R E V I E W S

exploit such habitats. The development of new fluor- 5 Fields,P.I., Groisman,E.A.and Heffron,F. (1989) Science 243,
escence-based reporters permits the analysis of live bac- 1059-1062
terial gene responses with single-cell resolution. 6 Meighen,E. (1993)FASEBJ. 7, 1016-1022
7 Contag,C.H. et al. (1995) Mol. MicrobioI. 18, 593-603
8 Pettersson,J. et al. (1996) Science 273, 1231-1233
Acknowledgements
9 Mahan,M.J. etal. (1995) Proc. Natl. Acad. Sci. U. S. A. 92,
We thank Evi Strauss,Nina Salamaand Tim McDanielfor their input 669-673
and editingexpertise, and J. Shea and D. Holdenfor sharingstrains 10 Heitboff,D.M. et al. (1997) Proc. Natl. Acad. Sci. U. S. A. 94,
and unpublished results. This work was supported by PHS grant 934-939
AI 26195 and by unrestrictedgifts from Lederle-PraxisBiologicals 11 Rhen,M., Riikonen,P. and Taira, S. (1993) Mol. Microbiol. 10,
and Bristol-MyersSquibb. 45-56
12 Kwaik,Y.A.and Pederson,Li. (1996) Mol. Microbiol. 21,
References 543-556
1 Garciadel Portillo,F. and Finlay,B.B.(1995) Trends MicrobioI. 13 Zhang,J.P. and Normark,S. (1996) Science 273, 1234-1236
3, 373-380 14 Cubitt,A.B.et al. (1995) Trends Biochem. Sci. 20, 448-455
2 Finlay,B.B.and Cossart,P. (1997) Science 276, 718-725 15 Valdivia,R.H. and Falkow,S. (1996) Mol. Microbiol. 22,
3 Garciadel Portillo,F. et al. (1992) Mol. Microbiol. 6, 3289-3297 367-389
4 GarcfaVdscovi,E., Soncini,F.C. and Groisman,E.A.(1996) Cell 16 Shea,J.E. et al. (1996) Proc. Natl. Acad. Sci. U. S. A. 93,
84, 165-174 2593-2597

Pore-forming proteins in
pathogenic protozoan parasites
M. Fbtima Horta

lthough pathogenic pro- Pore-forming proteins (PFPs) may play inside the cell generates osmotic

A tozoa differ in many

ways, many of them pro-
duce cytolytic proteins that dis-
important roles in pathogenesis by
protozoan parasites by either directly
damaging the plasma membrane of the
pressure, causing a water influx.
The cell disrupts through a pro-
cess known as colloid-osmotic
rupt target cell membranes by host cells or ensuring intraceUular survival lysis3,4.
forming discrete channels in of the parasites by promoting their exit Death of host cells mediated
the lipid bilayer. These pore- from lysosomal vacuoles. The L e i s h m a n i a by PFPs from bacterial patho-
forming proteins (PFPs) are amazonensis pore-forming cytolysin, gens has frequently been de-
thought to play a significant leishporin, may play a crucial role in the scribed (for reviews, see Refs
role in the pathogenesis of pathogenesis of leishmaniasis. 4-6). Protozoan pathogens can
many protozoan infections 1,2. now be added to the list of PFP-
Channel formation by PFPs is M.F. Horta is at the producing organisms, and the
Dept de Bioquimica e Imunologia, notion of a cause-and-effect re-
a well-defined mechanism of Instituto de CiSncias Biol6gicas,
membrane damage used in bio- Universidade Federal de Minas Gerais, lationship between cytolysins
logical systems ranging from 31270-010 Belo Horizonte, MG, Brazil. and host tissue damage can
bacteria to vertebrates: the C9 tel: +55 31 441 5 7 7 7 , fax: +55 31 441 5963, now be extended to parasitic
component of the complement e-mail: [email protected] diseases.
system and perforin from cyto-
lytic lymphocytes in vertebrates are good examples. The outsider's attack
PFPs bind to the plasma membrane of a target cell E n t a m o e b a histolytica, which is responsible for human
and insert into the lipid bilayer by changing their con- amoebiasis, was the first protozoan in which PFPs were
formation and exposing hydrophobic domains. A discovered 7,s. Amoebiasis is an enteric illness that may
chain reaction follows in which the altered molecules spread to multiple organs when the parasite invades
bind others, oligomerizing around a central axis to form the intestinal mucosa. The main manifestation of the
transmembrane pores that grow in diameter by the disease is the destruction of host tissues, which results
successive annexation of new monomers. Some PFPs from parasite contact-mediated cytolysis. The cytolytic
act as monomers that adopt circular or semicircular effect has been attributed to a family of pore-forming
structures without mutual association. Ions and small peptides, termed amoebapores, that are localized within
molecules can then pass freely across the lipid bilayer, cytoplasmic granular vesicles of the trophozoites 9.
and the high concentration of macromolecules trapped Amoebapores form transmembrane channels with an
Copyright © 1997 Elsevier Science Ltd. All rights reserved. 0966 842X/97/$17.00 PII: S0966-842X(97)01109-8

