- Split View
-
Views
-
Cite
Cite
Pascal Michon, Jamie R. Stevens, Osamu Kaneko, John H. Adams, Evolutionary Relationships of Conserved Cysteine-Rich Motifs in Adhesive Molecules of Malaria Parasites, Molecular Biology and Evolution, Volume 19, Issue 7, July 2002, Pages 1128–1142, https://doi.org/10.1093/oxfordjournals.molbev.a004171
- Share Icon Share
Abstract
Malaria parasites invade erythrocytes in a process mediated by a series of molecular interactions. Invasion of human erythrocytes by Plasmodium vivax is dependent upon the presence of a single receptor, but P. falciparum, as well as some other species, exhibits the ability to utilize multiple alternative invasion pathways. Conserved cysteine-rich domains play important roles at critical times during this invasion process and at other stages in the life cycle of malaria parasites. Duffy-binding–like (DBL) domains, expressed as a part of the erythrocyte-binding proteins (DBL-EBP), are such essential cysteine-rich ligands that recognize specific host cell surface receptors. DBL-EBP, which are products of the erythrocyte-binding–like (ebl) gene family, act as critical determinants of erythrocyte specificity and are the best-defined ligands from invasive stages of malaria parasites. The ebl genes include the P. falciparum erythrocyte-binding antigen-175 (EBA-175) and P. vivax Duffy-binding protein. DBL domains also mediate cytoadherence as a part of the variant erythrocytic membrane protein-1 (PfEMP-1) antigens expressed from var genes on the surface of P. falciparum-infected erythrocytes. A paralogue of the ebl family is the malarial ligand MAEBL, which has a chimeric structure where the DBL domain is functionally replaced with a distinct cysteine-rich erythrocyte-binding domain with similarity to the apical membrane antigen-1 (AMA-1) ligand domain. The Plasmodium AMA-1 ligand domain, which encompasses the extracellular cysteine domains 1 and 2 and is well conserved in a Toxoplasma gondii AMA-1, has erythrocyte-binding activity distinct from that of MAEBL. These important families of Plasmodium molecules (DBL-EBP, PfEMP-1, MAEBL, AMA-1) are interrelated through the MAEBL. Because MAEBL and the other ebl products have the characteristics expected of homologous ligands involved in equivalent alternative invasion pathways to each other, we sought to better understand their roles during invasion by determining their relative origins in the Plasmodium genome. An analysis of their multiple cysteine-rich domains permitted a unique insight into the evolutionary development of Plasmodium. Our data indicate that maebl,ama-1, and ebl genes have ancient origins which predate Plasmodium speciation. The maebl evolved as a single locus, including its unique chimeric structure, in each Plasmodium species, in parallel with the ama-1 and the ebl genes families. The ancient character of maebl, along with its different expression characteristics suggests that MAEBL is unique and does not play an alternative role in invasion to ebl products such as EBA-175. The multiple P. falciparumebl paralogues that express DBL domains, which have occurred by duplication and diversification, potentially do provide multiple functionally equivalent ligands to EBA-175 for alternative invasion pathways.
Introduction
A malaria parasite throughout its erythrocytic life cycle relies on adhesive molecules to recognize, invade, and develop in this host cell. Invasion of the erythrocytes by Plasmodium merozoites requires specific interactions between parasite surface ligands and erythrocyte receptors (Chitnis and Blackman 2000 ). This cascade of events represents potential targets to reduce or eliminate the blood-stages of malaria parasites and their pathological outcomes through the use of vaccine-blockade strategies (Oaks et al. 1991 , pp. 169–210). Different Plasmodium species within humans and other hosts have evolved distinct invasion pathways to utilize unique sets of erythrocyte receptors. Some species, like Plasmodium falciparum and P. yoelii, have the ability to use multiple alternative pathways of invasion, whereas other species, like P. vivax, lack this flexibility and are dependent on interaction with a single receptor during the critical junction-forming step early in the process of invasion. The ability to invade using alternative receptors is presumed to be dependent on a parasite's ability to express different types of functionally equivalent ligands.
