Molecular and Biochemical Parasitology 106 (2000) 199–212
The selectable marker human dihydrofolate reductaseenables sequential genetic manipulation of the Plasmodium
berghei genome
Tania F. de Koning-Ward a, David A. Fidock b, Vandana Thathy c,Robert Menard c, Rosalina M.L. van Spaendonk a, Andrew P. Waters a,*,
Chris J. Janse a
a Department of Parasitology, Leiden Uni6ersity Medical Centre, Postbox 9600, 2300 RC Leiden, The Netherlandsb Malaria Genetics Section, Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, NIH, Bethesda,
MD 20892-0425, USAc New York Uni6ersity Medical Center, Department of Pathology, Michael Heidelberger Di6ision of Immunology, New York 10016,
New York, USA
Received 9 July 1999; received in revised form 24 September 1999; accepted 15 October 1999
Abstract
Genetic transformation of malaria parasites has been limited by the number of selectable markers available. For therodent malaria parasite, Plasmodium berghei, only a single selection marker has been at hand, utilising thedihydrofolate reductase-thymidylate synthase gene from either P. berghei or Toxoplasma gondii to confer resistanceto the anti-malarial drug pyrimethamine. Here we report the use of the human dihydrofolate reductase (hDHFR) geneas a new selectable marker, which confers resistance to the antifolate inhibitor WR99210 upon both pyrimethaminesensitive and resistant isolates of P. berghei. Transfection with circular constructs containing the hDHFR generesulted in the generation of highly resistant parasites containing multiple copies of episomally-maintained plasmids.These parasites showed around a 1000-fold increase in resistance to WR99210 compared to the parental parasites. Wewere also able to generate and select transgenic parasites harbouring only a single copy of hDHFR targeted into theirgenome, despite the fact that these parasites showed only a fivefold increase in resistance to WR99210 compared tothe parental parasites. Importantly, and for the first time with malaria parasites, the hDHFR gene could be used inconjunction with the existing pyrimethamine selectable markers. This was demonstrated by reintroducing thecircumsporozoite (CS) gene into transgenic CS-knockout mutant parasites that contained the P. berghei DHFR-TSselectable marker. The development of hDHFR as a second selectable marker will greatly expand the use oftransformation technology in Plasmodium, enabling more extensive genetic manipulation and thus facilitating more
Genetic transformation of Plasmodium has pro-vided increased opportunities to study the biologyof malaria parasites with the aim of providing amore rational approach to vaccine and drug de-sign. To date, only one selectable marker has beenavailable for stable transfection of malaria para-sites, that being a modified form of the dihydrofo-late reductase-thymidylate synthase (DHFR-TS)gene obtained from either Plasmodium berghei [1]or Toxoplasma gondii [2]. These genes confer resis-tance to the anti-malarial drug pyrimethamine.Since Plasmodium is a haploid organism, a singleselectable marker has allowed the stable expres-sion of transgenes [3–5] and the deletion of non-essential genes by site-directed integration [6–8].Optimal use of the transfection technology in theabsence of selectable marker recycling, however,requires additional selectable markers. Theavailability of such markers would, for example,enable mutant parasites whose genes have beendisrupted to be genetically complemented with ananalogous or modified form of the gene. Multipleselectable markers would also permit the intro-duction or disruption of more than a single genewithin the same parasite clone. The identificationof new selectable markers has been difficult sincemalaria parasites are naturally resistant to most ofthe drugs that are used in selection systems fortransfection of other eukaryotic cells. An addi-tional complication for the transfection of rodentand non-human primate malaria parasites is thelack of reproducible in vitro culture systems and,therefore, drug selection of transfected parasiteshas to be performed in vivo. As a consequence,toxicity of the drugs used for selection of trans-fected parasites to the hosts may exclude addi-tional selectable marker genes.
The aim of this study was to evaluate thefeasibility of the human DHFR (hDHFR) gene asa new selectable marker for transfection of P.berghei, using the antifolate drug WR99210 for
the selection of transfected parasites. Previousstudies have demonstrated that P. falciparum iso-lates are highly sensitive to this drug [9,10] and ithas been reported to be active in vivo againstrodent and primate malarias [11,12]. Likepyrimethamine, WR99210 causes the selective andpotent inhibition of malarial DHFR [13–15].However, the binding interactions in the folatesubstrate pocket are distinct for these two drugsand WR99210 has been shown to be equallyeffective against pyrimethamine-resistant P. falci-parum [9,10,16] and P. berghei [12]. Importantly,hDHFR is innately resistant to bothpyrimethamine and WR99210 and recently it hasbeen demonstrated that transfection of P. falci-parum with the hDHFR gene is able to conferresistance to both drugs [16,17]. However, selec-tion of WR99210 resistance against apyrimethamine-resistant background was notdemonstrated.
In this study we show that transfection of P.berghei with the hDHFR gene confers resistanceto WR99210 and thus can be used as a newselectable marker. We also found that hDHFRcan be used in conjunction with the existingpyrimethamine resistance markers by reintroduc-ing the CS gene into a pyrimethamine-resistantcircumsporozoite protein (CS)-knockout mutantof P. berghei [7] and selecting for geneticallycomplemented transgenic parasites using hDHFRas the selectable marker in conjunction withWR99210.
2. Materials and methods
2.1. Parasites
The following clones of P. berghei were used inthis study: The non-transformed clone 15cy1 ofthe ANKA strain [18], the non-transformedNK65 strain [7], the pyrimethamine-resistantclone 120.3 of the ANKA strain that contains a
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212 201
single genomic copy of the pyrimethamine-resis-tant T. gondii DHFR-TS gene (tDHFR-TS) [C.Janse, unpublished], and the pyrimethamine-resis-tant CS knock-out clone CS(− )2 of the NK65strain that contains a single genomic copy of thepyrimethamine-resistant P. berghei DHFR-TSgene (bDHFR-TS) [7] integrated into the CSlocus.
