A 150-nM sample of HPIV2 F protein did not mobilize Morris, 1994) was added prior to the addition of purified [Ca2ϩ] under the same conditions (n ϭ 5) (data not SU (0.2 ␮M) or HPIV2 F (1.5 ␮M) proteins. SIV SU induces i Ϫ shown). Cl secretion in mouse ileal sections as measured by To further characterize the Ca2ϩ response, the kinetics Ussing chamber. Following the addition of Fsk, purified of SU- and neurotensin-treated HT-29 cells was exam- 2ϩ ined. The kinetics of the SU-induced [Ca ]i response was similar to neurotensin, i.e., time of onset was 15 s vs TABLE 2 2ϩ 20 s, maximum [Ca ]i was reached at 30–35 s vs 28 s, SIV SU-Induced Chloride Currents in Mouse Intestinal Mucosa 2ϩ and [Ca ]i returned to baseline by 140 s vs 147 s (Fig. 5). The 125 nM SU-treated cells were slower and variable in Agonista Concentration (␮M) ⌬Isc (␮A/cm2)b Ϯ SD n the onset of the response (40–70 s) and returned to Forskolin 10 23.73 Ϯ 4.48 24 baseline levels by 120 s (Fig. 5). HPIV2 F-treated cells did SU 0.2 6.32 Ϯ 1.65 7 not respond and the intensity remained at background HPIV2 F 1.5 3.44 Ϯ 2.28 7 levels (Fig. 5). a Forskolin was added to the apical surface. SU and HPIV2 F were added to both surfaces of the mucosa. SU increased IP3 levels in HT-29 colonic cells b The change in short-circuit current (⌬Isc) was calculated by sub- A competitive radioreceptor assay was utilized to mea- tracting the stimulated Isc from the Isc measured immediately prior to the addition of agonist. The mean tissue conductance immediately sure IP3 levels in HT-29 cells treated with SU, wt protein, following the buffer change (apical buffer changed from Krebs–Ringer neurotensin, HPIV2 F, or mock-treated. SU increased IP3 bicarbonate to N-methyl-D-glucamine-Krebs sodium-free) was 23.08 Ϯ levels in a dose-dependent manner (Fig. 6). At 75 and 150 5.96 mS/cm2 (n ϭ 29). SIV SU IS A VIRAL ENTEROTOXIN 255

SU was added apically and basolaterally and induced a et al., 1993). Other studies indicate that an increased ⌬Isc of 23.73 Ϯ 4.48 ␮A/cm2 (Table 2). Addition of 1.5 ␮M level of TNF-␣ or IL-1␤ can induce active ClϪ secretion in HPIV2 F (a 7.5-fold increase in molar concentration over the intestine (Stockman et al., 1998; Schmitz et al., 1996; SU), both apically and basolaterally, induced a smaller Chang et al., 1990). Immune mediators such as TNF-␣ ClϪ response than that seen upon the addition of SU and IFN-␥ are also implicated in altering tight-junction (Table 2). All responses to agonist were sensitive to permeability (Stockman et al., 1998; Madara and Stafford, bumetanide, an antagonist to the Na-K-Cl cotransporter 1989). Moreover, malabsorption, a common complication on the basolateral surface. associated with SIV/HIV infections, is also associated with diarrhea (Heise et al., 1991, 1993a,b, 1994). Our data DISCUSSION showing secretory currents across the intestinal mucosa suggest SIV SU induces a secretory diarrhea. However, SIVmac239 is a virus derived from an infectious mo- examination of duodenal forceps biopsies of HIV-in- lecular clone that causes a fatal disease in rhesus ma- fected individuals suffering from diarrhea show no etio- caques clinically similar to HIV-induced AIDS (Kodama et logical agent for active ion secretion in those patients al., 1989; Kestler et al., 1990). We examined the role of SU (Stockman et al., 1998). Instead, there is a decrease in derived from SIVmac239 in the GI dysfunction and enter- epithelial resistance, indicating a breakdown in the epi- opathy associated with this group of viruses. In both thelial barrier function or a leak flux mechanism. Addi- HIV/SIV-infected individuals there are cases in which no tional studies are needed to fully understand the mech- enteric pathogen can be identified that would account for anism(s) of HIV-induced diarrhea and wasting in the the diarrheal disease (Heise et al., 1991, 1993b; Ullrich et absence of an identifiable agent. al., 1989; Lambl et al., 1996). For these cases, it has been Our study examined the role of SU added to (i) the speculated that the virus or viral protein(s) may directly peritoneum of 6- to 8-day-old mouse pups, (ii) the lumen contribute to the GI dysfunction and diarrhea. In support of mouse intestinal loops, (iii) the unstripped mouse of this hypothesis, IP administration of purified HIV-1 SU intestinal mucosa, and (iv) the media of cultured HT-29 (n ϭ 4) into suckling mice was shown to induce diarrhea colonic cells. In a natural, productive infection with SIV, (J. M. Ball and M. K. Estes, unpublished data). SU would be present inside the cell before it is trans- In addition to the critical role of viral entry and patho- ported to the cell surface. The different cellular locations genesis of HIV/SIV infections, SU may be important in of SU could affect the overall role this protein plays in disease induction. The SU is noncovalently bound to the diarrhea induction. Cells treated with extracellular SU transmembrane (TM) protein and this weak bond is eas- may respond differently from those cells undergoing a ily broken, allowing SU to dissociate from TM (Robey et productive viral infection, in which SU would be pro- al., 1986). Free SU can induce immune-mediated killing duced intracellularly, transported to the cell surface, and of uninfected target cells (Hunter and Swanstrom, 1990) released with viral virions. However, the free SU found and bind to CD4ϩ cells, rendering them susceptible to outside the cell would mimic the SU that we added antibody-dependent cellular toxicity (Lyerly et al., 1987) extracellularly. The mechanism of diarrhea is likely to be or to class II restricted T-cell-mediated cytolysis (Lanza- multifactorial and may include factors such as the strain vecchia et al., 1988). of virus, route of transmission, viral load, and the immune To address the role of SU in SIV- and HIV-related competency of the individual host at the time of sam- enteropathy, we tested whether a purified SIVmac239 SU pling. Infected cells are subject to dynamic events asso- recombinant protein could cause diarrhea. The purified ciated with a productive infection that would vary greatly SU glycoprotein induced a dose-dependent diarrhea in from those cells treated only with SU. The purified SU 6- to 8-day-old