Among the best-known merozoite ligands involved in invasion are products of the erythrocyte-binding–like (ebl) family of adhesion molecules, which includes the P. falciparum erythrocyte-binding antigen-175 (EBA-175) and the P. vivax Duffy-binding protein (DBP). EBA-175 binds to the trypsin-sensitive, sialic acid–dependent residues on glycophorin A of human erythrocytes (Orlandi et al. 1992 ). Although the sialic acid–dependent pathway appears to be the dominant pathway in many laboratory-adapted parasitic lines (Mitchell et al. 1986 ; Perkins and Holt 1988 ), this may not be the case within natural parasite populations (Okoyeh et al. 1999 ). In contrast to EBA-175, the DBPs of P. vivax (Wertheimer and Barnwell 1989 ) and P. knowlesi (Haynes et al. 1988 ) interact with a chymotrypsin-sensitive, peptide epitope on the Duffy blood-group surface antigen. In spite of their differences in receptor specificity, an analysis of these three malarial adhesion molecules identified them as a part of a homologous family of malarial erythrocyte-binding proteins (Adams et al. 1992 ), which are now referred to as Duffy-binding–like erythrocyte-binding proteins (DBL-EBP) (Peterson et al. 1995 ) (fig. 1A ). The members of this ebl family have a similar exon-intron structure with conserved splicing boundaries, indicating a common evolutionary origin (Adams et al. 1992 ). Regions II and VI are the cysteine-rich regions of the DBL-EBP extracellular domains that have numerous conserved cysteine and hydrophobic amino acid residues, suggesting a conserved, functionally important three-dimensional structure. Most important are the conserved DBL domains of region II, which are confirmed as erythrocyte-binding ligand domains (Chitnis and Miller 1994 ; Sim et al. 1994 ). The DBP has only a single copy of the DBL domain, whereas the EBA-175 region II has two DBL domains (F1 and F2), although these tandem DBL domains appear to function as a single ligand domain (Sim et al. 1994 ). Region VI is a smaller, more highly conserved domain immediately preceding the transmembrane domain, but it has no known function. These data indicate that the ebl genes have evolved within Plasmodium species to have specificity for different types of erythrocyte receptors while still maintaining homologous functions in the invasion process.
Multiple ebl genes occur within a few species of Plasmodium (fig. 1 ), providing a genetic basis for ligand diversity that can facilitate alternative pathways of invasion. In a cloned line of P. knowlesi, two very similar ebl genes, encoding erythrocyte-binding proteins of unique binding specificity, were identified, in addition to the dbp (Adams et al. 1990 ; Chitnis and Miller 1994 ; Ranjan and Chitnis 1999 ). Differences in the erythrocyte-binding activity of these two P. knowlesi DBL-EBPs and that of the DBP are attributable to the differences in the DBL domains. Even though these three P. knowlesiebl genes share significant nucleotide identity, they are located on separate chromosomes (Adams et al. 1990 ). The organization pattern of the ebl genes in this species, closely related but on separate chromosomes, suggests a relatively recent dispersal of the duplicated genes (Ruvkun and Hobert 1998 ). Additional analysis of other P. knowlesi lines indicates that the diversity of the P. knowlesiebl was generated by intergenic recombination, including DBL domain exchange (Prickett et al. 1994 ). Another close homologue to the P. vivax DBP was characterized in the similar simian malaria parasite, P. cynomolgi (Okenu et al. 1997 ). Its gene structure and amino acid composition are very similar to the ebls found in P. vivax and P. knowlesi, including only a single DBL domain in region II. A second apparent cross-hybridizing gene was not isolated, but its presence suggests that duplication and diversification is an inherent property of the ebl.