2.2. Drug sensiti6ity assays
The drug WR99210 (BRL 6231 [11]) was kindlyprovided by David Jacobus (Jacobus Pharmaceu-ticals, Princeton, NJ) and was made as a 20 mgml−1 or 40 mg ml−1 solution in 70% DMSO.Pyrimethamine was dissolved in DMSO to give afinal concentration of 2.5 mg ml−1.
The in vitro sensitivity of the asexual blood-stage development of the different P. bergheiclones to WR99210 and pyrimethamine was deter-mined in drug assays as described previously [19].Briefly, synchronised ringforms were incubated inRPMI 1640 culture medium in 24-microwellplates under standard culture conditions in thepresence of different concentrations of drugs for aperiod of twenty hours. Serial dilutions of thedrug were made in culture medium and the con-centration of drug tested ranged from 3 pM to300 nM. At the end of the 20 h incubation periodGiemsa stained smears were made and blood sam-ples collected for FACS analysis [20]. IC50 valueswere calculated by determination of DNA synthe-sis by FACS analysis and/or determination ofschizont development in Giemsa stained smears[19]. All in vitro drug assays were performed induplicate on two or three separate occasions.
To examine the in vivo sensitivity of P. bergheito WR99210 and to select transgenic parasitesharbouring the hDHFR gene in vivo, infectedSwiss mice (25 g) were treated with WR99210 bysubcutaneous (s.c.) injection of a single dose onthree consecutive days. WR99210 was adminis-tered at concentrations ranging between 6 and 40mg kg−1 bodyweight as indicated in section Sec-tion 3. Blood smears were made daily to monitorthe course of parasitemia and the effectiveness ofthe treatment regime.
2.3. DNA constructs containing the humanDHFR gene
The hDHFR gene was PCR-amplified from theclone pHDWT [16] using the oligonucleotides hd-hfr5%B and hdhfr3%B (Table 1). The resulting PCRproduct was digested with BamHI and clonedinto the BamHI site of vector pDb.-.. Db [21], thuscreating pDb.Dh.. Db (Fig. 1A). In this latter con-struct, expression of hDHFR is under the regula-tory control of the 5% and 3% untranslated regions(UTRs) of the bDHFR-TS gene [4].
A fragment containing a truncated D unit-smallsubunit ribosomal RNA (D-SSU rRNA) gene,lacking the last 21 bp from the 3% end, was am-
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212202
Fig. 1. Transgenic P. berghei, containing the hDHFR gene onepisomally-maintained plasmids. A. Schematic representationof construct pDb.Dh.. Db for episomal expression of hDHFR.The hDHFR gene was placed under the regulatory control ofthe 5% and 3% UTR sequences from the bDHFR-TS gene (DT).The relative positions of oligonucleotides L379 and L431 areindicated. Dashed lines represent pUC19 plasmid sequences.B, BamHI; Bc, BclI; E, EcoRI; Ev, EcoRV; H, HindIII; S,SphI. B. Detection of plasmid pDb.Dh.. Db in transgenic para-sites by PCR. Oligonucleotides L379 and L431 were used in aPCR to amplify the hDHFR gene and flanking 3% UTR fromtemplate DNA of clone 15cy1 (parental parasite control, lane1); exp. 101.1 (transgenic parasites, lane 2); 101.2 (transgenicparasites, lane 3) and pDb.Dh.. Db (plasmid control, lane 4). C.Detection of episomally-maintained plasmids in transgenicparasites by plasmid rescue. Total parasite DNA was isolatedfrom exp.101.1, 101.2, 102 and 103, and used to transfectEscherichia coli. Plasmid DNA, isolated from 6 randomlypicked colonies from each transformation, was digested withHindIII and BamHI (lanes 1–5) and restriction enzyme pat-terns were compared to the pattern obtained from constructpDb.Dh.. Db (lane 6). D. Southern blot analysis of genomicDNA from transgenic parasites. Total parasite DNA wasdigested with BclI and probed with the 3% UTR of the bD-HFR-TS gene. The 1.4-kb fragment originates from the ge-nomic copy of the bDHFR-TS gene, while the 6-kb fragmentrepresents pDb.Dh.. Db, which is linearised by BclI digestion.Lane 1, clone 15cy1 (parental parasite control); lanes 2–4,parasites isolated from exp. 101.2 that had been treated witheither 6, 12 or 20 mg/kg bodyweight WR99210, respectively.The band indicated by an arrow is most likely undigestedplasmid DNA but may be the result of spurious integration ofthe plasmid construct into the genome. However, as we havenever observed the latter phenomenon previously we considerthis unlikely.
enzyme sites at the 5% and 3% ends of the PCRproduct to facilitate subsequent cloning into theunique EcoRI site of pDb.Dh.. Db to create con-struct pDb.Dh.. Db/D-SSU (Fig. 2A).
To create the hDHFR/CS construct, oligonu-cleotides L441 and L442 were annealed togetherand ligated into the unique HindIII site ofconstruct pCS [7]. This linker encompasses se-quence of the CS 3% UTR from nucleotides300–340, which in P. yoelii contains potentialsites for polyadenylation [22], and an EcoRIsite to facilitate cloning into pDb.Dh.. Db. Confir-mation that the linker had been cloned in thecorrect orientation was obtained by sequencing.The whole CS coding region and flanking 5% UTR(2.0 kb) and extended 3% UTR (0.34 kb) wasremoved from this construct on a 3.34-kb EcoRIfragment and cloned into the EcoRI site ofpDb.Dh.. Db (Fig. 1A), creating pDb.Dh.. Db/CS(Fig. 3A).
Plasmid DNA for transfection was purified us-ing Qiagen Maxiprep Columns (Chatsworth, CA).For episomal-based transfections, 30–50 mg ofcircular DNA was used. To obtain site-specificintegration of the construct into the D-SSUrRNA or the CS locus, 50 mg of pDb.Dh.. Db/SSUor pDb.Dh.. Db/CS was linearised at the uniqueSpeI or PacI sites, respectively, prior totransfection.