Balb/C mouse pups, caused fluid accu- alone may be sufficient to bind to another, as yet uniden- 2ϩ mulation in adult mouse intestinal loops in the absence tified, cell surface receptor that in turn mobilizes [Ca ]i 2ϩ of histological alterations, mobilized [Ca ]i in cultured and activates a signal transduction cascade. Shimizu colonic cells, increased IP3 levels in cultured cells, and and colleagues (1999) showed that an orphan G protein- induced ClϪ secretory currents in intact mouse mucosa. coupled receptor can serve as a functional coreceptor These data are consistent with the findings described for for HIV productive infection in brain-derived cells. Single- the first viral enterotoxin, rotavirus NSP4 (Dong et al., point mutations in the Gly-Pro-Gly-Arg tip sequence in 1997; Ball et al., 1996), and define SIV SU as the second the V3 loop domain enable these viruses to use a phy- described viral enterotoxin. logenetically distant coreceptor in the absence of CXC The mechanism(s) of SIV/HIV enteropathy is not fully and/or CC coreceptors (Shimizu et al., 1999, 2000). More- understood. Inflammatory mediators such as tumor ne- over, a different novel G protein-coupled receptor also crosis factor (TNF)-␣, interleukin (IL)-1␤, and interferon serves as a coreceptor for some HIV-1, HIV-2, and SIV (IFN)-␥ levels are increased in the intestinal mucosa strains (Shimizu et al., 2000). However, a specific cholin- upon HIV infection and may play a role in HIV enterop- ergic receptor that activates the phosphoinositide (PI) athy (Vyakarnam et al., 1991; McGowan et al., 1994; Kotler pathway upon SIV SU binding is not yet known. 256 SWAGGERTY ET AL.

Our data vary with an earlier study that examined the the SU enterotoxin and known enteric pathogens are ability of gp120 from M- and T-tropic strains of HIV and both contributing to the wasting syndrome seen in HIV/ SIV to initiate a signal upon binding to CCR5 (Weissman SIV-infected individuals. Consistent with other HIV/SIV et al., 1997). The latter study concluded that binding of pathogenic events, the mechanism(s) of viral enteropa- CCR5 by M-tropic strains causes a calcium flux, whereas thy is likely complex, combining Ca2ϩ flux, immune me- binding by the T-tropic strains such as SIVmac239 does diators, and the PI signaling pathway. not. The different results may be attributed to the differ- Identification of SU as a viral enterotoxin enables di- ent cells used in the experiments (lymphoid vs epithe- rect testing of this mechanism of SU-induced diarrhea lial). In a recent study, Arthos and colleages (2000) dem- and allows for the development of potential therapeutics onstrated recombinant HIV and SIV SU glycoproteins directed at preventing or controlling the devastating 2ϩ mobilized [Ca ]i in human macrophages; however, SU wasting/enteropathy associated with this group of vi- 2ϩ derived from SIVmac239 did not increase [Ca ]i in their ruses. system, whereas SU from SIVPBj1.9 did. The authors 2ϩ suggest the lack of [Ca ]i mobilization resulted from the MATERIALS AND METHODS inability of SIVmac239 to replicate in macrophages. Our 2ϩ data show [Ca ]i mobilization in HT-29 intestinal cells, Cell culture which were previously shown to support HIV-1 and HIV-2 Spodoptera frugiperda (Sf9) insect cells, a gift from Dr. infections (Yahi et al., 1992; Kagnoff and Roebuck, 1999). M. K. Estes (Baylor College of Medicine, Houston, TX) However, it is not known whether HT-29 cells support were maintained in suspension at 26°C in SF-900 II SIVmac239 replication. serum-free media (Gibco BRL Products, Gaithersburg, We demonstrated that SU induced a transient flux in ϩ MD) containing 100 units of penicillin (Sigma Chemical, [Ca2 ] and increased IP levels in cultured colonic cells. i 3 St. Louis, MO) and 10 mg/ml streptomycin (Sigma), and Neurotensin, a neuropeptide that causes a transient in- 2ϩ were rotated at 85–90 rpm. Sf9 cells were used to grow crease in [Ca ]i levels mediated by PI turnover, also 2ϩ baculovirus stocks and for baculovirus recombinant pro- mobilized [Ca ]i and increased IP3 levels in our system. 2ϩ tein expression. HT-29 cells, human colon adenocarci- Given that SIV SU induces [Ca ]i mobilization, increases Ϫ noma (ATCC, Rockville, MD), were maintained in 90% IP levels, and stimulates Cl secretion, we propose that 3 McCoy’s 5a medium supplemented with 2 mM L-glu- phospholipase C (PLC) and a PI signaling pathway are tamine, 0.1 mM nonessential amino acids, 100 units of activated. PI turnover as a mechanism of calcium and IP 3 penicillin, 10 mg/ml streptomycin, and 10% fetal bovine increase is in agreement with the NSP4 viral enterotoxin serum (FBS; HyClone Laboratories, Logan, UT) in a hu- (Dong et al., 1997). mid environment at 37°C with 5% CO Calcium is important in many aspects of the virus life 2. cycle and was shown to contribute to retrovirus-induced 2ϩ Baculovirus stocks disease. HIV-1 SU mobilizes [Ca ]i from caffeine-sensitive stores in HT-29-D4 cells (Dayanithi et al., 1995). Wildtype (wt) baculovirus, Autographa californica nu- Additional evidence to suggest that Ca2ϩ contributes to clear polyhedrosis virus (AcNPV), was provided by Dr. lentivirus pathogenesis is supported by the fact that M. K. Estes. Generation of the SU baculovirus recombi- HTLV-1- and SIV-infected cells display an increase in nant has previously been described (Yamshchikov et al., 2ϩ [Ca ]i (Wright and Olsen, 1989). Diop and colleagues 1995). Briefly, the env gene was placed under the control (1995) reported that HIV-1 SU is toxic for neurons in of a hybrid capsid/polyhedrin promoter (Pcap/polh) and culture in a dose-dependent manner and that toxicity introduced into Sf9 cells with AcNPV. Those recombi- 2ϩ corresponds with a rise in [Ca ]i. Taken together, these nants that expressed the env gene were selected by 2ϩ findings clearly suggest that alterations in [Ca ]i are plaque assay and subsequently analyzed by PCR. part of the mechanism of SIV and/or HIV disease. SIV The vector pBlueBac4.5 (Invitrogen, San Diego, CA) 2ϩ SU-induced [Ca ]i mobilization suggests a role for a was used for construction of the recombinant baculovi- calcium-dependent signal transduction pathway that, in rus expressing the soluble human parainfluenza-2 fusion