regulation of pfemp1-var2csa ... - uppsala university · cell and molecular biology, uppsala...

Regulation of PfEMP1–VAR2CSA translation by aPlasmodium translation-enhancing factorSherwin Chan1, Alejandra Frasch1, Chandra Sekhar Mandava2, Jun-Hong Ch’ng1,3,Maria del Pilar Quintana1,4, Mattias Vesterlund5, Mehdi Ghorbal6,7, Nicolas Joannin1,Oscar Franzén8, Jose-Juan Lopez-Rubio6,7, Sonia Barbieri9, Antonio Lanzavecchia9,10,Suparna Sanyal2 and Mats Wahlgren1*

Pregnancy-associated malaria commonly involves the binding of Plasmodium falciparum-infected erythrocytes to placentalchondroitin sulfate A (CSA) through the PfEMP1–VAR2CSA protein. VAR2CSA is translationally repressed by an upstreamopen reading frame. In this study, we report that the P. falciparum translation enhancing factor (PTEF) relieves upstreamopen reading frame repression and thereby facilitates VAR2CSA translation. VAR2CSA protein levels in var2csa-transcribing parasites are dependent on the expression level of PTEF, and the alleviation of upstream open reading framerepression requires the proteolytic processing of PTEF by PfCalpain. Cleavage generates a C-terminal domain that containsa sterile-alpha-motif-like domain. The C-terminal domain is permissive to cytoplasmic shuttling and interacts withribosomes to facilitate translational derepression of the var2csa coding sequence. It also enhances translation in aheterologous translation system and thus represents the first non-canonical translation enhancing factor to be found in aprotozoan. Our results implicate PTEF in regulating placental CSA binding of infected erythrocytes.

The PfEMP1 family of proteins is encoded by the var gene loci,which are present in ∼60 variant copies per genome1. The vargenes are highly polymorphic across parasite strains2 and

generally exhibit mutually exclusive expression, with each parasiteactively transcribing only one var gene at any one time3,4.However, the mechanisms that regulate exclusive allelic regulationin the transcription and translation of var genes remain elusive.One particular PfEMP1 variant, VAR2CSA, has been studied ingreat detail given that it plays an important role in the developmentof pregnancy-associated malaria (PAM). VAR2CSA mediates thebinding of infected erythrocytes (IEs) to the glycosaminoglycanchondroitin sulfate A (CSA) expressed at the surface of syncytio-trophoblasts in the placental lining5,6. The encoding var2csa geneexhibits a 5′ upstream regulatory element that is unique withinthe var gene family7. It also contains a conserved upstream openreading frame (uORF) spanning 360 nucleotides that represses thetranslation of the coding region of var2csa (ref. 8).

Mok et al. described a spontaneously generated mutantPlasmodium falciparum clone that failed to translate VAR2CSAdespite maintaining active var2csa transcription9. The uORF ofvar2csa was subsequently shown to repress the translation of thedownstream var2csa coding region, through insufficient translationre-initiation10. Notably, IEs that were selected for CSA binding andthus transcribed var2csa were capable of derepressing the effects ofthe uORF, even when presented upstream of a reporter gene. Thereporter failed to be translated, however, when transfected intonon-var2csa-transcribing parasites, despite the presence of the

reporter transcripts8. Trans-acting factors were therefore suggestedto be modulating the derepression and translation of var2csa (ref. 10).

Here, we show that the P. falciparum translation enhancing factor(PTEF) protein encoded by PF3D7_0202400 binds to ribosomesand stimulates efficient derepression and translation of VAR2CSA.

ResultsDefective var2csa translation coincides with low ptef expression.P. falciparum 3D7S8.4.2 is a mutant parasite defective in VAR2CSAtranslation, despite a high abundance of the var2csa transcript9. Toidentify the factors involved in regulation of VAR2CSA translation,we obtained the whole-genome sequence for 3D7S8.4.2 andcompared it with the reference 3D7 genome. Major chromosomaldeletions were observed in the left arm of chromosome 2 and inthe right arm of chromosome 9 of the 3D7S8.4.2 parasite,resulting in the deletion of 47 coding genes (SupplementaryTable 1). These deletions are known to occur spontaneouslyduring long-term in vitro propagation of P. falciparum11.

To complement the genome data, we also obtained the transcrip-tomes of 3D7S8.4.2 and NF54CSA, parasites that translateVAR2CSA, and the IEs of which bind to CSA. As 3D7 is a cloneof NF54, the related genetic background shared by these parasitesenabled phenotypic comparisons to be made. We compared thetranscriptomes of NF54CSA and 3D7S8.4.2 at 10, 20, 30 and 40 hpost-invasion (h.p.i.). Despite long-term in vitro divergence,pairwise comparisons of the transcriptomes at four time pointssuggested very similar life-cycle progression and close synchrony

1Department of Microbiology, Tumor and Cell Biology (MTC), Karolinska Institutet, Box 280, Nobels väg 16, 171 77 Stockholm, Sweden. 2Department ofCell and Molecular Biology, Uppsala University, Box-596, 751 24 Uppsala, Sweden. 3Department of Microbiology, National University of Singapore 117545,Singapore. 4Escuela de Medicina y Ciencias de la Salud, Facultad de Ciencias Naturales y Matemáticas, Universidad del Rosario, Calle 12C No. 6-25, Bogotá,Colombia. 5Cancer Proteomics, Department of Oncology-Pathology, Karolinska Institutet, 17176 Stockholm, Sweden. 6University of Montpellier, Faculty ofMedicine, Laboratory of Parasitology-Mycology, Montpellier F34090, France. 7CNRS – 5290, IRD 224 – University of Montpellier (UMR ‘MiVEGEC’),Montpellier, France. 8Department of Genetics and Genomic Sciences, Institute of Genomics and Multiscale Biology, Icahn School of Medicine at MountSinai, New York, New York 10029, USA. 9Institute for Research in Biomedicine, Università della Svizzera italiana, Bellinzona 6500, Switzerland. 10Institute ofMicrobiology, ETH Zurich, Zurich 8093, Switzerland. *e-mail: [email protected]

ARTICLESPUBLISHED: 8 MAY 2017 | VOLUME: 2 | ARTICLE NUMBER: 17068

NATURE MICROBIOLOGY 2, 17068 (2017) | DOI: 10.1038/nmicrobiol.2017.68 | www.nature.com/naturemicrobiology 1

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

mailto:[email protected]

http://dx.doi.org/10.1038/nmicrobiol.2017.68

http://www.nature.com/naturemicrobiology

between the two parasite populations (NF54CSA vs 3D7S8.4.2,R2 = 0.68–0.95, Fig. 1a). As described above, both parasites expresshigh levels of var2csa transcript (Fig. 1b). Notably, 3–8% of theirgenomes exhibited at least twofold expression differences, and withreads per kilobase of transcript per million mapped reads (RPKM)of >10 in at least one parasite population. The smallest differenceswas observed at the late trophozoite stage at 30 h.p.i. (148 genes)and the greatest differences at the schizont stage at 40 h.p.i.(459 genes, Fig. 1a and Supplementary Table 2).

We subsequently screened for potential var2csa translation-regulating gene candidates using the list of differentially expressedgenes, based on three major criteria. First, preference was given togenes previously shown to be associated with PAM. Second,candidate genes with PEXEL motifs predicted in their gene productswere excluded, as they would be exported out of the parasite

cytoplasmic compartment and are less likely to be of relevance totranslational regulation. Third, the temporal expression patternshould be permissive to potential translation regulation of var2csa.Seven genes were found to be differentially expressed between the3D7S8.4.2 and NF54CSA parasites at all time points(Supplementary Fig. 1b), but only one gene, ptef (PlasmoDB geneID PF3D7_0202400, or previously PFB0115w), located on the leftarm of chromosome 2, met all three predetermined criteria.3D7S8.4.2 constitutively expresses much lower levels of ptef thanNF54CSA, with a maximal difference being observed at 20 h.p.i.(Fig. 1b). Interestingly, this time point coincides with activevar2csa transcription, and the difference in ptef expression wasconfirmed by quantitative PCR (qPCR, Fig. 1d). According to thecriteria, high ptef transcript and protein levels have been reportedin parasites obtained from women with PAM in multiple

10 20 30 40

h.p.i.

3D7S8.4.2, log(RPKM)

NF5

4CSA

, log

(RPK

M)

10 h.p.i.

287

100

R2 = 0.95

d

ptef var2csa0

1

2

3

4

5

Rela

tive

expr

essi

on

NF54CSA 1× panNF54CSA 5× pan

***

****

e

a


NF5

4CSA

, log

(RPK

M)

20 h.p.i.

60

104

R2 = 0.91


NF5

4CSA

, log

(RPK

M)

30 h.p.i.

107

41

R2 = 0.90

−2 −1 0 1 2 3 4 5−2 −1 0 1 2 3 4 5−2 −1 0 1 2 3 4 5−2 −1 0 1 2 3 4 5−2

−1

0

1

2

3

4

5

−2

−1

0

1

2

3

4

5

−2

−1

0

1

2

3

4

5

−2

−1

0

1

2

3

4

5


NF5

4CSA

, log

(RPK

M)

40 h.p.i.

