Background
Antigenic variation is a major characteristic of malaria parasites of the genus
Plasmodium leading to severe and chronic infections in a variety of vertebrates. Malaria parasites have developed strategies to evade host immune responses by expressing a large and diverse repertoire of variant proteins on the surface of parasitized red blood cells (pRBCs) [
1‐
4]. By rapidly switching between these antigens, the parasites avoid antibody-mediated immunity of the host thus enabling the parasites to proliferate in the host without being completely eliminated by the adaptive immune response. Furthermore, these surface antigens were proposed to be involved in adherence to endothelial cells causing sequestration of late trophozoites and schizonts in post-capillary venules in specific inner organs, which considerably contributes to severe pathology in the host [
5‐
8].
In the most virulent human pathogen
Plasmodium falciparum, the PfEMP1 protein family encoded by widely studied
var genes was shown to be expressed at the surface of pRBC mediating the parasite binding to endothelial cells lining small blood vessels [
9,
10]. Several
in vitro assays with
P. falciparum revealed a tight regulation of the expression of the individual variant antigens by silencing, activation and mutually exclusive expression, thus only one variant protein is expressed in parasites at any given time [
10‐
13]. By switching the expression to another variant, antigenic properties of the surface of pRBCs changes and the parasites prevent their complete clearance [
10,
14]. Typical waves of parasitaemia in persistent
P. falciparum infections reflect the repeated switching between different members of the diverse repertoire of variant antigens [
10].
Further variant surface antigens encoded by multigene families were found in several other
Plasmodium species infecting humans, monkeys and rodents [
4,
15‐
17]. However, homologues of the
var genes are not present in
Plasmodium vivax or other malaria species [
18], with exception of several
var-like sequences from the chimpanzee parasite
Plasmodium reichenowi[
19], a close relative of
P. falciparum.
The largest multigene family in
Plasmodium genomes was described to be formed by the
Plasmodium interspersed repeat (
pir) genes [
20,
21]. The superfamily of
pir genes constitutes the major variant surface antigen family in most
Plasmodium species. They were found in the human pathogen
P. vivax (
vir) [
22], in the simian parasite
Plasmodium knowlesi (
kir) [
18] and in the rodent malaria species
Plasmodium yoelii (
yir),
Plasmodium berghei (
bir) and
Plasmodium chabaudi (
cir) [
15,
17,
23]. Transcriptional changes of
yir genes modulated by host immunity were reported in immunocompetent mice infected with
P. yoelii[
24]. In contrast to the exclusive expression of individual
var genes and consecutive activation of different genes in
P. falciparum, transcriptional profiling analyses in
P. yoelii showed a simultaneous expression of a broad range of different
yir genes within different intra-erythrocytic developmental parasite stages [
25]. In a single parasite, however, only one to three
yir transcripts were detectable. Cunningham and colleagues [
25] concluded that antigenic variation in
P. yoelii probably involves exposure of the immune system to many different YIR antigens and transcriptional switching takes place without any epigenetic memory. For the variant VIR proteins of the most widely distributed human malaria pathogen
P. vivax, a similar differential expression in natural infections could also be detected [
26] and for the
cir genes of the rodent malaria species
P. chabaudi antigenic switching in laboratory mice has already been described [
15].
Although the PIR proteins were localised close to the surface of pRBCs infected with
P. vivax, P. yoelii, P. berghei, and
P. chabaudi[
20,
22,
24,
27], less is known about the role of PIR proteins in host/parasite interactions. For instance, a supposed correlation of antigenic variation of the PIRs with sequestration in inner organs has not yet been analysed. In
P. falciparum, adherence of pRBCs to different host receptor such as CD36, ICAM-1 or chondroitin-sulfate A (CSA) is mediated by the major variant protein PfEMP1 [
6,
8,
28], but no homologues to these proteins were detected in most other malaria pathogens. Therefore, it is conceivable that the PIR proteins, which are also expressed on the surface membrane of pRBCs, are as well involved in adherence to endothelial cells and sequestration of pRBCs in the microvasculature of inner organs.
