Plasmodium ookinetes coopt mammalian plasminogento invade the mosquito midgutAnil K. Ghosha, Isabelle Coppensa, Henrik Gårdsvollb, Michael Plougb, and Marcelo Jacobs-Lorenaa,1

aDepartment of Molecular Microbiology and Immunology, Malaria Research Institute, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD21205; and bFinsen Laboratory, Rigshospitalet, Copenhagen Biocenter, 2200 Copenhagen N, Denmark

Edited* by Peter Agre, Johns Hopkins Malaria Research Institute, Baltimore, MD, and approved August 30, 2011 (received for review March 6, 2011)

Ookinete invasion of the mosquito midgut is an essential step forthe development of the malaria parasite in the mosquito. Invasioninvolves recognition between a presumed mosquito midgut re-ceptor and an ookinete ligand. Here, we show that enolase lines theookinete surface. An antienolase antibody inhibits oocyst develop-ment of both Plasmodiumberghei and Plasmodium falciparum, sug-gesting that enolase may act as an invasion ligand. Importantly, wedemonstrate that surface enolase captures plasminogen from themammalian blood meal via its lysine motif (DKSLVK) and that thisinteraction is essential for midgut invasion, because plasminogendepletion leads to a strong inhibition of oocyst formation. Althoughaddition of recombinant WT plasminogen to depleted serum res-cues oocyst formation, recombinant inactive plasminogen does not,thus emphasizing the importance of plasmin proteolytic activity forookinete invasion. The results support the hypothesis that enolaseon the surface of Plasmodium ookinetes plays a dual role in midgutinvasion: by acting as a ligand that interacts with the midgut epi-thelium and, further, by capturing plasminogen, whose conversionto active plasmin promotes the invasion process.

transmission-blocking vaccines | insect vectors of disease

Few weapons are available to fight the unbearable burden ofmalaria (1, 2). For transmission to occur, Plasmodium, the

causative agent of malaria, must complete a complex and in-completely understood developmental program in its vectormosquito (3). Plasmodium differentiation in the mosquito beginsin the midgut lumen with fertilization and differentiation of theresulting zygote into a motile ookinete. After crossing the midgut,the ookinete differentiates into an oocyst that, when mature,releases thousands of motile sporozoites that, in turn, invade thesalivary glands. Ookinete invasion of the midgut is a crucial step,the failure of which results in aborted development and un-successful transmission. Little is known about themolecular eventsthat lead tomidgut invasion. Circumstantial evidence suggests thatinvasion of the mosquito midgut by ookinetes requires specificinteractions between the parasite and the epithelial surface (4, 5).In an attempt to elucidate these interactions at themolecular level,we have previously screened a phage display library for peptidesthat bind to the Anopheles midgut epithelium. This screen led tothe identification of Salivary gland and Midgut peptide 1 (SM1), adodecapeptide that binds tightly to the midgut luminal surfaceand, importantly, efficiently inhibits Plasmodium berghei ookineteinvasion (4). Based on these results, we hypothesized that SM1mimics the domain of an ookinete surface protein ligand involvedin the recognition of a midgut receptor.Here, we show that SM1 is a mimotope of the ookinete surface

protein enolase. Moreover, enolase interacts with the abundantmammalian plasma protein plasminogen, and this interactionappears to be essential for progression of the Plasmodium lifecycle in the mosquito. The results suggest that in evolution, a closerelationship developed between the three relevant organisms,with the parasite having coopted plasminogen from its mamma-lian host to invade its mosquito vector.

ResultsAnti-SM1 Antibody Recognizes Ookinete Surface Component(s). Ourprevious phage display library screening led to the identificationof the SM1 dodecapeptide that not only binds to the luminalsurface of the mosquito midgut but, importantly, strongly inhibitsookinete invasion (4). These results led to the hypothesis thatSM1 mimics (mimotope) an ookinete surface ligand that interactswith a putative midgut receptor and that this interaction is re-quired for invasion. According to this premise, the peptide wouldbind to, and sterically shield, the putative mosquito receptor,precluding its interaction with the ookinete ligand. This hypoth-esis advocates for the similarity between SM1 and an unknownookinete invasion ligand. To test this prediction, we produced ananti-SM1 antibody and used it as a probe in immunofluorescenceassays to determine whether the antibody recognizes an ookinetesurface component. As shown in Fig. 1A, the antibody actuallybinds to the surface of both P. berghei and Plasmodium falciparumookinetes. Control experiments indicated that antibodies cannotrecognize cytoplasmic proteins of nonpermeabilized ookinetes(Figs. S1 and S2). To identify the protein(s) specifically recog-nized by the antibody, we analyzed ookinete proteins by Westernblotting using our anti-SM1 antibody for detection. As shown inFig. 1B, anti-SM1 antibodies recognized two major P. bergheiproteins of ∼65 kDa and ∼48 kDa. Other protein bands wereeither also present when incubated with control preimmune se-rum or were not detected reproducibly in repeat experiments.Additional fractionation of the ookinete extracts by 2D gel elec-trophoresis (Fig. S3), followed by mass spectrometric analysis ofthe excised proteins, revealed that the ∼65-kDa protein is anRNA helicase and the ∼48-kDa protein is enolase (EC 4.2.1.11).Because RNA helicase is a cytoplasmic protein, we focused ourefforts on establishing the possible functional significance ofenolase on the ookinete surface.