turn, may alter intestinal epithelial transport and contrib- protein (HPIV2 F) (Tong and Compans, 2000). The pBlue- ute to the diarrhea and wasting seen in SIV and/or HIV Bac4.5 plasmid contains the early-to-late promoter and infections. However, the precise role of calcium in enter- the very late polyhedrin promoter from AcNPV. A DNA opathy is yet to be determined. fragment encoding a soluble HPIV2 F protein was am- The presence of a viral enterotoxin in SIV and/or HIV plified by PCR from the pGEM-3-PI2 F template de- viruses offers one explanation for the GI dysfunction scribed by Hu et al. (1990). The sense primer consisted often associated with these viral infections. This expla- of an oligonucleotide corresponding to sequences up- nation is not intended to discount or replace those in- stream of the HPIV2 F protein coding sequence with a stances in which the GI disorders are associated with an modification to include a BamHI restriction site. The identifiable enteric pathogen. However, it is feasible that antisense primer corresponded to nucleotides SIV SU IS A VIRAL ENTEROTOXIN 257

1626–1612 of the pGEM-3-PI2 F plasmid with an introduction of a stop codon and an EcoRI restriction site. The resultant PCR fragment was digested with BamHI and EcoRI and was then ligated into the pBlueBac4.5 plasmid digested with BamHI and EcoRI. The recombinant clone was sequenced to ensure that the correct clone was obtained. Sf9 cells were cotransfected with Bac-N-Blue DNA and the vector pBlueBac4.5 encoding the soluble HPIV2 F protein using a MaxBac 2.0 Transfection kit (Invitro- gen). After a 5-day incubation, the media containing recombinant viruses were collected and a plaque assay FIG. 7. (A) SDS–PAGE of purified SU and wt proteins. Lane 1, molec- was performed according to the Invitrogen instruction ular weight markers; lane 2, wt protein; lane 3, SIV SU. Arrow indicates manual. Recombinant virus plaques that did not contain SU protein. The estimated purity of SU is 80%. (B) SDS–PAGE of purified HPIV2 F protein. Lane 1, molecular weight markers; lane 2, HPIV2 F occlusion bodies and showed blue color were selected protein. Arrow indicates HPIV2 F protein. and were screened by PCR amplification to determine which plaque had the expected size of the PCR products. The selected plaques were subjected to four rounds of (Pierce Chemical, Rockford, IL) according to the manu- plaque purification. Baculovirus recombinants express- facturer’s instructions (Fig. 7A). A 1-␮g sample of HPIV2 Ј ing soluble HPIV2 F are designated as rBV-HPIV2 F . F protein was loaded onto a 10% polyacrylamide gel SU and HPIV2 F recombinant and wt baculovirus containing 5 M urea, electrophoresed at 150 V for 1.5 h, ϫ 6 stocks were grown in Sf9 cells by infecting 3 10 and stained with Coomassie Brilliant Blue G-250 (Bio- cells/ml at a multiplicity of infection (m.o.i.) of 0.1. The Rad Laboratories, Hercules, CA) (Fig. 7B). The estimated supernatant from the infected cells was harvested 5–7 molecular weight of HPIV2 F was 54 kDa. The estimated days postinfection and the cell debris was removed by molecular weight of SU was determined to be 80 kDa centrifugation at 830 g for 10–15 min. Virus titers were and purity was estimated to be 80%. Purity of SU was determined by plaque assays as previously described by confirmed by HPLC analysis (Beckman System Gold; Volkman and Summers (Volkman and Summers, 1975; Beckman Instruments, Fullerton, CA), by use of a C4 Summers and Smith, 1976, 1988). All baculovirus stocks analytical column (Waters C4 Delta Pak column; Waters were stored at 4°C. & Associates, Milford, MA), a linear gradient of water and acetonitrile, and integration of the peaks (data not Protein expression, purification, and characterization shown). Contaminating proteins were controlled for in the wt and HPIV2 F controls. SU, wt baculovirus protein, SU or wt baculovirus proteins were expressed by in- 6 and HPIV2 F concentrations were determined by the fecting 2 ϫ 10 Sf9 cells/ml at a m.o.i. of 1. The super- Micro BCA Protein Assay Reagent Kit according to man- natant was harvested 24, 48, or 72 h postinfection. HPIV2 6 ufacturer’s instructions and were estimated to be 1.3, 0.4, F protein was expressed by infecting 2 ϫ 10 Sf9 cells/ml and 7 mg/ml, respectively (Pierce Chemical). The micro- at a m.o.i. of 2 and harvested after 72 h. Intact cells molar concentration of the wt baculovirus was unable to and/or cell debris were separated from soluble proteins be determined, since molecular weights of contaminat- by centrifugation at 830 g for 15–20 min. The baculovirus ing proteins were unknown; therefore, an equal amount virions were removed by ultracentrifugation at 75,000 g. (in micrograms) was administered. Purified protein and The supernatant containing soluble proteins was col- PBS were assayed for the presence of endotoxin using lected and passed through a Nalgene Maxi Centrifuge the Limulus Amebocyte Lysate (LAL) test (Associates of Filter (Nalge Nunc International, Rochester, NY) with a Cape Cod, Falmouth, MA) in endotoxin-free tubes ac- 30-kDa molecular weight cutoff. The filter apparatus was cording to the manufacturer’s protocol. A value of 0.5 transferred to a 50-ml conical tube (Corning, Oneonta, endotoxin units (EU)/ml or less was considered accept- NY) and centrifuged at 3000 g until 2–3 ml of solute able for use in our experiments. remained. A single filter was used to collect each time point. Once the protein was concentrated onto the filter, Mouse inoculation and diarrhea scoring the filter apparatus was removed, inverted into a new 50-ml conical tube, and centrifuged at 3000 g for 30 min Balb/C mice, acquired from Harlan Sprague–Dawley to remove the protein from the membrane of the filter (Indianapolis, IN), and CD-1 mice, from Charles River apparatus. A 1-␮g sample of SU and wt baculovirus Laboratories (Wilmington, MA), were housed at Texas protein was loaded onto a 12% polyacrylamide gel and A&M University Laboratory Animal Resources and Re- electrophoresed at 150 V for approximately 1.5 h. The SU search Facility. Purified SU was diluted to 50 ␮l in sterile, gel was stained with GelCode Blue Stain Reagent endotoxin-free PBS and was injected intraperitoneally 258 SWAGGERTY ET AL.