The complex phenomenon of sequestration is hitherto not completely understood as indicated by the common assumption that sequestration of mature pRBCs in the microvasculature of the host tissues, as found in
P. falciparum infections, does not occur for
P. vivax. Thus, it has been assumed that the human parasite
P. vivax must have developed different strategies e.g. adherence to barrier cells in the spleen to avoid spleen clearance [
29]. Recently, however,
in vitro assays provided new evidence for a cytoadherence of
P. vivax pRBCs to endothelial cells and placental cryosections suggesting a cytoadherence comparable to that of
P. falciparum pRBCs
in vivo[
30]. Direct or indirect evidences for an involvement of VIR antigens in binding to host receptors are currently still missing. In the most studied rodent model
P. berghei, real-time
in vivo imaging of transgenic parasites revealed CD36-mediated sequestration of schizonts in adipose tissues and lung as well as an accumulation of schizonts in the spleen [
31]. Accumulation of other blood stages was also observed in different tissues including the brain and placenta of pregnant mice [
32‐
34]. Nevertheless, identification of parasite ligands involved in binding to CD36 or other host receptors are still missing.
In vitro adherence of
P. chabaudi infected erythrocytes to purified human CD36 has been observed [
35]. Moreover, sequestration to microvascular endothelial cells was reported to appear in an organ specific manner in
P. chabaudi infections, predominantly in liver, but also in brain and spleen thus resembling at least partially the sequestration pattern in
P. falciparum infections.
In this study, the relationship between antigenic variation of the cir multigene family and accumulation of parasites localization in different host tissues was analysed at the molecular level. Bioinformatic characterization of the cir genes was used to identify different subfamilies within the multigene family. First indications for a tissue-specific expression of cir genes were obtained by amplification of a representative sample of transcribed cir genes from different host tissues. Transcriptional profiling analyses of cir genes in different host tissues using RFLP analysis of RT-PCR products provided further evidence of a tissue-specific expression of cir genes in P. chabaudi infections. These results provide first indications for a possible correlation of antigenic variation of the cir multigene family and sequestration in host tissues in P. chabaudi infections.
Methods
All annotated sequence data of putative
cir genes in the
Plasmodium Genome Database Resource version 8.0 [
36] (PlasmoDB; plasmodb.org) were compiled (Additional file
1). For confirmation of correct annotation, a CD-BLAST [
37‐
39] was performed to identify and locate the conserved domain of the CIR-BIR-YIR superfamily [
22,
23] (Pfam protein families database accession number [PF06022]). Sequences with a partial CIR-BIR-YIR conserved domain were excluded from all subsequent analyses.
For phylogenetic reconstruction, putative CIR proteins were aligned using ClustalW2 [
40] and a phylogenetic tree was calculated with PhyML 3.0 [
41] using maximum likelihood estimation and the JTT model [
42] for amino acid substitution. The gamma shape parameter and the proportion of invariable sites were estimated and the number of substitution rate categories was set to four. The implemented BIONJ algorithm was used to build the starting tree. Resulting trees (Newick format) were visualised and processed with MEGA4 [
43,
44].
For validation of structural protein motifs such as transmembrane domains, signal peptides and PEXEL motifs, analyses of the protein sequences were examined with Protscale [
45], TMHMM 2.0 [
46,
47] and SignalP [
48]. Results were compared with annotations extracted from PlasmoDB.
Mice
All experiments were performed with 5-8 weeks-old outbred female NMRI mice (Crl:NMRI(Han)) provided by Charles River (Sulzfeld, Germany). The mice were kept in cages with a maximum of five animals per cage and received food and water ad libitum. The experiments were planed according to all relevant guidelines for animal protection and approved by German authorities responsible for animal protection.