2.4. Transfection of parasites and drug selection
P. berghei schizonts (108) were prepared fortransfection and DNA constructs were introducedby electroporation as previously described [23,24].The resulting transfection mix was inoculated in-travenously (i.v.) into 2–4 Swiss mice. Thirtyhours after inoculation of the schizonts, the ani-mals were treated s.c. with WR99210 for 3 consec-utive days with a single dose of between 6 and 40mg kg−1 bodyweight to select for transformedparasites containing hDHFR. Parasites survivingthis drug treatment were mechanically passaged tonaive mice by intraperitoneal (i.p.) injection of 108
parasites. This new group of mice was treated at30 h post-infection with WR99210 as describedabove. When the parasitemia reached 10–20%,blood was collected by cardiac puncture and the
plified by PCR using oligonucleotides L78R and332R [R. van Spaendonk, unpublished]. Theseoligonucleotides incorporated EcoRI restriction
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212 203
leukocytes removed [25]. Total parasite DNA wasisolated as described [26] and genotype analysiswas performed using PCR and Southern analysis
of restriction fragments and chromosome fieldinversion gels according to standard methodolo-gies [27,28].
Fig. 2. Transgenic P. berghei, containing a single copy of the hDHFR gene targeted into the D-SSU rRNA gene. A. Schematicrepresentation of the construct pDb.Dh.. Db/SSU containing a truncated D-SSU as target sequence, the wildtype D-SSU rRNA locusand the resulting D-SSU rRNA disruptant locus, as indicated. Construct pDb.Dh.. Db/SSU was linearised within the target sequenceat the unique SpeI site (indicated by an asterisk) prior to transfection. Dashed lines represent pUC19 plasmid sequences. B, BamHI;E, EcoRI; H, HindIII; K, KpnI; Sp, SpeI. B. Detection of the hDHFR gene in transgenic parasites by PCR, using oligonucleotidesL379 and L426. Lane 1, clone 15cy1 (parental parasite control); lanes 2–4, transgenic parasites from exp. 122c, 122d and 122d clone4, respectively; lane 5, pDb.Dh.. Db plasmid control. C. Detection of integration of the hDHFR gene into the D-SSU rRNA gene byPCR, using oligonucleotides L92R and L431. Lane 1, clone 15cy1 (parental parasite control); lanes 2–4, transgenic parasites fromexp. 122c, 122d and 122d clone 4, respectively; lane 5, D-SSU rRNA gene integrant control [R. van Spaendonk, unpublished]. D.Southern blot hybridisation of HindIII-KpnI digested DNA purified from clone 15cy1 parasites. Lane 1, parental parasite control;lane 2, transgenic parasites from exp. 122d clone 4 using the D-ETS probe. E. Southern blot hybridisation of (I) HindIII or (II)KpnI digested parasite DNA with the hDHFR probe. Lane 1, construct pDb.Dh.. Db/SSU; lane 2, clone 15cy1 (parental parasitecontrol); lane 3, transgenic parasites from exp. 122d clone 4.
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212204
Fig. 3. Reintroduction of the CS gene into the CS-knockout mutant of P. berghei. A. Schematic representation of the constructpDb.Dh.. Db/CS, the CS locus of the CS knockout mutant (CS(− )2), and the complemented CS knockout generated by targetingconstruct pDb.Dh.. Db/CS into the CS knockout locus, as indicated. Construct pDb.Dh.. Db/CS, which contains the CS coding regionand flanking 5’ UTR and extended 3’ UTR on a 3.34 kb EcoRI fragment, was linearised at the unique PacI site (indicated by anasterisk) prior to transfection in order to target the hDHFR/CS construct into the CS knockout locus. Dashed lines representpUC19 plasmid sequences. E, EcoRI; H, HindIII; K, KpnI; P, PacI; Ps, PstI; X, XbaI. B. Hybridisation of P. berghei chromosomesseparated by FIGE with the CS or hDHFR probe. Lane 1, clone 15cy1 (parental parasite control), lanes 2 and 4, transgenic parasitesisolated from exp. 125c and 128, respectively; lane 3, transgenic parasites isolated from transfection exp. 125b in which onlyepisomally-replicating pDb.Dh.. Db/CS could be detected (data not shown). C. (I) Southern blot hybridisation of KpnI digestedparasite DNA with the hDHFR probe. Lane 1, pDb.Dh.. Db/CS (plasmid control); lane 2, clone 15cy1 (parental parasite control); lane3, transgenic parasites isolated from transfection exp. 125b in which only episomally-replicating pDb.Dh.. Db/CS could be detected(data not shown); lanes 4 and 5, transgenic parasites isolated from exp. 125c and exp. 128, respectively. (II) Southern blothybridisation of XbaI/PstI digested parasite DNA with a DHFR-TS probe from P. berghei. Lane 1, transgenic parasites from exp.128; lane 2, parasites isolated from exp. 139.
3. Results
3.1. Sensiti6ity of P. berghei to WR99210
As a first step to examine the applicability of
WR99210 in the P. berghei rodent model, wedetermined the in vitro sensitivity of P. berghei tothis drug in standard drug assays for asexualblood stage development [19]. Parasite cloneswere tested that were known to be either suscepti-
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212 205
ble or resistant to the drug pyrimethamine. Table2 shows that asexual blood stage development inall P. berghei clones is highly sensitive toWR99210 (IC50 values in the low nM range),including pyrimethamine-resistant clones contain-ing either the P. berghei (bDHFR-TS) or T.gondii DHFR-TS (tDHFR-TS) selectable markergene.
We next examined the in vivo sensitivity ofblood stage development of P. berghei toWR99210. In preliminary experiments we ob-served that i.p. injection of WR99210 was associ-ated with some known gastrointestinal intolerance[13] and, therefore, we chose to administer the
drug via the s.c. route. Mice with an initialparasitemia of between 1 and 3% were treatedwith a single dose of WR99210 on three consecu-tive days at a concentration of either 16, 20, 32 or40 mg kg−1 bodyweight. All mice were cured ofparasites and, furthermore, remained free ofparasites for at least 11 days after the final drugtreatment. If WR99210 was administered atlower concentrations of 6 or 12 mg kg−1 body-weight, the parasitemias dropped rapidly andwere undetectable levels by light-microscopy bythe time the final drug treatments weregiven. However, the parasitemias increased againbetween day 3 and 11 after the final injection.These results demonstrated that WR99210 is ef-fective in inhibiting blood stage developmentof P. berghei in mice and suggested that WR99210could also potentially be used in vivo to selectfor WR99210-resistant parasites expressing hD-HFR.