TRENI)S IN MICP,()BI()LOGY 363 W,L 5 NO. 9 SEPTEMBER 1 997

REVIEWS

internal diameter averaging 2 nm (Ref. 7) and are most

active at acid pH (Refs 10,11). The cDNA-deduced
amino acid sequence of amoebapores reveals a peptide
of 77 amino acids 12. Both pathogenic and nonpatho-
genic isolates of E. histolytica possess cytolytic and
pore-forming activities. However, the nonpathogenic
isolates are threefold less cytolytic, with 60-80% less
pore-forming activity than the pathogenic forms 1°,11.
Naegleria fowleri, a free-living amoeboflagellate that
can cause invasive meningoencephalitis in humans,
also expresses a PFP. As the disease is a fulminating
infection with massive tissue destruction 13, it has been
suggested that a parasite-induced cytolytic process
takes place. Indeed, N. fowleri has been reported to
be cytotoxic in vitro ~4-16, and multiple cytolytic activ-
ities have been described in the amoebae lysates T M . Fig. 1. Balb/c peritoneal macrophages infected with Leishmania
The major cytolytic component is a heat-resistant pro- amazonensis. Amastigotes live inside parasitophorous vacuoles
tein, called N-PFP (Ref. 19), that is associated with until the macrophages are disrupted. Parasites are attached to
the parasite surface membrane 2°. It is a 66-kDa or the inner membrane of the vacuoles, which are often extended
to the borders of the cell in contact with the inner plasma mem-
50-54-kDa molecule under reducing or nonreducing brane (arrow heads). Scale bar = 5 pm.
conditions, respectively, that produces channels of
3.6-5.2 nm diameter. N-PFP accounts for 70-90% of
the total cytolytic activity of N. fowleri lysates 2° seems to be responsible for vacuolar disruption by the
and, therefore, may account for the tissue injury seen parasite. It forms large pores of -10 nm, which can lyse
in vivo. erythrocytes and nucleated cells. Analogous to LLO,
How E. histolytica and N. fowleri cytolysins kill host Tc-TOX is optimally active at pH 5.5, but is inactive at
cells in vivo is still unclear. It has been hypothesized that neutral pH, suggesting that it acts in the acidic intra-
amoebapores are secreted by trophozoites in vitro in re- cellular vacuoles formed shortly after parasite invasion.
sponse to stimuli that mimic parasite-target contact s. In fact, it can be localized to the lumen of the parasite-
However, recent work 21 indicates otherwise: amoeba- containing phagosomes 24. Significantly, treatment of in-
pore release appears to be associated with disintegration fected cells with drugs that raise the pH of intracellular
of the parasites, and intact amoebae fail to secrete the compartments inhibits parasite escape into the cytosol2s.
peptides, even after stimulation with concanavalin A,
bacterial lipopolysaccharide (LPS) or the calcium iono- 'Porin' out
phore, A23187. Likewise, N-PFP is not secreted by The most recently discovered member of the PFP-
N. fowleri in vitro 17, and contact between the target cell producing protozoa is another trypanosomatid. Re-
and the parasite seems to be a requisite for cytolysis~4,15. cently, we have shown that extracts of Leishmania
In E. histolytica, PFP secretion may occur by exocytosis amazonensis contain a cytolytic activity that damages
of the cytolysin-containing granular vesicles. erythrocytes and nucleated cells, including macrophages
Contact-dependent cytotoxicity is also thought to be (its vertebrate host cells) 26,27. This activity is mediated
mediated by a protozoan PFP in Trichomonas vaginalis by a heat-labile protein that is found in promastigotes
infection. However, it has been shown that cytolysis by and amastigotes and is optimally active at pH 5.0-5.5.
T. vaginalis can also occur in a contact-independent In contrast to Tc-TOX and LLO, it is still active at neu-
fashion, during which a change in pH triggers the re- tral pH. L. amazonensis-mediated haemolysis is in-
lease of the lytic molecules =. hibited by macromolecules, such as polyethyleneglycol
6000, indicating its colloid-osmotic nature, which is
The enemy from within typical of pore formation 2v. Indeed, whole-cell patch-
Insertion of parasite PFPs into plasma membranes does clamp experiments have demonstrated that L. amazon-
not always result in the immediate death of the host cells. ensis extracts induce voltage-dependent nonselective
For example, the facultative intracellular bacterium channels in the macrophage membrane, indicating that
Listeria monocytogenes, initially trapped inside phago- L. amazonensis cytolysin is a PFP (F.S.M. Noronha et
cytic vacuoles, is unable to survive in this compartment. al., unpublished), which we have called leishporin.
To escape into the cytoplasm, the bacterium uses its Preliminary results indicate that the size of leishporin
pore-forming cytolysin, listeriolysin O (LLO; Ref. 4), is -55-65 kDa.
to disrupt the vacuolar membrane 23 without affecting The mechanism of pore formation by leishporin re-
the plasma membrane and killing host cells. sembles that of classical PFPs. Patch-clamp experiments
Trypanosoma cruzi, the causative agent of Chagas' and osmotic protection by molecules of various diam-
disease, may rely on an analogous mechanism to gain a eters have shown that the diameters of the pores in-
safe environment inside its host cell: a C9-related PFP, crease in size (from -1.1 nm to >6.1 nm) with time or
called Tc-TOX, which is secreted by the intracellular with the concentration of the cytolytic extract. This
stage amastigote 24. Appropriately, Tc-TOX, a molecule suggests that pores are formed by aggregation or pol-
of 60-66 kDa (nonreduced) or 70-75 kDa (reduced), ymerization of single protein units.