(IP) into 6- to 8-day-old pups using a 30G needle. To affinized, then rehydrated through graded alcohols (ab- determine whether the diarrheal response was dose- solute, 95%, 80%, 70%). Each section was stained with dependent, 0.01 to 0.45 nmol of SU was administered to instant hematoxylin (Shandon) and eosin (Sigma) and 54 pups. Wt baculovirus (28.9 ␮g) and HPIV2 F (1 nmol) coverslipped with Permount mounting media (Fisher Sci- supernatants were prepared identically to SU and served entific, Fairlawn, NJ) for microscopic examination utiliz- as negative controls. Following injection, the pups were ing an Olympus BH-2 microscope (Olympus Optical, monitored for diarrhea every 2 h for 10 h and then again Tokyo, Japan). at 24 h post injection. The pups were monitored by one person who scored the diarrhea on a scale of 1 to 4 as Loading HT-29 cells with Fluo-4 for calcium imaging previously described (Ball et al., 1996). HT-29 cells were seeded at 1.0 ϫ 10 6 cell/ml into LAB-TEK two-well chamber slides with coverglass slides Intestinal loops (2 ml/well; Nalge Nunc International) and allowed to To determine whether SU would induce fluid accumu- grow for 12–16 h. Cells were loaded using the following lation in closed intestinal segments, SU protein was modified protocol (Scheenen et al., 1998). The growth surgically introduced into mouse intestinal loops. Four- medium was removed and replaced with a physiological to 6-week-old Balb/C mice were fasted for 3–5 h, and calcium-free buffer (PCFB: 150 mM NaCl, 5 mM KCl, 10 anesthetized with IsoFlo isoflurane, USP (Abbott Labora- mM glucose, 20 mM HEPES, 1 mM MgCl) (Steinberg et tories, North Chicago, IL). An incision was made in the al., 1987) and incubated for 10–15 min at 37°C, 5% CO2. abdomen, the intestines exposed, and a short segment The buffer was removed and the cells were loaded with of the intestine (ileum and a portion of the jejunum) was 1 ml PCFB containing 3% FBS, 1 ␮M Fluo-4 (Molecular tied off using two coated 4-0 Vicryl polyglactin sutures Probes, Eugene, OR), and 0.02% Pluronic F-127 (Sigma). (Ethicon, Somerville, NJ). Eleven animals were adminis- The dye solution was allowed to incubate on the cells for tered 50 ␮l of 0.30 nmol SU dissolved in sterile, endo- 30 min at room temperature to minimize compartmental- toxin-free PBS. A total of 16 animals received either 50 ␮l ization of the dye. The dye was then removed from the PBS or 1 nmol HPIV2 F protein and served as the neg- cells, the cells were rinsed twice with PCFB, and 0.5 ml ative control. The incision was sutured closed and mice of the PCFB was added to the cells for an additional were isolated and kept warm for the 3-h recovery period, 30-min incubation at room temperature. The FBS was 2ϩ at which point they were humanely terminated; the in- found to inhibit SU [Ca ]i mobilization and was therefore testinal loops were removed and weighed, and their removed once the dye had been loaded into the cells. SU length was measured. A ratio of weight (g) to length (cm) and either neurotensin (positive control) or HPIV2 F (neg- was calculated for the SU-treated loops and then com- ative control) were added to 0.5 ml PCFB and added to pared to the g/cm ratio of control loops. the cells immediately prior to analysis.