Infection of mice
A non-clonal line of
P. chabaudi very similar but not identical to
P. chabaudi AS was used [
49‐
51]. Blood stages of
P. chabaudi were weekly passaged in female NMRI mice by intraperitoneal injection (i.p.) of a droplet of tail vein blood diluted in PBS. Parasitaemia was evaluated in Giemsa-stained blood smears and total erythrocytes number was counted in a Neubauer chamber.
For the experiments, blood of an infected NMRI mouse was collected by cardiac puncture under anesthesia. For each transcriptional profiling experiment, six mice were infected i.p. with approximately 100 parasitized red blood cells (pRBCs). Organs and blood were collected at about 30% parasitaemia, i.e. just before peak parasitaemia.
For transcriptional profiling during the course of an infection, tail vein blood of mice infected i.p. with 100 pRBCs was passaged i.p. into naïve female NMRI mice at days 7 (early infection), 14 (around peak parasitaemia), 21 (7 days after peak parasitaemia) and 35 (21 days after peak parasitaemia). Expression of cir genes was analysed when parasitaemia reached about 30%.
RNA and DNA extraction
Blood samples of P. chabaudi infected mice were collected by cardiac puncture, rapidly frozen and stored at -80°C. Small pieces of liver, spleen, kidneys, lung and brain were transferred into RNA Later (Sigma Aldrich) and kept at -80°C for long-term storage. Total RNA was extracted using NucleoSpin® RNA II kit (Macherey-Nagel) according to the manufactures instructions. Genomic DNA extraction was performed with the NucleoSpin® Blood kit (Macherey-Nagel).
Verification of complete cir gene structures
To verify the complete gene structure of selected
cir genes, RT-PCR and genomic PCR were performed in parallel. The primer pairs used for amplification and detailed information about PCR conditions can be found in Additional file
2.
Cloning of RT-PCR products and sequencing
Residual contaminating genomic DNA in total RNA preparations was digested with DNase I (Fermentas). First strand cDNA was synthesised using 1 μg RNA and the RevertAid™ Premium Reverse Transcriptase (Fermentas) with Oligo dT18 primers including reactions without reverse transcriptase as negative controls for amplification.
For transcription analysis, specific primers spanning the second exon encoding an essential part of the CIR-BIR-YIR domain (> 98% of its length) were designed for both major cir gene subfamilies. The primer pairs f1up (5'-AATACGCTATTTTATGGTTTAGTTATAAA-3')/f1lo (5'-TGAAATTCCTAAAATAATGGGTATTATTAAAA-3') and f2up (5'-TATGCTATTTTATGGTTAAGTTATATGCTA-3')/f2lo (5'-ATGAATACTTATAAGCAATTCCCAAGAAAA-3') were targeted to highly conserved sequence regions for amplification of a broad range of cir genes.
Following cDNA synthesis, amplification with the AccuPrime™ DNA polymerase (Invitrogen) was performed using subfamily-specific primers for cir subfamily 1 and 2. After an initial denaturation for 30 s at 94°C, 40 cycles of 10 s at 94°C, 30 s at 55°C, 30 s at 72°C followed by a final extension of 10 min at 72°C. RT-PCR products were then isolated from a 0.8% agarose gel and precipitated in the presence of glycogen.
In order to sample a first repertoire of transcribed P. chabaudi cir genes, RT-PCR products of each tissue (blood, liver, spleen, kidney, lung and brain) were gel-purified and cloned into the pCR™4-TOPO® TA vector (Invitrogen). Thirty-six clones for each tissue and subfamily were sequenced (GATC, Constance, Germany) resulting in sequences for 216 independent cDNA clones for cir subfamily 1 and 2.
Transcriptional profiling of RT-PCR products by RFLP
For transcriptional profiling, DNase digestion of RNA and cDNA synthesis was performed as described before and RT-PCR products were amplified with the Phusion® Hot Start II High-Fidelity DNA Polymerase (Fermentas) using subfamily-specific primers for cir subfamily 1 and 2. After an initial denaturation for 30 s at 98°C, 50 cycles of 10 s at 98°C, 30 s at 55°C, 30 s at 72°C were performed followed by a final extension for 10 min at 72°C.
