Proc. Natl. Acad. Sci. USAVol. 88, pp. 2807-2810, April 1991Microbiology

Malaria parasite chitinase and penetration of the mosquitoperitrophic membrane

(Plasmodium gallinaceum/Aedes aegypti/vector competence)

MARCEL HUBER*, ENRICO CABIBt, AND Louis H. MILLER*:*Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, and tLaboratory of Biochemistry and Metabolism, National Instituteof Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD 20892

Contributed by Louis H. Miller, December 31, 1990

ABSTRACT Malaria parasites (ookinetes) appear to digestthe peritrophic membrane in the mosquito midgut duringpenetration. Previous studies demonstrated that lectins specificfor N-acetylglucosamine bind to the peritrophic membrane andproposed that the membrane contains chitin [Rudin, W. &Hecker, H. (1989) Parasitol. Res. 75, 268-2791. In the presentstudy, we show that the peritrophic membrane is digested bySerratia marcescens chitinase (EC 3.2.1.14), leading to therelease of N-acetylglucosamine and fragmentation of the mem-brane. We also report the presence of a malaria parasitechitinase that digests 4-methylumbelliferyl chitotriose. Theenzyme is not detectable until 15 hr after zygote formation, thetime required for maturation of the parasite from a zygote toan ookinete, the invasive form of the parasite. At 20 hr, theenzyme begins to appear in the culture supernatant. Thechitinase extracted from the parasite and found in the culturesupernatant consists of a major band and two minor bands ofactivity on native polyacrylamide gel electrophoresis. Thepresence of chitin in the peritrophic membrane, the disruptionof the peritrophic membrane during invasion, and the presenceof chitinase in ookinetes suggest that the chitinase in ookinetesis used in the penetration of the peritrophic membrane.

The peritrophic membrane found in the midgut of arthropodsis defined as a membranous sac surrounding the ingestedblood (1). The function of the peritrophic membrane in thevarious arthropod orders is unclear, but there is speculationthat in bloodsucking insects it is one of the factors determin-ing the competence for the transmission of parasites (2, 3).These insects form a peritrophic membrane in response to themechanical distention resulting from the uptake of blood (4).In Aedes aegypti, formation starts immediately after theblood meal, but about 12 hr elapse before it attains a maturestructure (5, 6). Plasmodium gallinaceum, an avian malariaparasite that infects A. aegypti, develops motile ookinetes inthe blood meal in 16-20 hr (3, 7). Parasites begin to appear inthe midgut epithelial cells around 24 hr, but the majorinvasion occurs around 30 hr (3). As the peritrophic mem-brane is fully formed in 12 hr and does not dissolve until 48hr after the blood meal, the parasite must traverse theperitrophic membrane to move from the blood meal to theepithelial cells. A recent ultrastructural study (3) produceddirect evidence that the peritrophic membrane is a barrier toinfection. The peritrophic membrane surrounding the pene-trating ookinetes lost its laminated structure and was lesselectron dense. Electron-dense material extended from theanterior end of the parasite. A similar modification waspreviously observed with Babesia microti passing throughthe peritrophic membrane of Ixodes ticks (8). Because evi-dence has been presented that the peritrophic membranecontains chitin (9), we searched for chitinase activity in the

parasite that might be used to cross the peritrophic mem-brane. We report here that P. gallinaceum ookinetes produceand secrete chitinase (EC 3.2.1.14). The appearance of theenzyme in the parasite after complete maturation of ooki-netes and its secretion into the medium suggest that chitinasemay be one of the enzymes involved in the penetration of thechitin-containing peritrophic membrane.

MATERIALS AND METHODSParasite and Mosquito. The 8A strain of P. gallinaceum

was maintained by subpassage in White Leghorn chickens.The Liverpool/black eye strain of A. aegypti was raised at26°C and 80% relative humidity and fed on sugar solution adlibitum.