Histological analysis Confocal fluorescence microscopy To determine whether SU induced histological alter- The calcium imaging studies were performed on an ations within the intestinal loop, sections of each loop MRC-1024 laser scanning confocal imaging system con- were analyzed for any changes. Upon completion of the trolled by LaserSharp software (Bio-Rad). The system physical measurements, the intestinal loops were at- was based on the Zeiss Axiovert 135 microscope and tached to a corkboard and fixed by immersion in 10% equipped with three independent low-noise photo-multi- ␭ neutral buffered Formalin for 24 h. An unmanipulated plier tube channels. The excitation light ( 488 nm) from segment of intestine not included in the intestinal loop a 15-mW krypton-argon laser (5 mW all lines, measured was removed to serve as the animal control. Represen- at the microscope stage) was delivered to the sample tative sections of the jejunum and the ileal loop were through a 63X Zeiss Plan-Fluor oil immersion objective trimmed, placed in Path Cassette IV tissue cassettes with a numerical aperture of 1.45. A region of 10–20 cells (Curtis Matheson Scientific, Houston, TX), and processed was randomly selected and the objective positioned to in a Shandon Hypercenter 2 (Shandon, Pittsburgh, PA). view the median section of the cells. The calcium kinet- Samples were subjected to dehydration through graded ics was monitored on single cells by digitally acquiring alcohols (70%, 80%, 95%, absolute) to xylene, then infil- images at 7-s intervals and calculating the average pixel trated with Paraplast paraffin (Fisher, Pittsburgh, PA). intensity using the Time Course Software (Bio-Rad). After being processed the tissues were embedded on IP measurements end in paraffin blocks utilizing a Tissue-Tek console 3

(Miles Scientific, Elkhart, IN). Coronal sections (3–5 mi- The IP3 assay was performed as previously described crons) of each tissue were cut on a Leitz Model 1512 with the following modifications (Dong et al., 1997; Bozou rotary microtome (Ernst Leitz Wetylan GMBH, Austria), et al., 1989; Mutoh et al., 1999). Briefly, HT-29 cells were floated in a water bath, collected on glass slides, depar- seeded into six-well plates and allowed to grow to con- SIV SU IS A VIRAL ENTEROTOXIN 259

fluency. The media was removed from the cells and the cuited current (Isc) required to nullify the spontaneous cells were washed with 10 mM lithium chloride (LiCl) for transepithelial potential difference was monitored con- 30 s to remove excess media. The cells were then tinuously with an automatic voltage clamp (Physiological incubated with 10 mM LiCl at 37°C for 10 min with gentle Instruments, San Diego, CA). To minimize the effects on ϩ shaking to inhibit inositol-1-phosphatase. The LiCl was Isc of cAMP-stimulated, electrogenic Na -glucose co- discarded and the cells were either mock-treated or transport across the intestinal cells (Grubb, 1995), the treated for 45 s with 75 nM SU, 150 nM SU, wt baculo- apical (mucosal) bath was changed to N-methyl-D-gluca- virus protein (43.3 ␮g) or 150 nM HPIV2 F (negative mine (NMDG)–Krebs sodium-free buffer (120 mM NMDG, controls), or 10 nM neurotensin (positive control), pre- 1.13 mM MgCl2, 4.7 mM KCl, 10 mM glucose, 1 nM pared in Na/HEPES-buffered saline (140 mM NaCl, 4.7 KH2PO4, 25 mM choline bicarbonate ([2-hydroxyethyl] mM KCl, 1.13 mM MgCl2, 10 mM HEPES, 10 mM glucose, trimethyl-ammonium; Sigma) prior to challenge. After

1 mM CaCl2). Basal levels were determined by treating temperature and ionic equilibration were reestablished, the cells with Na/HEPES-buffered saline for 45 s. The tissues were challenged with a known agonist: 10 ␮M reaction was stopped with 0.2 to 0.3 volumes of ice-cold coleus forskohlii in dimethyl sulfoxide (forskolin [FSK]; 20% perchloric acid and the cells were removed from the Calbiochem, San Diego, CA). Once the response of the plate by scraping, transferred to a 1.5-ml tube, sonicated known agonist was established, the unstripped mucosa for 2 min, and incubated on ice for 20 min. The sample was treated with 0.2 ␮MSUor1.5␮M HPIV2 F protein. was then centrifuged at 16,000 g at 4°C for 15 min, and The changes in transepithelial electrical current that the supernatant was transferred to a new tube and neu- were induced by anion flux were measured as Isc.Bu- tralized with 10 M KOH. A precipitate formed, which was metanide (Calbiochem) sensitivity (400 ␮M prepared in removed by centrifugation at 16,000 g at 4°C for 15 min, dimethyl sulfoxide) confirmed that the anionic flux being and the supernatant was then transferred to a siliconized measured was caused by the chloride secretory re- tube and assayed for IP3 by a competitive radioreceptor sponse (Frizzell and Morris, 1994). assay (TRK1000; Amersham Pharmacia Biotech, Piscat- away, NJ) according to the manufacturer’s instructions. ACKNOWLEDGMENTS Briefly, standards were prepared fresh for each assay and used within1hofpreparation. Samples were diluted We thank Dr. Friedhelm Schroeder for expert advice with the fluo- in the assay buffer, to which diluted tracer and binding rescence imaging and Dr. John Williams for assistance with the statis- protein were added. The sample was vortexed and incu- tics. This work was supported in part by the Interdisciplinary Research Initiatives Program, Texas A&M University and the Center for AIDS bated on ice for 15 min. Samples and standards were Research Developmental Award, Baylor College of Medicine. then centrifuged at 4°C for 15 min at 16,000 g, at which point the supernatant was discarded and the pellet re- suspended in water and incubated at room temperature REFERENCES for 10 min. The samples were then decanted immediately Arthos, J., Rubbert, A., Rabin, R. L., Cicala, C., Machado, E., Wildt, K., into scintillation vials containing 5 ml CytoScint scintilla- Hanback, M., Steenbeke, T. D., Swofford, R., Farber, J. M., and Fauci, tion fluid (ICN Biomedicals, Irvine, CA) and mixed, and A. S. (2000). CCR5 signal transduction in macrophages by human the radioactivity was measured for 4 min in a 1600 TR immunodeficiency virus and simian immunodeficiency virus enve- lopes. J. Virol. 74, 6418–6424. Tri-Carb liquid scintillation analyzer (Packard Instrument Ball, J. M., Tian, P., Zeng, C. Q. Y., Morris, A. P., and Estes, M. K. (1996). Company, Downers Grove, IL). Age-dependent diarrhea induced by a rotaviral nonstructural glycoprotein. Science 272, 101–104. ClϪ secretory current measurements Bhat, S., Spitalnik, S. L., Gonzalez-Scarano, F., and Silberberg, D. H. (1991). Galactosyl ceramide or a derivative is an essential compo- To determine whether SU induced ClϪ secretory cur- nent of the natural receptor for human immunodeficiency virus type rents across intact mouse mucosa, 18- to 23-day-old I envelope glycoprotein gp120. Proc. Natl. Acad. Sci. USA 88, 7131– 7134. Balb/C mice were euthanized by overdosing with IsoFlo Bini, E. J., and Cohen, D. (1999). Impact of protease inhibitors on the isoflurane. The peritoneal cavity was opened and the outcome of human immunodeficiency virus-infected patients with ileum located. Approximately 4 cm of ileum was removed chronic diarrhea. Am. J. Gastroenterol. 94, 3553–3559. and cut longitudinally into segments 6–8 mm long, and Bozou, J. C., Rochet, N., Magnaldo, I., Vincent, J. P., and Kitabgi, P. (1989). Neurotensin stimulates inositol triphosphate-mediated calcium mo- mounted unstripped between pieces of Kimwipes EX-L bilization but not protein kinase C activation in HT29 cells: Involve- extra-low-lint wipers (Kimberly-Clark, Atlanta, GA) onto ment of a G-protein. Biochem. J. 264, 871–878. 2 micro-Ussing chamber sliders (aperture of 0.04 cm ). The Cerf-Bensussan, N., and Guy-Grand, D. (1991). Intestinal intraepithelial tissues were then equilibrated at 37°C for 30 min in lymphocytes. Gastroenterol. Clin. North Am. 20, 549–576. Krebs–Ringer bicarbonate buffer [115 mM NaCl, 4.7 mM Chang, E. B., Musch, M. W., and Mayer, L. (1990). Interleukins 1 and 3 stimulate anion secretion in chicken intestine. Gastroenterology 98, KCl, 1.13 mM MgCl2, 10 mM glucose, 1.15 mM Na2H2PO4 1518–1524. (anhydrous), 25 nM NaHCO3] and gassed with 95% Cook, D. G., Fantini, J., Spitalnik, S. L., and Gonzalez-Scarano, F. (1994). O2/5% CO2 throughout the experiment. The short-cir- Binding of human immunodeficiency virus type I (HIV-1) gp120 to 260 SWAGGERTY ET AL.