100

359

R2 = 0.68

0

5,000

10,000

15,000

Read

s pe

r mill

ion

ptef

FCR3S1.2(NR)IT4CD36/ICAM1FCR3

NF54CSA3D7S8.4.2

10 20 30 400

5,000

10,000

15,000

20,000

h.p.i.

Read

s pe

r mill

ion

var2csa


NF54CSA3D7S8.4.2

10 20 30 400

50

100

h.p.i.

Expr

essi

on p

erce

ntile

ptef


NF54CSA3D7S8.4.2

b c

3D7S8.4.2 NF54CSA0.0

0.1

0.2

0.3

Rela

tive

copy

num

ber (

ptef

/con

trol

)

**

Figure 1 | ptef expression is associated with the CSA-binding capacity of IEs. a, Pairwise comparisons of the whole transcriptomes of NF54CSA and3D7S8.4.2 at 10, 20, 30 and 40 h.p.i. Orange and blue data points denote genes upregulated and downregulated in NF54CSA, respectively, and the numbersdenote genes falling under the corresponding categories. Only genes with RPKM> 10 in at least one of the populations are considered. R2 values show thecorrelations between the transcriptomes. Arrows denote ptef expression, which was upregulated in NF54CSA at all time points. RPKM, reads per kilobase oftranscript per million mapped reads. b, Expression of ptef and var2csa transcripts in five P. falciparum strains as extracted from RNA-sequencedtranscriptomes. c, Respective expression percentile of ptef based on RPKM values. d, Differences in ptef expression between NF54CSA and 3D7S8.4.2 werevalidated with qPCR. The plot shows relative expression of ptef normalized to fructose biphosphate aldolase: 3D7S8.4.2 (n = 3) and NF54CSA (n = 4)biological replicates. Error bars indicate s.e.m., **P≤0.01, Student’s t-test. e, Differences in the expression of ptef and var2csa in two NF54CSAsubpopulations that were split after initial CSA-binding selection. One (5× pan) was continuously selected for CSA binding four more times at 4- to 6-weekintervals and the other (1× pan) was not selected over the same period of time. Expression was normalized to fructose biphosphate aldolase relative to theexpression of 1× pan; n = 2 biological replicates, error bars indicate s.e.m., ***P≤0.001, ****P≤0.0001, Student’s t-test.

ARTICLES NATURE MICROBIOLOGY

NATURE MICROBIOLOGY 2, 17068 (2017) | DOI: 10.1038/nmicrobiol.2017.68 | www.nature.com/naturemicrobiology2




independent studies12–16, supporting their possible role in theregulation of var2csa translation. Furthermore, the breakpoint onthe left arm of chromosome 2 of parasite 3D7S8.4.2 is located∼1.5 kb upstream from the start codon of ptef (SupplementaryFig. 1a). This implies possible perturbation of the transcriptionalregulation of ptef, resulting in its downregulation in 3D7S8.4.2,thus reinforcing the possibility that inefficient var2csa translationmay be linked to a loss of ptef.

The association of the expression of ptef with PAM could denotea function related to the regulation of the CSA-binding proteinVAR2CSA and, thus, the CSA binding capacity of IEs. Highexpression, however, could also be due to other requirements forplacental niche adaptation. To determine whether high expressionof ptef is a possible requisite restricted to CSA-binding parasites,we extended our transcriptomic comparison to three non-CSA-binding parasites (FCR3, FCR3S1.2 (non-rosetting) andIT4-CD36/ICAM1). The expression of ptef in these parasitepopulations was found to be 7- to 15-fold lower at 20 h.p.i.compared with NF54CSA, yet their expression percentile of ptefremained high, all with ptef expression among the top 20% of

expressed genes (Fig. 1b,c). One population of CSA-pannedNF54CSA was further divided into two subpopulations. Onesubpopulation was subjected to routine CSA panning (fouradditional CSA pannings at intervals of four to six weeks), andthe other was not enriched for the CSA-binding phenotype. Wefound that the CSA-binding phenotype co-selects for high var2csaand ptef transcript levels (Fig. 1e).

Sequence analysis of PTEF identifies an RNA-binding domain.Based on the primary sequence of PTEF, we demarcated the geneproduct to include an N-terminal domain (NTD), a linkerdomain (LD) and a C-terminal domain (CTD) (Fig. 2a). TheNTD contains two tandemly arranged bipartite nuclearlocalization signals (NLSs) with a hydrophobic stretch of aminoacids (HS, residues 1–20) (Fig. 2a and Supplementary Fig. 2a).The LD is highly negatively charged and contains imperfecttandem repeats with a simple amino-acid composition. There areat least two genotypes with variable numbers of repeats inlaboratory-adapted strains (Supplementary Fig. 2a). The CTDincludes a leucine-rich nuclear export signal (NES) and shares

b

α-hsp70

kDaNF5

4CSA

3D7S8.4.2

α-PTEF(α-CTD)

α-PTEF(α-CTD)

185

50

a 53 kDa

PTEF Predicted Mw = 142 kDa

SAM-like domainNTD (1008–1183)

N-terminal domainNTD (1–368)

Linker domainLD (369–756)

C-terminal domainCTD (757–1192)

Hydrophobicstretch

HS (1–20)

Nuclear localizationsignal

NLS (254–286)

Nuclear exportsignal

NES (1073–1080)

d

Time (h)

α-hsp70

α-PTEF(α-CTD)

α-PTEF(α-NTD)

CPI (dilution)

Epoxomicin (nM)

Time (h)

c

kDa

185

185

115

115

50

α-hsp70

0−ve Le

upeptin

Aprotinin

E641×

ALL

N

2× ALL

N

DSM1

4 4 4 4

0 4 4 4 4− − − − −

4 4 4 4 4

0 0 100 500 1,000− − − − −

0 1,000 500 100 50

4 4 4

Figure 2 | Proteolytic processing separates antagonistic targeting motifs in the PTEF protein. a, The primary structure of a PTEF protein, the N-terminaldomain (NTD, red), the linker domain (LD, white) and the C-terminal domain (CTD, blue) are demarcated by features in the primary sequence. HS, hydrophobicsequence; NLS, nuclear localization signal; NES, nuclear export signal. Numbers in parentheses denote corresponding residues in the primary structure. b,Parasite lysates were resolved in SDS–PAGE and analysed by immunoblotting with antibodies against the CTD of PTEF (α-CTD IgG), with the detection of hsp70serving as a loading control. The arrow denotes CTD cleavage products. Representative of five biological replicates. c, Immunoblotting of SDS–PAGE-resolvedlysates from parasite cultures incubated in cysteine protease inhibitor cocktail (CPI), representative of four biological replicates (left), and epoxomicin for 4 h(right), with both α-CTD IgG (upper) and α-NTD serum (lower), representative of two biological replicates. d, NF54CSA lysate was incubated with individualprotease inhibitors: –ve (no inhibitor), leupeptin (75 µM), aprotinin (300 nM), E64 (10 µM), 1× ALLN (40 µM), 2× ALLN (80 µM) and DSM1 (75 µM).Representative of two biological replicates. Demarcation is not based on the cleavage pattern, and the precise cleavage site has not been mapped.

NATURE MICROBIOLOGY ARTICLES





significant homology with the gametocyte-specific antigenPfg25/27, which encompasses residues 1008–1183 of PTEF(Fig. 2a). This resembles a sterile alpha motif-like (SAM-like)domain that is known to be involved in RNA binding in Pfg25/27(refs 17,18). Thus, the SAM-like domain in the CTD of PTEF ismost probably capable of binding to RNA.

Phylogenetic analysis revealed that the SAM-like domains canonly be found in primate Plasmodium spp. and in P. gallinaceum.Unlike most other Plasmodium spp. that harbour only oneSAM-like domain in the genome, two copies of the SAM-likedomains can be found in P. falciparum and P. reichenowi of theLaverania clade as well as in P. gallinaceum (Supplementary

b

Ring

Trophozoite

Schizont

DAPI Merge

a

PTEF

368

SAM-like domain(1008–1183)

NTD (1–368) LD (369–756) CTD (757–1192)

HS (1–20) NLS (254–286) NES (1073–1080)20 756 1,192

1. HS-GFP (1–15)

2. NLS-GFP (240–300)

3. NTD-GFP (1–368)

4. CTD-GFP (757–1192)

5. PTEF-GFP (1–1192)

6. GFP

GFP Hoechst BF Merge

GFP

GFP

GFP

GFP

GFP

GFP

DAPI Merge

Ring

Trophozoite

Schizont

α-PTEF(α-NTD)

α-PTEF(α-CTD)

Figure 3 | PTEF localizes to multiple cellular compartments. a, Direct live imaging on parasites transfected with GFP genes fused with HS (1), NLS (2),NTD (3), CTD (4), full-length PTEF (5) and GFP alone (6). The corresponding fusion strategy is shown on the left. Hoechst dye was used to stain the DNAof live parasites. BF, bright field. The upper and lower panels in (5) illustrate two distinct localization patterns consistently observed despite the same GFPfusion construct. Images are representative of three stable transfectant lines for (4), (5) and (6), two stable transfectant lines for (2) and (3), and one stabletransfectant line for (1). b, Immunofluorescence assay (IFA) using anti-NTD serum (left) and anti-CTD IgG (right). DAPI was used to stain parasite DNA.Images are representative of five IFA replicates. Scale bars, 5 µm.






