Article
Structure-Guided Identification of a Family of Dual
Receptor-Binding PfEMP1 that Is Associated with
Cerebral MalariaGraphical AbstractHighlightsd Structural basis for P. falciparum PfEMP1 binding to
endothelial receptor ICAM-1defined
d A sequence motif derived from structure predicts group A
PfEMP1 binding to ICAM-1
d These ICAM-1-binding PfEMP1s also all bind to endothelial
protein C receptor (EPCR)
d Expression of dual ICAM-1- and EPCR-binding PfEMP1 is
associated with cerebral malariaLennartz et al., 2017, Cell Host & Microbe 21, 403–414
March 8, 2017 ª 2017 The Author(s). Published by Elsevier Inc.
http://dx.doi.org/10.1016/j.chom.2017.02.009Authors
Frank Lennartz, Yvonne Adams,
Anja Bengtsson, ..., Lars Hviid,
Matthew K. Higgins, Anja T.R. Jensen
Correspondence
matthew.higgins@bioch.ox.ac.uk
(M.K.H.),
atrj@sund.ku.dk (A.T.R.J.)
In Brief
Plasmodium falciparum-infected
erythrocytes display PfEMP1 proteins
that bind various endothelial receptors,
including ICAM-1. Lennartz et al.
structurally characterize PfEMP1 binding
to ICAM-1, allowing them to identify a
PfEMP1 family that simultaneously binds
to both ICAM-1 and EPCR. Dual-binding
PfEMP1s display stronger endothelial
adhesion and are associated with
cerebral malaria.
Accession Numbers
KF984156
KJ866957
KJ866958
KJ866959
KM364031
KM364032
KM364033
KM364034
5MZA
Cell Host & Microbe
ArticleStructure-Guided Identification
of a Family of Dual Receptor-Binding PfEMP1
that Is Associated with Cerebral Malaria
Frank Lennartz,1,9 Yvonne Adams,2,3,9 Anja Bengtsson,2,3,9 Rebecca W. Olsen,2,3 Louise Turner,2,3 Nicaise T. Ndam,4,5
Gertrude Ecklu-Mensah,2,3,6 Azizath Moussiliou,5 Michael F. Ofori,6 Benoit Gamain,7 John P. Lusingu,8
Jens E.V. Petersen,2,3 Christian W. Wang,2,3 Sofia Nunes-Silva,7 Jakob S. Jespersen,2,3 Clinton K.Y. Lau,1
Thor G. Theander,2,3 Thomas Lavstsen,2,3 Lars Hviid,2,3 Matthew K. Higgins,1,10,* and Anja T.R. Jensen2,3,*
1Department of Biochemistry, University of Oxford, South Parks Road, OX1 3QU Oxford, UK
2Centre for Medical Parasitology, Department of Immunology and Microbiology (ISIM), Faculty of Health and Medical Sciences, University
of Copenhagen, 1165 Copenhagen, Denmark
3Department of Infectious Diseases, Copenhagen University Hospital (Rigshospitalet), 2100 Copenhagen, Denmark
4Faculté de Pharmacie, Institut de Recherche pour le Développement (IRD), COMUE Sorbonne Paris Cité, 75013 Paris, France
5Faculté des Sciences de la Santé (FSS), Université d’Aboméy Calavi, 01 BP 526 Cotonou, Benin
6Department of Immunology, Noguchi Memorial Institute for Medical Research, University of Ghana, Legon, Ghana
7UMR_S1134, Université Sorbonne Paris Cité, Université Paris Diderot, Inserm, INTS, Unité Biologie Intégrée du Globule Rouge, Laboratoire
d’Excellence GR-Ex, 75013 Paris, France
8National Institute for Medical Research, Tanga Centre, 11101 Dar es Salaam, Tanzania
9Co-first author
10Lead Contact
*Correspondence: matthew.higgins@bioch.ox.ac.uk (M.K.H.), atrj@sund.ku.dk (A.T.R.J.)
http://dx.doi.org/10.1016/j.chom.2017.02.009SUMMARY
Cerebral malaria is a deadly outcome of infection by
Plasmodium falciparum, occurring when parasite-in-
fected erythrocytes accumulate in the brain. These
erythrocytes display parasite proteins of the PfEMP1
family thatbindvariousendothelial receptors.Despite
the importance of cerebral malaria, a binding pheno-
type linked to its symptoms has not been identified.
Here, we used structural biology to determine how a
group of PfEMP1 proteins interacts with intercellular
adhesion molecule 1 (ICAM-1), allowing us to predict
binders from a specific sequence motif alone. Anal-
ysis of multiple Plasmodium falciparum genomes
showed that ICAM-1-binding PfEMP1s also interact
with endothelial protein C receptor (EPCR), allowing
infected erythrocytes to synergistically bind both
receptors. Expression of these PfEMP1s, predicted
to bind both ICAM-1 and EPCR, is associated with
increased risk of developing cerebral malaria. This
study therefore reveals an important PfEMP1-binding
phenotype that could be targeted as part of a strategy
to prevent cerebral malaria.
INTRODUCTION
Malaria affects hundreds of millions of people each year. While
most cases are not life threatening, a significant number of
infections by Plasmodium falciparum result in severe malaria,
manifested in one or more of three major syndromes: anemia,Cell Host & Microbe 21, 403–414, M
This is an open access article undrespiratory distress, and cerebral malaria (CM). These occur
as parasites infect and divide within human erythrocytes. The
infected erythrocytes adhere to blood vessels and tissue sur-
faces, allowing the parasite to avoid clearance by the spleen.
In addition, organ-specific sequestration can have major conse-
quences for development of specific malaria symptoms, partic-
ularly in CM, the most debilitating form of the disease. Here, in-
fected erythrocytes accumulate in the brain, occluding blood
flow, inducing inflammation, and leading to major neurological
complications (Hviid and Jensen, 2015). Even with antimalarial
treatment, mortality rates due to CM in children range between
15% and 20% (Dondorp et al., 2010; Seydel et al., 2015), and
survivors of CM often suffer from a wide variety of long-lasting
neurological damage, which can result in loss of motor function,
impairment in learning and language capability, or an increased
risk of epilepsy (Birbeck et al., 2010; Idro et al., 2005).
During blood stage infection, the parasite expressesmembers
of different variant surfaceprotein families.Of these,Plasmodium
falciparum erythrocyte membrane proteins 1 (PfEMP1) are best
understood. They are displayed on infected erythrocyte sur-
faces and tether the cells to various receptors. EachPlasmodium
falciparum genome contains 
60 PfEMP1-encoding var genes,
which can be classified according to their chromosomal context
into groups A–E. They have large ectodomains containing multi-
ple duffy binding-like (DBL) andcysteine-rich inter-domain region
(CIDR) domains that can interact with specific human endothelial
receptors. Sequence analysis allows the CIDR andDBL domains
to be divided into general subgroups associated with specific
binding phenotypes (Hviid and Jensen, 2015). In particular, sub-
classesofDBLbdomains found ingroupA,B, andCPfEMP1bind
ICAM-1 (Bengtsson et al., 2013; Howell et al., 2008; Janes et al.,
2011); CIDRa1 domains from group A PfEMP1 bind endothelial
protein C receptor (EPCR) (Lau et al., 2015; Turner et al., 2013);arch 8, 2017 ª 2017 The Author(s). Published by Elsevier Inc. 403
er the CC BY license (http://creativecommons.org/licenses/by/4.0/).
and group B and C PfEMP1 contain CIDRa2–6 domains, which
bind CD36 (Hsieh et al., 2016; Robinson et al., 2003).
With a number of receptors available to bind to PfEMP1, a
major goal has been to determine whether PfEMP1s with spe-
cific receptor-binding phenotypes are associated with cere-
bral and other forms of severe malaria. These studies mostly
focused on children, as adults from malaria endemic areas
have a lower risk of death from complications during infection.
Early studies showed that parasites that express group A
PfEMP1 or those that form rosettes are linked with severe ma-
laria (Doumbo et al., 2009; Jensen et al., 2004). The search
was further focused with the discovery that severe malaria is
associated with a subset of group A and B/A PfEMP1s (Avril
et al., 2012; Claessens et al., 2012; Lavstsen et al., 2012), which
were later shown to bind to EPCR (Turner et al., 2013). This
demonstrated a link between severe malaria and expression of
EPCR-binding PfEMP1 (Bernabeu et al., 2016; Jespersen
et al., 2016; Turner et al., 2013).