Gel-purified RT-PCR products (150 ng) were digested with Alu I or Xap I (Fermentas) in 5 μl for 3 h at 37°C followed by enzyme inactivation for 20 min at 65°C. The restricted fragments (30 ng) were analysed with the Agilent 2100 Bioanalyzer using the DNA 1000 LabChip® kit (Agilent) following the manufactures instructions. In order to ensure reproducibility of the RFLP profiles, both RT-PCR and RFPL were usually repeated at least twice resulting in 4 replicates.
Statistical analyses
In order to evaluate whether clones representing certain cir subfamilies were significantly more frequently recovered from one tissue than from others, frequencies were compared using a Z-test.
Discussion
In order to improve understanding of host/parasite interactions correlating antigenic variation and accumulation of parasite localization in different host tissues, expression of the P. chabaudi cir multigene family was analysed in vivo at the transcriptional level. Despite diverse in vitro or in vivo investigations regarding antigenic variation in Plasmodium species, the complete molecular mechanism of antigenic switching in the parasites is, hitherto, far away from being fully understood. The present study, therefore, examined P. chabaudi infections in immunocompetent mice to get further insight in the complex phenomenon of antigenic variation and sequestration.
In an initial phylogenetic analysis of the annotated repertoire of putative CIR proteins, two subfamilies and an unassigned group of very long CIRs could be identified. While the subfamily 1 and 2 exhibit the predicted primary structure of common PIR proteins, the unassigned CIR proteins show more divergence often due to large insertions within the conserved CIR-BIR-YIR domain.
It has already previously been proposed that different subsets within the
pir superfamily exhibit different functions. In
P. yoelii and
P. berghei, for example, a stage specific role was assumed since not all
pir genes are transcribed equally when different life cycle stages were compared [
24,
55]. Indeed, different roles for
P. chabaudi cir subfamily 1 (with changes in expression pattern) and subfamily 2 (with minor or no changes in expression pattern) genes are likely. The large insertions and additional transmembrane domains found within the group of unassigned CIR proteins might also be interpreted as hints for distinct functions of these proteins. Since none of them has been examined in any functional or even localization study, there is still no clue for the role these proteins might play in the
Plasmodium life cycle or in the pathogens immune evasion strategy. For the VIR proteins in
P. vivax diverse protein domain and structure predictions have been suggested to indicate different functions for different subsets of these proteins as for example those with PEXEL motif or those with additional transmembrane domains [
56]. However, the presence or absence of the PEXEL motifs might simply result in usage of alternative transport signals or pathways to the host cell membrane [
57] and currently no functional role has been suggested for any of the PIR proteins with more than one transmembrane domain. Since bioinformatic analysis of the CIR proteins revealed not a single strong PEXEL motif, it can be assumed that PEXEL motifs are not an important feature in the function of CIR proteins and that at least the majority of PIR proteins is able to be transported to the pRBC surface without an obvious PEXEL translocation signal.
A common ancestry of the gene families forming the
pir superfamily has been postulated due to e.g. conserved sequence motifs as well as structural predictions within the PIR sequences [
20]. By phylogenetic comparison of CIR and YIR proteins as performed here, this hypothesis can be further strengthened. The phylogenetic relationship of the CIR proteins with the widely studied YIRs in
P. yoelii shows that the YIR protein sequences only share high similarity to CIR proteins of subfamily 2 and neither to subfamily 1 nor the group of unassigned CIR proteins. Comparison to the YIR antigens of the close relative
P. yoelii suggests that most of the CIR sequences in the
P. chabaudi genome have evolved by diversification after separation of
P. yoelii and
P. chabaudi from a common ancestor. Such a dispersing evolution of distinctly evolved subfamilies has also been demonstrated by phylogenetic analysis of PIR proteins encoded in
P. yoelii and
P. berghei genomes [
52] and it is likely that the high variability of antigens in the individual malaria species distinctly evolved probably in response of the host immune pressure.