Enolase Occurs on the Ookinete Surface. To investigate whetherenolase is indeed present on the ookinete surface, we producedrecombinant P. falciparum enolase that was used to generate arabbit polyclonal antibody. Immunofluorescence assays providedevidence that enolase is on the surface of both P. berghei andP. falciparum ookinetes (Fig. 1C). Control experiments showedthat general antibody access to internal structures required per-meabilization (Figs. S1 and S2), thus supporting the conclusionthat the anti-SM1 and antienolase antibodies bind to a surfaceookinete component. Developmental studies indicated that sur-face enolase cannot be detected in gametocytes, gametes, andzygotes and that it is initially detected at the retort (immature

Author contributions: A.K.G., I.C., and M.J.-L. designed research; A.K.G. and I.C. per-formed research; H.G. and M.P. contributed new reagents/analytic tools; A.K.G., I.C.,and M.J.-L. analyzed data; and A.K.G., I.C., M.P., and M.J.-L. wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.1To whom correspondence should be addressed. E-mail: mlorena@jhsph.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1103657108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1103657108 PNAS | October 11, 2011 | vol. 108 | no. 41 | 17153–17158

MICRO

BIOLO

GY

ookinete, ∼12 h) and later stages of ookinete differentiation, bothfor P. berghei and P. falciparum (Fig. S4). Immunoelectron mi-croscopy confirmed the presence of enolase on the ookinete sur-face, especially on the apical pellicle complex that is involved ininvasion (Fig. 2A andB). Enolase is a cytosolic protein involved in

glycolysis, but a significant proportion (10%) of the protein wasdetected on the ookinete surface (Fig. 2C).

Antienolase Antibody Inhibits Oocyst Formation. The polyclonalanti-P. falciparum enolase antibodywas used to investigatewhethersurface enolase plays a role in midgut invasion. Oocyst formationwas used as a proxy for ookinete invasion of the mosquito midgut.We found that the anti-Plasmodium enolase antibody efficientlyinhibited P. berghei (Fig. 3A) and P. falciparum (Fig. 3B) oocystformation, consistent with the concept that surface enolase playsan important role in mosquito midgut invasion. Moreover, theseexperiments provided further in vivo evidence for the presence ofenolase on the ookinete surface.

Enolase Mediates the Binding of Plasminogen to the OokineteSurface. Enolase from various microorganisms has a unique 6-aamotif that is recognized by the lysine-binding Kringle domains ofplasminogen from various mammalian species (6, 7). Our bio-informatic analyses revealed that this lysine motif, DKSLVK, isconserved among enolases of several microbial pathogens and isidentical in allPlasmodium species sequenced todate (8–13) (TableS1). Based on this information, wepredicted that the lysinemotif ofenolase on the ookinete surface mediates the binding to themammalian plasminogen present in the ingested blood meal. Asshown in Fig. 4A, immunofluorescence assays revealed that plas-minogen binds to both P. berghei and P. falciparum ookinetes. Thisinteraction is specific, as indicated by the competitive inhibition ofplasminogen binding by the lysinemotif peptideDKSLVK (Fig. 4Band Fig. S5). Control experiments indicated that inhibition ofplasminogen binding by theDKSLVKpeptide is specific because itdid not affect binding of anti-Pbs21 and antienolase antibodies tothe ookinete surface (Fig. S6 A and B). Furthermore, a peptidederivative that had the two lysines replaced by alanines did notinhibit plasminogen binding (Fig. S6C).

Plasminogens Necessary for Ookinete Invasion of the MosquitoMidgut. Several experimental approaches were used to inves-tigate the importance of plasminogen binding to the ookinetesurface.Whenmosquitoes fed on P. berghei-infectedmice that hadbeen injected with the DKSLVK peptide, parasite development

A B

C

Fig. 1. Binding of the anti-SM1 and anti-Plasmodium enolase antibodiesto ookinetes. (A) Anti-SM1 antibody was incubated with fixed but non-permeabilized P. berghei (Upper) or P. falciparum (Lower) ookinetes, andbinding was detected with Alexa Fluor-488–labeled (green) secondaryantibody. Nuclei were detected by DAPI staining (blue). The last panel ineach row shows a differential interference contrast (DIC) image of thesame ookinete. (B) Western blot of purified P. berghei ookinetes (5 × 106

per lane). Lane 1 shows incubation with preimmune serum. Lane 2 showsincubation with anti-SM1 antibody from the same rabbit. The arrowspoint to proteins specifically and reproducibly recognized by the anti-SM1antibody. (C ) Binding of an anti-P. falciparum enolase antibody to ooki-netes. (Upper) Nonpermeabilized P. berghei ookinete incubated witha mixture of antienolase antibody (green), antibody against the surfacePbs21 protein (red), and the DAPI nuclear stain (blue). (Lower) Non-permeabilized P. falciparum ookinete incubated with a mixture of anti-enolase antibody (green), an antibody against the surface protein Pfs28(red), and the DAPI nuclear stain (blue). Between 72 and 110 ookineteswere analyzed for each of the staining protocols, and 100% of the ooki-netes displayed the pattern illustrated. A small proportion (<10%) of theookinetes did not stain with any of the reagents and may represent deadparasites.