In P. falciparum multiple diverse ebl, present mostly on separate chromosomes, have the consensus gene structure and cysteine-rich domains but lack any significant nucleotide identity with each other (Adams et al. 2001 ). ebl-1, present on chromosome 13, was the first paralogue of eba-175 that was identified (Peterson et al. 1995 ; Peterson and Wellems 2000 ). Interestingly, the ebl-1 was not found in the HB3 line of P. falciparum that is known to exhibit a slow proliferation phenotype in vitro, associated with the loss of a subtelomeric region of chromosome 13 (Hinterberg et al. 1994 ). Three other novel P. falciparumebl genes were identified through the Malaria Genome Project: baebl,pebl, and jesebl (Adams et al. 2001 ; Mayer et al. 2001 ; Thompson et al. 2001 ; Triglia et al. 2001 ). BAEBL (a.k.a. EBA-140), an ebl product recently characterized as a P. falciparum erythrocyte-binding protein, was identified as a potential ligand for the alternative pathway of merozoite invasion into the erythrocytes, recognizing a sialic acid–dependent erythrocyte surface epitope (Mayer et al. 2001 ; Thompson et al. 2001 ). Although there is no consensus as yet on the exact identity of the receptor for BAEBL, it is clearly different from the sialic acid–dependent receptor on glycophorin A recognized by EBA-175. Therefore, the data for P. knowlesi and P. falciparum demonstrate that gene duplication is a common characteristic of the Plasmodiumebl family, and genome amplification leads to diversification, providing the molecular basis for functionally redundant ligands for alternative invasion pathways.
In a separate line of evolution, extensive tandem duplications and diversifications have occurred with the DBL domains in the P. falciparumvar gene family, creating an enormous and diverse family of adhesion molecules (Cooke et al. 2000 ). Multiple tandem copies of the variant DBL are found in the adhesive molecule, erythrocytic membrane protein-1 (PfEMP-1), encoded by these var genes. PfEMP-1 expressed onto the surface of P. falciparum-infected erythrocytes mediates cytoadherence to endothelial and erythrocyte receptors (Baruch et al. 1995 ; Smith et al. 1995 ; Su et al. 1995 ). There are many var copies in each organism, but a parasite expresses only one at a time. Each PfEMP-1 has several variant copies of the DBL domain in tandem, and each can bind a different receptor. These variant domains were recently grouped as five different classes (α, β, γ, δ, and ε) according to their sequence similarity that also corresponds to shared receptor recognition phenotype (Smith et al. 2000b ), indicating again a pattern of gene duplication and diversification.
The maebl is a unique ebl paralogue that was initially identified in rodent malaria parasites, P. yoelii and P. berghei (Kappe et al. 1997 , 1998 ). The P. yoelii and P. bergheimaebl has the characteristics of the ebl family, except for its distinct ligand domains (fig. 1 ). It is a type-I transmembrane protein with a carboxyl cysteine-rich (C-cys) region that is homologous to region VI of the DBL-EBPs. However, the MAEBL amino cysteine-rich domain, which occurs as a tandem duplication (M1 and M2), has no similarity to the consensus DBL domain of the ebl family. Instead, the MAEBL ligand domain has a partial similarity with the cysteine domains 1 and 2 (domains 1/2) of the Plasmodium and Toxoplasma apical membrane antigen-1 (AMA-1). In erythrocyte cytoadherence assays, the M1 and M2 domains of the P. yoelii MAEBL bound mouse erythrocytes and so were shown to be functionally equivalent to the DBL as ligand domains.
The relationship of AMA-1 to the MAEBL ligand domain is intriguing because AMA-1 is a ubiquitous molecule in Plasmodium and in the distantly related apicomplexan T. gondii (Donahue et al. 2000 ; Hehl et al. 2000 ). AMA-1 is a type-I transmembrane protein located in the neck of the malaria merozoite rhoptries and later on the surface of the invasive merozoite (Peterson et al. 1989 ; Crewther et al. 1990 ; Narum and Thomas 1994 ). The putative ectodomain of AMA-1 is defined by three cysteine-rich domains elucidated by disulfide bond patterns (Hodder et al. 1996 ). It is just the portion of the AMA-1 ectodomain conserved with MAEBL, domains 1/2, which is conserved between the Plasmodium and Toxoplasma AMA-1 (Donahue et al. 2000 ; Hehl et al. 2000 ). Previously, a role for AMA-1 as a parasite ligand was suggested because monovalent antibody fragments could inhibit merozoite invasion into erythrocytes (Thomas et al. 1984 ). Recently, we confirmed that domains 1/2 of AMA-1 could mediate adhesion to erythrocytes (Fraser et al. 2001 ). In our study using P. yoelii, AMA-1 of this rodent malaria parasite bound mouse but not human erythrocytes. Interestingly, significant differences in the erythrocyte receptor-binding specificity of the P. yoelii MAEBL and AMA-1 were apparent in a preliminary comparative analysis of these ligand domains (J. Adams et al., unpublished data).