Fig. 3). Surprisingly, recent genome data suggest an expansion of thisdomain to 22 tandem copies in P. malariae19. However, only theLaverania clade harbours the unique expansion of the PTEF-NTDand -LD in one of the two SAM-like domain-containing proteins.We observed a possible association between this expansion and theemergence of the var gene family in the Laverania clade. In addition,some interesting sequence motifs are found along the protein(Supplementary Fig. 2b), although the presence of such motifs couldhave appeared by chance due to the high AT content of the genome.

Calpain cleaves the PTEF protein to generate the CTD fragment.The presence of both an NLS and an NES suggests that parts of PTEFmay be transported to different cellular compartments. Signal-transduced post-translational modifications or proteolytic cleavagecould resolve conflicting localization signals. Indeed,immunoblotting of lysates from both 3D7S8.4.2 and NF54CSA withan antibody generated against the CTD of PTEF showed minimaldetection of bands in 3D7S8.4.2, whereas two clusters of bands,with sizes of ∼185 and 55 kDa, were found in NF54CSA (Fig. 2b),suggesting that a cleavage event occurred that may have separatedthe NTD and CTD. To determine if such a cleavage event isrelevant in vivo, we treated cultures with a cocktail of cysteineprotease inhibitors (CPIs) and proteasome inhibitors (epoxomicin)for 4 h; this short incubation was intended to minimize effects oncell growth that could have confounded the results. Incubation withthe CPI cocktail drastically reduced the levels of cleavage products,as detected using antibodies against a short peptide on the NTD orthe CTD, whereas proteasome inhibition by epoxomicin resulted in

the accumulation of cleavage products (Fig. 2c). This resultsupports the notion that PTEF is cleaved in vivo. To bettercharacterize this event, we subsequently tested the effect using anarray of individual protease inhibitors specific to different classes ofproteases20. Although several of the cysteine protease inhibitorsaffected PTEF cleavage to some degree, only calpain inhibitor1 (ALLN) greatly reduced the cleavage products, suggestingPfCalpain-mediated processing of PTEF (Fig. 2d).

The CTD fragment of the PTEF protein localizes to thecytoplasm. To study the localization of the PTEF protein and itsdomains, we generated six green fluorescent protein (GFP) fusionconstructs using the pARL2-GFP vector and transfected them into3D7S8.4.2. These constructs are the HS-GFP (residues 1–15), theNLS-GFP with 15 residues flanking both ends (residues 240–300),the NTD-GFP (residues 1–368), the CTD-GFP (residues 757–1192)and the full-length PTEF-GFP (residues 1–1192). GFP alone wastransfected as a control. The HS-GFP displayed exclusivelocalization to the parasitophorous vacuole (PV)/parasite plasmamembrane (PPM), whereas the NLS-GFP construct faithfullyimported GFP into the nucleus (Fig. 3a). Interestingly, the NLSsignal was overridden when present in the whole NTD, resultingin diffuse localization of the NTD-GFP to the PV/PPM as well asto the erythrocyte cytoplasm and membrane (Fig. 3a). Bycontrast, CTD-GFP was only found to localize in the parasitecytoplasm. Again, the full-length PTEF-GFP displayed a diffusepattern in the entire parasite and host erythrocytes, with a smallfraction of IEs showing intense localization to the PV/PPM. We

a

37374373737444

−+−−−++−+

−−−−+−−−−

+−−+−−−−−

Temperature

RNaseA

Inhibitor (chem)

NF54CSABlank

3D7S8.4.2NF54CSA

kDa

1,236

242

146

20

*

**

†

PTEF-CTD-GFP

Temperature−+−

37374

+−−

Inhibitor (Ab)

Inhibitor (Ab) RNaseA

kDa

1,236

242

b

c

α-PTEF (α-CTD) Coomassie blue

GFP

InputInput

Flow th

rough

Final w

ash

Final w

ash

EluteElute

PTEF-CTD-GFP

α-GFP(27/80 kDa)

α-RPS14(16 kDa)

Figure 4 | CTD fragment of PTEF interacts with the ribosomes. a,b, Cytoplasmic fractions extracted from NF54CSA, 3D7S8.4.2 (a) and CTD-GFPtransfectant (b) were pretreated under different conditions and were resolved in blue native polyacrylamide gel and subsequently analysed byimmunoblotting using α-CTD IgG (left) and Coomassie blue staining (right). *Detection of high-molecular-weight complexes (>1.2 MDa). **An abundance ofproteins at 20 kDa after the addition of RNase. †Presence of RNase inhibiting heavy-chain antibodies. Representative of four and three biological replicates fora and b, respectively. c, Co-immunprecipitation using GFP magnetic beads was performed with cytoplasmic fractions extracted from GFP and CTD-GFPtransfected parasites. Successful immunoprecipitation of GFP proteins was confirmed (left) and RPS14 was detected using α-human RPS14 (right).Representative of three biological replicates.






reason that the NTD-containing full-length protein could give riseto the signal in the erythrocyte cytoplasm, whereas cleaved CTDfragments are retained inside the parasite (Fig. 3a).

Indirect immunofluorescence assays of NF54CSA parasites wereused to confirm the localization observed with the GFP constructs,using antibodies against the NTD and CTD of the protein. Duringring stages, CTD staining was localized to the cytoplasm of theparasite, whereas late-stage parasites displayed strong fluorescencesignals that covered both the parasitic and erythrocytic compart-ments in most IEs, resembling the pattern of full-length GFPfusion transfectants in the live imaging. However, heterogeneousstaining was observed for some late-stage IEs, with some showingpredominant localization within the parasite compartment,

whereas others exhibited preferential staining at the PV/PPM(Fig. 3b). Despite the lack of a PEXEL motif, the export of thePTEF protein was also observed when using antibodies againstthe NTD, although a lower erythrocytic-to-parasitic signal ratiowas observed (Fig. 3b). A region of hydrophobic residues haspreviously been shown to facilitate export in PEXEL-negativeexported proteins21, and our data suggest that the hydrophobicstretch of the PTEF protein is probably required, but not solelysufficient, to drive protein export.

We also performed a fractionation experiment in NF54CSA IEsusing a protocol described in ref. 22. We found that the cluster of55 kDa bands indicative of the CTD fragments of the PTEFprotein was preferentially, but not exclusively, partitioned into the

Untransfe

cted

NF54CSA

NF54CSA W

T

NF54CSA-

ptefKO

NF54CSA-ptef

KO

complemented

0

100

200

300

400

500

RLU

(bla

nk a

djus

ted)

P = 0.0228

V2uGU transient transfection

NF54CSA W

T

+ ALL

N

NF54CSA W

T

+ aprotinin

NF54CSA-ptef

KO

+ ALL

N

−20

0

20

40

60

80

100

Redu

ctio

n in

RLU

(rel

ativ

e to

unt

reat

ed c

ells

) (%

)

V2uGU stable transfection c d

NF54CSAWT

NF54CSA-ptefKO

NF54CSA-ptefKO

complemented

0

2

4

6

8

10

Rela

tive

copy

num

ber (

var2

csa/

cont

rol)

NS

P = 0.0544, NS

var2csa cDNA levela

α-hsp70

α-PTEF(α-CTD)

mAb PAM1.4(α-VAR2CSA)

kDa NF54CSA

NF54CSA-ptef

KO complemented

NF54CSA-ptef

KO

NF54CSA

NF54CSA-ptef

KO complemented

NF54CSA-ptef

KO

185

50

b

Figure 5 | PTEF is required for efficient VAR2CSA translation. a, qPCR analysis of var2csa expression in NF54CSAWT, ptefKO and ptefKO complemented,with expression relative to fructose biphosphate aldolase shown. NF54CSAWT (n = 4) and ptefKO and ptefKO complemented (n= 3) biological replicates. Errorbars indicate s.e.m. NS, not significant, Student’s t-test. b, Total lysates from NF54CSA, NF54CSA-ptefKO and NF54CSA-ptefKO complemented were resolved inSDS–PAGE and subsequently analysed via immunoblotting with antibodies to VAR2CSA (PAM1.4) (left) and to CTD-PTEF (right) illustrates successfulcomplementation; α-hsp70 blotting was performed as control. Representative of four biological replicates. c, A luciferase reporter assay with transienttransfection of the var2csa-promotor plasmid (V2uGU). Luciferase activity is presented as relative luciferase activity (RLU) subtracted from the blank value; n = 3technical replicates. Error bars represent s.d., Student’s t-test. This figure shows one representative experiment of three independent experiments showing asimilar trend. d, Luciferase activity of var2csa-promotor plasmid (V2uGU) in NF54CSAWT and ptefKO treated with protease inihbtors, 40 µMALLN or 300 nMaprotinin. Activity is presented as a percentage reduction in the corresponding untreated transfectants; n = 3 biological replicates. Error bars indicate s.e.m.