In this study a non-clonal P. chabaudi line very similar to the clonal P. chabaudi AS strain was used. The deduced protein sequences of the amplified cir transcripts of these parasites, however, were quite similar but not identical to the annotated putative CIR predicted from the genome sequence of the clonal P. chabaudi AS strain supposing a high variability of antigens even between closely related strains within the same species.
Cloning and sequencing of a first subset of
cir genes demonstrated that a broad range of subfamily 1 and subfamily 2
cir genes is transcribed during a
P. chabaudi infection in immunocompetent mice infected with a starter population of a minimal size. These findings are consistent with those described previously for other
pir multigene families, e.g. the
vir genes in
P. vivax, the
yir genes in
P. yoelii and the
bir genes in
P. berghei[
25,
26,
55]. In contrast to the mutually exclusive expression of only one
var gene found in
P. falciparum parasites cultured
in vitro[
58], it has been shown that in
P. vivax and in several rodents
Plasmodium species many different
pir genes were transcribed
in vivo within a parasite population in an individual host. Interestingly, examinations of natural
P. falciparum infected human samples has shown that - in contrast to the limited transcription pattern in cultured
P. falciparum parasites - many transcripts of
var and
stevor genes are also transcribed simultaneously
in vivo[
59,
60]. Such contradictory
in vivo and
in vitro findings make clear that
in vivo models such as
P. chabaudi or other rodent malaria parasites are essential for complex investigations of antigenic variation.
In accordance with previous studies of the
yir genes of
P. yoelii[
24], switching of
cir gene expression could be detected around peak parasitaemia in the course of infection suggesting that antigenic variation might be modulated by selective forces exercised by the host immune system. However, these transcriptional changes were not observed in all infected mice and, moreover, were not detectable to the same extent for all cir subfamilies. Most marked differences in mRNA expression patterns could be observed for cir subfamily 1 whereas no or obviously less transcriptional switching was detected for cir subfamily 2 during the infection.
Completely different cir gene expression patterns of the progenies derived from 100 pRBCs starter populations originating from the same parental population strongly suggest that a large parasite population can be extremely heterogeneous. In micromanipulated
P. yoelii pRBCs, it has been shown that only one to three different
yir transcripts were transcribed within an individual cell but many different transcripts were detected within a whole parasite population [
25]. In addition, rapid switching in the transcribed repertoires of
yir genes between different clonal host parasites populations and parasite developmental stages has been described. Although no infections with single pRBCs were performed in the present study, the initial diversity of parasites was apparently kept at a minimum using the minimal infectious dose resulting in patent infections by intraperitoneal infection as revealed by the substantial differences in RT-PCR RFLP patterns between mice infected with 100 pRBCs. Therefore, the number of transcribed
cir genes per pRBC is presumably also much lower than that found to be transcribed in a large population.
Previous studies have analysed switching between different
yir genes in
P. yoelii infected mice both in immunocompetent [
24] and highly immunodeficient [
25] mice. Efficient and frequent switching of expressed
yir genes could be observed even in the absence of any selecting force of an adaptive immune system [
25]. The fact that
cir gene subfamily 1 and subfamily 2 expression patterns apparently do not necessarily change in the course of a
P. chabaudi infection in immunocompetent mice even within three weeks suggests lower on-off switching frequencies at least for these groups of
cir genes. This is particularly surprising as adaptive immune responses are well known to effectively select for parasites expressing new variant antigens in other protozoan parasites [
61‐
63].
