Hansenula anomala killer toxin induces secretion and severe acute injury in the rat intestine

GASTROENTEROLOGY 1995;109:1900-1906

Hansenula anomala Killer Toxin Induces Secretion and Severe Acute Injury in the Rat Intestine

MASSIMO PETTOELLO-MANTOVANI,* '* AGOSTINO NOCERINO,* LUCIANO POLONELLI, §

GIULIA MORACE, Ir STEFANIA CONTI, § LUClO DI MARTINO, ~ GIORGIO DE RITIS, t

MICHELE IAFUSCO,* and STEFANO GUANDALINI* *Department of Pediatrics, Albert Einstein College of Medicine, Bronx, New York; *Department of Pediatrics, Federico II University, Naples, Italy; §Institute of Microbiology, University of Parma, Parma, Italy; and I[Catholic University of Sacred Hearth, Rome, Italy

Background & Aims: The yeast Hansenula anomala has been associated with gastrointestinal symptomatology and damage to the intestinal wall in humans, In vitro and in vivo, H. anomala secretes a toxin, killer toxin, which is lethal to other microorganisms. In view of the very high rate of killer phenotype expression recorded for H. anomala strains in nature, this study aimed to investigate the hypothesis that H. anomala killer toxin plays a role in the pathogenesis of H. anomala-induced enteritis. Methods: Effects of active and heat-inactivated H. anomala killer toxin on intestinal fluid homeostasis and electrolyte balance were investigated in rat small intestine using a standard intestinal perfusion technique. Sections of the perfused jejunum tracts were examined histologically. Results: H. anomala killer toxin induced a significant secretion of water and electrolytes. No significant change was observed when either heat-inactivated H. anomala killer toxin or control growth medium were tested, Histological analysis showed ischemic degeneration of villi and sloughing of surface epithelium in 50% of active H. anomala killer toxin-perfused jejuna. Conclusions: This paper pre- sents original observations compatible with the hypothesis that H. anomala killer toxin plays a role in the pathogenesis of H. anomala-induced enteritis.

ansenuIa anomala, a yeast belonging to the class scomycetes, 1-4 has been reported in the last decade

as being pathogenic in humans 5 after its isolation from either blood or body fluid cultures. 4'6 Although immuno- compromised and severely ill patients are the primary targets, 5 subjects with no discernible clinical immuno- logic abnormalities have been reported to have an H. anomala infection) '5'v Despite recognition of this yeast

as an emergent opportunistic pathogen, very little information is available on its pathogenicity, v's

H. anomala (also classified in the genus Pichia), 9 similar to a variety of other yeasts including Candida albicans, Saccharomyces cerevisiae, or Cryptococcus neoformans, 1°-14 has the ability to secrete a toxin, 1°a5 killer toxin (KT), *I, in

vitro that is lethal to different strains of the same species, 1° eucaryotic microorganisms, and bacteria) 1 We have shown previously that H. anomala produces KT in vivo in the tissue of normal and immunosuppressed experimentally infected mice, 12 suggesting that toxin production may be involved in the mechanism of infection associated with killer yeasts. A later clinical report

described, for the first time, invasion of human tissue by H. anomala, v This yeast was recognized as a causative

agent of severe gastrointestinal symptomatology after a finding of its infiltration of the intestinal wall and was

also associated with histological findings of necrotizing

enteritis] Colonization of human gastrointestinal tract by yeasts

and fungi including C. albicans or Histoplasma capsulatum is well documented, and related gastrointestinal symp- toms are well characterized) 6-18 Because the occurrence of killer characters in yeasts has been investigated and

because H. anomala showed a high frequency of killer phenotypes, ranging to > 7 0 % , l° we hypothesized that

H. anomala killer toxin (HaKT) plays a role in the pathogenesis ofH. anomala-induced enteritis. It is known that

toxins produced by enteric organisms can alter normal secretory and absorptive intestinal processes. 19'2° We

therefore investigated the effects of HaKT on the intestinal fluid homeostasis and electrolyte balance of the proximal intestine in a rat model. Histological examination was performed to show KT-associated damage.