Differences in the expression and location of merozoite molecules, on the surface or sequestered in one of the apical organelles (micronemes, rhoptries, dense granules), are considered to be indicators of functional differences during invasion. The ebl products, PkDBP and EBA-175, are expressed during the final stages of schizont development and are localized to the micronemes (Adams et al. 1990 ; Sim et al. 1992 ). Micronemes are the last of the three apical organelles to form and do not appear until the end of schizont development. In contrast, the P. yoelii and P. falciparum MAEBL appear to localize to the rhoptries and on the surface of mature merozoites (Noe and Adams 1998; Blair et al. 2002b ). This difference in localization is supported by an analysis of the P. yoelii MAEBL, which determined that MAEBL is expressed at the onset of schizont development (Noe et al. 2000 ). Soon after MAEBL is expressed, it is rapidly processed to remove the M1 domain, leaving only M2 as the functional ligand domain in the invasive merozoite. The differences in expression and localization between the MAEBL and the DBL-EBP are consistent with different functions for these erythrocyte-binding molecules.
The structure of MAEBL originally isolated from the rodent malaria parasites suggested a molecule likely to be involved in invasion. Therefore, we were interested to see whether MAEBL is present in Plasmodium infecting humans and primate models. In the present study, we describe the identification of maebl in the two closely related malarial species P. vivax and P. knowlesi, in addition to the P. falciparummaebl that we characterized separately. The characterization of maebl from these evolutionarily distinct species was important in order to include the phylogenetic analysis of other ebl genes as well as ama-1. This comparative phylogenetic study was made using deduced, conserved cysteine-rich domains. We sought to gain an insight into the roles of these adhesion molecules in parasite invasion through an analysis of the relative origins of maebl,ebl, and ama-1 and their intraspecies divergence, especially within P. falciparum. The maebl is of ancient origin and evolved as a single locus, including its unique chimeric structure, in each Plasmodium species, in parallel with the ama-1 and ebl genes families. The ancient character of maebl, along with its different pattern of expression in parasites, suggests that MAEBL does not have a redundant or alternative role in invasion similar to that of ebl products such as EBA-175. On the other hand, the multiple P. falciparumebl paralogues, which have occurred by duplication and diversification, potentially provide multiple, functionally equivalent ligands to EBA-175 for alternative invasion pathways.
Materials and Methods
Parasites and DNA Material
The genomic DNA of P. knowlesi (Malayan H) and P. vivax (Sal-1) (provided by W. E. Collins, Center for Disease Control, Atlanta, Ga.) was obtained by phenol-chloroform extraction following a standard protocol.
PCR Amplification and Cloning of P. knowlesi and P. vivaxmaebl
A region encoding the MAEBL amino cysteine-rich domain M2 was chosen as a target for PCR amplification from P. vivax and P. knowlesi because of its high degree of conservation in the rodent malarial species (Kappe et al. 1998 ). Oligonucleotide primers 5′-AATCCTCAAGCCGAATATATGGATAGGTTTGATAT-3′ and 5′-CTTTTTATTTATAAGACTTTTGCATTTTCC-3′, derived from the P. yoeliimaebl sequence, were used to PCR-identify this region using the P. knowlesi and P. vivax genomic DNA. The PCR reactions using standard reagents were subjected to 1 cycle of 45 s at 94°C and 35 cycles of 15 s at 94°C, 30 s at 42°C, 1 min at 72°C. A portion of the P. knowlesimaebl M1 domain was obtained by PCR using antisense internal primer (5′-CCCTCTCCCAATTTTCCTTCCTTTTTC-3′) with sense primer (5′-AGAGGTAGTTGTCCAGATTATGG-3′) deduced from the P. falciparummaebl M1 (P. Michon et al., unpublished data). PCR conditions were the same as mentioned earlier in this article, except for the annealing temperature (55°C).