3.2. Selection of transgenic P. berghei containingthe hDHFR gene on episomally-maintainedplasmids with WR99210
Four separate transfection experiments wereperformed whereby construct pDb.Dh.. Db contain-ing the hDHFR gene (Fig. 1A) was introducedinto the pyrimethamine-sensitive clone 15cy1. Par-asites were i.v. injected into mice (2–3 per experi-ment), which were treated 30 h post infection with16 (exp. 101.2), 20 (exp. 102), 32 (exp. 101.1) or40 mg kg−1 (exp. 103) WR99210. By day 3-6 afterdrug treatment, parasites could be detectedin all mice by light microscopy. When the para-sitemia reached 1–3%, parasites were mechani-cally passaged to naive mice, which were againtreated with WR99210 at a concentration of 20mg kg−1. Parasites from these mice were collectedwhen the parasitemia reached 10–20% and totalparasite DNA was extracted and analysed toconfirm if transfected parasites harboured hD-HFR.
PCR analysis using a hDHFR specific oligonu-cleotide (L379) and a pUC19 specific oligonucle-otide (L431) generated the expected product of1.1 kb, indicating that construct pDb.Dh.. Db was
Table 2WR99210 sensitivity of P. berghei clones transfected withvarious DHFR genes
line) HFR-TS in genomeconferringpyrimethamine resis-tance
Clone 139 3×10−6 Multiple copies ofhDHFR on episomeand single copy ofbDHFR-TS ingenomeSingle copy of hD-1×10−8Clone 128HFR and bDHFR-TS in the genome
a The WR99210 IC50 value is given as the mean resultobtained from standard in vitro drug assays to test the sensi-tivity of P. berghei blood stage development. These assayswere performed in duplicate on two or three separate occa-sions.
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212206
present in these parasites (Fig. 1B). Plasmid res-cue, through transformation of Escherichia coliwith parasite DNA [24] and restriction enzymeanalysis confirmed the presence of the construct inthe parasites in an unrearranged form (Fig. 1C).The copy number of the plasmid within the trans-fected parasites was established by digesting totalparasite DNA with BclI and probing with the 500bp SphI-EcoRV fragment of the 3% UTR of thebDHFR-TS gene (DT-3% probe) (Fig. 1D). BclIcuts once within the coding region of the bD-HFR-TS gene and once within the 5% UTR of thisgene and thus a 1.4-kb fragment of the endoge-nous bDHFR-TS gene and the entire 6.0-kb plas-mid hybridised to this probe. The relativeintensity of the hybridisation signals of the twofragments were analysed using ImageQuant Soft-ware (Molecular Dynamics, Sunnyvale, CA) Thiswas used to estimate the average copy number inthe four different experiments, which was calcu-lated to be about 20, 24, 40 and 42 in parasitesfrom exp. 101.2, 102, 101.1 and 103, respectively(data not shown). An additional selection in miceof the resistant parasites from exp. 101.2 witheither 6, 12, or 20 mg kg−1 bodyweight WR99210demonstrated that an increase in drug concentra-tion generates an increase in plasmid copy num-ber, with the average copy number observed being25, 27 and 51, respectively (Fig. 1D).
Parasites of exp. 101.2 (20 copies of the plas-mid/parasite) were also tested for their sensitivityto WR99210 in the standard in vitro drug assaysas described above. These parasites showed IC50
values of 1 mM, approximately 1000-fold higherthan the IC50 value for the non-transfectedparental clone 15cy1 (Table 2). These results indi-cate that the increase in resistance of transfectedparasites to WR99210 was due to the presence ofepisomally-maintained plasmids containing thehDHFR gene.
3.3. Selection of transgenic P. berghei containinga single copy of the hDHFR gene integrated intothe D-SSU rRNA gene with WR99210
Parasites transfected with construct pDb.Dh.. Db
maintained a high copy number of this plasmidduring selection with WR99210 and the copy
number appeared to be related to the drug con-centration. The question arose, therefore, whethera single copy of hDHFR would be sufficient togenerate a resistance level to WR99210 to allowfor selection of transgenic parasites harbouringonly one copy of the selectable marker. Conse-quently, we targeted one copy of the hDHFR geneinto the D-SSU rRNA gene locus, which is lo-cated on chromosome 6 of P. berghei parasites ofclone 15cy1 [A. Waters and C. Janse, unpub-lished]. This single copy gene was chosen as it hasbeen successfully targeted using the tDHFR-TSselectable marker gene [R. van Spaendonk, un-published]. Accordingly, a D-SSU rRNA target-ing vector containing a hDHFR expressioncassette (pDb.Dh.. Db/SSU) was constructed (Fig.2A). Parasites of clone 15cy1 were transfectedwith this vector after linearisation with SpeI andi.v. injected into two mice that were treated forthe following 3 consecutive days with either 6 or12 mg kg−1 WR99210 (exp. 122c and 122d, re-spectively). We chose to select with these lowerdrug concentrations in the expectation that para-sites with only a single copy of the hDHFR genemight be killed at higher drug concentrations.
In both mice drug treatment resulted in aninitial parasite clearance to below detectable lev-els, although by days 3 to 6, the parasitemiaincreased to levels of around 1%. These parasiteswere mechanically passaged to naive mice, whichwere treated again with 6 mg kg−1 bodyweightfor 3 consecutive days. When the parasitemia ofthese mice reached 10–20%, parasites were col-lected and their DNA extracted.