M a t e r i a l s a n d M e t h o d s

Production of HaKT

The yeast killer strain H, anomala UCSC 25F, previously isolated from a patient, was used. 11 It exerted killer activity against a wide range of potentially sensitive microor-

Abbreviations used in this paper: HaKT, Hansenula anomala killer toxin; KT, killer toxin; MAb, monoclonal antibody.

© 1995 by the American Gastroenterological Association 0016-5085/95/$3,00

December 1995 H. ANOMALA KILLER TOXIN PATHOGENICITY 1901

ganisms) 1 HaKT was generated as described previously. 21

Briefly, flasks (1-L volume) with 500 mL of Sabouraud broth, buffeted at pH 3.5 with 0.1 mol/L citric acid and 0.2 mol/L K2I-IPO4, were seeded with a loopful of a 48-hour culture of H. anomala and incubated at 22°C for 18 hours at 70 rpm on an orbital shaker. The cells were then removed by centrifuga- tion at 1500g at 4°C, and the supernatant was filtered and concentrated (50X) with a PM-10 membrane in a TCF10A ultrafiltration cell (Amicon Corp., Danvers, MA) under N2 pressure. The protein and polysaccharide concentrations in

I-IaKT extract were 6 and 2.5 mg/mL, respectively, determined using the Lowry method and Anthrone t e s t . 22-24 Accordingly,

a concentrated (50×) extract of the growth medium (Sabour- aud broth, modified; Difco Laboratories, Detroit, MI) was pre-

pared.

Separat ion of HaKT

HaKT consists of a large molecule (>80 kilodaltons) as shown by us previously 21 using a sodium dodecyl sulfate gel and Western blot assay performed on the HaKT and the anti-HaKT monoclonal antibodies (MAbs) also used in the present study. This size is compatible with that described by Sawant et al. 25 This large glycoproteic molecule has been indicated 22 as the precursor of small but still biologically active

fragments (from 15 to 25 kilodaltons) investigated in previous studies. 22 A volume of 4.5 mL of the concentrated KT was loaded twice onto a Sephadex GS0-fine column (Pharmacia-

LKB, Piscataway, NJ) previously equilibrated with 1 mmol/L sodium acetate buffer, pH 4.5. The column was calibrated using a set of proteins of known molecular weight, e2 bovine

serum albumin, ovalbumin, chymotrypsinogen A, and ribo- nuclease A (Sigma Chemical Co., St. Louis, MO). The detection system was a 2238 Uvicord SII (Pharmacia-LKB) equipped

with a flow cell and a 278-nm filter. The flow rate was 35 cm/ h, and 5-mL fractions were collected. Each fraction was tested for killer activity. 2~'26 Fractions of > 6 7 kilodaltons containing

active substance (large molecule) and fractions showing the presence of active substance recovered in the range from 15 to 25 kilodaltons (small fragments) were pooled to be used for the perfusion studies. The protein (4 mg/mL) and polysaccharide (1 mg/mL) concentrations of the final HaKT pooled extract were determined. The same pooled extract was also tested by a conventional well assay 2v using sensitive microorganisms and an MAb assay 21'26 to confirm the presence of killer activity.

The MAb KT4 against the killer toxin of H. anomala UCSC 25F was prepared according to a procedure described previously, e~ Ant i -HaKT MAb KT4 showed high sensitivity (1:4000) and did not react with controls and toxins produced by other species.

Animals

Wistar male rats weighing 230-280 g were used. They were fed Purina rat chow (Purina Mills Inc., St. Louis, MO) and water ad libitum. The rats were housed in a temperature- controlled room at 22°C with alternating 12-hour light-dark

cycles. Twelve hours before the experiment, food was withheld,

but water was given with no restriction.

Intestinal Perfusion Studies

After the fast period, rats were anesthetized with an intraperitoneal injection of sodium pentobarbital (6 mg/100 g body wt), and body temperature was kept at 37°C by a thermostatically controlled heating lamp. 28 The abdominal

cavity was opened by a midline incision, and the small intestine was cannulated below the suspensory muscle of the duodenum (also known as ligament of Treitz in humans) and again t5 cm below the proximal cannula. 29 This segment was rinsed

free of intestinal contents with 0.9% NaC1 solution warmed at 37°C. The proximal cannula was connected by a polyvinyl tube with a peristaltic pump (Unispense dispenser; Wheaton