While these studies established associations between se-
vere malaria and particular groups of PfEMP1, no connections
have been identified to any specific individual severe malaria
syndrome. In particular, attempts to link CM to expression of
certain groups of PfEMP1 remain inconclusive. Some studies
have correlated cerebral disease with ICAM-1 binding (Ochola
et al., 2011; Silamut et al., 1999; Turner et al., 1994), but
others found no such link (Newbold et al., 1997; Rogerson
et al., 1999). Indeed, ICAM-1-binding DBLb domains occur
in B- and C-type PfEMP1s that are associated with uncompli-
cated malaria as well as A-type PfEMP1s associated with se-
vere disease, suggesting that ICAM-1 binding alone is not a
driver of CM (Bengtsson et al., 2013; Howell et al., 2008;
Janes et al., 2011).
A significant obstacle to associating specific adhesion pheno-
types with disease outcomes has been the inability to directly
predict, using sequence information, adhesion traits of PfEMP1
expressed in patients. While EPCR- and CD36-binding CIDR do-
mains can be predicted from their sequences (Hsieh et al., 2016;
Robinson et al., 2003; Turner et al., 2013), ICAM-1-binding do-
mains cannot. We therefore aimed to understand the molecular
basis for ICAM-1 binding by A-type PfEMP1, to define a motif al-
lowing the identification of ICAM-1-binding DBLb domains from
sequence alone, and to determine whether the expression of
these ICAM-1-binding domains is associated with the develop-
ment of CM.
RESULTS
The Structural Basis for ICAM-1 Binding by PfEMP1
In the absence of a structure of a DBLb domain bound to
ICAM-1, we purified complexes of a diverse set of DBLb do-
mains from group A PfEMP1 bound to the N-terminal two do-
mains of ICAM-1 (ICAM-1D1D2), increasing the likelihood of
obtaining well-diffracting crystals. Of these, a complex of the
PF11_0521 (PlasmoDB: PF3D7_1150400) DBLb3_D4 domain
and ICAM-1D1D2 formed crystals that diffracted to 2.8 Å resolu-
tion (Table S1). We used molecular replacement, with previous
structures of ICAM-1 domains and a model derived from the
varO DBLa domain as search models, to determine the struc-
ture (Figures 1A and 1B). We also used small-angle X-ray scat-404 Cell Host & Microbe 21, 403–414, March 8, 2017tering to confirm that the arrangement of the complex was the
same in solution as that observed in the crystals, and that it
further matched the arrangement of a complex of the DBLb
domain bound to full-length ICAM-1D1D5 (Figure S1).
The PF11_0521 DBLb3_D4 domain adopts the classical DBL
domain fold, consisting of a core of a helices, decorated by
numerous loops (Figures 1, S2A, and S2B). In contrast to previ-
ously characterized DBL domains, an extended a helix and a
glycine and proline-rich linker form a unique feature that pro-
trudes from the domain and generates part of the ICAM-1 inter-
action site (Figures 1, S2A, and S2B). Indeed, a chimeric protein
containing this region of an ICAM-1-binding domain, trans-
planted to a related, non-binding DBLb domain, produced an
ICAM-1-binding domain (Figure 2A), and antibodies targeting
this region (either purified from peptide-immunized rat or from
human plasma) prevented infected erythrocytes from binding
to ICAM-1 (Figures 2B and 2C).
The DBLb domain interacts with ICAM-1 through a complex,
elongated binding interface that contains three distinct subsites
with a total surface area of 1,144 Å2 (Figure 1C). Site 1 is formed
from a patch of hydrophobic residues along one side of the
extended a helix of DBLb that interacts with a corresponding
hydrophobic patch on the surface of ICAM-1 D1. Site 2 is formed
by the subsequent linker of the DBLb domain and forms
hydrogen bonds with both ICAM-1 D1 and D2 domains. Site 3
is formed from part of a long loop that protrudes from the
DBLb domain and also contacts both ICAM-1 D1 and D2. The
parts of this loop not in contact with ICAM-1 are not observed
in the electron density, indicating flexibility (Figures 1B and
S2C). This loop is indeed predicted to be intrinsically disordered
(Figure S2D), but becomes partially ordered when bound to
ICAM-1. Such a flexible region of a protein, which forms part
of, or is immediately adjacent to, its ligand-binding site, is an im-
mune evasion strategy used by surface proteins from other
pathogens (Kim et al., 2004; Kwong et al., 1998), but had not
been previously observed in any PfEMP1 domain bound to its
receptor.
A Sequence Motif Allows for the Prediction of ICAM-1
Binding A-type PfEMP1
We next aimed to use the crystal structure to identify key mo-
lecular determinants used by the DBLb domains for ICAM-1
binding, and to thereby identify a sequence motif that can be
used to predict these ICAM-1 binders. For this, we mutated
residues in the DBLb domain that either directly contact
ICAM-1 or are responsible for the unusual architecture of the
binding site (Figure 2D; Table S2). By testing the eight residues
that directly interact with ICAM-1 (Figure S3; Table S2), we
found that each of the three subsites contains a critical residue
that, when mutated, disrupts ICAM-1 binding (Figure 2E; Table
S3). In addition, certain structural residues, in particular gly-
cines and prolines, are essential for ICAM-1 binding as they
allow sharp backbone bends that correctly present binding
residues (Figures 2 and S3; Table S3). To ensure that the
observed effects were not due to disruption in the overall
fold of the DBL domain, all mutants that showed an effect on
ICAM-1 binding were tested by circular dichroism spectros-
copy and found to be indistinguishable from the wild-type
domain (Figure S4).
Figure 1. The Structural Basis for ICAM-1 Binding by Group A PfEMP1s
(A) Front and side views of the PF11_0521 DBLb_D4 domain (green) bound to ICAM-1D1D2 (D1, light blue; D2, dark blue). Dashed boxes highlight three sites that
make direct contact with ICAM-1.
(B) Back view of PF11_0521 DBLb_D4 domain (green) bound to ICAM-1D1D2. Dashed lines represent residues not visible in the electron density map with the
disordered regions near site 3 (orange) of the ICAM-1-binding site highlighted.
(C) Three distinct sites in the PF11_0521 DBLb_D4 domain directly interact with ICAM-1 with residues that mediate this interaction indicated.
See also Figures S1 and S2 and Tables S1 and S2.Based on this analysis, coupledwith sequence analysis of pre-
viously identified A-type ICAM-1 binding domains (Bengtsson
et al., 2013), we defined a sequence motif (I[V/L]x3N[E]GG[P/A]
xYx27GPPx3H) that contains the determinants for ICAM-1
binding by these group A PfEMP1s. This motif was used to
search sequence databases, identifying a total of 145 ICAM-1-
binding DBL domains (mostly A-type PfEMP1s with a few B/A-
type PfEMP1s) that contain the motif and are therefore predicted
to bind to ICAM-1 (Table S4). To test these predictions, we used
ELISA to analyze the interaction between ICAM-1 and 16 motif-
containing DBLb domains, selected to represent sequence di-
versity (Table S4). All 16 DBLb domains with the motif bound
ICAM-1, while 10 group A DBLb domains without the motif did
not (Figures 2F and 3A; Table S4). In sequence analysis of
DBLb domains (n = 55) with known ICAM-1-binding properties
and DBLb S3 sequences (n = 1,823) from whole-genome
studies, the ICAM-1-binding group A and B/A domains with
the motif clustered separately from other domains, including
group B and C ICAM-1-binding DBLb domains lacking the motif
(Figures 3 and S5). Taken together, this suggests at least twogroups of ICAM-1-binding PfEMP1s with different evolutionary
histories and that the majority of ICAM-1-binding group A and
B/A PfEMP1s can be predicted from the presence of a highly
conserved binding site (Figures 2G and 2H).
Parasites Expressing ICAM-1-Binding DBLb Domains
Elicit Inhibitory Antibodies and Are Associated with CM
The high sequence identity of DBLbdomains containing themotif
(Table S4) encouraged us to determine whether these domains
might induce cross-reactive antibodies during natural infections.