PCR analysis on parasite DNA, using hDHFR-specific oligonucleotides (L379 and L426) oroligonucleotides specific for an integration eventinto the D-SSU rRNA gene (L92R and L431)demonstrated that hDHFR gene had been success-fully targeted into the target locus since the re-spective expected products of 0.56 kb and 2.4 kbwere obtained. (Fig. 2B and 2C). Furthermore,probing of genomic DNA digested with HindIIIand KpnI with a 750-bp D-ETS fragment (probe99C, [29]) that flanks the D-SSU rRNA gene andwhich hybridises to the external transcribedspacer (ETS) sequence from both the C and Dribosomal units, confirmed that integration of a
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212 207
single copy of the construct had occurred withinthe target locus (Fig. 2D).
Unexpectedly, plasmid rescue experiments usingundigested parasite DNA revealed that the para-sites from both exp. 122c and 122d additionallycontained episomal DNA on which the hDHFRgene also resided (data not shown). Parental para-sites were also present in these selected popula-tions, albeit at a low frequency since they couldonly be detected by PCR using oligonucleotidesL260R and L271R and not by Southern blotanalysis (Fig. 2D). To remove the episomal plas-mids, limiting dilution was performed [23], inwhich the cloning procedure and the subsequentexpansion of the clones in mice were performed inthe absence of drug selection. The 4 parasitesclones that were obtained from exp. 122d werehence named clone 122d.1-122d.4. PCR analysisof these clones revealed the presence of hDHFR(primers L379/L426) and integration into the D-SSU rRNA gene (primers L92R/L431) but nocircular plasmids or tandem repeats of the inte-grated construct (primers L425/L431). Further-more, Southern blot analysis of these clonesdemonstrated the hDHFR probe only hybridisedto a single HindIII fragment of 11.4 kb or to aKpnI fragment of 10.9 kb, which are the predictedsizes for a single integration event into the D-SSUrRNA gene, and not to an 8.0-kb fragment thatwould correspond to episomally-replicatingpDb.Dh.. Db/SSU or to tandemly repeated plasmidcopies in the genomic locus (Fig. 2E).
The in vitro sensitivity to WR99210 of clone122d.4, which contains only a single integratedcopy of the hDHFR gene, was also tested usingstandard drug assays. Interestingly, the IC50 ofthese parasites to WR99210 was 5 nM (Table 2),only slightly above the IC50 value for the untrans-fected parasites, demonstrating that a single copyof hDHFR under control of the promoter used inthis study does not generate high levels of resis-tance to WR99210. However, this single copy ofthe human DHFR gene does generate a 1000-foldincrease in resistance to pyrimethamine (data notshown). This suggests that hDHFR enzymaticactivity is significantly more sensitive to WR99210than to pyrimethamine at the concentrations ofdrug tested in this study.
3.4. Reintroduction of the CS gene into aCS-knockout mutant of P. berghei using thehDHFR/WR99210 selection system
To determine whether the hDHFR gene couldalso be used as a selectable marker in conjunctionwith the existing pyrimethamine resistance mark-ers, we introduced the hDHFR gene into trans-genic parasites (clone CS(− )2) that alreadycontained the pyrimethamine-resistant DHFR-TSgene of P. berghei. In these parasites the CS genehad been disrupted by targeting the resistant bD-HFR-TS gene into the CS locus [7]. The parasitesof clone CS(− )2 are sensitive to WR99210 asdetermined by in vitro drug assays, with an IC50
value comparable to the non-transformed clone15cy1 (Table 2). Furthermore, CS(− )2 is sensi-tive to WR99210 in vivo at a concentration of 12mg kg−1 bodyweight (data not shown).
Construct pDb.Dh.. Db/CS, which contains boththe hDHFR and CS expression cassettes (Fig.3A), was introduced into CS(− )2 parasites aseither episomes (exp. 126 and 139) or as linearDNA (exp. 125 and 128) in order to target theconstruct into the disrupted CS locus (Fig. 3A).Transfected parasites were i.v. injected into mice,which were then treated with WR99210 at either 6mg kg−1 (exp. 125a) or 12 mg kg−1 (exp. 125band 125c, 126, 128 and 139). When the para-sitemia of these mice reached 5-10%, parasiteswere mechanically passaged to naı̈ve mice, whichwere given the same WR99210 treatment regime.Parasites were collected when parasitemiasreached 10–20% and parasite DNA was extractedfor analysis.
PCR analysis of parasite DNA using the hD-HFR-specific oligonucleotides (L379 and L426) orthe CS-specific oligonucleotides (L476 and L477)gave the expected product size, demonstratingthat the construct was present in parasites fromall transfection experiments (data not shown).Plasmid rescues on parasite DNA extracted fromexp. 126 and 139 confirmed the presence of theconstruct in these parasites in an unrearrangedform, and probing BclI digests with the 3% UTRof the bDHFR-TS gene revealed that the plasmidcopy number was approximately 10 per parasite(data not shown).
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212208
To determine if the construct had integratedinto the disrupted CS locus located on chromo-some 4 [23] in parasites from exp. 125b, 125c and128, chromosomes from these parasites were sepa-rated by field inversion gel electrophoresis (FIGE)and then probed with either a 1.1-kb HincII frag-ment from the CS open reading frame (ORF) orwith the hDHFR ORF. Both probes hybridised tochromosome 4 of parasites isolated from exp.125c and 128, indicating that linear pDb.Dh.. Db/CS had been successfully targeted to the disruptedCS locus (Fig. 3B). To confirm these results,Southern blots of KpnI-digested genomic DNA ofexp. 125c and 128 were also probed with hDHFR(Fig. 3C). This blot demonstrates that the hD-HFR probe hybridised to a 13.8 kb KpnI frag-ment, which is the predicted fragment size for anintegration event into the disrupted CS locus.However, this probe also recognised a fragment of8.8 kb in both experiments, which corresponds tothe presence of intact, episomally-maintainedpDb.Dh.. Db/CS and/or to tandemly repeated plas-mid copies in the genomic locus. Plasmid rescueexperiments confirmed the presence of this con-struct in an unrearranged form (data not shown).The presence of episomes in addition to the de-sired integration event into the CS-locus isanalogous to the presence of episomes in theexperiments in which we targeted the hDHFR intothe D-SSU rRNA locus (see above). A Southernblot of XbaI/PstI digested DNA from exp. 128probed with the bDHFR-TS probe did not revealthe presence of the parental CS-knockout geno-type of clone CS(− )2, indicating that if theseepisomally-replicating plasmids arose by reversionor recombination between the several identicalsequences present at the integration locus wherebythe plasmid looped back out of the genome, thensuch events must have occurred at a low fre-quency (Fig. 3C). Nonetheless, the maintenanceof parasites from exp. 128 in mice for a period of3 weeks in the absence of drug pressure resulted inthe loss of these additional plasmids, while asingle integrated copy of hDHFR was maintained.