Preparation of Parasites and Leukocytes (WBCs). Sexualstage parasites of P. gallinaceum were isolated as described(10). Parasites were incubated at 26°C in complete M199medium (GIBCO) supplemented with 1 mM L-glutamine, 50,g of streptomycin per ml, 50 units of penicillin per ml, 40 ,ugof gentamycin per ml, and 20 units of nystatin per ml. Thewhole procedure was carried out under sterile conditions.The parasite cultures were tested for bacterial and fungalcontamination on chocolate agar II slants (25°C and 37°C),thioglycollate medium (37°C), and Sabouraud dextrose agarslants (25°C) (Becton Dickinson). They were evaluated forbacterial and fungal contaminations after 2 wk.To test whether the chitinase was from host-cell contam-

ination, WBCs were isolated from noninfected chickensfollowing the same protocol, omitting wheat germ agglutina-tion and filtration through glass wool. They were then grownunder the same conditions as the parasites.

Extraction of Parasites and WBCs. Parasites and WBCswere extracted in the following buffer by four cycles offreeze/thaw: 0.1 M sodium phosphate, pH 6.8/1 mM phe-nylmethylsulfonyl fluoride/1 mM EGTA/50 ug of antipainper ml/0.5 ,ug of pepstatin per ml/0.5 ,g of leupeptin per ml.After centrifugation (13,000 x g at 4°C for 5 min), thesupernatant was used for measurement of chitinase activity.

Culture Supernatants. Culture supernatants were concen-trated about 10-fold in a centrifugal microconcentrator (Mi-crosep, 10-kDa cutoff; Filtron Technology, Northborough,MA) and then assayed for chitinase activity.Measurement of Chitinase Activity. Chitinase activity was

assayed in two ways: (i) 4-Methylumbelliferyl-N,N',N"-triacetyl-/-chitotrioside (Calbiochem) was used at a concen-tration of 125 uM in 0.1 M sodium phosphate (pH 6.8) (11).After incubation at 26°C, the reaction was terminated by theaddition of 950 ul of 0.5 M glycine-NaOH (pH 10.5). Theamount of liberated product was determined on a Perkin-

Abbreviation: WBC, leukocyte.*To whom reprint requests should be addressed at: Laboratory ofParasitic Diseases, National Institute of Allergy and InfectiousDiseases, Building 4, Room 126, Bethesda, MD 20892.

2807

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

June

10,

202

1

Proc. Natl. Acad. Sci. USA 88 (1991)

Table 1. Chitinase digestion of peritrophic membranepH 4.8 pH 6.3

With Without With WithoutExp. chitinase chitinase chitinase chitinase

1 0.120 0.003 0.060 0.0072 0.068 0.000 0.052 0.0003 0.072 0.008 0.009 0.0144 0.093 0.004 ND NDND, not done. Peritrophic membranes were incubated with S.

marcescens chitinase (with chitinase) or in buffer alone (withoutchitinase). One peritrophic membrane was used per assay. Thesupernatant after centrifugation was then exposed to glusulase. Thenumbers give the absorption at 585 nm in an assay for N-acetylglu-cosamine.

Elmer fluorescence spectrophotometer (excitation at 350 nm,emission at 440 nm) using a standard curve of 4-methylum-belliferone. (ii) Chitinase activity of cell extracts and culturesupernatants was analyzed in a nondenaturing 4-15% gradientpolyacrylamide gel (Integrated Separation Systems, HydePark, MA) using a glycol chitin overlay gel in 0.1 M sodiumphosphate (pH 6.8) and Calcofluor white for detection (12).

Lectin Binding to Peritrophic Membrane. A. aegypti werefed on a noninfected chicken. Twenty to 30 hr later, theperitrophic membranes were removed and incubated with a1: 200 dilution of fluorescein isothiocyanate-labeled lectins(Vector Laboratories) in 10 mM Hepes/150 mM NaCl, pH7.5, for 30 min. They were then washed for 15 min in the samebuffer and mounted on slides. For inhibition studies, 0.5 MN-acetylglucosamine was used.

Chitinase Digestion of Peritrophic Membrane. Peritrophicmembranes were removed from A. aegypti 24-30 hr after anoninfectious blood meal. They were incubated in either 0.1M sodium acetate (pH 4.8) or 0.1 M potassium phosphate (pH6.3) and 0.5 unit per ,lI of purified chitinase from Serratiamarcescens (13). After incubation of 12 hr at 26°C, thesamples were centrifuged (13,000 x g, 5 min) and thechitinase was inactivated by incubating at 65°C for 30 min.The pH was adjusted by the addition of 5 vol of 0.1 M sodiumphosphate (pH 6.8), and any di- or oligosaccharide in thereaction mix was further digested (60 min, 37°C) with glusu-lase (1 1.l; New England Nuclear) as a source of ,-N-acetylglucosaminidase. The amount of released N-acetylglu-cosamine residue was determined as described (14).