We therefore screened samples from Liberian adults for the
presence of immune plasma that inhibited DBLb domains from
binding to ICAM-1. Plasma IgG from a pool of these inhibitory
samples was then affinity purified separately on three different
DBLb domains containing the motif (Dd2var32, KM364031, and
PFD1235w [PlasmoDB: PF3D7_0425800] DBLb_D4) and tested
for the ability to inhibit ICAM-1 binding of DBLb domains
(Figure 4A). The IgG showed the potential to broadly inhibit
ICAM-1 binding by a panel of motif-containing DBLb domains
from group A PfEMP1, but did not prevent ICAM-1 binding byCell Host & Microbe 21, 403–414, March 8, 2017 405
Figure 2. Structure-Guided Identification of a Motif that Predicts ICAM-1 Binding
(A) ICAM-1 binding (ELISA OD 490 nm; ± SD) of recombinant 3D7 PFD1235wDBLb3_D4, PFD1235wDBLb3_D5, and a chimeric DBLb3_D5 containing the ICAM-
1-binding motif region of DBLb3_D4 (D5_motif).
(B and C) Inhibition of 3D7 PFD1235W infected erythrocyte adhering to ICAM-1 under flow in the presence of anti-ICAM-1 Abs, affinity-purified anti-PFD1235w
DBLb3_D4 (anti-D4) IgG, anti-PFD1235w_motif IgG, anti-ITVAR13, and control IgG. IgG antibodies affinity purified from (B) rat anti-serum and (C) human serum. ±
SD of a minimum of three independent experiments done in triplicate. Data were analyzed using one-way ANOVA. Asterisks indicate significance (p = 0.0004).
(D) The three sites within the PF11_0521 DBLb3_D4-binding site, indicating residues that directly contact ICAM-1 (site 1, yellow; site 2, red; site 3, orange).
Structural residues important for positioning of interacting residues are highlighted in teal.
(E) Surface plasmon resonance curves for injection of 2-fold dilution series of DBLb wild-type and binding site mutants over ICAM-1D1D5.
(F) ICAM-1 binding (ELISA OD 490 nm; ± SD three replicates) of 26 recombinant group A DBLb domains (4 DBLb1, 14 DBLb3, 2 DBLb6, 2 DBLb7, 2 DBLb11,
1 DBLb12, and 1 DBLb unknown sub-class), confirming prediction of binding domains. ‘‘DC’’ indicates in which domain cassette the domain is found. ‘‘ND’’
indicates that the DC type is unknown as only the DBLb sequence is available. ‘‘No’’ indicates that the domain is not part of a knownDC. Presence of DC13 prior to
DBLb is indicated.
(G andH) Sequence logo showing conservation of (G) residues that contact ICAM-1 and (H) residues important for the unusual architecture of the ICAM-1-binding
site, based on 145 DBLb domains predicted to bind ICAM-1. Red triangles, residues critical for direct interaction with ICAM-1.
See also Figures S3–S5 and Tables S2, S3, and S4.
406 Cell Host & Microbe 21, 403–414, March 8, 2017
Figure 3. Phylogeny of ICAM-1- and Non-binding DBLb Domains
(A)Maximum likelihood tree of 55 completeDBLbdomains. The tree is drawn to
scale, with branch lengths measured in substitutions per site. Red, ICAM-1
binders; blue, ICAM-1 non-binders (Avril et al., 2016; Bengtsson et al., 2013;
Janes et al., 2011; Lau et al., 2015). Filled circles, DBLb domains tested in this
study; open squares, DBLbdomains tested, but data not shown.UKN indicates
unknown group ID. Shaded areas indicate DBLb with the ICAM-1 motif.
(B) Maximum likelihood tree of 1,823 complete DBLb S3 sequences from the
seven published genomes and 226 annotated Sanger genomes. Filled circles,
DBLb experimentally shown as ICAM-1 binders. Open circles, DBLb experi-
mentally shown as ICAM-1 non-binders. Colors indicate the CIDR domain
N-terminal to DBLb: red, EPCR binders with ICAM-1-binding motif; orange,
EPCR binders with no ICAM-1-binding motif; green, non-EPCR binders
(group A); blue, CD36 binders; magenta, VAR1.
See also Figure S5 and Table S4.a DBLb domain from a group B PfEMP1 (IT4VAR13). In contrast,
IgG that was affinity purified on a closely related non-ICAM-1-
binding DBLb domain that lacks the motif (PFD1235w DBLb_D5)did not significantly inhibit binding of any DBLb domain tested
(Figure 4A). Therefore, plasma from adults in a malaria endemic
region contains IgG that shows the potential to broadly inhibit
ICAM-1 binding of motif-containing PfEMP1.
We also tested whether children develop antibodies that spe-
cifically target the ICAM-1 binding site of the motif-containing
DBLb domains. To this end, we screened plasma samples from
Plasmodium falciparum-exposed non-hospitalized Tanzanian
children living in an area of high malaria transmission against a
peptide containing the motif. A high proportion (69 of 76) of sam-
ples contained motif-reactive antibodies, and the DBLb:ICAM-1-
binding inhibition was significantly higher in plasma with high
reactivity (ELISA optical density [OD] > 1) against the motif
when compared with those with low motif reactivity (ELISA
OD < 1) (Figure 4B).
Next, to correlate the expression of the group A ICAM-1-bind-
ing PfEMP1with disease outcomes, we used the sequencemotif
to design primers to selectively amplify motif-encoding var gene
transcripts. We used these to assess transcript levels in children
with CM and severe malarial anemia (SA), and in other pediatric
malaria patients (other malaria, OM; non-CM and non-SA) (Fig-
ure 4C; Table S5). The abundance of var gene transcripts was
calculated relative to the transcript of an endogenous control
gene and was translated into transcript units (Lavstsen et al.,
2012). This analysis showed higher transcription of motif-encod-
ing, ICAM-1-binding PfEMP1 in children with CM than in children
with SA or in those admitted to hospital with malaria without
these signs of severity (OM). Due to the challenges in acquiring
a sufficiently large number of samples from single locations, pa-
tients were recruited at different sites and logistic regression
modeling was used to correct for this. Comparison of the three
groups showed that levels of transcripts encoding ICAM-1-
binding PfEMP1 were significantly associated with risk of CM
(p = 0.02) when comparing CM patients to malaria patients
without CM (SA and OM). In contrast, transcript levels reported
by primers targeting loci encoding EPCR-binding PfEMP1
(Mkumbaye et al., 2017) was not higher in CMpatients compared
to SA plus OM patients (p = 0.751). These data suggest a direct
link between CM and expression of this particular subset of
ICAM-1-binding PfEMP1s (Figure 4C).
Motif-Containing ICAM-1-Binding DBLb Domains Are
Associated with EPCR-Binding Domains, Allowing
Simultaneous Dual Receptor Binding
The correlation between CM and expression of group A PfEMP1
containing the conserved ICAM-1-binding site raised the ques-
tion of why ICAM-1 binding was not unambiguously associated
with CM in previous studies. With ICAM-1 binding alone unlikely
to be the driver of cerebral disease, we therefore examined the
broader context of the motif-containing ICAM-1-binding do-
mains. Wemade the striking observation that all tested PfEMP1s
containing these domains also contain a neighboring EPCR-
binding CIDRa1 domain (Table S4; Figures 3B and S5). These
PfEMP1s form a separate cluster from PfEMP1s that contain
an EPCR-binding domain, but no motif containing ICAM-1-
binding domain (Figures 3B and S5), and we find that 14% of all
EPCR binders have such a motif-containing ICAM-1-binding
domain. In contrast to this, B- and C-type ICAM-1-binding
PfEMP1s are frequently associated with CIDRa domains thatCell Host & Microbe 21, 403–414, March 8, 2017 407
Figure 4. The Conserved ICAM-1 Binding Site Is Recognized by Patient Plasma and Linked to CM
(A) The inhibition of DBLb domains binding to ICAM-1 by IgG from a plasma pool from Liberian P. falciparum-exposed adults purified on ICAM-1-binding
DBLb_D4 domain from Dd2var32, KM364031, and PFD1235w or on a closely related, but non-ICAM-1-binding DBLb_D5 domain from PFD1235w. ICAM-1
inhibitory capacity is >74% (black), 51%–74% (dark gray), 21%–50% (light gray), and 0%–20% (white).