S.P. Inc., Millville, NJ). The jejunal segment was perfused at a rate of 0.2 mL/min with the following isotonic perfusion

fluid: NaC1, 115 mmol/L; KC1, 15 mmol/L; NaHCO3, 15 mmol/L (for a final osmolarity of 290 mOsm/L); [14C]

polyethylene glycol (PEG; mol wt, 4000), 2.5 }_tCi/100 mL; and unlabeled PEG, 3 glL, as a nonabsorbable water marker (Amity P.G., Milan, Italy). ~°'31 The pH was fixed at 7.4 by gassing with 95% 02 and 5% CO2. 32'33 The eluate was col-

lected in 20-minute aliquots from the distal cannula. An equilibrium period of 30 minutes was allowed, followed by three

consecutive 20-minute collection periods for baseline measure- ments of water and electrolyte net transport. Subsequently, according to the criteria adopted in studies on Vibrio chole,~e enterotoxin, 34 the jejunal segment was injected with 1 mL of

(1 ×) HaKT extract at a final concentration of 80 ~g protein and 20 btg polysaccharide. To our knowledge, no information is currently available on the amount of HaKT produced during active infection. The jejunal segment was exposed to the yeast toxin for 120 minutes, arid a new perfusion followed by a period of equilibrium and cycle of eluate collections was performed as described.

In a second set of experiments, intestinal fluid homeostasis and electrolyte balance were analyzed after exposing the jejunal segment to inactivated HaKT extract. HaKT was inactivated by heating the extract at 100°C for 15 minutes, 34 and the inactivation was confirmed using a killer activity test. 26 Fi-

nally, a set of control experiments was performed using the growth medium extract.

In preliminary blocking experiments, HaKT and anti- HaKT MAb KT4 were incubated at 4°C for 24 hours. The supernatant obtained was perfused in the rat jejuna according to a procedure described previously. No histological and transport changes were observed (data not shown).

Water and electrolyte transport were measured according to criteria described by us previously. 29'31 Briefly, the following formula was used to calculate net water transport. The efflux volume was calculated on the basis of its weight assuming that 1 g = 1 mL. In each sample, the concentrations of Na + and K + were determined using a flame photometer and the concentration of CI was determined using an $26 Coming chlorido-

1902 PETTOELLO-MANTOVANI ET AL. GASTROENTEROLOGY Vol. 109, No. 6

meter (Coming Corp., Cambridge, MA). Aliquots of 0.5 mL were counted in a Beckman £S-7500 ]3-liquid scintillation counter (Beckman Instruments Inc., Somerset, NJ) to assess the concentration of [I~C]PEG 4000 and, hence, of PEG 4000. Water absorption was calculated according to the formula by Sladen and Dawson 3I as follows: Water Absorption = [Fx(S1-S2)PEGI/PEG2]/W, where F is flow velocity in mL/ min; Se is solute concentration in the perfusing solution in ~Eq/mL; Se is the solute concentration in the effluent solution in ~tEq/mL; PEG1 is PEG 4000 concentration in the perfusing solution; PEG2 is PEG 4000 concentration in the effluent solution; and W is the weight of the intestinal segment under consideration. Positive results indicate net absorption and negative results indicate net secretion into the lumen. All results are expressed as ~tL/min or bLEq" min - l "g wet wt -1. Under similar experimental conditions, we have previously repeatedly measured transport data expressing the results also per centi- meter of perfusion segment, finding no sizeable difference in any parameter.

Histo logica l S t u d i e s

At the end of the experiment, the animal was killed and a fragment of the segment perfused was isolated and fixed in 10% formol saline. A second fragment from the contiguous nonperfused jejunal segment was also isolated and fixed, serving as control. The samples were then embedded in paraffin wax, sectioned, and stained with H&E for examination by light microscopy. 35

Immunofluorescent Studies

The indirect fluorescent antibody procedure 12 was used for detection of HaKT in the tissue sections. As reported previously, an MAb against the KT of H. anomala UCSC 25F was prepared according to a previously described procedure 21 and used in this study. The anti-HaKT MAb KT4 was tested and proved to be reactive in vitro either by double immunodiffu- sion with KT or immunofluorescent techniques with KT-producing ceils. Anti-HaKT MAb KT4 proved to be reactive with yeast strains belonging to the same or other species of genus Hansenula (H. mrakii) but not with the ones obtained by recognized killer yeasts of other genera (Candida, Kluyvero- myces, or Saccharomyces).