RESULTSPresence of Chitin in A. aegypt Peritrophic Membrane. It

has been previously shown that the peritrophic membrane ofA. aegypti reacts with lectins specific for N-acetylglu-

O hr 10 hr

cosamine (9). We confirmed the binding of three fluoresci-nated lectins, wheat germ agglutinin, the succinylated form ofwheat germ agglutinin, and Datura stramonium lectin, to theperitrophic membrane and the blocking of the binding byN-acetylglucosamine (data not shown). To test specificallyfor the presence of chitin in the peritrophic membrane, wedigested the peritrophic membrane with a highly purifiedchitinase from S. marcescens. After chitinase digestion, thesample was centrifuged and the supernatant was digestedwith glusulase as a source of P3-N-acetylglucosaminidase,because the chitinase does not degrade chitin to a N-acetyl-glucosamine monomer. By this procedure, N-acetylglu-cosamine was released at pH 4.8 and pH 6.3 (Table 1). Thereason that one of seven samples released no N-acetylglu-cosamine is unknown. The release of N-acetylglucosamineby chitinase indicated that the peritrophic membrane indeedcontains chitin. The presence of chitin was further confirmedby fluorescence after exposure to Calcofluor white, whichhas a high affinity for chitin (15).As the parasite appears to digest the peritrophic membrane

as it passes through the membrane (3), we determined theeffect of chitinase on the integrity of the membrane. Apurified chitinase from S. marcescens had no protease ac-tivity at pH 4.8 (13). Peritrophic membranes exposed to 1 unitof this enzyme per ul disintegrated in 36 hr at room temper-ature (data not shown). The control membrane in the bufferalone remained intact.

Identification of Parasite Chitinase. Since the membranecontains chitin, we investigated whether the parasite synthe-sized a chitinase and secreted it in order to pass through themembrane. After gametocytes are taken into the mosquitomidgut during a blood meal, gametes emerge and fertilize.The zygotes develop in 10 hr into retort forms and in 15 hr intoookinetes (7), the stage that penetrates the peritrophic mem-brane. In in vitro culture, zygotes also develop into retortforms in 10 hr and into ookinetes in around 15 hr (Fig. 1). Thein vitro forms were studied for chitinase activity. The activitywas compared at 0 hr, when only zygotes, unfertilizedgametes, and around 10% WBCs were seen, and at 20 hr,when around 50%o of the parasites were mature ookinetes.The 0 time extract contained no chitinase activity (Table 2).At 20 hr, activity was in the parasite extract, and at 25 hr,chitinase activity was also found in the culture supernatant,presumably secreted by the parasite, as no parasite lysis wasobserved by light microscopy.The amount of 4-methylumbelliferyl-N, N', Nn-triacetyl-,B-

chitotrioside cleaved by the parasite enzyme from the culturesupernatant increased linearly with time and enzyme con-centration (data not shown). Boiling the enzyme before theassay destroyed its activity.

1 I- hr

FIG. 1. Development of P. gallinaceum with time after fertilization: zygote (Left), retort (Center), and ookinete (Right). The numbersindicate the time after fertilization. (Bar = 10 ,Lm.)

2808 Microbiology: Huber et al.

Dow

nloa

ded

by g

uest

on

June

10,

202

1

Proc. Natl. Acad. Sci. USA 88 (1991) 2809

Table 2. Enzymatic activity with 4-methylumbelliferyl-N,N',N"-triacetyl-8-chitotriosideSpecific

activity,* pmolTime, per 106 cells

Date Sample hr % ookinetes per hr Contaminationt11/1/90 P.g. extract 0 0 +11/1/90 P.g. extract 20 61 1025 +7/25/90 P.g. extract 20 50 12609/4/90 P.g. extract 20 50 175010/16/90 P.g. medium 25 85 1790 +10/29/90 P.g. medium 25 55 1660 +10/17/90 WBC extract 0 4 +10/17/90 WBC extract 20 9 +8/2/90 WBC extract 20 117/31/90 WBC extract 20 1010/17/90 WBC medium 25 8 +

P.g., P. gallinaceum.*Specific activities were calculated with respect to the number ofookinetes for P. gallinaceum samples.tCultures were scored as contaminated (+) when they showed growth in at least one of the test mediadescribed in the text and negative (-) for contamination when no bacteria or fungi grew.