cytoplasmic fraction of the parasite. By contrast, the full-lengthprotein predominantly appeared in the nuclear SDS-soluble fraction(Supplementary Fig. 4). In addition, fractionation of NTD-GFP-transfected 3D7S8.4.2 confirmed that only the intact NTD-GFPfusion was exported (Supplementary Fig. 4). As the CTD fragmentof PTEF harbours an RNA-binding SAM-like domain, we examinedwhether the CTD fragment can bind to RNA in vivo. To corroboratethis hypothesis, we performed native gel electrophoresis with theparasite cytoplasmic fraction of NF54CSA IEs and found that theCTD fragment of the PTEF protein formed a complex that migratedto a position above 1.2 MDa, in addition to a distinct complex at∼240 kDa (Fig. 4a). These complexes were found to be absent inthe 3D7S8.4.2 parasite, which expresses a low level of the PTEFprotein. Interestingly, the complex above 1.2 MDa was sensitive toRNase treatment and could be stabilized with RNase-inhibitingantibodies, suggesting the presence of RNA components in thecomplex. Conversely, a duplicate Coomassie-stained gel showed ahigh proteinous content within the given migration range(Fig. 4a). This complex was again detectable when 3D7S8.4.2 wastransfected to expressCTD-GFP fusion proteins (Fig. 4b). One possibleabundant RNA and protein-containing complex with a highmolecularweight is the ribosomecomplex.To identifywhether ribosomes actuallyco-migrate with the CTD fragment of the PTEF protein, we excised gelbands at the correspondingmigration distance, followed by analysis viamass spectrometry (MS/MS). We found that up to 81% (26/32) and55% (28/51) of all ribosomal proteins in the 40S and 60S subunits,respectively, were detected in the 1.2 MDa region (SupplementaryFig. 5). By contrast, only one ribosomal protein was detected in thesamearea followingRNase treatment, butmost non-ribosomal proteins

remained (Supplementary Fig. 5 and Supplementary Table 3). As co-migration is insufficient to verify the physical interaction between theCTD fragment and the ribosomes, we performed a co-immunoprecipi-tation assay to validate their physical interaction.We chose an antibodyagainst human RPS14 that can cross-react with the Plasmodium hom-ologue (there are relatively low levels of similarity betweenP. falciparumand Homo sapiens ribosomal protein homologues, SupplementaryTable 4). Co-immunoprecipitation of the IE lysate of the CTD-GFP-transfected 3D7S.8.4.2 using a GFP antibody pulled down RPS14,whereas no RPS14 was pulled down by GFP proteins in the controltransfectant (Fig. 4c), providing direct evidence of interactionsbetween the CTD fragments and ribosomes.

PTEF promotes efficient translation of var2csa. To elucidate thefunctions of the PTEF protein, we generated NF54CSA ptef geneknockout (KO) mutant (NF54CSA-ptefKO), which completelyabolished PTEF expression, using the CRISPR method(Supplementary Fig. 6a,b)23. Consequently, the CSA-binding capacitywas also lost upon knocking ptef out (Supplementary Fig. 6d).Interestingly, we consistently discovered concomitant deletions ofpfemp3 and kahrp in the left arm of chromosome 2 upon theselection of KO parasites. Such deletion patterns were not evidentwhen negative selection pressure was not applied and whenepisomal maintenance of the plasmid was allowed (SupplementaryFig. 6c). These findings may suggest that the loss of PTEF proteincan cause chromosome instability. Because both pfemp3 and kahrpare major components of the knob structure and are well implicatedin the export and surface presentation of PfEMP1 protein to the redblood cell24,25, the loss of CSA-binding capacity could be an

0 10,000 20,000 30,000 40,0000

200

490 500 510

Wavelength (nm)

520 530 540

400

600

800

2,000

4,000

6,000

tGFP

fluo

resc

ence

(a.u

.)tG

FP fl

uore

scen

ce (a

.u.)

0

2,000

4,000

6,000

a b

c d

tGFP

fluo

resc

ence

(a.u

.)

Time (s)

0 10,000 20,000 30,000 40,000

Time (s)

No templateDNA template+ SAM-like domain+ CTD

No templateRNA template+ SAM-like domain+ CTD

DN

A +

SA

M-li

ke d

omai

n

No

tem

plat

e

DN

A te

mpl

ate

RNA

+ S

AM

-like

dom

ain

RNA

+ C

TD

Pixel ratio0 1 1.2 2 1 1.5 1.2

tGFP

DN

A +

CTD

RNA

tem

plat

e

tGFPtGFP + CTDtGFP + SAM-like domain

Figure 6 | tGFP synthesis in RTTF in the presence of PTEF domains. a,b, Real-time synthesis of active turboGFP (tGFP) in E. coli-based RTTF (reconstitutedtranscription-translation-folding) system starting from DNA (a) or mRNA (b) as the template, with or without the CTD and SAM-like domain of PTEF(0.1 µM). c, Autoradiograph of the gel with tGFP produced in RTTF supplemented with 35S-Met. The relative pixel quantification of the tGFP band is listedbelow each line. The results in a–c are representative of three individual experiments. d, Diagram showing the fluorescence spectrum of tGFP whenincubated with recombinant CTD and the SAM-like domain of PTEF. Each spectrum was recorded twice in two experimental replicates.






off-target effect unrelated to ptef deletion. We therefore complementedthe KO parasites with the CTD coding sequence of ptef using the pLNvector backbone selectable by blasticidin. Although the resultantcomplemented parasite (NF54CSA-ptefKO complemented) has aslightly higher multiplication rate than the KO parasites, knockoutor over-expression of ptef in NF54 or 3D7S8.4.2 parasites did notresult in any aberrant morphological phenotypes, suggesting thatPTEF proteins are non-essential in asexual-stage parasites under invitro culture conditions (Supplementary Fig. 6e). Immunoblottingusing a human monoclonal antibody generated from a PAMpatient demonstrated a reduced VAR2CSA protein level inNF54CSA-ptefKO than in NF54CSA parasites26. Importantly,CTD-ptef complementation restored VAR2CSA protein levels(Fig. 5b). var2csa transcription is active in all three parasites withcomparable levels of var2csa transcript detected (Fig. 5a), whichindicates that the reduced VAR2CSA protein level is independent ofthe transcription dynamics. Therefore, the CTD of PTEF cansufficiently enhance endogenous VAR2CSA protein levels.

We also quantitatively measured the effects of PTEF on var2csatranslation using a reporter assay. For this experiment, parasiteswere transfected with a construct (V2uGU) containing theGaussia luciferase reporter under the control of the 2.8 kb 5′untranslated region of the var2csa regulatory sequence, a yeast dihy-droorotate dehydrogenase (DHODH) gene allows the selection ofstable transfectants10, which will concomitantly activate the epi-somal var2csa promoter through a switch-like mechanism7.During transcient transfection, the NF54CSA-ptefKO parasiteshave markedly lower luciferase activities than the wild-type or ptefcomplemented parasites (Fig. 5c), suggesting that PTEF is essentialfor efficient derepression of the reporter gene when under theregulation of the uORF of the var2csa 5′ UTR.

As PfCalpain processing of the PTEF proteins generates CTDfragments that interact with the ribosomes, we sought to testwhether this cleavage process facilitates the translational regulationof PTEF. When stably transfected wild-type NF54CSA was incu-bated with a calpain inhibitor, we observed a >65% reduction inreporter activity. By contrast, neither calpain inhibition in theptefKO parasite nor aprotinin treatment of the wild-type parasiteaffected the reporter activity (Fig. 5d). These results indicate thatPfCalpain inhibition hampers PTEF-dependent translation of thereporter, and reinforces the hypothesis that efficient translation ofvar2csa is PTEF-dependent.

Because PTEF interacts with ribosomes and is able to enhancethe translation of var2csa, we wondered if the presence of PTEFmay identify a heterogeneous ribosome population with an‘enhanced feature’. Although the translation mechanism is substan-tially different between eukaryotes and prokaryotes, their ribosomesmaintain some highly conserved features. We therefore tried to testthe property of PTEF (CTD and the SAM-like domain) in anin vitro reconstituted-transcription-translation-folding (RTTF)system of Escherichia coli27. Synthesis of the reporter turbo-GFP(tGFP) proteins using both DNA and mRNA as the templateswas followed by an increase in their intrinsic fluorescence. Uponaddition of CTD or SAM to the RTTF reaction, a significantincrease in the production of the tGFP proteins was observed,with comparable enhancement with both the CTD and SAM-likedomain (Fig. 6a,b). The enhancement was similar when mRNAwas used as the template (Fig. 6b) compared to DNA (Fig. 6a), sup-porting our claim that the CTD and the SAM-like domain influencetranslation but not transcription. Enhancement in protein productionwas also visible from an increase in the band intensity of the reporterproteins in the autoradiographed gel when 35S-methionine wasadded to the RTTF reactions (Fig. 6c). These experiments demonstratethat theCTDandSAM-likedomains are general translationenhancers,even in a heterologous system. As a control experiment, we testedwhether the CTD or SAM of PTEF could influence the intrinsic

fluorescenceof the reporterproteins. For that, purified tGFPwas incu-bated with CTD and SAM, and tGFP fluorescence was recorded in aspectrofluorimeter (excitation 480 nm) before and after incubation.No increase in tGFP fluorescence was observed (Fig. 6d). Takentogether, these results imply that the CTD and SAM of PTEF aregeneral translation enhancers.