Changes in the expression patterns of variant antigens appear to occur more frequently in cir subfamily 1 than in subfamily 2. No data are currently available for switching frequencies within the highly heterogeneous group of unassigned
cir genes. Fonager
et al[
52] already speculated that within the
yir gene family those genes belonging to highly heterogeneous groups would be more important for antigenic variation than members of more conventional subfamilies. The phylogenetic analysis including members of all previously defined
yir groups [
52] shows that even the most divergent YIR proteins cluster together with the CIR subfamily 2. This suggests that subfamily 2 is ancient among the CIR proteins while subfamily 1 and the unassigned CIR proteins have probably evolved from ancient genes by diversification. The higher apparent switching frequency observed for cir subfamily 1 than subfamily 2 genes is a first experimental hint corroborating the hypothesis that those
pir family members that evolved relatively recently might play a more prominent role for antigenic variation than those showing ancient properties [
52]. Future work in both
P. yoelii and
P. chabaudi should, therefore, no longer neglect these unusual
pir members from the analyses but should pay special attention to them.
Analysis of changes in
cir gene expression patterns in the course of one asexual round of multiplication in 24 h revealed only minor changes in gene expression for subfamily 1 and virtually no changes for subfamily 2. In particular, no reduction in the complexity of the
cir genes in late parasite stages was observed in comparison to early rings as has been described for
var genes [
58]. Since Cunningham
et al[
25] found evidence that in
P. yoelii even schizonts transcribe at least up to three different
yir genes, a mutually exclusive expression of
pir genes in late individual parasites or parasite populations appears to be highly unlikely. The same was also shown for
vir genes in
P. vivax with more than one antigen expressed in a single parasite and different expression patterns between parasites [
26]. Transcriptomic analysis of
P. vivax intraerythrocytic developmental cycle stages also revealed that many
vir genes are expressed early during the ring stage and turned off later while others are expressed predominantly in schizonts [
64]. These authors could not find any linkage between the position of the
vir gene within the phylogenetic tree and its predominant expression time. In contrast to these results, only minor changes during the intraerythrocytic cycle were found in the present study. Possible explanations for this discrepancy include the fact that of course the RFLP analysis is less sensitive with regard to the detection of changes in expression of individual genes when compared to the microarray method used by Bozdech
et al[
64]. In addition, only for about 60% of the
vir genes consistent temporal expression patterns could be observed for three different
P. vivax isolates. If the highly expressed genes show no temporal expression pattern, such a pattern would clearly not be detectable with the RFLP method used here since it is not able to detect minor cir transcripts at all. Finally, the different experimental designs with only 100 pRBCs as founders in the
cir gene study and non-selected parasites from naturally infected patients for the
P. vivax transcriptome study might well contribute to the observed differences.
Remarkable differences in the mRNA expression patterns between pRBCs in blood, liver, spleen, kidney, lung and brain could be observed suggesting a host-tissue specific expression of
cir genes. It is believed that the PIR proteins, like the Pfemp1 protein family in
P. falciparum, are possibly involved in adhesion to host receptors thus mediating accumulation and sequestration in different host tissues. For the BIR or YIR proteins, for example, an expression of these molecules on the surface of pRBCs has already been demonstrated [
24,
27] but neither direct nor indirect evidence for a correlation of an adhesion of PIR proteins to host endothelial cells has yet been found. Accumulation of pRBCs in different tissues expressing different
cir genes, however, can be considered to be a first experimental hint that CIR and maybe also other PIR proteins might indeed be involved in adhesion and sequestration. Whether rapid and dramatic changes in tissue distribution of
P. chabaudi parasites at peak parasitaemia, i.e. exclusion from the splenic red pulp [
65], has effects on
cir gene expression patterns would also be interesting to analyse in the future. Since
P. chabaudi is the only frequently used experimental malaria model with synchronous development and a strong sequestration phenotype, these results suggest that further studies of CIR proteins will provide important new insights into the interaction of non-
P. falciparum malaria pRBC with host epithelia.
Authors' contributions
JK designed the study, planned and supervised all experiments. PE performed the experiments. JK and PE did bioinformatic and statistical analyses and wrote the manuscript. All authors read and approved the final version.