galactosylceramide (GalCer): Relationship to the V3 loop. Virology cloning and sequence analysis of the fusion glycoprotein gene of 201, 206–214. human parainfluenza virus type 2. Virology 179, 915–920. Coue¨del-Courteille, A., Le Grand, R., Tulliez, M., Guillet, J. G., and Venet, Hunter, E., and Swanstrom, R. (1990). Retrovirus envelope glycopro- A. (1997). Direct ex vivo simian immunodeficiency virus (SIV)-specific teins. Curr. Top. Microbiol. Immunol. 157, 187–253. cytotoxic activity detected from small intestine intraepithelial lympho- Kagnoff, M. F., and Roebuck, K. A. (1999). Human immunodeficiency cytes of SIV-infected macaques at an advanced stage of infection. virus type 1 (HIV-1) infection and expression in intestinal epithelial J. Virol. 71, 1052–1057. cells: Role of protein kinase A and C pathways. J. Infect. Dis. 179, Daniel, M. D., Letvin, N. L., King, N. W., Kannagi, M., Sehgal, P. K., Hunt, S444–S447. R. D., Kanki, P. J., and Essex, M. (1985). Isolation of T-cell tropic Kaup, F. J., Ma¨tz-Rensing, K., Kuhn, E. M., Hunerbein, P., Stahl-Hennig, HTLV-III-like retrovirus from macaques. Science 228, 1201–1204. C., and Hunsmann, G. (1998). Gastrointestinal pathology in rhesus Dayanithi, G., Yahi, N., Baghdiguian, S., and Fantini, J. (1995). Intracel- monkeys with experimental SIV infection. Pathobiology 66, 159–164. lular calcium release induced by human immunodeficiency virus Kestler 3rd, H. W., Kodama, T., Ringler, D., Marthas, M., Pedersen, N., type 1 (HIV-1) surface envelope glycoprotein in human intestinal Lackner, A. A., Regier, D., Sehgal, P. K., Daniel, M. D., and King, N. W. epithelial cells: A putative mechanism for HIV-1 enteropathy. Cell (1990). Induction of AIDS in rhesus monkeys by molecularly cloned Calcium 18, 9–18. simian immunodeficiency virus. Science 248, 1109–1112. Desrosiers, R. C., and Ritter, G. D., Jr. (1989). Use of simian immuno- Kewenig, S., Schneider, T., Hohloch, K., Lampe-Dreyer, K., Ullrich, R., deficiency viruses for AIDS research. Intervirology 30, 301–312. Stolte, N., Stahl-Hennig, C., Kaup, F. J., Stallmach, A., and Zeitz, M. Diop, A. G., Lesort, M., Esclaire, F., Dumas, M., and Hugon, J. (1995). (1999). Rapid mucosal CD4ϩ T-cell depletion and enteropathy in Calbindin D28K-containing neurons, and not HSP70-expressing neu- simian immunodeficiency virus-infected rhesus macaques. Gastro- rons, are more resistant to HIV-1 envelope (gp120) toxicity in cortical enterology 116, 1115–1123. cell cultures. J. Neurosci. Res. 42, 252–258. Kodama, T., Wooley, D. P., Naidu, Y. M., Kestler 3rd, H. W., Daniel, M. D., Dong, Y., Zeng, C. Q. Y., Ball, J. M., Estes, M. K., and Morris, A. P. (1997). Li, Y., and Desrosiers, R. C. (1989). Significance of premature stop The rotavirus enterotoxin NSP4 mobilizes intracellular calcium in codons in env of simian immunodeficiency virus. J. Virol. 63, 4709– human intestinal cells by stimulating phospholipase C-mediated 4714. inositol 1,4,5-phosphate production. Proc. Natl. Acad. Sci. USA 94, Kotler, D. P., Gaetz, H. P., Lange, M., Klein, E. B., and Holt, P. R. (1984). 3960–3965. Enteropathy associated with the acquired immunodeficiency syn- Ehrenpreis, E. D., Ganger, D. R., Kochvar, G. T., Patterson, B. K., and drome. Ann. Intern. Med. 101, 421–428. Craig, R. M. (1992). D-xylose malabsorption: Characteristic finding in Kotler, D. P., Reka, S., and Clayton, F. (1993). Intestinal mucosal inflam- patients with the AIDS wasting syndrome and chronic diarrhea. J. mation associated with human immunodeficiency virus infection. Acquir. Immune Defic. Syndr. Hum. Retrovirol. 5, 1047–1050. Dig. Dis. Sci. 38, 1119–1127. Fantini, J., Cook, D. G., Nathanson, N., Spitalnik, S. L., and Gonzalez- Kraehenbuhl, J. P., and Neutra, M. R. (1992). Molecular and cellular Scarano, F. (1993). Infection of colonic epithelial cell lines by type 1 basis of immune protection of mucosal surfaces. Physiol. Rev. 72, human immunodeficiency virus is associated with cell surface ex- 853–879. pression of galactosylceramide, a potential alternative gp120 recep- Kuhn, E. M., Ma¨tz-Rensing, K., Stahl-Hennig, C., Makoschey, B., tor. Proc. Natl. Acad. Sci. USA 90, 2700–2704. Hunsmann, G., and Kaup, F. J. (1997). Intestinal manifestations of Foudraine, N. A., Weverling, G. J., van Gool, T., Roos, M. T., de Wolf, F., experimental SIV-infection in rhesus monkeys (Macaca mulatta): A Koopmans, P. P., van den Broek, P. J., Meenhorst, P. L., van Leeuwen, histological and ultrastructural study. J. Vet. Med. B 44, 501–512. R., Lange, J. M., and Reiss, P. (1998). Improvement of chronic diarrhea Kurtz, R. C., Lightdale, C. J., and Winawer, C. J. (1985). Malabsorption in patients with advanced HIV-1 