Nondenaturing polyacrylamide gels run at pH 8.9 resolvedthree bands of activity, with band 2 always having thegreatest activity (Figs. 2 and 3). We could not determinewhether the various bands resulted from multiple chitinaseproteins or were the result of a single protein interacting withitself or other proteins that do not have chitinase activity.Attempts to renature the enzyme activity after electropho-resis on a denaturing gel (SDS/PAGE) were unsuccessful.The activity was not present at 0, 5, or 10 hr, times when

no mature ookinetes were observed (Figs. 2A and 3). Theactivities in the native gels were first evident in parasiteextracts at 15 hr (Fig. 2), when mature ookinetes were firstobserved (Fig. 1). The activity was still present in parasitesat 20 and 25 hr, when no further development was observed.At 15 hr, when the parasites were first found to have chitinaseactivity, none was found in the culture supernatant. Thechitinase first appeared in culture supernatant at 20 hr andwas also observed at 25 hr (Fig. 2A).

A

Cell Extrac'

B\UrJ e'N ed urn

Exclusion of Nonparasite Origin of the Chitinase. Althoughsterile technique was used to isolate and culture parasites, in-30% of the isolates contamination was found in the thio-glycollate medium or on chocolate II agar slants at 37°C;however, the numbers of organism were small, -5 coloniesper 100 Al of culture on chocolate II agar. More importantly,chitinase activity was identical in cultures that were sterilewhen compared to those that had the minor contaminants(Table 2). The qualitative characteristics of the bands innative gels (bands 1-3) were identical in samples with andwithout contaminants (Fig. 2 and data not shown).As chicken serum contains chitinase (16), it was also

important to exclude production of chitinase by contaminat-ing host cells. The chitinase activity in chicken serum has aband of slower mobility in native gels than band 1 and anotherone between parasite bands 2 and 3 (data not shown). Theparasite preparations had at most 10% contaminating WBCs.WBCs equal in number to the entire parasite preparationproduced -100-fold less chitinase activity than parasiteswhen measured with chitotriose (Table 2). In addition, low

Parasi te White cells

E E EM E E E M

Time:(hours)

*:. ..: ..

.. .: .. :!4rS t-;K.,

Proc. Natl. Acad. Sci. USA 88 (1991)

chitinase activity in the WBC preparations was present at 0time, when there was no activity in the parasite preparations(Table 2). Furthermore, no activity was seen in native gels ofWBC extracts taken at 0, 5, and 22 hr or in the culturesupernatant (Fig. 3). We conclude that the activity in theparasite culture could not have been derived from contami-nating host WBCs.

DISCUSSIONThe peritrophic membrane forms after a blood meal inblood-sucking arthropods and may act as a barrier for inva-sion of ingested microorganisms. There are three strategiesthat parasites use in blood-sucking arthropods to move fromthe blood meal to the midgut epithelium. (i) Parasites such asfilaria invade before the formation of the peritrophic mem-brane (17). (ii) Parasites such as leishmania may persist in theremnants of the blood meal until the peritrophic membranedissolves and then attach to epithelium (18). (iii) The parasitemay penetrate the membrane after it is formed. Malariaparasites fall into the third category. As the parasite was seenin ultrastructural studies to disrupt the membrane (3), thebiochemical basis of penetration would depend on the chem-ical composition of the membrane. The peritrophic mem-brane had been suggested, from indirect studies, to consist ofchitin (9). The demonstration in the present study of therelease of N-acetylglucosamine by the action of chitinaseprovides direct evidence for the presence of chitin in A.agypti peritrophic membrane. Furthermore, the membranedisintegrated in a purified chitinase preparation that had noprotease activity under the conditions ofthe study. Thus, oneparasite enzyme that might play a role in the disruption of themembrane would be a chitinase.