Splinkerette used with P. knowlesi Genomic DNA
A P. knowlesi genomic splinkerette library was made as follows: total P. knowlesi genomic DNA was fragmented with HindIII and ligated with a 15 × molar excess of HindIII splinkerette (Devon et al. 1995 ) in 20 μl for 5 h at room temperature. Oligonucleotide primer VII (5′-CGAATCGTAACCGTTCGTACGAGAA-3′), specific to the splinkerette adapter, was used with P. knowlesimaebl sequence-specific primers 5′-GTTCCTTCCTTATGAGCGAACTATGATTTTCGTCC-3′ and 5′-GGATTGGACATATGTTACATCATTTATTAGGCCTG-3′ to amplify the flanking regions of the maebl M2 domain (upstream and downstream, respectively) using Long-range PCR (Expand™ Long Template PCR System, Roche Molecular Biochemicals) in a 50-μl reaction under conditions recommended by the manufacturer. PCR conditions combining Hot Start (Chou et al. 1992 ) and Touchdown (Don et al. 1991 ) were as follows: 10 cycles of 10 s at 92°C, 30 s at 67°C, 5 min at 68°C with a 1°C decrease in annealing temperature at each cycle; 15 cycles of 10 s at 92°C, 30 s at 63°C, 5 min at 68°C; 1 cycle of 7 min at 68°C.
Inverse PCR
The genomic DNA of P. vivax and P. knowlesi was fragmented with EcoRI and HindIII, respectively, phenol-chloroform extracted, ethanol precipitated, and resuspended in H2O. Five hundred nanograms was ligated with 1 μl T4 DNA ligase (Life Technologies, Inc.) overnight at 14°C in 100 μl total volume. After phenol-chloroform extraction and ethanol precipitation, the pellet was resuspended in 10 μl H2O. One microliter was used as template for inverse PCR (IPCR) using primers 5′-AAATTCACCCACTTCATTGCC-3′ and 5′-GTTTGAAGAAGCTACCTTTGATGAG-3′ for P. vivax and primers 5′-GTTCCTTCCTTATGAGCGAACTATGATTTTCGTCC-3′ and 5′-TTTATTTTTCTACCTAGGTTTAATGAAGCTAC-3′ for P. knowlesi. The reaction mixture contained 1.5 mM MgCl2, 200 μM each dNTP, 0.5 units PLATINUM Pfx DNA polymerase (Life Technologies, Inc.), 300 nM of each oligonucleotide primer in Pfx buffer to a final volume of 50 μl. PCR conditions were 1 cycle of 3 min at 94°C; 10 cycles of 15 s at 94°C, 30 s at 65°C, 7 min at 65°C with a 1°C decrease in annealing temperature at each cycle; 30 cycles of 15 s at 94°C, 30 s at 55°C, 7 min at 65°C; 1 cycle of 7 min at 65°C.
Cloning and DNA Sequencing Analysis
All PCR products were cloned in pCRII (TA or TOPO-TA cloning kits, Invitrogen) according to the manufacturer's protocols. Cloned DNA was sequenced by the dideoxy chain termination method (Sanger et al. 1977 ), using the Thermo Sequenase sequencing kit on an ALFexpress™ (Amersham Pharmacia Biotech). The nucleic acid and deduced amino acid sequences were aligned using ALIGNMENT (Geneworks 2.1, Intelligenetics). Sequence similarities were searched online using BLAST (Altschul et al. 1990 ) (http://www.ncbi.nlm.nih.gov/BLAST/). Leader sequences and transmembrane domains were predicted using online software programs SignalP V1.1 (Nielsen et al. 1997 ) and HMMTOP 2.0 (Tusnady and Simon 1998 ), respectively.