DiscussionIn this study, we have identified PTEF as a trans-acting factor requiredfor efficient translation of var2csa. CSA binding of IE via VAR2CSA tothe syncytiotrophoblasts in the placenta is a principal feature of PAM.The surface expression of VAR2CSA on IEs requires three successivesteps mediated by multiple molecular players, from switching tovar2csa transcription, translational derepression of var2csa andexport to the host cell surface. Although surface VAR2CSA expressionwas confoundedby the concomitant loss ofkahrp and pfemp3whenwegenerated the Ptef-knockout parasites, which are implicated in thesurface presentation of exported proteins24,25, our subsequent comple-mentation experiment clearly demonstrated that PTEF can regulateVAR2CSA translation. Importantly, a strict association between highptef expression and PAM isolates has been reported previously12–16.Similarly, higher ptef transcription has also been observed in CSA-panned parasites when compared to the isogenic non CSA bindingor CD36 binding parasites of genetically distinct backgrounds28,29.Our result, combined with these previous reports, strongly suggestsPTEF to be an important virulence factor of P. falciparum. Currentviews emphasize a selective advantage for parasites that have switchedto express var2csa following placental development. However, this failsto account for how the uORF repression is overcome. Importantly,var2csa transcription can also be detected in non-pregnant individ-uals30,31, and a switching hierarchy frequently drives the switching-on of its locus32. This is evolutionarily unsound, as it illicits anunnecessary immune response, although the binding niche is notavailable.Ourdata suggest amodel inwhich theuORFof var2csa effec-tively prevents ‘leaky expression’ ofVAR2CSA,while translational acti-vation of VAR2CSA is achieved through PTEF regulation(Supplementary Fig. 7). The degree of repression of the var2csauORF is mediated by length, but not sequence, dependent on boththe uORF and the inter-cistronic region (ICR), favouring re-initiationas the derepression mechanism8. Notably, PTEF also augments trans-lation in E. coli in vitro, suggesting that PTEF interacts with a highlyconserved component of the ribosome and interferes withtranslation at a conserved step. In in vitro eukaryotic translationsystems, efficient splitting of the post-termination complex canpromote translation re-initiation at downstream ORFs33. Increasedribosome recycling would also facilitate translation in vitro by increas-ing the availability of ribosomes. Therefore, PTEF may have a role indetermining post-termination dynamics during translation, an impor-tant mechanism to be explored further (Supplementary Fig. 8). Onaverage, P. falciparum transcribes longer 5′ UTRs than other eukary-otic organisms34,35, and ribosome profiling reveals frequent occupancyon the 5′ UTR; this may imply widespread translation initiationupstream for most coding regions36. Therefore, PTEF may regulate asubset of uORF-containing RNAs required for adaptation to the pla-cental niche. Drosophila DNER-MCT1 is so far the only trans factorknown to promote translation re-initiation37. DNER-MCT1 bindsdirectly to a subset of mRNAwith uORFs and promotes re-initation.Meanwhile, PTEF binds ribosomes and promotes derepression ofvar2csa. Mechanistically, this alludes to the idea of the existence ofsubsets of ‘specialized ribosomes’, where a heterogeneous ribosomepopulation may purposely adopt different features andproperties38,39. Notably, all Plasmodium ribosomal RNA (rRNA)gene copies are non-identical, and the expression of each rRNAcopy is stage-dependent40,41, suggesting the presence of heterogeneousribosome populations to be a highly plausible biological feature ofthe parasites. A similar property may also underlie the homologous






Pfg25/27 protein,whichhas been shown tobindRNA inanon-specificmanner and to predominantly pulldown rRNA in vivo17,18.

With few specific transcription factors identified, the transcriptionalcascade of P. falciparum is mainly dependent on developmentalcycle progression42,43. By contrast, the general delay between peaktranscript and protein expression44,45 and an over-representationof RNA-binding proteins in the genome highlight the importanceof post-transcriptional and translational regulation46. PTEF is thefirst non-canonical ribosome-binding protein that can promotetranslation in Plasmodium spp., exposing a new paradigm of generegulation. The need for the PfCalpain processing step in assertingPTEF function also raises potential therapeutic opportunities.

MethodsParasite culture, CSA-binding selection and transfection. P. falciparum clone3D7S8.4.2, FCR3S1.2 (non-rosetting), strain NF54CSA, FCR3 and IT4CD36/ICAM1were established and cultured using standard methods47. We routinely maintainedthe cultures at 3% haematocrit to support higher levels of parasitaemia growth.We used 5% sorbitol to regularly synchronize the cultures. Cultures were checkedquarterly for mycoplasma contamination using a MycoAlert Mycoplasma detectionkit (Lonza).

To enrich for CSA-binding IEs, a Petri dish (BD Falcon) was coated with100 µg ml−1 chondroitin 4-sulfate from bovine trachea (Sigma) overnight, followedby blocking with 3% BSA in PBS. Mid-to-late trophozoite IEs were resuspendedin complete medium and incubated in the coated dish for 30 min, and unboundIEs were washed away with RPMI until no binding was observed in theuncoated area.

The RBC preloading protocol was used to transfect the parasites48. Briefly, 400 µlof fresh RBCs were washed with cytomix and then mixed with 150 µg plasmid DNAand resuspended in cytomix to a total volume of 800 µl. This mixture wassubsequently transferred to a 0.2 cm cuvette (Bio-Rad). Electrophoration wasperformed using a Gene Pulser Electroporation System (Bio-Rad) with anexponential decay program set to 0.31 kV and 950 µF, and the resulting timeconstant ranged from 12 to 20 ms. The transfected RBCs were then inoculated withschizont-infected IEs. Either pyrimethamine (50 nM, Sigma), DSM1 (1.5 µM, MR4)or blasticidin (2.5 µg ml−1, Invivogen) was applied depending on the drug selectioncassette used 24 h post transfection. 5-Fluorocytosine (40 µM) was only used for thenegative selection of KO recombinants.

Genomic DNA extraction. A 200 µl aliquot of IEs (∼10% parasitaemia) was lysedwith 0.1% saponin in PBS. The supernatant was then removed, and DNA wasextracted from the pellet using a DNeasy Blood and Tissue kit (Qiagen).

Whole-genome sequencing and assembly. The 3D7S8.4.2 genome was sequencedwith a 454 Titanium shotgun sequencing system (Roche) using standard protocols.Reads of up to 25× genome coverage were recovered. We used a modified protocolfor the library and titration procedures49,50. All raw sequences were subjected toquality filtering procedures involving the trimming of low-quality nucleotides fromreads with Trimmomatic51 and mapping to the 3D7 reference genome usingBowtie52. Visualization was carried out using the Artemis/ACT software program53.

RNA extraction, cDNA synthesis and qPCR analysis. All parasites were tightlysynchronized for three consecutive cycles before total RNA extraction. Then, 1 ml ofthe Trizol reagent (Ambion) was used to lyse 100 µl of IEs, and following themanufacturer’s recommendations, a second round of phase separation wasperformed by mixing an acid phenol/chloroform solution with the aqueous phaserecovered from first separation at a 1:1 vol/vol ratio. cDNA was synthesized from500 ng of total RNA using iScript reverse transcriptase (Bio-Rad). qPCR wasperformed using iQ SYBR Green Supermix (Bio-Rad). The reaction was runfor 40 cycles of 95 °C for 10 s and at 60 °C for 1 min, and data were recordedduring the elongation step. The obtained threshold cycle values were usedto analyse the results via the 2−ΔCt or 2−ΔΔCt method. Fructose biphosphatealdolase was used for normalization. All primers used are listed inSupplementary Table 5.

RNA sequencing. Total RNA extracted at 10, 20, 30 and 40 h.p.i. from the differentparasite strains was analysed in a Bio-analyzer (Agilent) for quality and quantity.Then, 2 µg of the RNA was used to construct a sequence library using the TruSeqStranded mRNA Library Prep kit (Illumina). Indexed sequence libraries were pooledfor multiplexing. Sequencing was performed on an Illumina HiSeq 2000 platform toobtain 2× 100 bp paired-end reads. Reads were aligned to the correspondingreference genomes using Star and Htseq54. 3D7S8.4.2 and NF54CSAwere aligned tothe 3D7 genome, and FCR3S1.2 (non-rosetting), FCR3 and IT4CD36/ICAM1 werealigned to the IT4 genome obtained from plasmodb. We obtained 7.26–12.56 millionuniquely mapped reads from the samples. The results are presented either as readsmapped per million or as RPKM values.

Cloning. All clonings were performed either with restriction enzymes digestion andsubsequent ligation with T4 DNA ligase (NEB), or with infusion cloning (Clontech).All GFP constructs were obtained by cloning into pARL2-GFP using Xho1 andKpn1 sites. For the construct to generate the ptef KO parasite, the same strategy wasused as described in ref. 23. To generate the complementation parasite, the codingsequence of the C-terminal domain was cloned into pLN-Ty1-FKBP vector usingAvrII and AflII sites. All primers are listed in Supplementary Table 5.