infection during potent antiretroviral and mucosal abnormalities of the small intestine in the acquired therapy. AIDS 12, 35–41. immunodeficiency syndrome. Ann. Intern. Med. 102, 619–622. Frizzell, R. A., and Morris, A. P. (1994). Chloride conductances of salt- Lambl, B. B., Federman, M., Pleskow, D., and Wanke, C. A. (1996). secreting epithelial cells. In “Current Topics in Membranes,” Vol. 42, Malabsorption and wasting in AIDS patients with microsporidia and pp. 173–214. Academic Press, New York. pathogen-negative diarrhea. AIDS 10, 739–744. Grubb, B. R. (1995). Ion transport across the jejunum in normal and Lanzavecchia, A., Roosnek, E., Gregory, T., and Berman, P. (1988). T cystic fibrosis mice. Am. J. Physiol. Gastrointest. Liver Physiol. 268, cells can present antigens such as HIV gp120 targeted to their own G505–G513. surface. Nature 334, 530–532. Hammache, D., Yahi, N., Dele´zay, O., Koch, N., Lafont, H., Tamalet, C., Letvin, N. L., Daniel, M. D., Sehgal, P. K., Desrosiers, R. C., Hunt, R. D., and Fantini, J. (1998). Specific interaction of HIV-1 and HIV-2 surface Waldron, L. M., MacKey, J. J., Schmidt, D. K., Chalifoux, L. V., and King, envelope glycoproteins with monolayers of galactosylceramide and N. W. (1985). Induction of AIDS-like disease in macaque monkeys ganglioside GM3. J. Biol. Chem. 273, 7967–7971. with T-cell tropic retrovirus STLV-III. Science 230, 71–73. Heise, C., Dandekar, S., Pradeep, K., Duplantier, R., Donovan, R. M., and Lyerly, H. K., Matthews, T. J., Langlois, A. J., Bolognesi, D. P., and Halsted, C. H. (1991). Human immunodeficiency virus infection of Weinhold, K. J. (1987). Human T-cell lymphotropic virus IIIB glyco- enterocytes and mononuclear cells in human jejunal mucosa. Gas- protein (gp120) bound to CD4 determinants on normal lymphocytes troenterology 100, 1521–1527. and expressed by infected cells serves as target for immune attack. Heise, C., Miller, C. J., and Dandekar, S. (1994). Primary acute simian Proc. Natl. Acad. Sci. USA 84, 4601–4605. immunodeficiency virus infection of intestinal lymphoid tissue is Madara, J. L., and Stafford, J. (1989). Interferon-gamma directly affects associated with gastrointestinal dysfunction. J. Infect. Dis. 169, 1116– barrier function in cultured intestinal epithelial monolayers. J. Clin. 1120. Invest. 83, 724–727. Heise, C., Vogel, P., Miller, C. J., and Dandekar, S. (1993b). Distribution McGowan, I., Radford-Smith, G., and Jewell, D. P. (1994). Cytokine gene of SIV infection in the gastrointestinal tract of rhesus macaques at expression in HIV-infected intestinal mucosa. AIDS 8, 1569–1575. early and terminal stages of AIDS. J. Med. Primatol. 22, 187–193. Morris, A. P., Kirk, K. L., and Frizzell, R. A. (1990). Simultaneous analysis Heise, C., Vogel, P., Miller, C. J., Halsted, C. H., and Dandekar, S. of cell Ca2ϩ and Ca2(ϩ)-stimulated chloride conductance in colonic (1993a). Simian immunodeficiency virus infection of the gastrointes- epithelial cells (HT-29). Cell Regul. 1, 951–963. tinal tract of rhesus macaques. Am. J. Pathol. 142, 1759–1771. Mutoh, T., Kumano, T., Nakagawa, H., and Kuriyama, M. (1999). Role of Henning, S. J., and Leeper, L. L. (1984). Duodenal uptake of lead by tyrosine phosphorylation of phospholipase C g1 in the signaling suckling and weanling rats. Biol. Neonate 46, 27–35. pathway of HMG-CoA reductase inhibitor-induced cell death of L6 Hu, X., Ray, R., Matsuoka, Y., and Compans, R. W. (1990). Molecular myoblasts. FEBS Lett. 446, 91–94. SIV SU IS A VIRAL ENTEROTOXIN 261

Nelson, J. A., Wiley, C. A., Reynolds, K. C., Reese, C. E., Margaretten, W., Summers, M. D., and Smith, G. E. (1988). “A Manual of Methods for and Levy, J. A. (1988). Human immunodeficiency virus detected in Baculovirus Vectors and Insect Cell Culture Procedures.” Texas Ag- bowel epithelium from patients with gastrointestinal symptoms. Lan- ricultural Experiment Station Bulletin No. 1555, College Station, TX. cet 1, 259–262. Tong, S., and Compans, R. W. (2000). Oligomerization, secretion, and Reed, L. J., and Muench, H. (1938). A simple method of estimating fifty biological function of an anchor-free parainfluenza virus type 2 (PI2) per cent endpoints. Am. J. Hyg. 27, 493–497. fusion protein. Virology 270, 368–376. Robey, W. G., Arthur, L. O., Matthews, T. J., Langlois, A., Copeland, T. D., Ullrich, R., Zeitz, M., Heise, W., L’age, M., Hoffken, G., and Riecken, E. O. Lerche, N. W., Oroszlan, S., Bolognesi, D. P., Gilden, R. V., and (1989). Small intestinal structure and function in patients infected Fischinger, P. J. (1986). Prospect for prevention of human immuno- with human immunodeficiency virus (HIV): Evidence for HIV-induced deficiency virus infection: Purified 120-kDa envelope glycoprotein enteropathy. Ann. Intern. Med. 111, 15–21. induces neutralizing antibody. Proc. Natl. Acad. Sci USA 83, 7023– Veazey, R. S., DeMaria, M., Chalifoux, L. V., Shvetz, D. E., Pauley, D. R., 7027. Knight, H. L., Rosenzweig, M., Johnson, R. P., Desrosiers, R. C., and Scheenen, W. J. J. M., Hofer, A. M., and Pozzan, T. (1998). Intracellular Lackner, A. A. (1998). Gastrointestinal tract as a major site of CD4ϩ measurement of calcium using fluorescent probes. In “Cell Biology: T cell depletion and viral replication in SIV infection. Science 280, A Laboratory Handbook” (J. E. Celis, Ed.), pp. 363–374. Academic Press, San Diego. 