Southern Blot Analysis of maebl
The genomic DNA of P. knowlesi fragmented with restriction enzymes (EcoRI, HindIII, BamHI, NsiI, and PstI) was separated by agarose gel electrophoresis, depurinated (0.25 M HCl), denatured (0.5 M NaOH-1.5 M NaCl), blotted onto nylon membranes (GeneScreen Plus, DuPont) in 20 × SSC (3 M NaCl, 300 mM sodium citrate, pH 7.0), and the membrane baked for 2 h at 80°C in a vacuum oven. A PCR-amplified fragment of the P. knowlesimaebl amino cysteine-rich domain M2 (described previously in this article) was radiolabeled by random priming reaction with (γP32-dCTP) using the Klenow fragment of DNA polymerase I according to the manufacturer's instructions (BRL). The membrane was prehybridized in 6 × SSC, 20 mM Na2PO4, pH 6.8, 5 × Denhardt's solution, 0.5% SDS, 100 mg/ml heparin for 2 h at 40°C, and then probed at 60°C overnight. The blot was then washed 5 times with 2 × SSC, 0.5% SDS, and incubated in 0.2× SSC, 0.1% SDS at 60°C for 30 min. The membrane was exposed to a X-ray film (NEF-496, Dupont) overnight at −70°C.
The genomic DNA of P. vivax was fragmented with restriction enzymes (EcoRI, BamHI, NsiI), a Southern blot made and processed as described previously in this article. The membrane was hybridized by a (γP32-dCTP) radiolabeled fragment of the P. vivaxmaebl C-cys region, PCR-amplified using primers 5′-AAATTAGATAAGAATGAGTATATAAAGAGAG-3′ and 5′-TGAAGCTTTGCTCTTCCG-3′ as described previously in this article.
Amino Acid Sequences
The deduced amino acid sequences from different malarial species were used for the comparison of the amino cysteine-rich domains of MAEBL (M1 and M2) with AMA-1 (domains 1 and 2): MAEBL—P. berghei ANKA (AAC05367), P. cynomolgi (AAF61431), P. falciparum (AY042084), P. vivax (AY042083), and P. knowlesi (AY042082), P. y. yoelii YM (AAC05366); AMA-1—P. berghei ANKA (U45969), P. c. chabaudi (U49743), P. fragile (M29898), P. cynomolgi (X86099), P. knowlesi (M37854), P. falciparum 3D7 (U65407), P. reichenowi (AJ252087), P. vivax (L27503), P. y. yoelii YM (U45970), and Toxoplasma gondii (AAB65410).
Sixteen amino acid sequences were used for the comparison of the C-cys domains: P. y. yoelii YM MAEBL (AAC05366), P. berghei ANKA MAEBL (AAC05367), P. falciparum MAEBL (AY042084), P. knowlesi MAEBL (AY042082) and P. vivax MAEBL (AY042083); P. vivax Duffy-binding protein (P22290), P. cynomolgi erythrocyte-binding protein (Y11396), P. knowlesi DBP alpha (P22545), P. knowlesi EBP beta (P50493), P. knowlesi EBP gamma (P50494), P. falciparum (3D7) EBA-175 (AAA75179), P. reichenowi EBP (CAB96159), P. falciparum (3D7) EBL-1 (AAD33018). Sequence data for JESEBL (AB080796, AB080797), PEBL (AB080994), and BAEBL (AAK49521, AAK55485) were deduced from the genomic sequence produced by the P. falciparum genome consortium group, for chromosome 1, 4, and 13, respectively, at The Sanger Centre (web site http://www.sanger.ac.uk/Projects/P_falciparum).
The deduced amino acid sequences of five DBL types (α, β, γ, δ, and ε) from four var genes, Dd2var1 (AAA75396), FCR3var3 (AAA75397), It-R29 (CAA73831), and MCvar1 (AAB60251) were used for comparison with DBL domains of 11 DBL-EBPs (fig. 1C ).
Phylogenetic Tree Construction
Protein sequence alignments were made using Clustal W (1.81) (Thompson et al. 1994 ) online (http://www.ebi.ac.uk/clustalw/) and transferred to SeqApp1.9 for Macintosh (obtained at http://iubio.bio.indiana.edu/soft/molbio/seqapp/) to allow manual adjustments. Regions of the protein sequences where it was impossible to produce a single reliable alignment across all taxa were excluded from these analyses via an exclusion set defined using MacClade 3.08. Phylogenetic trees were constructed by neighbor-joining (NJ) method and maximum parsimony (MP) using PAUP*4b6 (*Phylogenetic Analysis Using Parsimony and other methods) for Macintosh (Swofford 2000 ). The NJ method was used because computer simulations have shown that this distance measure is more efficient at reconstructing the true tree when sequences are relatively distantly related (Saitou and Nei 1987 ); NJ distances were calculated on the basis of mean character difference. The MP method was also used because a recent study recommended the use of parsimony instead of distance-based methods to compare paralogues (i.e., molecules originating from gene duplication) (Thornton and Desalle 2000 ). A consensus tree was obtained from the best trees in the MP method. Bootstrap values were calculated with 1,000 pseudoreplicates for both methods and were expressed as percentages on the tree branches. Nodes receiving bootstrap support <50% were collapsed and presented as unresolved polytomies.