Using standardised drug assays, we next testedthe WR99210 in vitro sensitivity of parasites fromexp. 128 that had lost their episomal plasmids andfrom exp. 139. These parasites showed an IC50 of
10 nM and 3 mM, respectively, while the parentalclone CS(− )2 had an IC50 value of 3 nM (Table2). This slight increase in WR99210 resistanceafter introduction of a single copy of the hDHFRgene in the CS locus corresponds with the smallincrease in the IC50 after targeting this gene in theD-SSU rRNA locus.
4. Discussion
Amongst other reasons, genetic transformationof malaria parasites has been limited by the num-ber of selectable markers that are available forselection of transgenic parasites. Up until now theonly selectable markers reported for the rodentparasite P. berghei have been the pyrimethamine-resistant DHFR-TS genes of T. gondii and of P.berghei, which both confer resistance against theantimalarial drug pyrimethamine. Here we reportthe use of the hDHFR gene as a selectablemarker, which confers resistance to the potentmalaria DHFR inhibitor WR99210 uponpyrimethamine-sensitive as well aspyrimethamine-resistant lines of P. berghei. Thisis possible since the binding interaction ofWR99210 with the DHFR substrate pocket isdistinct to that of pyrimethamine.
The concentration of WR99210 used to s.c.treat mice infected with P. berghei in this studywas around 20-fold lower than that used in aprevious study conducted by Canfield et al [13].However, in this earlier investigation WR99210was administered at very high concentrations inorder to establish the curative dose, based upon acriterion of parasite clearance for a total of 60days. We found that a significantly reduced drugchallenge was sufficient to effectively select forresistant parasites transfected with and expressinghDHFR.
WR99210 selection of P. berghei parasitestransfected with circular constructs containing thehDHFR gene resulted in the generation of highlyresistant parasites containing high copy numbersof episomally-maintained plasmids. Interestingly,the plasmid copy number per parasite increasedwith increasing concentrations of WR99210 usedfor selection of transformed parasites, more obvi-
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212 209
ously so than plasmids encoding pyrimethamineresistance [1]. The possibility to control plasmidcopy numbers would be beneficial in experimentswhere the level of expression of transgenes isimportant. We also found that transfection oftransgenic parasites harbouring thepyrimethamine-resistant T. gondii or P. bergheiDHFR-TS genes with circular constructs contain-ing the hDHFR gene rendered parasites 1000-times more resistant to WR99210. The fact thatthese pyrimethamine-resistant transgenic parasitesharbouring multiple copies of the hDHFR genecould be selected for using WR99210 is signifi-cant, since hDHFR will now be a powerful toolfor the episomal complementation of knock-outmutant parasites and for the expression of a sec-ond transgene in already transgenic parasites.
In order to use the hDHFR selectable marker asa means to knock-out a second gene or to targeta second transgene into the genome of transgenicparasites, however, a single genomic copy of thismarker must also confer an increase in resistanceto WR99210. Based on the high resistance level ofparasites containing episomally-maintained hD-HFR constructs, it was surprising to find thatparasites harbouring a single copy of the hDHFRgene were only slightly more resistant toWR99210 than the parental parasites. It seemsunlikely that this is due to a decrease in transcrip-tion efficiency of the hDHFR gene when targetedinto the genome, since the same single copy ofhDHFR is sufficient to generate around a 1,000-fold increase in resistance to pyrimethamine. Asimilar increase in pyrimethamine resistance isobserved when a single copy of the bDHFR-TSselectable marker gene is placed under control ofthe same regulatory regions used to drive expres-sion of hDHFR and targeted into the same rRNAD-SSU or CS loci of the parasite genome [7,30].These results indicate that while hDHFR is highlyresistant to pyrimethamine, it is not as effective inproviding sufficient tetrafolate pools at concentra-tions of WR99210 that are able to block theparasite’s DHFR enzyme. This could be due toWR99210 binding more efficiently thanpyrimethamine to the hDHFR enzyme at theconcentrations of drug used in this study and maypartially explain the toxic side effects of WR99210
when administered to humans [31]. Alternatively,binding of WR99210 to the parasite DHFR maysterically block the function of the TS moiety thatis complexed to DHFR or WR99210 may betargeting an additional parasite enzyme involvedin the folate synthesis and metabolism pathway ashas been previously proposed [32]. Our results atthis stage, however, cannot support either of thesespeculations.