A B

y

74%80

Ca

PM

IMC

elin

g d

en

sity

60

PMmi

% g

old

la

be

20

40

1011%

DV10%

11%

<3%

Fig. 2. Ultrastructural detection of enolase in Plasmodium ookinetes. (Aand B) Two examples of enolase localization (gold particles) using a poly-clonal anti-P. falciparum enolase antibody and cultured P. berghei ooki-netes. Surface localization is indicated by arrows. (C) Stereological analysis ofimmunogold-labeled structures. The density (gold particles per mm2) of la-beled structures was determined from 20 cryosections. The percentage ofindividual intracellular compartment density was determined from the sumof gold density normalized for the variation in expression of enolase. Al-though the majority of enolase localized to the cytoplasm, a significantproportion was located on the surface and in the nucleus, consistent withthe enolase distribution in blood forms (20). Gold particles on single- ordouble-membrane–bound compartments excluding the nucleus (organelles)correspond to negligible background. DV, digestive vacuole; IMC, innermembrane complex; mi, secretory microneme organelle; PM, plasma mem-brane. (Scale bar: 200 nm.)

A B

/ g

ut

sts / g

ut

Oo

cysts

Before After

Oo

cys

Before After

Preimmune Immune

N 54 53Range 0-63 0-22

I hibiti 78%Mean 30.3 6.8

Inhibition - 8% - 85%

Range 2-310 0-312 0-310 1-128

Preimmune Immune

Before After

N 52 52 52 54

Mean 87.8 82.0 126 19.6

Before After

P value - .0001Inhibition - 78%Inhibition - 8% - 85%

P value - - - < .0001

Fig. 3. Inhibition of P. berghei and P. falciparum oocyst formation byantienolase antibody. (A) Inhibition of P. berghei oocyst formation byantienolase antibody. For each experiment, a group of control An. gambiaemosquitoes was fed on a P. berghei-infected mouse (before). The mouse wasthen injected i.v. with 1 mg (total protein) of preimmune or antienolaseserum, and after 10 min, a group of experimental mosquitoes was fed on thesame mouse (after). Data for three independent experiments were pooled.N represents the number of mosquitoes analyzed, range represents theminimum and maximum number of oocysts per gut, mean indicates themean number of oocysts per gut, and inhibition indicates the percent de-crease of oocyst numbers. The P value was calculated by the Mann–Whitneytest. (B) Inhibition of P. falciparum oocyst formation by antienolase anti-body. P. falciparum gametocytes were mixed with preimmune or immunesera (1 mg/mL total serum protein) and fed to An. gambiaemosquitoes. Datafor three separate experiments were pooled.

17154 | www.pnas.org/cgi/doi/10.1073/pnas.1103657108 Ghosh et al.

into oocysts was strongly inhibited compared with that in mos-quitoes fed on mice receiving a similar dose of an unrelatedpeptide (Figs. 4B and 5A). Similar results were obtained by use ofthe lysine analog 6-amino hexanoic acid (6HA) (Fig. S7). Theseresults suggest that the plasminogen-enolase interaction plays animportant role in mosquito infection. This role was further testedby feeding mosquitoes on a P. falciparum culture suspended inplasminogen-free medium (plasminogen-depleted human serum,

Fig. S8). As shown in Fig. 5B, removal of plasminogen had a stronginhibitory effect on parasite development. This inhibitory effect isindeed caused by the lack of plasminogen, per se, because theability of the depleted medium to support parasite developmentcould be restored by the addition of recombinant plasminogen(Fig. 5B). Plasminogen is the zymogen form of the broad-spec-trum serine protease plasmin. To investigate the importance ofthe catalytic activity of plasmin for oocyst differentiation, weadded a recombinant plasminogen carrying a point mutation(S741A) in the catalytic triad. In contrast to the WT protein, theinactive plasminogen was unable to rescue the impaired oocystformation phenotype of the plasminogen-depleted serum (Fig.5B), providing direct evidence for the requirement of unimpairedcatalytic activity of the surface-bound plasminogen to completeookinete invasion. We therefore conclude that mammalian plas-minogen plays a crucial role in the ability of the parasite to com-plete its developmental cycle in the mosquito.

DiscussionSM1 Mimics Enolase. SM1 is a dodecapeptide that binds tightly tothe luminal surface of the mosquito midgut epithelium, and bydoing so, it efficiently inhibits ookinete invasion (4, 14). Theseobservations suggest that the peptide binds to a midgut receptor,and by doing so, it competitively inhibits binding to an ookineteligand, thus interfering with invasion (Fig. 6, Center). In this sce-nario, peptide conformation resembles that of an ookinete ligandinvolved in receptor binding. Bioinformatic searches for Plasmo-dium proteins containing sequences similar to the 8-aa loop wereunsuccessful. To address the possible similarity of the SM1 struc-ture to that of an ookinete protein involved in invasion, we gen-erated an anti-SM1 antibody that recognizes enolase, leading to thediscovery that this protein is present on the ookinete surface. Theseobservations constitute circumstantial evidence supporting thehypothesis that one function of surface enolase is to interact witha putative mosquito midgut receptor for invasion (Fig. 6). The factthat antienolase antibody inhibits ookinete invasion (Fig. 3 A andB) is consistent with this hypothesis. Our previous findings revealedthat SM1 also binds to the outer surface of the salivary glands (4).We found that SM1mimics the sporozoite ligand thrombospondin-related anonymous protein (TRAP) because this protein is alsorecognized by the anti-SM1 antibody. As expected, TRAP andSM1 competitively bind to the salivary gland receptor saglin (15).TRAP is not expressed by ookinetes, and saglin is not expressed inmidgut epithelial cells, however. It seems that enolase is theookinete SM1 mimotope (it is recognized by the anti-SM1 anti-body) and that it binds to a midgut receptor different from saglin.