Nucleotide sequence data for the 3′cys region, aligned according to the protein sequence alignment, were also analyzed by NJ, using genetic distances based on a general time-reversible model and ML analysis, also using a general time reversible model, with substitution rates being estimated from the data.
RESULTS
Identification and Cloning of the P. knowlesi and P. vivaxmaebl
The maebl was identified in the two closely related malarial species P. vivax and P. knowlesi. Various PCR-based methods were used to isolate the full-length coding and untranslated regions (UTR) of the maebl from P. vivax and P. knowlesi. The deduced amino acid sequences of these cloned PCR products showed identity with the P. yoelii and P. berghei MAEBL, especially in the highly conserved cysteine-rich domains. The 5′ UTR of P. knowlesimaebl was obtained from a HindIII splinkerette (Devon et al. 1995 ) library, whereas exons 4, 5, and the 3′ UTRs in both species were obtained by IPCR. The remaining sequences of P. knowlesi and P. vivaxmaebl were obtained by PCR amplification of internal regions (data not shown). A small portion of the sequence in the middle of the repeat region of P. knowlesimaebl was not confirmed. Southern blot hybridizations of P. vivax and P. knowlesi genomic DNA identified a single copy of the maebl in each species (fig. 2 ).
Gene Structure of P. knowlesi and P. vivaxmaebl
Five exons were identified in the P. vivax and P. knowlesimaebl on the basis of the homology to the P. yoelii YM maebl and other Plasmodiumebl (Adams et al. 1992 , 2001 ; Kappe et al. 1998 ). These exons included a putative signal peptide, a large extracellular domain, a putative transmembrane domain, and a cytoplasmic tail (fig. 3 ). M1 and M2, the duplicated amino cysteine-rich ligand domains with similar identity to AMA-1 domains 1/2, spanned the end of exon 1 through the first half of exon 2. The remaining portion of exon 2 encoded a tandem repeat region of charged amino acids (fig. 3 ) and ended with the C-cys domain homologous to region VI of the ebl family (Adams et al. 1992 ). The combined lengths of the exons and introns of the P. knowlesi and P. vivaxmaebl genes were different (7.0 and 6.5 kb, respectively). The different lengths of these maebl genes were the result of the repeats region and different intron sizes. The deduced amino acid sequences had a high degree of identity even outside the conserved cysteine-rich domains (fig. 3 ).
Comparison of MAEBL M1 and M2 Domains with AMA-1 Domains 1/2
The deduced amino acid sequences of the MAEBL domains M1 and M2 were aligned for all Plasmodium species known so far to possess maebl (data not shown). All cysteine residues (16) were conserved in number and position. The MAEBL domains M1 and M2, and AMA-1 domains 1/2 in P. vivax and T. gondii were aligned with manual adjustment (fig. 4 ). Because only 10 of the 16 MAEBL cysteine residues were conserved in AMA-1, it was possible to align just a part of the length of MAEBL M1 and M2 domains to the entire AMA-1 domains 1/2 as shown in figure 5 .