(B) ICAM-1-binding inhibition by plasma with low (ELISA OD < 1) and high (ELISA OD > 1) anti-motif-DBLb IgG in P. falciparum-exposed Tanzanian children
(1–17 years) (Lusingu et al., 2004). Boxplot with median. Whiskers, 5% and 95% percentiles.
(C) Transcript levels (in arbitrary transcription units, Tu) of var gene subtypes in Ghanaian, Tanzanian, and Beninese children hospitalized with malaria, shownwith
medians and 25% and 75% percentiles.
See also Table S5.interact with CD36 and are distributed across the phylogenetic
tree (Figure 3B; Table S4). This suggests that the group A
ICAM-1-binding PfEMP1s are a subset of the ICAM-1-binding
PfEMP1s, separate from group B and C PfEMP1s, and that
they also have the capacity to bind to EPCR.
The link between the presence of both ICAM-1- and EPCR-
bindingdomainswithin a singlePfEMP1 raised the intriguingpos-
sibility that these PfEMP1s might simultaneously engage both
receptors. A few PfEMP1s have previously been identified that
contain bothEPCR-bindingCIDRadomains and ICAM-1-binding
DBLb domains (Avril et al., 2016; Oleinikov et al., 2009; Turner
et al., 2013). However, it was unknown whether they could
interact with both receptors simultaneously, and whether such
dual binding would have consequences for infected erythrocyte
adhesion. We therefore produced a set of three different re-
combinant PfEMP1s and used surface plasmon resonance to
test their potential for binding to the two receptors simulta-
neously. Indeed, we found that protein constructs containing
the ICAM-1- and EPCR-binding domains of these PfEMP1s all
bound simultaneously to ICAM-1 and EPCR with nanomolar
affinity (Figures 5 and S6; Table S6).
Next, we examined the effect of dual binding to ICAM-1 and
EPCR in the context of infected erythrocytes. For this, erythro-
cytes infected with parasites expressing PfEMP1 (PFD1235w
or HB3VAR03) predicted to bind both receptors were selected
in vitro (Figure S7) and binding to both receptors was tested
under flow conditions using recombinant receptors (Figures
6A–6C) and in an immunofluorescence assay (Figures 6G–6I).
These studies were performed at a range of flow rates from 0.5
to 5.0 dyn/cm2, chosen to induce the most physiologically rele-408 Cell Host & Microbe 21, 403–414, March 8, 2017vant shear stresses. Infected erythrocytes were capable of at-
taching between 0.5 and 1.0 dyn/cm2 in the presence of either
recombinant receptor alone. When ICAM-1 and EPCR were
simultaneously available at comparable molar concentrations,
erythrocytes infected with parasites expressing dual-binding
PfEMP1 demonstrated an increased ability to bind at a higher
shear stress, i.e., 2 dyn/cm2, compared to when either receptor
was available alone, though this adhesion did not exceed
that measured at 1 dyn/cm2 (Figures 6A and 6B; Table S7A). In
contrast, erythrocytes infected with parasites known to bind
ICAM-1 and CD36 (IT4VAR13; Figures S7E and S7F) showed
increased adhesion in the presence of both ICAM-1 and CD36
compared to CD36 binding alone at lower shear stress, but did
not confer the ability to withstand higher shear stresses, and
the adhesion decreased in response to increasing shear stress
(Figure 6C). The final parasite, known to bind to PECAM-1
(PF11_0008 [PlasmoDB: PF3D7_1100200]; Figures S7G and
S7H), failed to bind any of the receptors tested (Figure S7L).
To further test the physiological relevance of our observations,
we assessed the binding of infected erythrocytes to HBMEC
with and without tumor necrosis factor a (TNFa) treatment (Fig-
ures 6D–6F and S7I–S7K). To determine the contribution of indi-
vidual receptors to binding, antibodies against either ICAM-1
or EPCR alone or a combination of both were used. With
HB3VAR03, at 0.5 dyn/cm2 the greatest inhibition was observed
on cells treated with both anti-receptor antibodies simulta-
neously (80%, p = 0.0361; Figure 6D), while HBMEC treated
with either anti-ICAM-1 or anti-EPCR antibodies demonstrated
a non-significant reduction in adhesion (Table S7B). A similar ef-
fect was observed at 0.75 dyn/cm2.With increasing shear stress,
Figure 5. PfEMP1 Can Simultaneously Bind to Both ICAM-1
and EPCR
ICAM-1D1D5 was immobilized at fixed concentration, followed by injection of
the respective head structure, (A) PF11_0521 DBLa1.7-CIDRa1.4-DBLb3,
(B) Dd2var32 DBLa1.7-CIDRa1.4-DBLb3, and (C) PFD1235w CIDRa1.6-
DBLb3, also at fixed concentration. A 2-fold dilution series of EPCR was in-
jected over the resulting ICAM-1::head structure complex. Arrows show start
points of protein injection. The insets show data fitted to a one-site kinetic
model for binding of EPCR to the respective ICAM-1::head structure complex.
Data (black lines) are modeled to a one-site model (red lines). See also Fig-
ure S6 and Table S6.the activity of single-antibody treatment was enhanced, and at
1 dyn/cm2 this adhesion inhibition was significant for each indi-
vidual anti-receptor antibody (Figure 6D; Table S7B), showing
that the ability of HB3VAR03 to bind to both receptors is impor-
tant at higher shear stresses. In comparison, 3D7PFD1235w
infected erythrocytes were significantly inhibited by the pres-
ence of either antibody at all shear stresses tested (Figure 6E;
Table S7B). As expected, IT4VAR13 only bound ICAM-1 andnot EPCR expressed on the surface of HBMEC (Figure 6F).
When the same experiments were conducted on activated cells,
all three parasite lines showed increased reliance on ICAM-1 for
adhesion, with HB3VAR03- and 3D7PFD1235w-infected eryth-
rocytes switching to a single-receptor phenotype due to down-
regulation of EPCR expression upon cell activation by TNFa
(Menschikowski et al., 2009) (Figures S7I–S7K). In summary,
the ability to bind to both ICAM-1 and EPCR gives parasites
significantly greater versatility, allowing for an increased capac-
ity to adhere at physiological higher shear stresses, and enables
them to continue to adhere despite changes in receptor expres-
sion upon cell activation.
Association of ICAM-1 and EPCR Dual-Binding PfEMP1
with the Development of CM
Our ability to predict dual-binding PfEMP1 from sequence next
allowed us to assess the link between the presence of parasites
expressing such PfEMP1 and the development of different se-
vere malaria syndromes in patient isolates, and to compare
this with the disease outcomes associated with PfEMP1s that
bind EPCR alone. For this, we analyzed a dataset containing
the full sequences of the six most abundant PfEMP1 transcripts
expressed in 45 children hospitalized with malaria (on average
corresponding to 75% of total var transcripts; see Supplemental
Experimental Procedures), the biggest available dataset of its
kind (Jespersen et al., 2016). These children presented with
CM or SA, or were admitted to hospital with malaria, but without
these signs of severity, and were immediately treated with anti-
malarial drugs (Table S5). We identified dual-binding PfEMP1
from the sequence data by presence of a DBLb domain contain-
ing the group A ICAM-1-binding motif adjacent to an EPCR-
binding CIDRa1 domain. EPCR-binding PfEMP1s were identi-
fied by the presence of CIDRa1 domains alone. In this dataset,
parasites expressing PfEMP1 predicted to bind EPCR, but not
ICAM-1, were more likely to be found in severe malaria cases
(CM plus SA) (p = 0.040) (Figure 7B). However, there was no as-
sociation between EPCR binding and either of the specific se-
vere malaria syndromes (with p = 1.000 for comparison of
CM with all other types of malaria). In contrast, parasites ex-
pressing dual ICAM-1- and EPCR-binding PfEMP1 were specif-
ically (p = 0.016) associated with CM, when compared with
non-cerebral disease (SA plus OM) (Figure 7A). These data sup-
port our finding that the expression of ICAM-1-binding DBLb
domains containing the motif, which to date has been 100%
associated with dual ICAM-1- and EPCR-binding PfEMP1 (Fig-
ures 3 and S5; Table S4), is associated with the development
of cerebral symptoms (Figures 4C and 7). Therefore, while
expression of PfEMP1s that bind to EPCR alone is not linked
to any specific severe malaria syndrome, the presence of dual-
binding motif-containing PfEMP1s is associated with the devel-
opment of CM.