Enolase Is Exported to the Ookinete Surface. Several pathogenicmicroorganisms from bacteria to Leishmania and Plasmodiummerozoites also express enolase on their surface (10, 16–20). Wefound that enolase is also exported to the surface of P. bergheiand P. falciparum ookinetes. None of the enolases identified todate has a signal sequence or a transmembrane domain, and themechanisms of enolase export and adhesion to the cell surfaceare unknown (20, 21). In yeast, the N-terminal 169-aa residues ofenolase are required for export to the surface, but the exact motifor mechanism has not been identified (22).Our experiments (Fig. 3 A and B) suggest that enolase may

serve as an antigen for transmission-blocking vaccines. There isevidence that enolase also occurs on the surface of merozoitesand that antienolase antibodies inhibit the growth of blood-stageparasites in vitro (20). Interestingly, high-titer antienolase anti-bodies are found in malaria patients in Asia, but the significanceof this finding has not yet been assessed (23).

Surface Enolase Captures Plasminogen from the Blood Meal. Plas-modium enolase contains a C-terminal 6-aa lysine motif that isidentical in all Plasmodium species sequenced to date and is also

A Plg Pbs21 DAPI Merge DIC

P bergheiP. berghei

DICMergeDAPIPfs28Plg

P. falciparum

DICMergeDAPIPfs28Plg

Mµ 001editpepoN 1 µM 10 µMB

Fig. 4. Immunolocalization of plasminogen on the ookinete surface andcompetition with the lysine-motif peptide (100%). (A) (Upper) Non-permeabilized cultured P. berghei ookinetes were incubated with a mixtureof antiplasminogen (Plg; green) and anti-Pbs21 (red) antibodies. Nuclearstaining with DAPI is shown in blue. DIC, differential interference contrastmicroscopy. (Lower) Similar experiment with cultured P. falciparum ooki-netes, except that anti-Pfs28 (red) antibodies were used instead of anti-Pbs21. One hundred percent of the ookinetes that stained with the anti-Ps21or anti-Pfs28 antibodies also stained with plasminogen. (B) (Upper) Plas-minogen binding to the P. berghei ookinete surface is competed by thelysine-motif peptide DKSLVK. Peptide concentrations were as indicated ineach panel. (Lower) DIC images of the ookinetes shown.

A B

Fig. 5. Ookinete dependence on its surface enolase and on interaction withmammalian plasminogen for invasion of the mosquito midgut. (A) Inhibitionof P. berghei oocyst formation by the DKSLVK peptide. For each experiment,a group of control An. gambiae mosquitoes was fed on a P. berghei-infectedmouse (before). The mouse was then injected i.v. with 200 μg of the DKSLVKpeptide or of an unrelated control peptide (QPQHFR), and after 10 min,a group of experimental mosquitoes was fed on the same mouse (after).Data for three independent experiments were pooled. N indicates thenumber of mosquitoes analyzed, range indicates the minimum and maxi-mum number of oocysts per gut, mean indicates the mean number ofoocysts per gut, and inhibition indicates the percent decrease of oocystnumbers. The P value was calculated by the Mann–Whitney test. (B) P. fal-ciparum dependence on mammalian plasminogen for oocyst formation. An.gambiae mosquitoes were fed on P. falciparum gametocytes suspended innormal human serum in plasminogen-depleted human serum (No PLG) or inplasminogen-depleted serum with added recombinant WT plasminogen[200 μg/mL; +PLG (wt)] or with added recombinant mutant plasminogencarrying a point mutation in the active site (S741A) [200 μg/mL; +PLG (mut)],as indicated. Data for three independent experiments were pooled.

Ghosh et al. PNAS | October 11, 2011 | vol. 108 | no. 41 | 17155

MICRO

BIOLO

GY

conserved in other pathogenic microorganisms (7, 13) (TableS1). This motif endows enolase with plasminogen-binding ca-pacity, thus enabling the capture of plasminogen from the sur-rounding environment. Binding of plasminogen to the lysinemotif of enolase is governed by five Kringle domains in the A-chain of plasminogen (24). Experiments showing that the lysinemotif peptide competitively inhibits plasminogen binding (Fig.4B and Fig. S6C) strongly support the concept that this in-teraction provides the basis for plasminogen binding to theookinete surface. 6HA is a lysine analog that has been usedin vitro to inhibit clot lysis (25). The Kringle domains of plas-minogen bind to the terminal lysine residue of fibrin, triggeringits degradation. 6HA binds with high affinity to Kringle domains,thus inhibiting plasminogen activation and fibrin degradation(26, 27). Similarly, 6HA had a strong inhibitory effect on oocystformation, providing additional evidence that plasmin plays animportant role in ookinete invasion of the mosquito midgut.

Plasminogen Is Necessary for Midgut Invasion. Plasminogen is themost abundant serine protease in mammalian blood (1–2 μM).Several pathogens hijack host plasminogen to assist in tissue andorgan invasion (6, 7, 28–32). The present experiments indicatethat plasminogen, and its conversion into active plasmin, is es-sential for invasion of the mosquito midgut. Thus, in evolution,Plasmodium seems to have coopted mammalian plasminogen toinvade its invertebrate host. The substrate of the activated plas-min in the mosquito midgut is presently unknown. The luminalsurface of the mosquito midgut is lined by a network of proteinsand carbohydrates (33) that the ookinete must traverse to reachthe surface of the midgut epithelium. Plasmin may degradecomponents of this network to facilitate invasion of the midgutepithelium (Fig. 6).