Despite the low increase in WR99210 resistanceof parasites containing a single copy of the hD-HFR gene, we were able to select for transgenicparasites that had a single copy of hDHFRtargeted into the genome on multiple occasions.Interestingly, using the plasmid rescue method, wefound in all experiments that parasites harbouringthe desired integration event also contained episo-mally-maintained plasmids containing the hD-HFR gene. This was despite the fact that purified,linear DNA was used in these transfection experi-ments. Considering that integration of hDHFRgene into the genome occurred by a singlecrossover event, it is possible that a reversionevent resulted in the plasmid looping back outfrom the genome, generating plasmids that couldsegregate during schizogony. Reversion must haveoccurred at a low frequency since the parentalgenotype could only be detected by PCR and notby Southern blot analysis. The presence ofepisomes in the parasite population would explainwhy a band the size of the plasmid constructhybridised to the hDHFR probe in the Southernblots. This size fragment would also be releasedfrom the genome, however, if tandemly repeatedcopies of the construct had integrated into thegenomic locus, although this is contradictory tothe results from the plasmid rescue experimentsthat had clearly indicated the presence of intact,episomally-replicating plasmids. Due to the com-plexity of the recombinant genomic locus we havebeen unable to resolve this discrepancy, althoughreversion has been previously observed when con-structs designed for a single crossover event con-taining the tDHFR-TS selection cassette havebeen used to direct integration into the genome[8]. In either situation, the WR99210 treatmentwould probably select for these parasites contain-ing extra copies of the hDHFR gene, since they
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212210
will be more resistant to WR99210. However,despite the fact that it was more difficult to selectfor transgenic parasites containing a single copyof the hDHFR gene due to the small increase ofresistance to WR99210, it was still possible toselect for the desired genetic events by loweringthe concentration of WR99210 for selection tobelow that required for the total cure of miceinfected with non-transformed parasites andcloning the parasites in the absence of drug toremove parasites containing additional copies ofthe plasmid construct.
By using hDHFR as a second selectable marker,we report here for the first time in Plasmodium thereintroduction of an intact gene into a knockoutmutant parasite which contains a disrupted formof that gene. In our case, we reintroduced the CSgene into the CS deletion mutant [7]. This wasachieved by introducing either multiple copies ofthe CS gene in trans (on episomes) or by targetinga single copy of the gene back into the disruptedCS locus. We anticipated that both of thesestrategies would result in the restoration of the CSwildtype phenotype since we have previouslyshown that genes located on episomes can beexpressed in the mosquito stages of the parasite[33]. However, in a number of controlled experi-ments, in which Anopheles stephensi mosquitoeswere fed on mice infected with transgenic para-sites, we could not detect CS transcription orexpression at days 10 and 18 post-infection byeither reverse transcriptase-PCR or indirect im-munofluorescence assays. Furthermore, thesetransgenic parasites were unable to form matureinfective sporozoites in the oocysts and salivaryglands of infected mosquitoes (data not shown).Since the completion of this study, it has beensubsequently discovered that the CS transcriptionunit isolated from pCS [7] harbours a 600–700 bpIS1 insertion sequence approximately 200 bp up-stream of the ATG start codon [R. Menard andV. Thathy, unpublished]. Given that in P. yoellithe CS transcriptional start site(s) has beenmapped around base −270 [22], this would placethe insertion sequence around the start site of CStranscription in our CS expression cassette if the5% UTRs between the P. berghei and P. yoelii areconserved. This would, therefore, explain the lack
of CS expression in our genetically-restored trans-genic parasites and contribute to the lack of ex-pression of the expected truncated form of CS inthe CS knockout parasites [7]. Nonetheless, thedevelopment of hDHFR as a selectable marker,together with the recent addition of the NEO(encoding neomycin phosphotransferase II fromtransposon Tn 5) and BSD (encoding blasticidin Sdeaminase of Aspergillus terreus) genes as positiveselectable markers for P. falciparum transfection[34], will open new avenues for the application oftransformation technology in Plasmodium. Thisincludes possibilities for genetic restoration ofgene knockout mutants as well as the introductionor disruption of more than a single gene withinthe same parasite clone.
Acknowledgements
The authors would like to acknowledge DrRaymond Blakley (St Jude’s Hospital, Memphis,TN) for providing the hDHFR cDNA, Dr DavidJacobus for invaluable discussions and supplyingWR99210 and Dr Alister Ward for critical read-ing of this manuscript. This research was sup-ported in part by the INCO-DC programme ofthe DGXII Department of the European Com-mission (Contract Number ERBIC18CT960052)and by the Life Sciences Foundation (SLW),which is subsidised by the Netherlands Organisa-tion for Scientific Research (NWO).
[2] Donald RGK, Roos DS. Stable molecular transformationof Toxoplasma gondii : a selectable dihydrofolate reduc-tase-thymidylate synthase marker based on drug resis-tance mutations in malaria. Proc Natl Acad Sci USA1993;90:11703–7.
[4] Kocken CHM, van der Wel AM, Dubbeld MA, NarumDL, van de Rijke FM, van Gemert G-J, van der Linde X,Bannister LH, Janse CJ, Waters AP, Thomas AW. Pre-
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212 211
cise timing of expression of a Plasmodium falciparum-derived transgene in Plasmodium berghei is a criticaldeterminant of subsequent subcellular localization. J BiolChem 1998;273:15119–24.
[5] Margos G, van Dijk MR, Ramesar J, Janse CJ, WatersAP, Sinden RE. Transgenic expression of a mosquito-stage malarial protein, Pbs21, in blood stages of trans-formed Plasmodium berghei and induction of an immuneresponse upon infection. Infect Immun 1998;66:3884–91.
[6] Crabb BS, Cooke BM, Reeder JC, Waller RF, CaruanaSR, Davern KM, Wickham ME, Brown GV, Coppel RL,Cowman AF. Targeted gene disruption shows that knobsenable malaria-infected cells to cytoadhere under physio-logical shear stress. Cell 1997;89:287–96.
[7] Menard R, Sultan AA, Cortes C, Altszuler R, van DijkMR, Janse CJ, Waters AP, Nussenzweig RS, Nussen-zweig V. Circumsporozoite protein is required for devel-opment of malaria sporozoites in mosquitoes. Nature1997;385:336–40.
[8] Sultan AA, Thathy V, Frevert U, Robson KJ, ChrisantiA, Nussenzweig V, Nussenzweig RS, Menard R. TRAP isnecessary for gliding motility and infectivity of Plasmod-ium sporozoites. Cell 1997;90:511–2.
[9] Rieckmann KH. The in vitro activity of experimentalantimalarial compounds against strains of P. falciparumwith varying degrees of sensitivity to pyrimethamine andchloroquine. Chemotherapy of malaria and resistance toantimalarials. WHO Tech Rep Ser 1973;529:58.