DISCUSSION
A major challenge in associating malaria disease phenotypes
to the expression of PfEMP1s with specific binding properties
has been the ability to predict ligand-binding capability from
sequencealone. In this study,weusedstructural studies todefine
a sequence motif that allowed the identification of a large set ofCell Host & Microbe 21, 403–414, March 8, 2017 409
Figure 6. Binding to ICAM-1 and EPCR Increases Adhesion of Infected Erythrocytes
(A–F) Erythrocytes infected with (A and D) HB3VAR03 (predicted ICAM-1 and EPCR binder), (B and E) 3D7PFD1235w (predicted ICAM-1 and EPCR binder), and
(C and F) IT4VAR13 (ICAM-1 and CD36 binder) binding to (A–C) receptor-coated microslides under flow conditions using recombinant ICAM-1, rEPCR, rICAM-1,
and rEPCR; rCD36; or rICAM-1 and rCD36, and to (D–F) resting human brain microvascular endothelial cells (HBMEC). To demonstrate specific adhesion,
channels coated with HBMEC were pre-incubated with anti-ICAM-1 (40 mg/mL), anti-EPCR (10 mg/mL), or anti-ICAM1 and EPCR combined (40 and 10 mg/mL,
respectively). Flow experiments were done in parallel with the same conditions for immobilizing proteins or cell-coated chips. Values are bound infected
erythrocyte per mm2 ± SEM. Data were also analyzed by two-way ANOVA with Tukey’s multiple comparison (Tables S7A and S7B).
(G–I) Immunofluorescence images of erythrocytes infected with parasite strains (G) HB3VAR03, (H) 3D7PFD1235w, and (I) IT4VAR13. Complexes of ICAM-1 or
EPCR were added to measure binding by immunofluorescence. Main panels show overlays of ICAM-1 (green), EPCR (red), and nuclear (blue) staining. Inserts
show single channels. Scale bar, 2 mm.
See also Figure S7 and Table S7.ICAM-1-binding groupAPfEMP1s fromsequence.We also show
that these ICAM-1-binding PfEMP1s all contain an additional
EPCR-binding CIDRa1 domain and that this allows binding to
both ligands. Finally, we show that the presence of parasites ex-
pressing PfEMP1s that bind to both ICAM-1 and EPCR is associ-
ated with increased risk of developing cerebral symptoms to
P. falciparum infections.410 Cell Host & Microbe 21, 403–414, March 8, 2017This link between dual binding and CM is supported by two in-
dependent studies (Figures 4C and 7). However, while signifi-
cant, it is not absolute. This may partly be because the data
show a snapshot of the parasites present in the peripheral blood
of the children at the time of hospital admission, and the children
involved were treated with antimalarial medication immediately
upon diagnosis, preventing further development of the disease.
Figure 7. Dual-Binding PfEMP1s Correlate with CM
Percentages of specific var gene transcripts in infected erythrocytes from 45
hospitalized Tanzanian children.
(A) Percentage of var gene transcripts encoding PfEMP1s predicted to bind
both ICAM-1 and EPCR.
(B) Percentage of transcripts that encode PfEMP1 predicted to bind EPCR and
not containing the ICAM-1-binding motif. CM, cerebral malaria; SA, severe
anemia; OM, children without the signs of severity found in the other groups.
See also Table S5.For example, in the analysis of whole PfEMP1 sequences (Fig-
ure 7), we cannot exclude the possibility that the small number
of patients (3/21) expressing dual-binding PfEMP1 and present-ing with SA, but not CM, may have developed cerebral symp-
toms if left untreated, causing the association detected here to
be an underestimate. There are also risks in overestimating the
link between CM and expression of this group of PfEMP1s due
to limitations in the sample size available for each of these
studies. However, this is mitigated by the fact that we observe
the same association between dual-binding PfEMP1s and CM
in two independent studies with samples from different patient
populations from three African countries (Figures 4C and 7),
suggesting that the presence of these PfEMP1s on the infected
erythrocyte surface is a risk factor for the development of CM.
There are several possible reasonswhy dual-binding PfEMP1s
might contribute to the development of CM. One consequence
of dual binding is increased endothelial adhesion under condi-
tions of high shear stress (Figure 6). CM is associated with
increased sequestered parasite biomass throughout the circula-
tory system, and this appears to have disproportionate effects on
the regulation of the homeostasis of the brain, making increased
adhesion a risk factor (Cunnington et al., 2013; Moxon et al.,
2009; Storm and Craig, 2014). In addition to these general ef-
fects, specific consequences for the brain may occur due to
modulation of signaling mediated by ICAM-1 and EPCR, as
EPCR-mediated signaling and ICAM-1 expression are linked
by a complex interplay that might also contribute to cerebral dis-
ease. Infected erythrocytes prevent EPCR from interacting with
its natural ligands, inhibiting signaling pathways that induce
anti-inflammatory responses and protect the integrity of the
endothelium (Petersen et al., 2015; Turner et al., 2013). Indeed,
inflammation and brain swelling both occur in CM and are impor-
tant in disease development. In addition, blockage of EPCR
signaling increases expression of several PfEMP1 adhesion re-
ceptors, including ICAM-1 (Moxon et al., 2013). Paradoxically,
parasite sequestration in the brain is also linked to loss of
EPCR, removing an important regulator of pro-coagulation re-
sponses (Moxon et al., 2013). The capacity to bind to both recep-
tors may give the parasite more potential to resist endothelial
changes in response to binding (i.e., downregulation of EPCR)
and to remain sequestered in the brain. Together with the
increased expression of ICAM-1 on endothelial cells, this could
lead to a vicious cycle that further exacerbates disease.
CM is a complex disease, and in addition to dual-binding
PfEMP1, various other aspects of both host and parasite biology
will influence the development of cerebral symptoms. Other
surface protein families, including STEVORs, RIFINs, and the
PfEMP1 involved in rosetting (the clustering of infected and unin-
fected erythrocytes), have been implicated in severe malaria
(Doumbo et al., 2009; Goel et al., 2015; Niang et al., 2014;
Rowe et al., 1997) and further study is required to assesswhether
they too have a significant impact on the development of cerebral
symptoms. In addition, host genetic factors, such as erythrocyte
surface protein polymorphisms, could affect the capacity of the
parasite to invade erythrocytes, altering parasitemia and there-
fore severity (Band et al., 2015). Nevertheless, the identification
of dual-binding PfEMP1 as a risk factor for CM highlights these
molecules as a promising and tractable candidate for the devel-
opment of a vaccine to contribute to the prevention of this formof
the disease.
A major challenge in targeting PfEMP1s through vaccination is
their high sequence variability, driven by a diversifying selectionCell Host & Microbe 21, 403–414, March 8, 2017 411
pressure to evade immune detection. This is demonstrated by
structural studies of CIDRa domains bound to EPCR or CD36
that revealed their interaction sites to retain their overall chemical
nature, while significantly diversifying in sequence (Hsieh et al.,
2016; Lau et al., 2015). This reduces the likelihood of the natural
development of cross-inhibitory antibody responses that target
these critical surfaces of CIDR domains. In contrast, the ICAM-
1-binding site of the group A PfEMP1 is remarkably conserved,
with a number of absolutely or predominantly conserved amino
acid residues that make critical, direct interactions with ICAM-1
(Figure 2G). Indeed, we find that plasma from P. falciparum-
exposed individuals contains IgG with the capacity to inhibit
ICAM-1 binding by a broad range of motif-containing DBLb do-
mains (Figure 4). This encourages future efforts to raise broadly
reactive antibody responses against these domains. This study
therefore highlights stronger endothelial adhesion capacity to
both ICAM-1 and EPCR as a risk factor for development of CM
and identifies a group of PfEMP1s with a remarkably conserved
ICAM-1-binding site as potential vaccine immunogen for use as
part of a strategy to prevent death due to CM.
EXPERIMENTAL PROCEDURES
More detailed methods and references are provided in the Supplemental
Experimental Procedures.