TRENDS IN MICROBIOLOGY 364 voL 5 NO. 9 SEPTEMBER 1997

 R E V I E W S

But what is the functional role

of the L. amazonensis PFP in vivo ? Removed
Once inoculated into the vertebrate Pro-PFP peptide
host, the promastigotes (the extra-
cellular form in the sandfly vec-
tor) are ingested by macrophages
and transform into amastigotes. Limited /
In contrast to T. cruzi, the phago- proteolysis
lysosome environment does not of precursor
intimidate L. amazonensis, and
the amastigotes not only survive
but also replicate inside these
\
acidic parasitophorous vacuoles
without escaping into the cyto- • Lipid
plasm (see Fig. 1). However, at
later stages of the parasite life cycle, ~ ~ bilayer
both the vacuolar and plasma
membranes are disrupted 28 to re-
lease the amastigotes, which then
infect adjacent cells. All descrip-
tions of the disruption of infected )teolysis
cells are vague: the cells are said inhibitor
Functional
to 'burst', an event that is sup- pore
posed to be purely mechanical.
Considering that the parasito- PFP + inhibitor ~ ' l l k ~ V ~ Degraded
phorous vacuole can harbour nu- ~ A • inhibitor
merous parasites, thus swelling
and/or fusing with other phago- Fig. 2. Possible mechanisms of activation of Leishmania amazonensis pore-forming protein (PFP).
cytic vesicles as the parasites multi- A cytolytically inactive pro-PFP may be cleaved by limited proteolysis, removing a peptide and yield-
ply28,29 (see Fig. 1), purely me- ing the active PFP. Alternatively, the PFP may initially be bound to an inhibitor, which is subsequently
degraded by proteolysis, releasingthe active PFP. In both cases, the active PFP can polymerize into
chanical rupture seems unlikely. the lipid bilayer, forming a transmembrane functional pore.
Our recent findings of an acid-
active PFP in L. amazonensis sug-
gest that amastigotes may use this molecule to disrupt neutral pH. Another possibility could be that the con-
host cells26,27. nection between the phagolysosome outer membrane
and the plasma inner membrane of L. amazonensis-
Launching the missile: aiming at two targets infected macrophages (see Fig. 1) favours the interac-
Leishporin does not exist freely in the cytosol of L. tion of the PFP with the inner membrane. It is also
amazonensis: it co-sediments with the membrane frac- possible that these cytolysins are tightly regulated and
tion but appears to be trapped inside membranous act only at specific stages of the parasite life cycle, in
vesicles in a soluble form 27. However, promastigotes do which case it would be reasonable to assume that the
not secrete the active molecule, even upon stimulation plasma membrane of T. cruzi-infected macrophages
with calcium ionophores (A23187 or ionomicin) z7 or could also be disrupted by Tc-TOX at later stages.
concanavalin A (F.S.M. Noronha and M.F. Horta, Our most recent data on leishporin concerns its mode
unpublished). We are currently investigating whether of activation. We have found that the cytolytic activity
amastigotes are able to secrete leishporin. Amastigotes of parasite extracts increases approximately fivefold
lodge in the macrophages by firmly adhering to the in- when kept at 4°C for 7d or at 37°C for 24h. This in-
ner surface of the phagolysosomes, which, in turn, can crease is totally blocked by protease inhibitors, and
become distended as the parasites multiply to the bound- both parasite-derived and exogenous soluble proteases
aries of the cell (Fig. 1). In vivo, it is possible that this can generate cytolytic activity (F.R. Almeida-Campos
intimate membrane contact triggers a secretion mecha- and M.F. Horta, unpublished). These and other results
nism. As the lysin seems to be a luminal protein, a ves- have led us to believe that the activity of leishporin
icle exocytosis mechanism for delivering the cytolysin arises from proteolysis. We are investigating two hy-
to its target can be envisaged. potheses: (1) that the active PFP is produced by limited
In contrast to T. cruzi, L. amazonensis amastigotes proteolysis of an inactive precursor and (2) that an in-
are faced with the task of disrupting two membranes: hibitory molecule bound to the cytolysin is proteolyti-
first the phagolysosomal membrane and then the plasma cally degraded to release the active cytolysin (Fig. 2).
membrane. A puzzling question is why T. cruzi PFP
only lyses the phagosomal membrane, leaving the host The war is just beginning
plasma membrane intact. The pH dependence of the Much investigation is needed to substantiate our as-
two PFPs may offer one explanation: whereas Tc-TOX sumptions and to determine whether the production
is active only at acid pH, leishporin is also lytic at of PFPs is a general feature of pathogenic protozoans.