Concluding RemarksOur experiments demonstrate that enolase occurs on the surfaceof Plasmodium ookinetes, where we hypothesize it acts in twoindependent capacities: (i) to mediate recognition of the mos-quito midgut epithelium during invasion and (ii) to mediate thebinding of mammalian plasminogen that is subsequently con-verted to active plasmin. These two functions may be performedby different domains of the enolase protein (Fig. 6). Identifica-tion of the putative midgut receptor will allow a direct test ofthis hypothesis.

Materials and MethodsMosquitoes. The Anopheles gambiae Keele and Anopheles stephensi strainswere maintained as described (15).

Parasites. P. berghei strain ANKA 2.34 was maintained by passage in SwissWebster mice (34, 35). P. falciparum NF54 strain was maintained accordingto standard methods (36). Gametocyte cultures were started at 0.5% para-sitemia, and the medium was changed daily until mature gametocytes wereobtained (usually on day 18) (37).

Peptide Synthesis. All peptides were obtained from Genmed Synthesis, Inc.

Ookinete Culture. In vitro differentiation of gametocytes into P. berghei or P.falciparum ookinetes was as described elsewhere (38). P. falciparum ooki-netes were produced by incubating a gametocyte-enriched culture for 24 h,after which they were purified (34). Ookinete numbers were determined bymicroscopic examination with a hemocytometer. For peptide competitionassays (Fig. 4B), gametocytes were cultured in ookinete transformationmedium with FBS. After 6 h, the culture was centrifuged at 659 × g for 5 min,washed with ookinete medium without serum, and resuspended in ookinetemedium containing plasminogen-depleted 10% (vol/vol) FBS, and parasiteswere allowed to differentiate further. Plasminogen was depleted with ly-sine-Sepharose beads (GE Healthcare, Inc.) per the manufacturer’s protocol.Depletion was verified by Western blot analysis (Fig. S5).

Fig. 6. Hypothetical model for the role of enolase in ookinete midgut invasion. (Left, No SM1 peptide) About midway in their development in the mosquitomidgut, ookinetes start expressing enolase on their surface (Fig. S4). (Right) Plasminogen from the surrounding blood meal (Figs. 4 and 5A and Figs. S6 and S7)is captured by the 6-aa C-terminal lysine domain of enolase (lysine motif). Plasminogen is subsequently converted into active plasmin (Fig. 5B), a broad-spectrum serine protease, thus promoting the hydrolysis of glycocalyx components. Local breach of glycocalyx integrity facilitates ookinete access to themidgut epithelium and interaction of a separate enolase domain (SM1-like domain) with a putative midgut receptor, thus promoting invasion. SM1 peptide(+SM1 peptide) mimics the SM1-like domain of enolase [anti-SM1 antibody recognizes enolase (Fig. 1B and Fig. S3) and competes with it for binding to theputative midgut receptor]. When the midgut receptors are occupied by the SM1 peptide, interaction with enolase on the surface of the ookinete is abrogatedand invasion is aborted.

17156 | www.pnas.org/cgi/doi/10.1073/pnas.1103657108 Ghosh et al.

Western Blotting. About 5 × 106 purified ookinetes were resuspended in20 μL of Laemmli buffer (39) and lysed by boiling. After centrifugation,the supernatant was fractionated by electrophoresis on a 10% (wt/vol)polyacrylamide-SDS gel, followed by transfer to a PVDF membrane (ISEQ00010; Millipore Corporation). Remaining procedures were as describedelsewhere (15).

2D Gel Electrophoresis and Western Blot Analysis. Purified P. berghei ooki-netes were separated by 2D gel electrophoresis as described elsewhere (40,41). Three identical gels were run in parallel, two for Western blotting (onewith immune serum and another with preimmune serum) and the third forrecovering protein spots for sequencing by tandem MS (15, 42).

Immunofluorescence Assays. Immunofluorescence assays were performed asdescribed elsewhere (4). Frozen samples were prepared as follows. Purifiedookinetes were smeared on slidesmaintained over ice, followed by immersioninto 4% (wt/vol) paraformaldehyde solution for 1 h at 4 °C and threewashes incold PBS for 15 min each. Slides were stored at −80 °C until use. Frozen slidescontaining fixed ookinetes were thawed at room temperature for 30 min anddipped in blocking buffer [4% (wt/vol) BSA in PBS] for 1 h at room tempera-ture. Some samples were permeabilized by incubation in 0.05% Triton X-100for 30 min at room temperature before dipping in blocking buffer. This wasfollowed by incubation with rabbit anti-SM1 antibody (1:1,000), rabbit anti-enolase antibody (1:1,000), or mouse anti-Pfs25 monoclonal antibody (1:2,000) in blocking buffer for 1 h at room temperature. Slides were washedthree times in PBS for 30 min each time and incubated in blocking buffer asdescribed above. Thiswas followedby incubation in the darkwithAlexa Fluor-488–labeled anti-rabbit IgG (1:1,000, A1108; Invitrogen, Inc.) or with AlexaFluor-586–labeled anti-mouse IgG (1:1,000, A21043; Invitrogen, Inc.) for 45min at room temperature. Slides were washed as before and mounted inProLong Gold antifade reagent with DAPI (P36935; Invitrogen, Inc.), and pic-tures were taken with a Leica DMLB microscope. Procedures for the experi-ments illustrated in Fig. S2 followed a similar protocol, except that freshookinetes (never frozen) were used and the entire experiment was conductedon the same day. The anti–α-tubulin antibody (T9026; Sigma) was used ata 1:1,000 dilution.