Statistics

The data are expressed as mean -+- SE with 8 rats per group. The data were analyzed using standard statistical analyses, and the significance of the differences was calculated using Student's t test for paired samples, 29 with each animal serving as its own control.

Resul ts

Intestinal Perfusion

As shown in Figure 1, intraluminal administration of H a K T induced a significant secretion of water

16

14

12

A 10

8

6

o 4

.-~ 2

0

-2

0! -8 j

W a t e r Na CI K

p < 0.01 p < 0.01 p < 0.01 p < 0.05

Figure 1. Effects of intraluminal administration of 1 mL of HaKT extract (B) on fluid and electrolyte transport in the rat perfused jejunum compared with baseline (11). Each column represents the mean + SE of 3 periods of collections, each lasting 20 minutes, of 8 different animals. Values above the 0 line represent net absorption, and values below it represent net secretion. HaKT clearly induced fluid and ion secretion.

(from 11.93 + 1.94 to - 3 . 9 3 + 1.78 b t l ' m i n - l " g -1)

and electrolytes (Na +, 1.86 + 0.51 to - 2 . 4 5 + 0.66;

CI- , 1.35 + 0.51 to - 1 . 2 5 + 0.32; and K +, 0.44 + 0.26 to - 0 . 3 6 + 0.18 ~tEq" min -1" g-l) . No secretive

effects and no significant difference in absorption rates

were observed by intraluminal administration of either

heat-inactivated H a K T (Figure 2A) or growth medium

extract (Figure 2B). Ha l f of HaKT-perfused jejuna showing water and elec-

trolyte secretion had damage to the mucosa. A significant

difference (P < 0.001) in water transport was observed

by comparing the data from the group of damaged and

undamaged mucosa (Figure 3). Because the recovery of

PEG was > 8 0 % in all experiments, we can reasonably conclude that the mucosa disruption did not influence the formula used to calculate the water transport.

Histology

Sections from 50% of HaKT-perfused jejuna had severe morphological damage (Figure 421), which was

not observed in the contiguous segments that were not perfused. In all examined sections, there was evidence of early severe injury. These changes included ischemic degeneration of villi, sloughing of surface epithelium, and vascular congestion. At this early stage of injury, the crypts had no apparent epithelial changes and the lamina propria was not yet morphologically affected. These findings are usually observed after acute toxic or ischemic injury. No morphological changes were observed in all

December 1995 H. ANOMALA KILLER TOXIN PATHOGENICITY 1903

segments perfused with heat-inactivated HaKT (Figure

4B) or growth medium extract.

I m m u n o f l u o r e s c e n c e

None of the tissue sections from the HaKT-perfused jejuna showed specific fluorescence using anti-HaKT MAb,

indicating a binding of KT to the tissue cells. Specific fluo-

rescence was also absent in all tissue sections from the intestine perfused with heat-inactivated HaKT. Finally, tissue sections from the growth medium extract-perfused intestine, used as an internal control, were negative.

Discussion

The data of this study indicate for the first time that KTs, HaKT in particular, could play a pathogenetic

role in the intestine. In the rat perfused jejunum, HaKT proved to have a secretory effect on water, Na +, CI-, and K + transport (Figure 1), which was not observed for heat-inactivated HaKT. Recent studies have suggested a

16

A 14

12

10

6

6

4

2

0 - -

-2 Water N a C l K

NS NS N S NS

16

,,1 B

10

8

6

4 -

2 *

0 - -

-2 Water Na CI K

N S NS NS NS

Figure 2. Effects of intraluminal administration of 1 mL of (A) heat- inactivated HaKT extract (11) and (B) growth medium (11) on fluid and electrolyte transport in the rat perfused jejunum compared with baseline (D). Each column represents the mean _+ SE of 3 periods of collections, each lasting 20 minutes, of 8 different animals. Values above the 0 line represent net absorption, and values below it represent net secretion. Heat-inactivated HaKT and growth medium did not influence fluid and ion transport.