427–431. Schmitz, H., Fromm, M., Bode, H., Scholz, P., Riecken, E. O., and Volkman, L. E., and Summers, M. D. (1975). Nuclear polyhedrosis virus Schulzke, J. D. (1996). Tumor necrosis factor-a induced ClϪ and Kϩ detection: Relative capabilities of clones developed from Trichoplu- sia ni ovarian cell line TN-368 to serve as indicator cells in a plaque secretion in human distal colon driven by prostaglandin E2. Am. J. Physiol. Gastrointest. Liver Physiol. 271, G669–G674. assay. J. Virol. 16, 1630–1637. Sears, C. L., Guerrant, R. L., and Kaper, J. B. (1995). Enteric bacterial Vyakarnam, A., Matear, P., Meager, A., Kelly, G., Stanley, B., Weller, I., toxins. In “Infections of the Gastrointestinal Tract” (M. J. Blaser, P. D. and Beverley, P. (1991). Altered production of tumor necrosis factors Smith, J. I. Ravdin, H. B. Greenberg, and R. L. Guerrant, Eds.), pp. ␣ and ␤ and interferon ␥ by HIV-infected individuals. Clin. Exp. 617–634. Raven Press, New York. Immunol. 84, 109–115. Shimizu, N., Soda, Y., Kanbe, K., Liu, H. Y., Jinno, A., Kitamura, T., and Weissman, D., Rabin, R. L., Arthos, J., Rubbert, A., Dybul, M., Swofford, Hoshino, H. (1999). An orphan G protein-coupled receptor, GPR1, R., Venkatesan, S., Farber, J. M., and Fauci, A. S. (1997). Macrophage- acts as a coreceptor to allow replication of human immunodeficiency tropic HIV and SIV envelope proteins induce a signal through the virus types 1 and 2 in brain-derived cells. J. Virol. 73, 5231–5239. CCR5 chemokine receptor. Nature 389, 981–985. Shimizu, N., Soda, Y., Kanbe, K., Liu, H. Y., Mukai, R., Kitamura, T., and Wright, K. A., and Olsen, R. G. (1989). In vitro infection of cell lines with Hoshino, H. (2000). A putative G protein-coupled receptor, RDC1, is a HTLV-1 and SIVmac results in altered intracellular free calcium con- novel coreceptor for human and simian immunodeficiency viruses. centration and increased membrane depolarization. Int. J. Cancer 44, J. Virol. 74, 619–626. 753–756. Smit-McBride, Z., Mattapallil, J. J., McChesney, M., Ferrick, D., and Yahi, N., Baghdiguian, S., Bolmont, C., and Fantini, J. (1992). Replication Dandekar, S. (1998). Gastointestinal T lymphocytes retain high po- and apical budding of HIV-1 in mucous-secreting colonic epithelial tential for cytokine responses but have severe CD4ϩ T-cell depletion cells. J. Acquir. Immune Defic. Syndr. Hum. Retrovirol. 5, 993–1000. at all stages of simian immunodeficiency virus infection compared to Yahi, N., Sabatier, J. M., Baghdiguian, S., Gonzalez-Scarano, F., and peripheral lymphocytes. J. Virol. 72, 6646–6656. Fantini, J. (1995). Synthetic multimeric peptides derived from the Steinberg, T. H., Newman, A. S., Swanson, J. A., and Silverstein, S. C. principal neutralizing domain (V3 loop) of human immunodeficiency (1987). Macrophages possess probenecid-inhibitable organic anion transporters that remove fluorescent dyes from the cytoplasmic ma- virus type 1 (HIV-1) gp120 bind to galactosylceramide and block trix. J. Cell Biol. 105, 2685–2702. HIV-1 infection in a human CD4-negative mucosal epithelial cell line. Stockman, M., Fromm, M., Schmitz, H., Schmidt, W., Riecken, E. O., and J. Virol. 69, 320–325. Schulzke, J. D. (1998). Duodenal biopsies of HIV-infected patients Yamshchikov, G. V., Ritter, G. D., Jr., Vey, M., and Compans, R. W. (1995). with diarrhoea exhibit epithelial barrier defects but no active secre- Assembly of SIV virus-like particles containing envelope proteins tion. AIDS 12, 43–51. using a baculovirus expression system. Virology 214, 50–58. Summers, M. D., and Smith, G. E. (1976). Occluded and nonoccluded Zeitz, M., Greene, W. C., Peffer, N. J., and James, S. P. (1988). Lympho- nuclear polyhedrosis virus growth in Trichoplusia ni: Comparative cytes isolated from the intestinal lamina propria of normal nonhuman neutralization, comparative infectivity, and in vitro growth studies. primates have increased expression of genes associated with T-cell J. Virol. 19, 820–832. activation. Gastroenterology 94, 647–655.