Recombinant protein synthesis and antibody production. The sequences of theCTD (residues 757–1192) and SAM-like domain (1008–1192) of PTEF were codonoptimized and cloned into the pJ414 vector (DNA2.0), with a 6× His tag added tothe C terminus, for E. coli expression. Bacterial cultures harbouring the plasmid weregrown overnight and then inoculated into fresh Luria-Bertani medium. Freshlyinoculated cultures were allowed to reach an optical density at 600 nm of 0.6. Thecultures were subsequently induced with 0.25 mM isopropyl-β-D-thiogalactoside for3 h at 37 °C, followed by pelleting and lysing the cells via sonication in lysis buffer(10 mM HEPES pH 7.5, 10% glycerol, 150 mM NaCl, and a complete proteaseinhibitor cocktail, Roche). The lysate was incubated in Talon resins (Clontech) for3 h at 4 °C and then washed with 500 mM NaCl and 20 mM sodium phosphatebuffer until the absorbance at 260 nm spectrophotometer reading reached a value ofzero. The recombinant proteins were then eluted with increasing concentrations ofimidazole, up to 150 mM. The protein solutions were dialysed for the removalof imidazole.

All immunization procedures were performed by Agrisera AB (www.agrisera.se).The recombinant CTD protein and a synthetic peptide from the NTD (residues55–75, STINTPKTQSQENKDINKETK) were used to immunize six rats usingFreund’s incomplete adjuvants (Agrisera). The sera were collected after fourimmunizations. Sera from the CTD immunizations were purified with a protein Gcolumn to obtain total IgG.

Production of recombinant PAM1.4. VSAPAM-specific monoclonal IgG1antibody (PAM1.4) was generated as described elsewhere26,55. cDNA wassynthesized from selected B cell cultures and both heavy chain and light chainvariable regions (VH and VL) were sequenced as previously described56. Antibodyheavy and light chains were cloned into human IgG1 and Igκ expression vectors56,and expressed by transient transfection of Expi293F cells (ThermoFisher Scientific)using polyethylenimine. Cell lines were routinely tested for mycoplasmacontamination. The antibody was affinity purified by protein A chromatography(GE Healthcare).

Immunofluorescence microscopy. For indirect immunofluorescence assay, IEs atring or trophozoite stages were attached to glass slides pretreated with a poly-L-lysinesolution (Sigma). The attached IEs were then treated with 4% paraformaldehyde inPBS. The cells were permeabilized with PBS containing 0.25% Triton X-100 for10 min and were subsequently washed with PBS three times. The slides were blockedwith 10% goat serum overnight at 4 °C and then incubated with α-CTD IgG(50 µg ml−1) or α-NTD immune serum (1:400) diluted in PBST for 1 h at roomtemperature. Finally, a 30 min incubation procedure with the correspondingsecondary antibodies was performed using goat α-rat IgG antibodies conjugatedwith Alexa 488 (1:200).

For live imaging, transfected parasite cultures were first incubated with Hoechst33342, trihydochloride trihydrate (10 µg ml−1, Invitrogen), for 30 min. Allfluorescence images were captured using a Nikon Eclipse 80i microscope.

SDS–PAGE and immunoblotting. IEs (>1 × 107) were lysed using 0.1% saponin inPBS and were then washed three times with PBS. When necessary, crosslinking wasperformed by 10 min incubation in 1% formaldehyde in PBS. The resulting parasitepellet was solubilized in NuPAGE LDS loading buffer (Invitrogen) with theNuPAGE sample-reducing agent (Invitrogen) at a 1:15 vol/vol ratio. The extractedproteins were resolved on NuPAGE Novex 4–12% Bis-Tris gels with MOPS runningbuffer (Invitrogen). PageRuler Plus prestained protein ladder (Fermentas) wasalways used as the molecular weight marker. The resolved proteins on the gel weresubsequently transferred to a nitrocellulose membrane in Tris-glycine transfer buffer(25 mM Tris, 192 mM glycine, 20% methanol, 0.025% SDS), and the membrane wasblocked in 1% western blocking reagent (Roche) in TBS overnight at 4 °C. α-CTD(10 µg ml−1), α-NTD (1:400), mouse α-GFP (1:500, Roche #11814460001),recombinant PAM1.4 (10 µg ml−1) and rabbit α-PfHsp70 (1:2,000, BioSiteSPC-186C/D) were diluted in 0.5% blocking buffer in TBS, followed by incubationfor 1 h at room temperature. The membrane was then washed with TBST three timesand incubated with the corresponding horseradish peroxidase (HRP)-conjugatedsecondary antibodies (1:5,000, GE Healthcare) for 45 min at room temperature.ECL prime western blotting detection reagent was added to the membrane,followed by development using ECL hyperfilm (GE Healthcare).

Cellular fractionation. The cellular fractionation protocol was adopted andmodified from a previous report22. Briefly, pelleted IEs were treated with 0.1%saponin in PBS. The parasite pellet was subsequently extracted with different buffersin the following order: cytoplasmic lysis buffer (20 mM HEPES, 10 mM KCl, 1 mMEDTA, 1 mM EGTA, 0.65% NP-40 and 1 mM dithiothreitol (DTT)) for 10 min on




http://www.agrisera.se



ice; low-salt buffer (20 mM HEPES, 100 mM KCl, 1 mM EDTA, 1 mM EGTA and1 mM DTT) for 20 min at 4 °C; DNase digestion buffer (20 mM Tris-HCl, 15 mMNaCl, 60 mMKCl, 1 mMCaCl2, 5 mMMgCl2, 5 mMMnCl2, 300 mM sucrose, 0.4%NP-40, 1 mMDTT and 100 U ml−1 DNase1) for 20 min at 37 °C; high-salt buffer (20mMHEPES, 1 MKCl, 1 mMEDTA, 1 mMEGTA and 1 mMDTT) for 20 min at 4 °C;and SDS extraction buffer (2% SDS and 10 mM Tris-HCl) for 20 min at roomtemperature. A complete protease inhibitor cocktail (Roche) was added to all thebuffers. All of the extraction buffers were adjusted to 10 times the volume of the pellet,and the pellet was washed three times with a 500 µl volume of the same buffer afterextraction. The extracted fractions were then run on SDS–PAGE gels, and theimmunoblotting procedure described above was performed.

Blue native gel electrophoresis. IEs (>1 × 108) were treated with 0.1% saponin andthen lysed with 10 volumes of cytoplasmic lysis buffer for 10 min on ice. Thesupernatants were then divided evenly and treated with 0.2 units µl−1 of Stop RNaseInhibitor RX (5 PRIME) or 2 µg µl−1 of RNaseA (Sigma) and subsequentlyincubated for 30 min at 37 °C. Next, the samples were diluted in NativePAGEsample buffer and loaded onto a NativePAGE Novex Bis-Tris gel (Invitrogen).Electrophoresis was run using separated anode buffer (50 mM Bis-Tris and pH 7.0)and cathode buffer (50 mM Tricine, 15 mM Bis-Tris and 0.02% Coomassie blueG250). After electrophoresis, the gel was soaked for 30 min in transfer buffer with1% β-mercaptoethanol. The gel was subsequently immunoblotted using a PVDFmembrane. When necessary, staining was performed on some of the gels using theSilverQuest staining kit (Invitrogen) or 0.1% Coomassie Blue R250 in 10% aceticacid and 50% methanol.

MS analysis. Gel bands were excised from silver-stained denaturing gels and thencut into small pieces, which were shrunk in acetonitrile. The samples were reducedby incubating the gel pieces for 30 min at 56 °C in 100 mM ammonium bicarbonatecontaining 10 mM DTT. The samples were shrunk again, and the supernatant wasaspirated. The samples were then alkylated by incubating in the dark for 20 min atroom temperature in 100 mM ammonium bicarbonate containing 55 mMiodoacetamide. Again, the gel pieces were shrunk and the supernatant aspirated.The gel pieces were destained by 30 min incubation in a 1:1 mixture of 100 mMammonium bicarbonate:acetonitrile, followed by shrinking and aspiration of thesupernatant. The gel pieces were subsequently incubated on ice for 2 h with13 ng µl−1 trypsin (sequencing grade modified, Pierce) in a buffer containing10 mM ammonium bicarbonate and 10% acetonitrile. Digestion was carried outovernight at 37 °C. On the following day, peptides were extracted by adding 2volumes of 1.67% formic acid in 67% acetonitrile with 15 min of incubation at 37 °C.The extracts were dried in a SpeedVac and resuspended in 3% acetonitrile and 0.1%formic acid. Online liquid chromatography–mass spectrometry was carried outusing a hybrid LTQ-Orbitrap Velos mass spectrometer (Thermo Scientific), andsamples were injected by an autosampler (HPLC 1200 system, AgilentTechnologies). The samples were trapped on a C18 guard desalting column(Agilent) and then separated on a 15-cm-long C18 picofrit column (100 µm internaldiameter, 5 µm bead size, Nikkyo Technos) installed at the nano-electrosprayionization source. Solvent A was 97% water, 3% acetonitrile and 0.1% formic acid,and solvent B was 5% water, 95% acetonitrile and 0.1% formic acid. At a constantflow rate of 0.4 µl min−1, the curved gradient was increased from 2%B to 40%B over45 min, followed by a steep increase to 100%B over 5 min. The survey scan wasperformed in the Orbitrap at a resolution of 30,000 (and mass range of300–2,000 m/z), followed by data-dependent MS/MS (centroid mode) over twostages: (1) the top five ions from the master scan were selected for collision-induceddissociation (CID) using 35% normalized collision energy via ion trap massspectrometry (ITMS) detection and (2) the same five ions were subjected to higher-energy collision dissociation (HCD) using 32.5% normalized collision energy withdetection in the Orbitrap (FTMS). The automatic gain control (AGC) targets were1 × 106 ions for MS, whereas for MS/MS, values of 3 × 104 (CID) and 5 × 104 (HCD)were used. The Proteome Discoverer v. 1.4, including the Sequest software program,was used to search the P. falciparum 3D7 UniProt database for protein identification,limited to a false discovery rate of 1%.