change in virulence of H. anomala. 7 This microorganism,

like other yeasts previously considered nonpathogenic in humans, is now increasingly reported as the cause of a wide spectrum of diseases, <36-39 including enteritis] The

predisposing circumstances for infection by H. anomala have been suggested to be similar to those indicated for Candida species, 7'4° a very well-recognized pathogenic

yeast in humans. In addition, as is the case with Candida,

Hansenula organisms have the ability to pass unharmed through the gastrointestinal tract of animals. 41 Under

certain circumstances, this property may enable Han-

senula to colonize the intestine and invade the mucosa as described] We have previously shown the production of KT after infection with H. anomala in a mouse tissue model. 12 The same sequence of events can be reasonably

postulated in humans. Although, to our knowledge, no other data are available on the pathogenicity and mechanism of action of HaKT in the intestine, we can hypothe-

size that it has a mechanism similar to that described for bacterial toxins. Bacterial toxins may alter normal bowel function by selective destruction of normal absorptive ceils, as in the case of Shiga-like toxin. 42 Alterna-

tively, they may stimulate normal secretory pathways or inhibit normal absorptive pathways to produce abnormal net secretion. ~9'43 In our study, HaKT seems to act pri- marily by an enterotoxic pathway similar to one we have reported for bacteria, with overstimulation of normal mechanisms leading to hypersecretion of electrolytes and

water. Nevertheless, in the jejuna of half of the animals perfused by active HaKT, the cytotonic effect was combined with a cytotoxic effect associated with the presence

tu

== o -2

E v

1 6 - .

1 4 -

1 2 -

1 0 -

8 -

6 -

4 .

2 .

0 -

- 2 .

- 4 -

- 6 .

- 6 .

Water Na C] K p < 0.001 NS NS NS

Figure 3. Fluid and ion transport after intraluminal administration of 1 mL of active HaKT extract. Comparison between jejuna showing damaged (11) and undamaged ([]) mucosa. E:], Baseline. Each column represents the mean _+ SE of 3 periods of collections, each lasting 20 minutes, of 8 different animals. Values above the 0 line represent net absorption, and values below it represent net secretion. A significant difference (P < 0.001) in water transport was observed between the two groups.


Figure 4. H&E-stained sections of rat jejunum after perfusion with (A) active and (B) heat-inactivated HaKT. (A) The surface epithelium of the villi show ischemic degeneration and sloughing, and the vessels within the lamina propria show a marked congestion. (B) In contrast, mucosa shows no visible changes from a normal morphology; the clear areas within the lamina propria of the two villi in the center of the photograph are as secondary to fixation artifact (original magnifi- cation: high-power, 2 4 0 x ; low-power inserts, 60x) .

of detectable gut damage. A significant difference in water transport could be observed once the data from jejuna showing a combined cytotonic-cytotoxic effect were compared with the data from those showing only cytotonic effect (Figure 3). The significance of the difference between these two groups and its relation to the presence or absence of a cytopathic effect needs to be further investigated. However, a few observations can be made. The HaKT protein used was from the same stock in all perfusion experiments; therefore, it could be hypothesized that host factors play a determining role among the variables, influencing the ability of the toxin to generate a cytotoxic effect. For instance, the presence of cell wall ~-glucan receptors is known to be necessary for the lethal action of KT to cellular microorganisms. 22'44 It has been shown

that once bound to these receptors, KTs are able to per- turb the membrane either by inhibiting some component of the proton pump 45 or acting more directly by forming

ion channels. 46 Whatever the mechanism, the physiologi-

cal changes induced by KTs to the wall lead the target cell to death depending on the number of KT receptors expressed on the cell surface or the stability of the bonds. 44 These well-established general concepts in the

mechanism of action of KT that induces toxicity to cellular microorganisms may apply to the observations of the present study. The presence of intestinal mucosa receptors for KT, HaKT in particular, has not been investigated to date. Because we did not explore this possibility, we can only speculate on the presence of receptors on the epithelium due to their need for KTs to be active. We

are currently investigating the presence of KTs epithelial receptors, given the hypothesis that their expression on mucosa, number, and stability of bonds with KT may have an influence on HaKT ability to be cytotoxic, 44 similarly to the interaction of KTs with microorganisms.