In the present study, we observed chitinase activity inmalaria ookinetes, the stage of the parasite found in themosquito midgut. Multiple bands of activity were found onnative gels when glycol chitin was used as substrate. It isunknown whether there are multiple genes that encodemultiple chitinases or whether the enzyme associates withitself or other proteins in a native gel. The activity did notresult from bacterial or fungal contamination, as parasitesthat were grown under sterile conditions in vitro containedthe three chitinase bands. Chitinase is found in chickenserum, but chicken serum was not used in our cultures.Although WBCs were seen in culture, little chitinase activityderived from WBCs when their numbers were equal to thenumbers of parasites in culture, and no activity was seen onthe native gels of WBC extracts.The chitinase appeared at the time when the parasite in

vitro was morphologically similar to the form of the parasitein vivo that penetrates the peritrophic membrane. Further-more, the parasite secretes the activity into the surrounding

medium. After attachment to the peritrophic membrane, thesecretion of chitinase may be a mechanism by which theparasite passes through the membrane so that it gains accessto the gut epithelium.There have been few biochemical studies that describe the

host-parasite interaction of malaria in the mosquito. Thefocus on the parasite in the mammalian host in part reflectsthe feeling that little could be done to modify the infection bysuch studies. Now, two avenues of attack are evident in themosquito: (i) modification ofthe mosquito through moleculartechniques that would block the life cycle in the mosquito and(ii) vaccines against the sexual stages of the parasite thatinduce host antibodies that are ingested and act in the bloodmeal. Such approaches to block the life cycle in the mosquitomay be developed through knowledge of the barriers toinfection and how the parasite deals with these barriers. Forexample, a vaccine against parasite chitinase could block theability of the parasite to cross the peritrophic membrane andto infect mosquitoes.

We thank Dr. Klaus-Peter Sieber for the initial observation leadingto this work, Andre Laughinghouse and Lynn Lambert for experttechnical assistance, Douglas C. Seeley for raising the mosquitoes,and David Keister for helpful discussions and encouragement. Spe-cial thanks go to June Park for her help during the summer.

1. Richards, A. G. & Richards, P. A. (1977) Annu. Rev. Entomol.22, 219-240.

2. Ponnudurai, T., Billingsley, P. F. & Rudin, W. (1988) Parasi-tol. Today 4, 319-321.

3. Sieber, K.-P., Huber, M., Kaslow, D., Banks, S. M., Torii, M.,Aikawa, M. & Miller, L. H. (1991) Exp. Parasitol., in press.

4. Chiang, R. G. & Davey, K. G. (1988) Science 241, 1665-1667.5. Freyvogel, T. A. & Staubli, W. (1965) Acta Trop. 22, 118-147.6. Perrone, J. B. & Spielman, A. (1988) Cell Tissue Res. 252,

473-478.7. Stohler, H. (1957) Acta Trop. 14, 302-352.8. Rudzinska, M. A., Spielman, A., Lewengrub, S., Piesman, J.

& Karakashian, S. (1982) Cell Tissue Res. 221, 471-481.9. Rudin, W. & Hecker, H. (1989) Parasitol. Res. 75, 268-279.

10. Kaushal, D. C., Carter, R., Rener, J., Grotendorst, C. A.,Miller, L. H. & Howard, R. J. (1983) J. Immunol. 131, 2557-2662.

11. Kuranda, M. J. & Robbins, P. W. (1987) Proc. Natl. Acad. Sci.USA 84, 2585-2589.

12. Trudel, J. & Asselin, A. (1989) Anal. Biochem. 178, 362-366.13. Roberts, R. L. & Cabib, E. (1982) Anal. Biochem. 127, 402-

412.14. Reissig, J. L., Strominger, J. L. & Leloir, L. F. (1955) J. Biol.

Chem. 217, 959-966.15. Maeda, H. & Ishida, N. (1967) J. Biochem (Tokyo) 62, 276-278.16. Lundblad, G., Hederstedt, B., Lind, J. & Steby, M. (1974) Eur.

J. Biochem. 46, 367-376.17. Perrone, J. B. & Spielman, A. (1986) J. Parasitol. 72, 723-727.18. Walters, L. L., Modi, G. B., Tesh, R. B. & Burrage, T. (1987)

Am. J. Trop. Med. Hyg. 36, 294-314.

2810 Microbiology: Huber et al.

Dow

nloa

ded

by g

uest

on

June

10,

202

1