Co-immunoprecipitation. 3D7S8.4.2 transfected with GFP or CTD-GFP constructswas used. IEs (>1 × 108) were treated with 0.1% saponin and then lysed with10 volumes of immunoprecipitation lysis buffer (50 mM Tris-HCl, 150 mM NaCl,1% NP40, 0.5% deoxycholate, protease inhibitor), with incubation on ice for 20 minto obtain soluble cytoplasmic proteins. The supernatant was then recovered andincubated with GFP-trap magnetic agarose beads (Chromotek). The subsequentprocedures followed the manufacturer’s instructions. The various fractions collectedwere resolved via SDS–PAGE and analysed via immunoblotting using rabbitα-human RPS14 (1:1,000, Abcam #ab174661).

Luciferase reporter assay. The V2uGU plasmid was a kind gift from K. Deitsch.For transient transfection, transfected ring-stage IEs were collected 72 h aftertransfection; a 100 µl volume of packed IEs was treated with 0.1% saponin and theresultant pellets were processed using the Gaussia Luciferase Flash Assay Kit(Pierce), and the difference in parasitaemia was adjusted by diluting the lysates.For stable transfection, NF54CSAWT and ptefKO were transfected and stable

transfectants were selected usingDSM1. A 1 ml aliquot of 1%parasitaemia cultureswasfirst treated with 0.1% saponin. The resultant parasite pellet was then processed usingthe Gaussia Luciferase Flash Assay Kit (Pierce) according to the manufacturer’srecommendations. FLUOstar Omega (BMG Labtech) was used to detect the luciferaseactivity level as colorimetric emission, with a 2 s delay and 15 s of integration. The blanksubtracted reading was represented as relative luciferase units (RLUs).

Synthesis of tGFP in the RTTF system. The RTTF system is composed of highlypure and active translation components from E. coli, which include 70S ribosomes(1 µM), all translation factors (1–10 µM), fMet-tRNAfMet, bulk tRNA (100 µM),20 different amino acids and 20 amino acyl tRNA synthetases. It also includesT7 RNA polymerase for efficient transcription, and an optimized energyregenerating system containing phosphoenol pyruvate (PEP), creatine phosphate,rNTPs, mayokinase, pyruvate kinase and creatine phosphokinase. The reactionswere performed in HEPES polymix buffer (pH 7.5) at 37 °C (ref. 57). To test theeffect of the CTD and SAM domain of PTEF on protein synthesis, 0.1 µM of eachwere added to the RTTF reaction. tGFP proteins were synthesized in RTTF usingplasmid DNA (cloned in pET24b) and in vitro transcribed mRNA58 as the template.The real-time synthesis of these proteins was followed in a TECAN Infinite 200 PROmultimode plate reader. For quantification of the proteins produced, the RTTFreaction was run with 35S-Met supplement. The reaction products were run directlyinto a 12% SDS–PAGE gel after boiling at 95 °C for 5 min. The gel was dried at 80 °Cfor 2 h on Whatman filter paper in a gel dryer and exposed overnight. Radioactivelylabelled protein was visualized using a phosphorimager (Bio-Rad personalmolecular imager) and quantified by Bio-Rad image lab software. The protein bandin the reaction without recombinant protein domains was taken as reference forthe quantification.

Statistical analysis. All experiments were performed two to five times. Data areplotted as mean values with variation calculated either as standard error of means,standard deviation or minimum to maximum values. Two-tailed Student’s t-test orPearson correlation were calculated using GraphPad Prism software. P < 0.05 wasconsidered statistically significant.

Data availability. All data that support the findings of this study are available fromthe corresponding author upon reasonable request. Raw sequencing readsfor 3D7S8.4.2 whole-genome sequencing (accession no. PRJNA377901), 3D7S8.4.2RNA sequencing (accession no. PRJNA377896) and NF54CSA RNA sequencing(accession no. PRJNA374979) are accessible from the Sequence Read Archive.

Received 11 January 2017; accepted 29 March 2017;published 8 May 2017

References1. Rao, Y. S. et al. Mutation bias is the driving force of codon usage in the

Gallus gallus genome. DNA Res. 18, 499–512 (2011).2. Rask, T. S., Hansen, D. A., Theander, T. G., Pedersen, A. G. & Lavstsen, T.

Plasmodium falciparum erythrocyte membrane protein 1 diversity in sevengenomes—divide and conquer. PLoS Comput. Biol. 6, e1000933 (2010).

3. Chen, Q. et al. Developmental selection of var gene expression inPlasmodium falciparum. Nature 394, 392–395 (1998).

4. Scherf, A. et al. Antigenic variation in malaria: in situ switching, relaxed andmutually exclusive transcription of var genes during intra-erythrocyticdevelopment in Plasmodium falciparum. Embo J. 17, 5418–5426 (1998).

5. Salanti, A. et al. Selective upregulation of a single distinctly structured var gene inchondroitin sulphate A-adhering Plasmodium falciparum involved inpregnancy-associated malaria. Mol. Microbiol. 49, 179–191 (2003).

6. Salanti, A. et al. Evidence for the involvement of VAR2CSA in pregnancy-associated malaria. J. Exp. Med. 200, 1197–1203 (2004).

7. Dzikowski, R. et al. Mechanisms underlying mutually exclusive expression ofvirulence genes by malaria parasites. EMBO Rep. 8, 959–965 (2007).

8. Amulic, B., Salanti, A., Lavstsen, T., Nielsen, M. A. & Deitsch, K. W.An upstream open reading frame controls translation of var2csa, a geneimplicated in placental malaria. PLoS Pathog. 5, e1000256 (2009).

9. Mok, B. W. et al. Default pathway of var2csa switching and translationalrepression in Plasmodium falciparum. PLoS One 3, e1982 (2008).

10. Bancells, C. & Deitsch, K. W. A molecular switch in the efficiency of translationreinitiation controls expression of var2csa, a gene implicated in pregnancy-associated malaria. Mol. Microbiol. 90, 472–488 (2013).

11. Kemp, D. J. et al. A chromosome-9 deletion in Plasmodium falciparum results inloss of cytoadherence. Mem. Inst. Oswaldo Cruz. 87, 85–89 (1992).

12. Francis, S. E. et al. Six genes are preferentially transcribed by the circulating andsequestered forms of Plasmodium falciparum parasites that infect pregnantwomen. Infect. Immun. 75, 4838–4850 (2007).

13. Fried, M. et al. The distinct proteome of placental malaria parasites.Mol. Biochem. Parasitol. 155, 57–65 (2007).

14. Ndam, N. T. et al. Plasmodium falciparum transcriptome analysis revealspregnancy malaria associated gene expression. PLoS One 3, e1855 (2008).




https://www.ncbi.nlm.nih.gov/bioproject/PRJNA377901/





15. Vignali, M. et al. NSR-seq transcriptional profiling enables identification of agene signature of Plasmodium falciparum parasites infecting children. Eur. J.Clin. Invest. 121, 1119–1129 (2011).

16. Bertin, G. I. et al. Differential protein expression profiles between Plasmodiumfalciparum parasites isolated from subjects presenting with pregnancy-associatedmalaria and uncomplicated malaria in Benin. J. Infect. Dis. 208, 1987–1997 (2013).

17. Camarda, G. et al. Regulated oligomerisation and molecular interactions of theearly gametocyte protein Pfg27 in Plasmodium falciparum sexual differentiation.Int. J. Parasitol. 40, 663–673 (2010).

18. Sharma, A., Sharma, I., Kogkasuriyachai, D. & Kumar, N. Structure of agametocyte protein essential for sexual development in Plasmodium falciparum.Nature Struct. Biol. 10, 197–203 (2003).

19. Rutledge, G. G. et al. Plasmodium malariae and P. ovale genomes provideinsights into malaria parasite evolution. Nature 542, 101–104 (2017).

20. Coppi, A., Pinzon-Ortiz, C., Hutter, C. & Sinnis, P. The Plasmodiumcircumsporozoite protein is proteolytically processed during cell invasion. J. Exp.Med. 201, 27–33 (2005).