Crystal Structure Determination
The PF11_0521 DBLb3_D4 domain was produced as described in the Supple-
mental Experimental Procedures and was mixed with a 1.5 molar excess of
ICAM-1D1D2 before size exclusion chromatography. Crystals were grown us-
ing sitting-drop vapor diffusion with a well solution of 10% (w/v) PEG 20000,
20% (v/v) PEG 500, and 0.1 M Tris-BICINE (pH 8.5) and cryo-cooled in well so-
lution with an increased PEG 500 concentration of 32%.
Data were collected on beamline I02 (Diamond Light Source), indexed and
refined using iMosflm, and scaled using SCALA (Evans and Murshudov,
2013). Molecular replacement was performed using Phaser-MR (McCoy et al.,
2007) with a trimmed DBLa domain (PDB: 2XU0) and individual D1 and D2 do-
mains from ICAM-1D1D2 (PDB: 1IC1) as searchmodels. This identified one copy
of the complex in the asymmetric unit. The structure was built and refined
through iterative cycles of refinement using Refmac5 (Murshudov et al., 2011),
Buster (Bricogne et al., 2016), and model building in Coot (Emsley et al., 2010).
Surface Plasmon Resonance
Surface plasmon resonance experiments were carried out in a Biacore T200
instrument (GE Healthcare). ICAM-1D1D5-Fc was bound to a CM5 chip (GE
Healthcare) pre-coupled to protein A while EPCR was biotinylated by BirA
and coupled to a biotin CAPture chip (GE Healthcare). Binding partners were
injected for 240 s with a dissociation time of 660 s. To analyze dual binding
to ICAM-1 and EPCR, ICAM-1D1D5-Fc was immobilized. Two- and three-
domain PfEMP1 constructs were then injected, followed by a concentration
series of EPCR. Kinetic sensorgrams were fitted to a 1:1 interaction model
to determine kinetic rate constants and dissociation constant.
Small-Angle X-Ray Scattering
Small-angle X-ray scattering data were collected on beamline BM29 at the Eu-
ropean Synchrotron Radiation Facility and processed using the ATSAS pro-
cessing suite. The resulting ab initio model was converted into an envelope
before manual model docking.
Patient Samples
Plasmodium falciparum isolates were collected frommalaria patients at hospi-
tals in Ghana, Tanzania, and Benin. Clinical manifestations of malaria were
classified according to the definitions and associated criteria of the World
Health Organization. CM was identified by a positive blood smear of the412 Cell Host & Microbe 21, 403–414, March 8, 2017asexual form of P. falciparum and unrousable coma (Blantyre coma score,
BCS % 2) with exclusion of other causes of coma and severe illness. SA
was identified by hemoglobin < 5g/dL and BCS > 2. OM was indicated by
BCS > 2 and hemoglobin > 5 g/dL. These studies were approved by the
respective ethics committees. Additional plasma samples were collected
from P. falciparum-exposed children during a cross-sectional village study in
Tanzania and from falciparum-exposed adult Liberians (Lusingu et al., 2004;
Theisen et al., 2000).
Antibody Production and Purification
Rat antibodies were obtained by immunization with DBLb domains, followed
by affinity purification of IgG. Human antibodies against DBLb domains or
the ICAM-1-binding motif were affinity purified from pooled plasma from
nine Liberian adults.
ELISA
ICAM-1 binding was assessed by ELISA using DBLb domain-coatedMaxisorp
plates, with detection using ICAM-1D1D5-Fc and rabbit anti-human IgG-HRP
(Dako). A similar assay was used to assess the reactivity of human plasma
samples to the DBLb domains or to assess their efficacy at blocking ICAM-1
binding.
Assessment of Infected Erythrocyte Adhesion to Receptors
Plasmodium falciparum parasites were maintained in in vitro blood culture and
selected for specific PfEMP1 expression using antibodies. Microslides (Ibidi)
were coated with recombinant receptors (ICAM-1D1D5-Fc, EPCR, or CD36)
or seeded with HBMEC (Sciencell). Infected erythrocytes were exposed to
the microslides at a range of physiological shear stresses and bound infected
erythrocytes were quantified by microscopy.
Immunofluorescence
Infected erythrocytes were incubated with complexes of ICAM-1D1D5-Fc:anti-
human IgG-Dylight 488 (Abcam) and/or EPCR:anti-his IgG (QIAGEN):anti-
mouse IgG Alexa 568 (Abcam) before washing and analysis in a Zeiss
AxioImager Z1.
Real-Time qPCR
Parasite RNA from clinical samples was reverse transcribed into DNA and sub-
jected to real-time qPCR using primer pairs (Table S7C) designed to amplify
ICAM-1 motif-containing var genes or var gene subclasses encoding EPCR-
binding CIDRa1 domains.
Sequence Analysis
DBLb domain sequences (Rask et al., 2010) were used to blast extract DBLb
sequences from assemblies of Illumina whole-genome sequencing data,
including MalariaGEN data (Miotto et al., 2015) and assembled amplicons
from eight Ghanaian isolates. Those predicted to bind ICAM-1 were identified
using search term ASNGGPGYYNTEVQK. A database containing sequences
of the six most expressed var genes in 45 Tanzanian children with severe or
uncomplicated malaria (Jespersen et al., 2016) was used to correlate relative
expression of motif-containing var genes when compared with disease
outcome. Var gene expression and disease severity were analyzed using the
Wilcoxon-Mann-Whitney rank-sum test. To determine whether study site
affected the outcome, a logistical regression model was used.
ACCESSION NUMBERS
Sequences originating from the present study are deposited in GenBank with
accession numbers GenBank: KF984156, KJ866957, KJ866958, KJ866959,
KM364031, KM364032, KM364033, and KM364034. Structure files have
been deposited in the PDB under ID code PDB: 5MZA.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures,
seven figures, and seven tables and can be found with this article online at
http://dx.doi.org/10.1016/j.chom.2017.02.009.
AUTHOR CONTRIBUTIONS
F.L., Y.A., A.B., T.L., M.K.H., and A.T.R.J. conceived and planned the study.
F.L., Y.A., L.H., M.K.H., and A.T.R.J. wrote the manuscript. F.L., A.B.,
R.W.O., L.T., B.G., S.N.-S., and A.T.R.J. contributed with recombinant pro-
teins and antibodies. F.L. purified and crystallized proteins and collected
and analyzed small-angle X-ray scattering data. F.L. and M.K.H. prepared
crystals, collected data, and solved the structure. F.L. and C.K.Y.L. performed
and analyzed surface plasmon resonance experiments. Y.A. performed adhe-
sion assays and IFA. A.B. and J.E.V.P. did parasite culturing and flow cytom-
etry analysis. R.W.O. and G. E.-M. performed ELISA studies. R.W.O. and
A.T.R.J. did real-time qPCRwork. J.S.J., T.L., and A.T.R.J. did the sequencing
and bioinformatics analysis. N.T.N., G.E.-M., A.M., M.F.O., J.P.L., C.W.W.,
and T.G.T. organized clinical work and processed clinical samples. All authors
read and commented on the manuscript. F.L., Y.A., A.B., M.K.H., and A.T.R.J.
contributed equally to the work.
ACKNOWLEDGMENTS
M.K.H. is a Wellcome Trust Investigator. A.B., A.T.R.J., C.W.W., G.E.-M.,
J.E.V.P., J.S.J., L.H., M.F.O., R.W.O., T.L., and Y.A. were supported by
Novo Nordisk Fonden (NNF13OC0006249), Hørslev-Fonden, Augustinus Fon-
den (12-0378 and 15-3000), Lundbeckfonden (R180-2014-3098 and R140-
2013-13448), Aase and Ejnar Danielsens Fond, Oda og Hans Svenningsens
Fond (BOF-10109), Axel Muusfeldts Fond, Grosserer L.F. Foghts Fond, Uni-
versity of Copenhagen UCPH2016 Programme of Excellence, the Consultative
Committee for Development Research (DANIDA), the Danish Council for Inde-
pendent Research (4004-00032, T1333-00220, 1331-00089B, and Sapere
Aude program DFF–4004-00624B), and the NIH (R01HL130678). Fieldwork
in Benin was supported by PEERS-EPIVAC grant from AIRD/IRD. We thank
the Beninese, Ghanaian, and Tanzanian donors; the MalariaGen community
project for the use of data (Miotto et al., 2015); Michael Theisen for immune
plasma from Liberia; and Mette Ulla Madsen and Maiken Høwning Visti for
excellent technical assistance. We acknowledge the Core Facility for Inte-
grated Microscopy, Faculty of Health and Medical Sciences, University of Co-
penhagen for use of the Zeiss AxioImage Z1 and Zachary Barbati for doing
initial experiments not included in this paper. We thank David Staunton for
assistance with biophysical experiments and Ed Lowe and the beamline scien-
tists at I02 at Diamond Light Source for assistance with data collection. The
accession codes for the sequences originating from Jespersen et al. (2016)
are GenBank: KX154876, KX154960, KX154911, KX154945, KX154859,
KX154816, KX154936, KX154871, KX154901, KX154818, KX154872,
KX154935, and KX154836.