TRENDS IN MICROBIOLOGY 365 VOL. 5 NO. 9 SEPTEMBER 1997

REVIEWS

References
Questions for future research 10jcius, D. and Young,J.D-E. (1990) Parasitol. Today 6,
• Is the expression of pore-forming proteins (PFPs) a common fea- 163-165
ture of all pathogenic protozoa? 2 Andrews, N.W. and Portnoy, D.A. (1994) Trends Microbioi. 2,
• What signals trigger protozoan PFP-mediated disruption of plasma 261-263
and/or vacuolar membranes in vivo? 3 Young,J.D-E and Cohn, Z.A. (1987) Adz,. ImmunoL 41,
• Do all protozoan cytolysins need to be previously activated to 269-332
create pores? 4 Bhakdi, S. and Tranun-Jensen, J. (1988) Prog. Allergy 40, 1-43
• How do PFP-producing parasites avoid self-lysis? Do they secrete 5 Ludwig,A. (1996) MicrobioIogia 12, 281-296
a lyrically inactive PFP that becomes activated outside the para- 6 Tweten, R.K. (1995)in Virulence Mechanisms of Bacterial
site? Does the parasite express membrane molecules that abort Pathogens (2nd edn) (Roth, J.A. et aI., eds), pp. 207-229, ASM
channel formation? Press
° How would PFP-knockout parasites behave? Would they still be 7 Lynch, E.C., Rosenberg, I.M. and Gitter, C. (1982) EMBO J. 1,
virulent? 801-804
° Are all protozoan PFPs structurally related? 8 Young,J.D-E et al. (1982)J. Exp. Med. 156, 1677-1690
9 Do&on, J.M. and Petri, W.A. (1994) Parasitol. Today 10, 7-8
10 Keller, F. et ai. (1988)J. Protozool. 35, 359-365
11 Leippe, M. et al. (1993) Mol. Biocbem. Parasitol. 59,
Nevertheless, recent data make the hypothesis of proto- 101-110
zoan PFPs as virulence factors quite attractive. 12 Leippe, M. et al. (1994) Mol. Microbiol. 14, 895-904
Two other protozoans should be considered as po- 13 Marciano-Cabral, F. (1988) Microbiol. Rev. 52, 114-133
tential PFP-producers: P l a s m o d i u m f a l c i p a r u m and 14 Brown, T. (1979)J. Med. Microbiol. 12, 363-371
T o x o p l a s m a gondii. These species reside in parasito- 15 Marciano-Cabral, F. et at. (1982) J. Parasitol. 68, 1110-1116
phorous vacuoles within erythrocytes and nucleated 16 Fullford, D.E., Bradley, S.G. and Marciano-Cabral, F. (1985)
cells, respectively, where the presence of channels 3° or ]. Protozoot. 32, 176-180
putative channels 31 has been reported. Although it is not 17 Lowrey, D.M. and McLaughlin,j. (1984) Infect. lmmun. 45,
certain whether these pores are of parasitic origin, re- 731-736
18 FuIford, D.E. and Marciano-Cabral, F. (1986)J. ProtozooI. 33,
cent data are consistent with this hypothesis 32,33.
498-502
Identifying the leishporin gene, disclosing its mecha- 19 Young,J.D-E. and kowrey, D.M. (1989)J. Biol. Chem. 264,
nism of action and defining its function(s) are primary 1077-1083
goals. Although it does not preclude other roles for leish- 20 Lowrey,D.M. and McLaughlin,J. (1985) Infect. Immun. 50,