21. Heiber, A. et al. Identification of new PNEPs indicates a substantial non-PEXELexportome and underpins common features in Plasmodium falciparum proteinexport. PLoS Pathog. 9, e1003546 (2013).

22. Oehring, S. C. et al. Organellar proteomics reveals hundreds of novel nuclearproteins in the malaria parasite Plasmodium falciparum. Genome Biol. 13,R108 (2012).

23. Ghorbal, M. et al. Genome editing in the human malaria parasitePlasmodium falciparum using the CRISPR-Cas9 system. Nature Biotechnol.32, 819–821 (2014).

24. Rug, M., Prescott, S. W., Fernandez, K. M., Cooke, B. M. & Cowman, A. F.The role of KAHRP domains in knob formation and cytoadherence ofP falciparum-infected human erythrocytes. Blood 108, 370–378 (2006).

25. Waterkeyn, J. G. et al. Targeted mutagenesis of Plasmodium falciparumerythrocyte membrane protein 3 (PfEMP3) disrupts cytoadherence ofmalaria-infected red blood cells. EMBO J. 19, 2813–2823 (2000).

26. Soerli, J. et al. Human monoclonal IgG selection of Plasmodium falciparumfor the expression of placental malaria-specific variant surface antigens.Parasite Immunol. 31, 341–346 (2009).

27. Mandava, C. S. et al. Bacterial ribosome requires multiple L12 dimers forefficient initiation and elongation of protein synthesis involving IF2 and EF-G.Nucleic Acids Res. 40, 2054–2064 (2012).

28. Ralph, S. A. et al. Transcriptome analysis of antigenic variation in Plasmodiumfalciparum—var silencing is not dependent on antisense RNA. Genome Biol. 6,R93 (2005).

29. Goel, S. et al. Targeted disruption of a ring-infected erythrocyte surface antigen(RESA)-like export protein gene in Plasmodium falciparum confers stablechondroitin 4-sulfate cytoadherence capacity. J. Biol. Chem. 289,34408–34421 (2014).

30. Duffy, M. F. et al. Transcribed var genes associated with placental malaria inMalawian women. Infect. Immun. 74, 4875–4883 (2006).

31. Tuikue Ndam, N. G. et al. High level of var2csa transcription by Plasmodiumfalciparum isolated from the placenta. J. Infect. Dis. 192, 331–335 (2005).

32. Ukaegbu, U. E. et al. A unique virulence gene occupies a principal position inimmune evasion by the malaria parasite Plasmodium falciparum. PLoS Genet.11, e1005234 (2015).

33. Skabkin, M. A., Skabkina, O. V., Hellen, C. U. T. & Pestova, T. V. Reinitiationand other unconventional posttermination events during eukaryotic translation.Mol. Cell. 51, 249–264 (2013).

34. Kumar, M., Srinivas, V. & Patankar, S. Upstream AUGs and upstream ORFscan regulate the downstream ORF in Plasmodium falciparum. Malar. J. 14,512 (2015).

35. Watanabe, J., Sasaki, M., Suzuki, Y. & Sugano, S. Analysis of transcriptomes ofhuman malaria parasite Plasmodium falciparum using full-length enrichedlibrary: identification of novel genes and diverse transcription start sites ofmessenger RNAs. Gene 291, 105–113 (2002).

36. Caro, F., Ahyong, V., Betegon, M. & DeRisi, J. L. Genome-wide regulatorydynamics of translation in the Plasmodium falciparum asexual blood stages. Elife3, e04106 (2014).

37. Schleich, S. et al. DENR-MCT-1 promotes translation re-initiation downstreamof uORFs to control tissue growth. Nature 512, 208–212 (2014).

38. Xue, S. F. & Barna, M. Specialized ribosomes: a new frontier in generegulation and organismal biology. Nature Rev. Mol. Cell. Biol. 13,355–369 (2012).

39. Dinman, J. D. Pathways to specialized ribosomes: the Brussels lecture. J. Mol.Biol. 428, 2186–2194 (2016).

40. Velichutina, I. V., Rogers, M. J., McCutchan, T. F. & Liebman, S. W. ChimericrRNAs containing the GTPase centers of the developmentally regulatedribosomal rRNAs of Plasmodium falciparum are functionally distinct. RNA 4,594–602 (1998).

41. Li, J., McConkey, G. A., Rogers, M. J., Waters, A. P. & McCutchan, T. R.Plasmodium: the developmentally regulated ribosome. Exp. Parasitol. 78,437–441 (1994).

42. Ganesan, K. et al. A genetically hard-wired metabolic transcriptome inPlasmodium falciparum fails to mount protective responses to lethal antifolates.PLoS Pathog. 4, e1000214 (2008).

43. Bozdech, Z. et al. The transcriptome of the intraerythrocytic developmentalcycle of Plasmodium falciparum. PLoS Biol. 1, e5 (2003).

44. Florens, L. et al. A proteomic view of the Plasmodium falciparum life cycle.Nature 419, 520–526 (2002).

45. Lasonder, E. et al. Analysis of the Plasmodium falciparum proteome byhigh-accuracy mass spectrometry. Nature 419, 537–542 (2002).

46. Coulson, R. M. R., Hall, N. & Ouzounis, C. A. Comparative genomics oftranscriptional control in the human malaria parasite Plasmodium falciparum.Genome Res. 14, 1548–1554 (2004).

47. Trager, W. & Jensen, J. B. Human malaria parasites in continuous culture.Science 193, 673–675 (1976).

48. Deitsch, K., Driskill, C. &Wellems, T. Transformation of malaria parasites by thespontaneous uptake and expression of DNA from human erythrocytes. NucleicAcids Res. 29, 850–853 (2001).

49. Zheng, Z. et al. Titration-free 454 sequencing using Y adapters. Nature Protoc. 6,1367–1376 (2011).

50. Zheng, Z. et al. Titration-free massively parallel pyrosequencing using traceamounts of starting material. Nucleic Acids Res. 38, e137 (2010).

51. Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer forIllumina sequence data. Bioinformatics 30, 2114–2120 (2014).

52. Langmead, B., Trapnell, C., Pop, M. & Salzberg, S. L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol.10, R25 (2009).

53. Carver, T. et al. Artemis and ACT: viewing, annotating and comparingsequences stored in a relational database. Bioinformatics 24,2672–2676 (2008).

54. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29,15–21 (2013).

55. Traggiai, E. et al. An efficient method to make human monoclonal antibodiesfrom memory B cells: potent neutralization of SARS coronavirus. Nature Med.10, 871–875 (2004).

56. Tiller, T. et al. Efficient generation of monoclonal antibodies from single humanB cells by single cell RT–PCR and expression vector cloning. J. Immunol.Methods 329, 112–124 (2008).

57. Koripella, R. K. et al. A conserved histidine in switch-II of EF-G moderatesrelease of inorganic phosphate. Sci. Rep. 5, 12970 (2015).

58. Pang, Y. H. et al. The antiprion compound 6-aminophenanthridine inhibits theprotein folding activity of the ribosome by direct competition. J. Biol. Chem. 288,19081–19089 (2013).

AcknowledgementsThe authors thank the Scilifelab in Stockholm for RNA-seq and mass spectrometryexperiments, K.W. Deitsch for providing luciferase reporter plasmids and for guidance onour experimental design, BEI Resources and MR4 for supplying DSM1, and BILS forproviding support for the transcriptomic analysis. This study was supported by the SwedishResearch Council (VR/2012-2014/521-2011-3377 toM.W.), the Söderberg Foundation andSwedish Academy of Sciences (to M.W.), the Swedish Strategic Foundation (to M.W.), theEU EviMalar Network of Excellence (to M.W.), a Distinguished Professor Award fromKarolinska Institutet (to M.W.) and the Swedish Research Council (VR 2013-8778,2014-4423 and 2016-06264 to S.S.) and the Knut and Alice Wallenberg Foundation (KAW2011.0081, RiboCORE to S.S.). The funders had no role in the study design, data collectionand analysis, decision to publish, or preparation of the manuscript.

Author contributionsS.C. and M.W. conceived the study. S.C., A.F., J.H.C. and M.P.Q. performed allexperiments. C.S.M. and S.S. performed the RTTF assay. M.V. performed the MSexperiment. J.J.L.R. and M.G. assisted in CRISPR design. S.B. and A.L. generated therecombinant PAM1.4. N.J. and O.F. performed computational analysis. S.C., A.F., C.S.M.,S.S. and M.W. interpreted data and wrote the manuscript.

Additional informationSupplementary information is available for this paper.

Reprints and permissions information is available at www.nature.com/reprints.

Correspondence and requests for materials should be addressed to M.W.

How to cite this article: Chan, S. et al. Regulation of PfEMP1–VAR2CSA translation by aPlasmodium translation-enhancing factor. Nat. Microbiol. 2, 17068 (2017).

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Competing interestsM.W. is a co-founder and member of the board of directors of Modus Therapeutics AB, acompany developing drugs for severe malaria. All other authors declare no financialconflicts of interest.





http://www.nature.com/reprints



regulation of pfemp1-var2csa ... - uppsala university · cell and molecular biology, uppsala...

Documents