Received: October 4, 2016
Revised: January 20, 2017
Accepted: February 10, 2017
Published: March 8, 2017
REFERENCES
Avril, M., Tripathi, A.K., Brazier, A.J., Andisi, C., Janes, J.H., Soma, V.L.,
Sullivan,D.J., Jr.,Bull, P.C.,Stins,M.F., andSmith, J.D. (2012).A restrictedsub-
set of var genesmediates adherence of Plasmodium falciparum-infected eryth-
rocytes tobrainendothelial cells. Proc.Natl. Acad.Sci.USA109, E1782–E1790.
Avril, M., Bernabeu, M., Benjamin, M., Brazier, A.J., and Smith, J.D. (2016).
Interaction between endothelial protein C receptor and intercellular adhesion
molecule 1 to mediate binding of Plasmodium falciparum-infected erythro-
cytes to endothelial cells. MBio 7, e00615-16.
Band, G., Rockett, K.A., Spencer, C.C., and Kwiatkowski, D.P.; Malaria
Genomic Epidemiology Network (2015). A novel locus of resistance to severe
malaria in a region of ancient balancing selection. Nature 526, 253–257.
Bengtsson, A., Joergensen, L., Rask, T.S., Olsen, R.W., Andersen, M.A.,
Turner, L., Theander, T.G., Hviid, L., Higgins, M.K., Craig, A., et al. (2013). A
novel domain cassette identifies Plasmodium falciparum PfEMP1 proteins
binding ICAM-1 and is a target of cross-reactive, adhesion-inhibitory anti-
bodies. J. Immunol. 190, 240–249.Bernabeu, M., Danziger, S.A., Avril, M., Vaz, M., Babar, P.H., Brazier, A.J.,
Herricks, T., Maki, J.N., Pereira, L., Mascarenhas, A., et al. (2016). Severe adult
malaria is associated with specific PfEMP1 adhesion types and high parasite
biomass. Proc. Natl. Acad. Sci. USA 113, E3270–E3279.
Birbeck, G.L., Molyneux, M.E., Kaplan, P.W., Seydel, K.B., Chimalizeni, Y.F.,
Kawaza, K., and Taylor, T.E. (2010). Blantyre Malaria Project Epilepsy Study
(BMPES) of neurological outcomes in retinopathy-positive paediatric cerebral
malaria survivors: a prospective cohort study. Lancet Neurol. 9, 1173–1181.
Bricogne, G., Blanc, E., Brandl, M., Flensburg, C., Keller, P., Paciorek, W.,
Roversi, P., Sharff, A., Smart, O.S., Vonrhein, C., and Womack, T.O. (2016).
BUSTER version 2.10.2 (Global Phasing).
Claessens, A., Adams, Y., Ghumra, A., Lindergard, G., Buchan, C.C., Andisi,
C., Bull, P.C., Mok, S., Gupta, A.P., Wang, C.W., et al. (2012). A subset of group
A-like var genes encodes the malaria parasite ligands for binding to human
brain endothelial cells. Proc. Natl. Acad. Sci. USA 109, E1772–E1781.
Cunnington, A.J., Riley, E.M., and Walther, M. (2013). Stuck in a rut?
Reconsidering the role of parasite sequestration in severe malaria syndromes.
Trends Parasitol. 29, 585–592.
Dondorp, A.M., Fanello, C.I., Hendriksen, I.C., Gomes, E., Seni, A.,
Chhaganlal, K.D., Bojang, K., Olaosebikan, R., Anunobi, N., Maitland, K.,
et al.; AQUAMAT group (2010). Artesunate versus quinine in the treatment of
severe falciparummalaria in African children (AQUAMAT): an open-label, rand-
omised trial. Lancet 376, 1647–1657.
Doumbo, O.K., Thera, M.A., Koné, A.K., Raza, A., Tempest, L.J., Lyke, K.E.,
Plowe, C.V., and Rowe, J.A. (2009). High levels of Plasmodium falciparum ro-
setting in all clinical forms of severe malaria in African children. Am. J. Trop.
Med. Hyg. 81, 987–993.
Emsley, P., Lohkamp, B., Scott, W.G., and Cowtan, K. (2010). Features and
development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501.
Evans, P.R., and Murshudov, G.N. (2013). How good are my data and what is
the resolution? Acta Crystallogr. D Biol. Crystallogr. 69, 1204–1214.
Goel, S., Palmkvist, M., Moll, K., Joannin, N., Lara, P., Akhouri, R.R., Moradi,
N., Öjemalm, K., Westman, M., Angeletti, D., et al. (2015). RIFINs are adhesins
implicated in severe Plasmodium falciparum malaria. Nat. Med. 21, 314–317.
Howell, D.P., Levin, E.A., Springer, A.L., Kraemer, S.M., Phippard, D.J., Schief,
W.R., and Smith, J.D. (2008). Mapping a common interaction site used by
Plasmodium falciparum Duffy binding-like domains to bind diverse host recep-
tors. Mol. Microbiol. 67, 78–87.
Hsieh, F.L., Turner, L., Bolla, J.R., Robinson, C.V., Lavstsen, T., and Higgins,
M.K. (2016). The structural basis for CD36 binding by the malaria parasite.
Nat. Commun. 7, 12837.
Hviid, L., and Jensen, A.T. (2015). PfEMP1—a parasite protein family of key
importance in Plasmodium falciparum malaria immunity and pathogenesis.
Adv. Parasitol. 88, 51–84.
Idro, R., Jenkins, N.E., and Newton, C.R. (2005). Pathogenesis, clinical fea-
tures, and neurological outcome of cerebral malaria. Lancet Neurol. 4,
827–840.
Janes, J.H., Wang, C.P., Levin-Edens, E., Vigan-Womas, I., Guillotte, M.,
Melcher, M., Mercereau-Puijalon, O., and Smith, J.D. (2011). Investigating
the host binding signature on the Plasmodium falciparum PfEMP1 protein
family. PLoS Pathog. 7, e1002032.
Jensen, A.T., Magistrado, P., Sharp, S., Joergensen, L., Lavstsen, T.,
Chiucchiuini, A., Salanti, A., Vestergaard, L.S., Lusingu, J.P., Hermsen, R.,
et al. (2004). Plasmodium falciparum associated with severe childhood malaria
preferentially expresses PfEMP1 encoded by group A var genes. J. Exp. Med.
199, 1179–1190.
Jespersen, J.S., Wang, C.W., Mkumbaye, S.I., Minja, D.T., Petersen, B.,
Turner, L., Petersen, J.E., Lusingu, J.P., Theander, T.G., and Lavstsen, T.
(2016). Plasmodium falciparum var genes expressed in children with severe
malaria encode CIDRa1 domains. EMBO Mol. Med. 8, 839–850.
Kim, J.H., Singvall, J., Schwarz-Linek, U., Johnson, B.J., Potts, J.R., and
Höök, M. (2004). BBK32, a fibronectin binding MSCRAMM from Borrelia burg-
dorferi, contains a disordered region that undergoes a conformational change
on ligand binding. J. Biol. Chem. 279, 41706–41714.Cell Host & Microbe 21, 403–414, March 8, 2017 413
Kwong, P.D., Wyatt, R., Robinson, J., Sweet, R.W., Sodroski, J., and
Hendrickson, W.A. (1998). Structure of an HIV gp120 envelope glycoprotein
in complex with the CD4 receptor and a neutralizing human antibody. Nature
393, 648–659.
Lau, C.K., Turner, L., Jespersen, J.S., Lowe, E.D., Petersen, B., Wang, C.W.,
Petersen, J.E., Lusingu, J., Theander, T.G., Lavstsen, T., and Higgins, M.K.