Expression of Recombinant P. falciparum Enolase. Total RNA was extractedfrom an asynchronous P. falciparum NF54 culture using the TRIzol Reagent(15596-018; Invitrogen, Inc.). cDNA was synthesized using an oligo-dT18primer, and the full-length enolase coding sequence was amplified with thefollowing gene-specific primers: forward primer bearing the EcoR1 site 5′-CCGGAATTCATGGCTCATGTAATAACTCGTATTAATGCC-3′ and reverse primer5′-GCCGCCGTCGACATTTAATTGTAATCTAAATTTTTCAGC-3′ with a Sal1 site.The PCR amplification products were purified with the QIA quick PCR puri-fication kit (28104; Qiagen), followed by digestion with EcoR1 (RO101L; NewEngland Biolabs) and Sal1 (RO138S; New England Biolabs). The digestionproduct was gel-purified, cleaned, and ligated to the PGEX-4T vector(Amersham Pharmacia Biotech, Inc.) at its EcoR1 and Sal1 sites. Transformedcells were grown at 25 °C, and expression was induced with isopropyl-β-galactosidase (15529-019; Invitrogen, Inc.) to a final concentration of 10mM (OD 0.6 at 600 nm). Cells were harvested after 3 h of induction, andprotein was purified as described elsewhere (43).

Antibodies and Purified Proteins. Rabbit polyclonal anti-SM1 antibody and apolyclonal rabbit antienolase antibody were produced as described else-where (15). Recombinant human plasminogen and its active site-mutatedvariant (Ser741Ala) were expressed in Drosophila S2 cells and purified bylysine-Sepharose chromatography (44). The eluted product was buffer-ex-changed into PBS.

Feeding P. falciparum Gametocytes to Mosquitoes and Measurement of OocystNumbers. Gametocyte feeding was done as described elsewhere (38). A ga-metocyteor ookinete culturewas added to each feeding chamber,making surethat the culture covered the entire feeding surface. Mosquitoes were allowedto feed on these feeders, and 8 (P. falciparum) or 15 (P. berghei) d later, themidguts were dissected for microscopic determination of oocyst numbers.

Competition Assays. Slides containing fixed ookinetes (stored at −80 °C) werethawed at room temperature for 30 min and blocked in 4% (wt/vol) BSA

solution in PBS (blocking buffer) for 1 h at room temperature. Each slide wasincubated for 1 h at room temperature in blocking buffer containing plas-minogen (2 μg/mL; Innovative Research, Inc.) or in blocking buffer containingthe same concentration of plasminogen and increasing concentrations (1 μM,10 μM, or 100 μM) of the lysine motif (DKSLVK) or the mutated (DASLVA)peptides. The plasminogen/peptide mixtures were preincubated for 1 h atroom temperature before addition to the slides. Slides were washed with PBSthree times for 30 min each time, followed by 30 min in blocking buffer.Slides were washed and blocked again as described above and were in-cubated for 1 h at room temperature with rabbit anti-mouse plasminogenantibody (1:1,000 dilution; Innovative Research, Inc.). Slides were washed andmounted in ProLong Gold antifade, and the images were recorded with theaid of a fluorescence microscope as described above. For experiments illus-trated in Fig. S6 A and B, ookinete-containing slides were incubated with 100μM lysine motif peptide (DKSLVK) for 1 h at room temperature and slideswere washed and blocked as before. Slides were then incubated with anti-Pbs21 antibody (1:1,000) or antienolase antibody (1:1,000) for 1 h andwashed, and staining was completed using Alexa Fluor-488–labeled anti-mouse IgG (1:1,000) for anti-Pbs21 antibody and Alexa Fluor-488–labeledanti-rabbit IgG (1:1,000) for antienolase antibody. Pictures were taken asdescribed before.

Peptide Inhibition Assays. Mouse infection with P. berghei was followed bydaily measurement of exflagellation events. When exflagellation reachedone to two events per field at a magnification of 40×, mice were anes-thetized and a group of ∼50 starved An. gambiae females (control group)was allowed to feed for 10 min at 19 °C. A total of 200 μg of the lysine motifpeptide (DKSLVK) dissolved in PBS was then injected into the tail vein of thesame mouse. After an interval of 10 min to allow dispersal of peptidethrough the circulation, another group of ∼50 starved An. gambiae females(experimental group) was allowed to feed. Mosquitoes that did not feed oronly partially fed were removed from the group, and the remaining mos-quitoes were kept in an insectary at 19 °C with a 10% (wt/vol) sugar solution.On day 15, mosquitoes were dissected and the oocyst number per midgutwas determined.