Finally, a mechanism similar to that seen in bacterial toxins should also be mentioned among the possible causes of cytotoxicity. A number of bacterial toxins (e.g.,

C. difficile toxin A) induce disease by stimulating host defense mechanisms, which in turn cause the epithelial damage. 47,48

None of the histol0giqa! sections from either damaged or undamaged mucosa showed a specific fluorescence compatible with the detection of HaKT by ant i-HaKT MAb. This observation is in line with a previous in vivo study using a mouse model, in which HaKT could be detected by indirect immunofluorescence in areas proximal to the producing yeasts but not in areas where histological damage ascribed to the toxin took place. 12 In the present study, it is possible that HaKT was not available

for ant i -HaKT MAb detection because toxin was removed by the subsequent fluid perfusions. However, it cannot be excluded that HaKT developed structural changes after binding to the receptor that altered its recognition property. In fact, it is recognized that KTs develop structural changes once bound to the cell receptors, as shown for the interaction of KTs with microorganisms. 44 For example., in_the toxic process leading to ion channel formation, killer 1-ki l ler toxin from Saccha- romyces cerevisiae after binding to its receptor is known to form oligomers of up to eight 0~ dimers starting from two primitive 0~ subunits. 44'46 A similar modification in HaKT structure could be postulated to explain the failure of a well-tested ant i -HaKT MAb in detecting the toxin in our animal model.

In conclusion, the data presented in this paper are

December 1995 H, ANOMALA KILLER TOXIN PATHOGENICITY 1905

original observations compatible with the hypothesis of a role played by HaKT in the pathogenesis of H. anomala- induced enteritis. The data may also have important im- plication in the pathogenesis of infections caused by KT- producing yeasts.

References

1. Sekhon AS, Kowalewska-Grochowska K, Garg AK, Vaudry W. Han- senula anomala fungemia in an infant with gastric and cardiac complications with a review of the literature. Eur J Epidemiol 1992 ;8 :305-308 .

2. Hoj J, Hsu KC, Rolston K, Hopfer RL, Luna M, Bodey PG. Tri- chosporon beigelii infection: a review. Rev Infect Dis 1986;8: 959 -967 .

3. Kassamali H, Anaissie EROJ, Rolston K, Kantarjian H, Fainstein V, Bodey GP. Disseminated Geotricum candidum infection. J Clin Microbiol 1987; 25 :1782-1783 .

4. Klein AS, Tortora GT, Malowitz R, Greene WH. Hansenula anomala: a new fungal pathogen. Arch Intern Med 1988; 148 :1210- 1213.

5. Dickensheets DL. Hansenula anomala infection. Rev Infect Dis 1989; 11 :507-508 .

6. Murphy N, Damjanovic V, Hart CA, Buchanan CR, Whitaker R, Cooke RWI. Infection and colonization of neonates by Hansenula anomala. Lancet 1986; 1 :291-293 .

7. Moses A, Maayan S, Shvil Y, Dudin A, Ariel I, Thalji A, Poiacheck i. Hansenula anomala infections in children: from asymptomatic colonization to tissue invasion. Pediatr Infect Dis J 1991;10: 400 -402 .

8. Gold JWM. Opportunistic fungal infections in patients with neo- plastic disease. Am J Med 1984 ;76 :458-463 .

9. Polonelii L, Conti S, Campani L, Gerloni M, Morace G, Chezzi C. Differential toxinogenesis in the genus Pichia detected by an ant i -yeast killer toxin monoclonal antibody. Antonie van Leeu- wenhoek 1991; 59 :139 -145 .

10. Rosini G. The occurrence of killer characters in yeasts. Can J Microbiol 1983 ;29 :1462 -1464 .

11. Polonelli L, Morace G. Reevaluation of the yeast killer phenome- non. J Clin Microbiol 1986 ;24 :866-869 .

12. Polonelli L, Conti S, Gerloni M, Campani L, Pettoello-Mantovani M, Morace G. Production of yeast killer toxin in experimentally infected animals. Mycopathologia 1990; 110 :169-175 .

13. Hayman GT, Bolen PL. Linear DNA piasmid of Pichia inositovora are associated with a novel killer toxin activity. Curr Genet 1991 ;19 :389 -393 .