(2015). Structural conservation despite huge sequence diversity allows
EPCR binding by the PfEMP1 family implicated in severe childhood malaria.
Cell Host Microbe 17, 118–129.
Lavstsen, T., Turner, L., Saguti, F., Magistrado, P., Rask, T.S., Jespersen, J.S.,
Wang, C.W., Berger, S.S., Baraka, V., Marquard, A.M., et al. (2012).
Plasmodium falciparum erythrocyte membrane protein 1 domain cassettes 8
and 13 are associated with severe malaria in children. Proc. Natl. Acad. Sci.
USA 109, E1791–E1800.
Lusingu, J.P., Vestergaard, L.S., Mmbando, B.P., Drakeley, C.J., Jones, C.,
Akida, J., Savaeli, Z.X., Kitua, A.Y., Lemnge, M.M., and Theander, T.G.
(2004). Malaria morbidity and immunity among residents of villages with
different Plasmodium falciparum transmission intensity in North-Eastern
Tanzania. Malar. J. 3, 26.
McCoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C.,
and Read, R.J. (2007). Phaser crystallographic software. J. Appl. Cryst. 40,
658–674.
Menschikowski, M., Hagelgans, A., Eisenhofer, G., and Siegert, G. (2009).
Regulation of endothelial protein C receptor shedding by cytokines is medi-
ated through differential activation of MAP kinase signaling pathways. Exp.
Cell Res. 315, 2673–2682.
Miotto, O., Amato, R., Ashley, E.A., MacInnis, B., Almagro-Garcia, J.,
Amaratunga, C., Lim, P., Mead, D., Oyola, S.O., Dhorda, M., et al. (2015).
Genetic architecture of artemisinin-resistant Plasmodium falciparum. Nat.
Genet. 47, 226–234.
Mkumbaye, S.I., Wang, C.W., Lyimo, E., Jespersen, J.S., Manjurano, A.,
Mosha, J., Kavishe, R.A., Mwakalinga, S.B., Minja, D.T., Lusingu, J.P., et al.
(2017). The severity of Plasmodium falciparum infection is associated with
transcript levels of var genes encoding EPCR-binding PfEMP1. Infect.
Immun. Published online January 30, 2017. http://dx.doi.org/10.1128/IAI.
00841-16.
Moxon, C.A., Heyderman, R.S., and Wassmer, S.C. (2009). Dysregulation of
coagulation in cerebral malaria. Mol. Biochem. Parasitol. 166, 99–108.
Moxon, C.A., Wassmer, S.C., Milner, D.A., Jr., Chisala, N.V., Taylor, T.E.,
Seydel, K.B., Molyneux, M.E., Faragher, B., Esmon, C.T., Downey, C., et al.
(2013). Loss of endothelial protein C receptors links coagulation and inflamma-
tion to parasite sequestration in cerebral malaria in African children. Blood 122,
842–851.
Murshudov, G.N., Skubák, P., Lebedev, A.A., Pannu, N.S., Steiner, R.A.,
Nicholls, R.A., Winn, M.D., Long, F., and Vagin, A.A. (2011). REFMAC5 for
the refinement of macromolecular crystal structures. Acta Crystallogr. D
Biol. Crystallogr. 67, 355–367.
Newbold, C., Warn, P., Black, G., Berendt, A., Craig, A., Snow, B., Msobo, M.,
Peshu, N., and Marsh, K. (1997). Receptor-specific adhesion and clinical dis-
ease in Plasmodium falciparum. Am. J. Trop. Med. Hyg. 57, 389–398.
Niang,M., Bei, A.K., Madnani, K.G., Pelly, S., Dankwa, S., Kanjee, U., Gunalan,
K., Amaladoss, A., Yeo, K.P., Bob, N.S., et al. (2014). STEVOR is a Plasmodium
falciparum erythrocyte binding protein that mediates merozoite invasion and
rosetting. Cell Host Microbe 16, 81–93.414 Cell Host & Microbe 21, 403–414, March 8, 2017Ochola, L.B., Siddondo, B.R., Ocholla, H., Nkya, S., Kimani, E.N., Williams,
T.N., Makale, J.O., Liljander, A., Urban, B.C., Bull, P.C., et al. (2011).
Specific receptor usage in Plasmodium falciparum cytoadherence is associ-
ated with disease outcome. PLoS ONE 6, e14741.
Oleinikov, A.V., Amos, E., Frye, I.T., Rossnagle, E., Mutabingwa, T.K., Fried,
M., and Duffy, P.E. (2009). High throughput functional assays of the variant an-
tigen PfEMP1 reveal a single domain in the 3D7 Plasmodium falciparum
genome that binds ICAM1 with high affinity and is targeted by naturally ac-
quired neutralizing antibodies. PLoS Pathog. 5, e1000386.
Petersen, J.E., Bouwens, E.A., Tamayo, I., Turner, L., Wang, C.W., Stins, M.,
Theander, T.G., Hermida, J., Mosnier, L.O., and Lavstsen, T. (2015). Protein
C system defects inflicted by themalaria parasite protein PfEMP1 can be over-
come by a soluble EPCR variant. Thromb. Haemost. 114, 1038–1048.
Rask, T.S., Hansen, D.A., Theander, T.G., GormPedersen, A., and Lavstsen, T.
(2010). Plasmodium falciparum erythrocyte membrane protein 1 diversity in
seven genomes—divide and conquer. PLoS Comput. Biol. 6, e1000933.
Robinson, B.A., Welch, T.L., and Smith, J.D. (2003). Widespread functional
specialization of Plasmodium falciparum erythrocyte membrane protein 1
family members to bind CD36 analysed across a parasite genome. Mol.
Microbiol. 47, 1265–1278.
Rogerson, S.J., Tembenu, R., Dobaño, C., Plitt, S., Taylor, T.E., andMolyneux,
M.E. (1999). Cytoadherence characteristics of Plasmodium falciparum-in-
fected erythrocytes from Malawian children with severe and uncomplicated
malaria. Am. J. Trop. Med. Hyg. 61, 467–472.
Rowe, J.A., Moulds, J.M., Newbold, C.I., and Miller, L.H. (1997). P. falciparum
rosetting mediated by a parasite-variant erythrocyte membrane protein and
complement-receptor 1. Nature 388, 292–295.
Seydel, K.B., Kampondeni, S.D., Valim, C., Potchen, M.J., Milner, D.A.,
Muwalo, F.W., Birbeck, G.L., Bradley, W.G., Fox, L.L., Glover, S.J., et al.
(2015). Brain swelling and death in children with cerebral malaria. N. Engl. J.
Med. 372, 1126–1137.
Silamut, K., Phu, N.H., Whitty, C., Turner, G.D., Louwrier, K., Mai, N.T.,
Simpson, J.A., Hien, T.T., and White, N.J. (1999). A quantitative analysis of
the microvascular sequestration of malaria parasites in the human brain.
Am. J. Pathol. 155, 395–410.
Storm, J., and Craig, A.G. (2014). Pathogenesis of cerebral malaria—inflam-
mation and cytoadherence. Front. Cell. Infect. Microbiol. 4, 100.
Theisen, M., Soe, S., Jessing, S.G., Okkels, L.M., Danielsen, S., Oeuvray, C.,
Druilhe, P., and Jepsen, S. (2000). Identification of a major B-cell epitope of
the Plasmodium falciparum glutamate-rich protein (GLURP), targeted by hu-
man antibodies mediating parasite killing. Vaccine 19, 204–212.
Turner, G.D., Morrison, H., Jones, M., Davis, T.M., Looareesuwan, S., Buley,
I.D., Gatter, K.C., Newbold, C.I., Pukritayakamee, S., Nagachinta, B., et al.
(1994). An immunohistochemical study of the pathology of fatal malaria.
Evidence for widespread endothelial activation and a potential role for intercel-
lular adhesion molecule-1 in cerebral sequestration. Am. J. Pathol. 145,
1057–1069.
Turner, L., Lavstsen, T., Berger, S.S., Wang, C.W., Petersen, J.E., Avril, M.,
Brazier, A.J., Freeth, J., Jespersen, J.S., Nielsen, M.A., et al. (2013). Severe
malaria is associated with parasite binding to endothelial protein C receptor.
Nature 498, 502–505.