Transmission Blocking Assays. P. berghei-infectedmice (1–2 exflagellations perfield) were anesthetized and exposed for 10 min at 19 °C to a group (control)of ∼50 starved An. gambiae females. Antienolase antibody, preimmune se-rum (1 mg), or peptides dissolved in 250 μL of PBS were injected through thetail vein. After a 10-min pause to allow for the dispersal of the injected ma-terial, the same mouse was exposed to another group (experimental) of ∼50starved An. gambiae females. Remaining procedures and oocyst counts weredone as described in the previous section. P. falciparum gametocytes wereproduced as described above. Gametocytes were mixed with the antienolaseantibody (final concentration of 1 mg/mL), or the preimmune serum of thesame rabbit and membrane was fed to a group of ∼50 starved An. gambiaefemales. Alternatively, 6HA (400 μg/mL PBS) was mixed with P. falciparumgametocytes and fed to mosquitoes in the same way, using PBS as a control.Mosquitoes were kept in a humidified chamber at 24 ± 2 °C, and midgutoocysts were counted on day 8 as described above.

Immunoelectron Microscopy. P. berghei ookinetes were fixed in 4% (wt/vol)paraformaldehyde (ElectronMicroscopy Sciences) in 0.25MHepes (pH 7.4) for1 h at room temperature, followed by 8% (wt/vol) paraformaldehyde in samebuffer overnight at 4 °C. The fixed ookinetes were frozen and sectioned aspreviously described. Sections were immunolabeled with rabbit antienolaseantibodies (1:250 in PBS/1% fish skin gelatin) and then with anti-rabbit IgG,followed by 10 nm of protein A-gold particles. Sections were examined witha Philips CM120 Electron Microscope at 80 kV (45).

ACKNOWLEDGMENTS. We thank the Johns Hopkins Malaria Research In-stitute mosquito and P. falciparum core facilities for help with mosquito rear-ing and parasite cultures. Receipt of the Pfs28 antibody from the MalariaResearch and Reference Reagent Resource Center (MR4) is gratefully acknowl-edged. This work receivedfinancial support fromNational Institutes of HealthGrant AI031478. Additional support was provided by the Johns HopkinsMalaria Research Institute and the Bloomberg Family Foundation. Supply ofhuman blood was supported by National Institutes of Health Grant RR00052.

1. Breman JG, Egan A, Keusch GT (2001) The intolerable burden of malaria: A new look

at the numbers. Am J Trop Med Hyg 64(Suppl 1–2):iv–vii.

2. Snow RW, Guerra CA, Noor AM, Myint HY, Hay SI (2005) The global distribution of

clinical episodes of Plasmodium falciparum malaria. Nature 434:214–217.

Ghosh et al. PNAS | October 11, 2011 | vol. 108 | no. 41 | 17157

MICRO

BIOLO

GY

3. Ghosh AK, Edwards MJ, Jacobs-Lorena M (2000) The journey of the malaria parasite inthe mosquito: Hopes for the new century. Parasitol Today 16:196–201.

4. Ghosh AK, Ribolla PE, Jacobs-Lorena M (2001) Targeting Plasmodium ligands onmosquito salivary glands and midgut with a phage display peptide library. Proc NatlAcad Sci USA 98:13278–13281.

5. Dinglasan RR, et al. (2007) Disruption of Plasmodium falciparum development byantibodies against a conserved mosquito midgut antigen. Proc Natl Acad Sci USA 104:13461–13466.

6. Lottenberg R, Minning-Wenz D, Boyle MD (1994) Capturing host plasmin(ogen): Acommon mechanism for invasive pathogens? Trends Microbiol 2:20–24.

7. Ehinger S, Schubert WD, Bergmann S, Hammerschmidt S, Heinz DW (2004) Plasmin(ogen)-binding alpha-enolase from Streptococcus pneumoniae: Crystal structure andevaluation of plasmin(ogen)-binding sites. J Mol Biol 343:997–1005.

8. Bergmann S, Rohde M, Chhatwal GS, Hammerschmidt S (2001) alpha-Enolase ofStreptococcus pneumoniae is a plasmin(ogen)-binding protein displayed on the bac-terial cell surface. Mol Microbiol 40:1273–1287.

9. Rios-Steiner JL, Schenone M, Mochalkin I, Tulinsky A, Castellino FJ (2001) Structureand binding determinants of the recombinant kringle-2 domain of human plasmin-ogen to an internal peptide from a group A Streptococcal surface protein. J Mol Biol308:705–719.

10. Jong AY, et al. (2003) Binding of Candida albicans enolase to plasmin(ogen) results inenhanced invasion of human brain microvascular endothelial cells. J Med Microbiol52:615–622.

11. Jolodar A, et al. (2003) Molecular cloning of an alpha-enolase from the human filarialparasite Onchocerca volvulus that binds human plasminogen. Biochim Biophys Acta1627:111–120.

12. Bernal D, et al. (2004) Identification of enolase as a plasminogen-binding protein inexcretory-secretory products of Fasciola hepatica. FEBS Lett 563:203–206.

13. Read M, Hicks KE, Sims PF, Hyde JE (1994) Molecular characterisation of the enolasegene from the human malaria parasite Plasmodium falciparum. Evidence for ancestrywithin a photosynthetic lineage. Eur J Biochem 220:513–520.

14. Ito J, Ghosh A, Moreira LA, Wimmer EA, Jacobs-Lorena M (2002) Transgenic anoph-eline mosquitoes impaired in transmission of a malaria parasite. Nature 417:452–455.

15. Ghosh AK, et al. (2009) Malaria parasite invasion of the mosquito salivary gland re-quires interaction between the Plasmodium TRAP and the Anopheles saglin proteins.PLoS Pathog 5:e1000265.

16. Candela M, et al. (2009) Bifidobacterial enolase, a cell surface receptor for humanplasminogen involved in the interaction with the host. Microbiology 155:3294–3303.