14. Wickner RB, icho T, Fujimura T, Winder WR. Expression of yeast L-A double-stranded RNA virus proteins produces derepressed replication: a sky- phenocopy. J Virol 1991 ;65 :155-161 .

15. Woods DR, Bevan EA. Studies on the nature of the killer factor produced by Saccharomyces cerevisiae. J Gen Microbiol 1968 ;81 :115-126 .

16. Eras P, Goldstein MN, Sherlock P. Candida infection of the gastrointestinal tract. Medicine 1972; 51 :367-379 .

17. Cappell MS, Mandell W, Grimes MM, Neu HC. Gastrointestinal histoplasmosis. Dig Dis Sci 1988 ;33 :353-360 .

18. Ekenna O, Sheretz RJ. Factor affecting colonization and dissemi- nation of Candida albicans from the gastrointestinal tract of mice. Infect Immun 1987 ;55 :1558-1563 .

19. Field M, Rao MC, Chang EB. Intestinal electrolyte transport and diarrheal disease (first of two parts). N Engl J Med 1989; 321 :800-806 .

20. Field M, Rao MC, Chang EB. Intestinal electrolyte transport and diarrheal disease (second of two parts). N Engl J Med 1989; 321 :879-883 .

21. Polonelli L, Morace G. Production and characterization of yeast killer toxin monoclonal antibodies. J Clin Microbiol 1987;25: 460 -462 .

22. Polonelli L, Castagnola M, Rossetti DV, Morace G. Use of killer toxins for computer-aided differentiation of Candida albicans strains. Mycopathologia 1985 ;91 :175-179 .

23. Lowry OH, Rosebrough N J, Farr AL, Randal RJ. Protein measure- ment with the Folin phenol reagent. J Biol Chem 1951; 193 :265- 275.

24. Dimler R J, Shaefer WC, Wise CS, Rist CE. Quantitative paper chromatography of glucose and its oligosaccharides. Ann Chem 1951; 24 :1411-1421 .

25. Sawant AD, Aheran DG. Involvement of a cell wall receptor in the mode of action of an anti-Candida toxin of Pichia anomala. Antimicrob Agents Chemother 1990; 34 :1331-1335 .

26. Polonelii L, Archibusacci C, Sestito M, Morace G. Killer system: a simple method for differentiating Candida albicans strains. J Clin Microbiol 1983; 17 :774-780 .

27. Woods DR, Bevan EA. Studies on the nature of the killer factor produced by Saccharomyces cerevisiae. J Gen Microbiol 1968; 51 :115-126 .

28. Mezoff AG, Giannella RA, Eade MN, Choen MB. Escherichia coil enterotoxin (STy) binds to receptors, stimulates guanyl cyclase, and impairs absorption in rat colon. Gastroenterology 1992; 102: 816 -822 .

29. Guandaiini S, Fasano A, Albini F, Marchesano G, Nocerino A, De Curtis M, Rubatelli FF, Pettenazzo A, Rubino A, Unconjugated bilirubin and the bile from light exposed Gunn rats inhibit intestinal water and electrolyte absorption. Gut 1988 ;29 :366-371 .

30. Harris MS, Dobbins JW, Binder HJ. Argumentation of neutral sodium chloride absorption by increased flow rate in rat ileum in vivo. J Clin Invest 1986 ;78 :431-438 .

31. Nocerino A, lafusco M, Guandalini S. Cholera toxin-induced small intestinal secretion has a secretory effect on the colon of the rat. Gastroenterology 1995; 108 :34-39 .

32. Guarino A, Tarallo L, Bemi-Canani R, Nocerino A, lafusco M, Raimondi F, Rubino A. Comparative effects of atrial natriuretic peptide and E. coil heat-stable toxin on rat intestinal transport. J Pediatr Gastroenterol Nutr 1992 ;15 :105-111 .

33. Hyun CS, Ahn J, Minhas BS, Cragoe E J, Field M. Ion transport in rabbit proximal colon: effects of sodium, amiloride, cAMP, and epinephrine. Am J Physiol 1994;266:G1071-G1082.