porin (see Ref. 27), the assumption that this molecule 478-482
is involved in macrophage rupture leads to a shift in the 21 Leippe, M. et at. (1995) Parasitology 111,569-574
current thinking that macrophage bursting is a direct 22 Fiori, P.L. et aI. (1996) Microb. Pathog. 20, 109-118
result of excessive parasite burden. This PFP may be a 23 Bide&i, J., Youngman, P.C. and Pormoy, D.A. (1990) Nature
crucial molecule for the pathogenesis of leishmani- 345, 175-176
asis, acting as both a tissue-damaging and an infection- 24 Andrews, N.W. el al. (1990) Cell 61, 1277-1287
25 Ley, V. et al. (1990)]. Exp. Med. 171,401-413
spreading factor. As leishmaniasis encompasses self- 26 Noronha, F.S.M., Ramalho-Pinto, F.J. and Horta, M.F. (1994)
healing skin ulcers and mucocutaneous lesions to fatal Braz. ]. Med. Biol. Res. 27, 477-482
visceral forms 28, it would be interesting to search for 27 Noronha, F.S.M., Ramalho-Pinto, F.J. and Horta, M.F. (1996)
a correlation between the pathologies caused by the Infect. Immun. 64, 3975-3982
different species and the presence of PFPs. Extending 28 Chang, K-P., Fong, D. and Bray, R.S. (1985) in Human Parasitic
this correlation to other protozoan parasites would Diseases (Vol. 1)(Chang, K-P. and Bray, R.S., eds), pp. 1-30,
open a whole new field of investigation. The scenario Elsevier
looks promising. 29 Veras, P.S.T. et al. (1995) Infect. Immun. 63, 3502-3506
30 Desai, S.A., Krogstad, D.J. and McCleskey, E.W. (1993) Nature
Acknowledgements 362, 643-646
I thank F. Juarez Ramalho-Pinto, F~itimaS.M. Noronha, Flfivia R.A. 31 Schwab,J., Beckers, C.J.M. and Joiner, K.A. (1994) Proc. Natl.
Campos, Jane Lima dos Santos and Santuza M. Teixeira for their Acad. Sci. U. S. A. 91,509-513
invaluable assistance. The research in my laboratory is supported by 32 Desai, S.A. and Rosenberg, R.L. (1997) Proc. Natl. Acad. Sci.
Financiadora de Estudos e Proletos (FINEP), Fundaq~o de Amparo 5 U. S. A. 94, 2045-2049
Pesquisa de Minas Gerais (FAPEMIG)and Conselho Nacional de 33 Ossorio, P.N., Dubremetz, J-F. and Joiner, K.A. (1994) J. Biol.
Desenvolvimento Cientffico e Tecnoldgico (CNPq). Chem. 269, 15350-15357

In the other Trends journals

A selection of recently published articles of interest to TIM readers.

• HIV and chemokines: ligands sharing cell-surface receptors, by P.R. Clapham - Trends in Cell Biology 7 , 2 6 4 - 2 6 8
• Salicylic acid and disease resistance in plants, by J. Durner, J. Shah and D.F. Klessig- Trends in Plant Science 2 , 2 6 6 - 2 7 4
• Bacterial surface display: trends and progress, by S. Stfihl and M. U h l 6 n - Trends in Biotechnology 15, 1 8 5 - 1 9 2

TRENDS IN MICROBIOLOGY 366 VOL. 5 NO. 9 SEPTEMBER 1997