17. Carneiro CR, Postol E, Nomizo R, Reis LF, Brentani RR (2004) Identification of enolaseas a laminin-binding protein on the surface of Staphylococcus aureus. Microbes Infect6:604–608.

18. Agarwal S, Kulshreshtha P, Bambah Mukku D, Bhatnagar R (2008) alpha-Enolasebinds to human plasminogen on the surface of Bacillus anthracis. Biochim BiophysActa 1784:986–994.

19. Quiñones W, et al. (2007) Leishmania mexicana: Molecular cloning and character-ization of enolase. Exp Parasitol 116:241–251.

20. Pal-Bhowmick I, Mehta M, Coppens I, Sharma S, Jarori GK (2007) Protective propertiesand surface localization of Plasmodium falciparum enolase. Infect Immun 75:5500–5508.

21. Pancholi V (2001) Multifunctional alpha-enolase: Its role in disease. Cell Mol Sci 58:902–920.

22. López-Villar E, et al. (2006) Genetic and proteomic evidences support the localizationof yeast enolase in the cell surface. Proteomics 6(Suppl 1):S107–S118.

23. Sato K, et al. (2000) Application of yeast enolase as antigen for immunodiagnosis ofmalaria. Southeast Asian J Trop Med Public Health 31(Suppl1):79–84.

24. Ponting CP, Marshall JM, Cederholm-Williams SA (1992) Plasminogen: A structuralreview. Blood Coagul Fibrinolysis 3:605–614.

25. Krishnamurti C, Vukelja SJ, Alving BM (1994) Inhibitory effects of lysine analogues ont-PA induced whole blood clot lysis. Thromb Res 73:419–430.

26. Rejante MR, Llinás M (1994) Solution structure of the epsilon-aminohexanoic acidcomplex of human plasminogen kringle 1. Eur J Biochem 221:939–949.

27. Kupcsik L, Alini M, Stoddart MJ (2009) Epsilon-aminocaproic acid is a useful fibrindegradation inhibitor for cartilage tissue engineering. Tissue Eng 15:2309–2313.

28. Bugge TH, et al. (1996) Loss of fibrinogen rescues mice from the pleiotropic effects ofplasminogen deficiency. Cell 87:709–719.

29. Fuchs H, Wallich R, Simon MM, Kramer MD (1994) The outer surface protein A of thespirochete Borrelia burgdorferi is a plasmin(ogen) receptor. Proc Natl Acad Sci USA91:12594–12598.

30. Coleman JL, et al. (1997) Plasminogen is required for efficient dissemination of B.burgdorferi in ticks and for enhancement of spirochetemia in mice. Cell 89:1111–1119.

31. Hajjar KA, Krishnan S (1999) Annexin II: A mediator of the plasmin/plasminogen ac-tivator system. Trends Cardiovasc Med 9:128–138.

32. Ghosh AK, Jacobs-Lorena M (2011) Surface-expressed enolases of Plasmodium andother pathogens. Mem Inst Oswaldo Cruz 106(Suppl 1):85–90.

33. Zieler H, Garon CF, Fischer ER, Shahabuddin M (2000) A tubular network associatedwith the brush-border surface of the Aedes aegypti midgut: Implications for patho-gen transmission by mosquitoes. J Exp Biol 203:1599–1611.

34. Ranawaka GR, Fleck SL, Blanco AR, Sinden RE (1994) Characterization of the modes ofaction of anti-Pbs21 malaria transmission-blocking immunity: Ookinete to oocystdifferentiation in vivo. Parasitology 109:403–411.

35. Rodríguez MC, et al. (2002) Plasmodium berghei: Routine production of pure ga-metocytes, extracellular gametes, zygotes, and ookinetes. Exp Parasitol 101:73–76.

36. Trager W, Jensen JB (1976) Human malaria parasites in continuous culture. Science193:673–675.

37. Ifediba T, Vanderberg JP (1981) Complete in vitro maturation of Plasmodium falci-parum gametocytes. Nature 294:364–366.

38. Ghosh AK, Dinglasan RR, Ikadai H, Jacobs-Lorena M (2010) An improved method forthe in vitro differentiation of Plasmodium falciparum gametocytes into ookinetes.Malar J 9:194–200.

39. Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature 227:680–685.

40. O’Farrell PH (1975) High resolution two-dimensional electrophoresis of proteins. J BiolChem 250:4007–4021.

41. O’Farrell PZ, Goodman HM, O’Farrell PH (1977) High resolution two-dimensionalelectrophoresis of basic as well as acidic proteins. Cell 12:1133–1141.

42. Perkins DN, Pappin DJ, Creasy DM, Cottrell JS (1999) Probability-based protein iden-tification by searching sequence databases using mass spectrometry data. Electro-phoresis 20:3551–3567.

43. Jethwaney D, et al. (2005) Fetuin-A, a hepatocyte-specific protein that binds Plas-modium berghei thrombospondin-related adhesive protein: A potential role in in-fectivity. Infect Immun 73:5883–5891.

44. Nilsen SL, Castellino FJ (1999) Expression of human plasminogen in DrosophilaSchneider S2 cells. Protein Expr Purif 16:136–143.

45. Fölsch H, Pypaert M, Schu P, Mellman I (2001) Distribution and function of AP-1 cla-thrin adaptor complexes in polarized epithelial cells. J Cell Biol 152:595–606.

17158 | www.pnas.org/cgi/doi/10.1073/pnas.1103657108 Ghosh et al.