34. Fasano A, Baudry B, Pumplin DW, Wasserman SS, Tall BD, Ketley JM, Kaper JB. Vibrio cholerae produces a second enterotoxin, which affects intestinal tight junctions. Proc Natl Acad Sci USA 1991; 88:5242 -5246.

35. Guandalini S, Fasano A, Migliavacca M, Verga MC, Mastrantonio- Gianfrilli P, Ferrara A, Alessio M, Malamisura B, Galati P, Pantosti A, Rubino A. Pathogenesis of postantibiotic diarrhea caused by Clostridium difficile: an in vitro study in the rabbit intestine. Gut 1988; 29 :598-602 .

36. Wang CJK, Schwarz J. The etiology of interstitial pneumonia. Identification as Hansenula anomala of a yeast isolated from lung of infants. Mycopathol Mycol Applic 1958 ;9 :299-306 .

37. Haron E, Anaissie E, Dumphy F, McCredie K, Fainstein V. Han- senula anomalafungemia. Rev Infect Dis 1988; 10 :1182-1186 .

38. McGinnis MR, Walker DH, Folds JD. Hansenula polymorpha infection in a child with granulomatous disease. Arch Pathol Lab Med 1980; 104 :290-292 .

39. Nohinek B, Zee-Cheng CS, Barnes WG, Dail L, Gibbs HR. Infective endocarditis of a bicuspid aortic valve caused by Hansenula anomala. Am J Med 1987 ;82 :165-168 .

40. Klein AS, Tortora GT, Malowitz R, Greene WH. Hansenula anomala: a new fungal pathogen. Arch Intern Med 1988; 148 :1210- 1213.


41. Cook AH. The chemistry and biology of yeasts. Orlando, FL: Aca- demic, 1958.

42. KarmaliMA. Infectionbyverocytotoxin-producingEscherichiacoli. Clin Microbiol Rev 1989;2:15-38.

43. Gyles CL. Escherichia coli cytotoxins and enterotoxins. Can J Microbiol 1992;38:734-746.

44. Bussey H. K1 killer toxin, a pore-forming protein from yeast. Mol Microbiol 1991;5:2339-2343.

45. de la Pena P, Barros F, Gascon S, Ramos S, Lazo PS. Primary effects of yeast killer toxin. Biochem Biophys Res Commun 1980; 96:544-550.

46. Martinac B, Zhu H, Kubalsky A, Zhou X, Culbertson M, Bussey H, Kung C. Yeast K1 killer toxin forms ion channels in sensitive yeast spheroplasts and in artificial liposomes. Proc Natl Acad Sci USA 1990;87:6228-6232.

47. Castagliuolo I, LaMont JT, Letoumeau R, Kelly C, O'Keane JC, Jaffer A, Theoharides TC, Pothoulakis C. Neuronal involvement in the intestinal effects of Clostridium difficile toxin A and Vibrio cholerae enterotoxin in rat ileum. Gastroenterology 1994; 107:657-665.

48. Kurose I, Pothoulakis C, LaMont JT, Anderson DC, Paulson JC,

Miyasaka M, Wolf R, Granger DN. Clostridium difficile toxin A- induced microvascular dysfunction. Role of histamine. J Clin In- vest 1994; 94:1919-1926.

Received March 15, 1995. Accepted August 22, 1995. Address requests for reprints to: Massimo Pettoello-Mantovani,

M.D., Ph.D., Department of Pediatrics, Albert Einstein College of Medicine, Room 401, Forcheimer Building, 1300 Morris Park Ave- nue, Bronx, New York 10461. Fax: (718) 430-8925.

The first two authors contributed equally to the study. The authors thank Dr. Todd R. Olson, Director of Gross Anatomy

at Albert Einstein College of Medicine, New York, for advice on comparative anatomy; Dr. Thobias Kollmann, Department of Microbiol- ogy, Albert Einstein College of Medicine, for revising statistics; Dr. Marsha Goldstein, Department of Pathology, New York University Medical Center, for comments on histology; Dr. Pam Smarnworawong and Dr. Phil Rosenbloom, Pennsylvania Hospital, Thomas Jefferson University, Philadelphia, for critical reading of the manuscript; and Gaetano Polito and Marcello Porretta for dedicated work and techni- cal contribution.

Documents

Hansenula anomala killer toxin induces secretion and severe acute injury in the rat intestine