[10] Childs GE, Lambros C. Analogues of N-benzyloxydihy-drotriazines: in vitro antimalarial activity against Plas-modium falciparum. Ann Trop Med Parasitol1986;80:177–81.
[11] Knight DJ. Babesia rodhaini and Plasmodium berghei. Ahighly active series of chlorophenoxyalkoxy-substituteddiamino-dihydrotriazines against experimental infectionsin mice. Ann Trop Med Parasitol 1981;75:1–6.
[12] Knight DJ, Mamalis P, Peters W. The antimalarial activ-ity of N-benzyl-oxydihydrotriazines. III. The activity of4,6-diamino-1,2-dihydro-2,2-dimethyl-1-(2,3,5-trichloro-propyloxy)-1,3,5-triazine hydrobromide (BRL51084) andhydrochloride (BRL6231). Ann Trop Med Parasitol1982;74:393–404.
[13] Canfield CJ, Milhous WK, Ager AL, Rossan RN,Sweeney TR, Lewis NJ, Jacobus DP. PS-15: a potent,orally active antimalarial from a new class of folic acidantagonists. Am J Trop Med Hyg 1993;49:121–6.
[14] Hekmat-Nejad M, Rathod PK. Plasmodium falciparum :Kinetic interactions of WR99210 with pyrimethamine-sensitive and pyrimethamine-resistant dihydrofolate re-ductase. Exp Parasitol 1997;87:222–8.
[15] Wooden JM, Hartwell LH, Vasquez B, Sibley CH. Anal-ysis in yeast of antimalarial drugs that target the dihydro-folate reductase of Plasmodium falciparum. Mol BiochemParasitol 1997;85:25–40.
[16] Fidock DA, Nomura T, Wellems TE. Cycloguanil and itsparent compound proguanil demonstrate distinct activi-ties against Plasmodium falciparum malaria parasites
transformed with human dihydrofolate reductase. MolPharmacol 1998;54:1140–7.
[17] Fidock DA, Wellems TE. Transformation with humandihydrofolate reductase renders malaria parasites insensi-tive to WR99210 but does not affect the intrinsic activityof proguanil. Proc Natl Acad Sci USA 1997;94:10931–6.
[18] Van der Kaay HJ, Boorsma EG. A susceptible andrefractive strain of Anopheles atropar6us to infection withPlasmodium berghei. Acta Leidensia 1977;45:13–9.
[19] Janse CJ, Waters AP, Kos J, Lugt CB. Comparison of invivo and in vitro antimalarial activity of artemisinin,dihydroartemisinin and sodium artesunate in the Plas-modium berghei–rodent model. Int J Parasitol1994;24:589–94.
[20] Janse CJ, Van Vianen PH. Flow cytometry in malariadetection, Methods Cell Biol 1994;Pt B:295–318.
[21] de Koning-Ward TF, Speranca MA, Waters AP, JanseCJ. Analysis of stage specificity of promoters in Plasmod-ium berghei using luciferase as a reporter. Mol BiochemParasitol 1999;100:141–6.
[22] Ruvolo V, Altszuler R, Levitt A. The transcript encodingthe circumsporozoite antigen of Plasmodium berghei uti-lizes heterogeneous polyadenylation sites. Mol BiochemParasitol 1993;57:137–50.
[23] Menard R, Janse CJ. Gene targeting in malaria parasites.METHODS: A Companion to Methods Enzymol1997;13:148–59.
[24] Waters AP, Thomas AW, van Dijk MR, Janse CJ. Trans-fection of malaria parasites. METHODS: A Companionto Methods Enzymol 1997;13:134–47.
[25] Janse CJ, Camargo A, del Portillo H, Herrera S, KumlienS, Mons B, Thomas A, Waters AP. Removal of leuko-cytes from Plasmodium 6i6ax-infected blood. Ann TropMed Parasitol 1994;88:213–6.
[26] Vervenne RAW, Dirks RW, Ramesar J, Waters AP,Janse CJ. Differential expression in blood stages of thegene coding for the 21-kilodalton surface protein ofookinetes of Plasmodium berghei as detected by RNA insitu hybridisation. Mol Biochem Parasitol 1994;68:259–66.
[27] Ausubel FM, Brent R, Kingston RE, Moore DD, Seid-man JG, Smith JA, Struhl K. Current Protocols inMolecular Biology. New York: Wiley, 1998.
[28] Janse CJ, Boorsma EG, Ramesar J, Van Vianen PH, Vander Meer R, Zenobi P, Casaglia O, Mons B, Van der BergFM. Plasmodium berghei : gametocyte production, DNAcontent and chromosome size polymorphisms duringasexual multiplication in vivo. Exp Parasitol1989;68:274–82.
[29] Thompson J, van Spaendonk RML, Choudhuri R, Sin-den RE, Janse CJ, Waters AP. Heterogenous ribosomepopulations are present in Plasmodium berghei duringdevelopment in its vector. Mol Microbiol 1999;31:253–60.
[30] van Dijk MR, Janse CJ, Waters AP. Expression of aPlasmodium gene introduced into subtelomeric regions ofPlasmodium berghei chromosomes. Science 1996;271:662–5.
T.F. de Koning-Ward et al. / Molecular and Biochemical Parasitology 106 (2000) 199–212212
[32] Yeo AE, Seymour KK, Rieckmann KH, ChristophersonRI. The activity of triple combinations of antifolatebiguanides, with and without folinic acid, against Plas-modium falciparum in vitro. Ann Trop Med Parasitol1997;91:17–23.
[33] de Koning-Ward TF, Thomas AW, Waters AP, Janse CJ.Stable expression of green fluorescent protein in bloodand mosquito stages of Plasmodium berghei. MolBiochem Parasitol 1998;97:247–52.
[34] Mamoun CB, Gluzman IY, Goyard S, Beverley SM,Goldberg DE. A set of independent selectable markers fortransfection of the human malaria parasite Plasmodiumfalciparum. Proc Natl Acad Sci USA 1999;96:8716–20.