A multiple alignment was produced for all MAEBL M1 and M2 domains and AMA-1 domains 1/2 from 9 Plasmodium species and T. gondii. Phylogenetic analysis used only regions that could be aligned. Unrooted phylogenetic trees were constructed using both the NJ and MP methods with bootstrap analysis (1,000 pseudoreplicates). Both methods gave almost identical tree topologies (i.e., almost identical branching patterns and clade structure). The MAEBL M1 and M2 domains clustered separately from AMA-1 domains 1/2 (bootstrap value 100%) (fig. 5 ). All MAEBL M1 domains grouped together separately from the M2 domain cluster (bootstrap values 99% and 100%, respectively). Clustering patterns within M1, M2, and AMA-1 clades reflected the phylogeny of Plasmodium species proposed earlier (Escalante and Ayala 1994 ). Similarly, the grouping profile in the AMA-1 clade, which includes more Plasmodium taxa (and T. gondii), is consistent with previous studies on AMA-1 from Plasmodium spp. (Verra and Hughes 2000 ).
Comparison of the ebl C-cys Regions
The deduced C-cys regions of known MAEBL were aligned with 11 DBL-EBP family members, which include proteins of known function like the P. vivax and P. knowlesi DBP and P. falciparum EBA-175. The ebl homologues have been described in other species, P. cynomolgi (Okenu et al. 1997 ) and P. reichenowi (Ozwara et al. 2001 ), as well as the five P. falciparumebl that have duplicated DBL domains (Adams et al. 2001 ; Mayer et al. 2001 ; Thompson et al. 2001 ; Triglia et al. 2001 ) (fig. 1 ). Multiple alignments of these C-cys showed amino acid conservation within this domain among different paralogues and across species (data not shown). All cysteine residues were conserved in number (eight) and position, except for EBL-1, which has only four cysteines in this region. Other amino acids frequently conserved have charged (Arg, Glu) and hydrophobic (Leu, Ile, Tyr, Phe) side groups.
Phylogenetic trees constructed for C-cys showed almost identical branching patterns and clade structure using either the MP or NJ method. All DBL-EBPs having a single DBL domain clustered together with both methods (bootstrap value 98% and 100%, respectively). EBA-175 was found associated with an EBP from P. reichenowi (bootstrap value 100%). This clade grouped with JESEBL but with less bootstrap support (MP 69%, NJ 57%). MAEBL C-cys regions of different species clustered together using the NJ method (bootstrap 80%), but the consensus tree using MP clustered in two groups (fig. 6 ). Nevertheless, an examination of the three most parsimonious trees (MPTs) indicated that two out of the three MPTs resolved MAEBL as a monophyletic clade (data not shown); in the third MPT, long-branch attraction (Felsentein 1978 ; Hendy and Penny 1989 ) between EBL-1, PEBL, and two of the more divergent MAEBL sequences may have disrupted the topology of the MAEBL clade. Both methods were unable to unequivocally group EBL-1, BAEBL, or PEBL either together or with other clades or taxa.
NJ and MP analyses of nucleotide sequence data for the 3′-cys region, aligned according to the protein sequence alignment, produced trees which reconfirmed MAEBL as a monophyletic clade. However, despite this, internal genetic distances within the MAEBL clade were considerable, with PkMAEBL and PvMAEBL being almost as genetically distinct from the other MAEBL taxa as were the majority of other taxa from each other (data not shown). Such a result indicates that even within the MAEBL clade, considerable and possibly ancient diversity exists.
Comparison of the DBL Domains
DBL domains are found in products of the ebl family, encoding DBL-EBP as well as PfEMP-1 from the P. falciparumvar superfamily (fig. 1 ). An analysis of the DBL domains' relatedness provides a useful comparison for the phylogenetic trees created for the corresponding C-cys domains. The P. vivax DBP and its homologues in P. knowlesi and P. cynomolgi have a single DBL domain, whereas the P. falciparumebl (e.g., EBA-175) and the homologue in P. reichenowi encode tandem copies of the DBL domain (fig. 1 ). The var genes, comprising nearly 50 copies scattered across all 14 chromosomes, encode from 2 to 7 DBL domains. Five DBL domain types (α, β, γ, δ, and ε) (Smith et al. 2000b ) deduced from four var genes were compared in this study with the ebl family DBL domains mentioned previously in this article.
A multiple alignment showed that among the 12 conserved cysteines of the DBL-EBPs region II (Adams et al. 1992 ), only 10 of them aligned with cysteine residues of PfEMP-1 as already described (Smith et al. 2000b ). Some of these 10 common cysteines are not found in certain var types (e.g., the β type is missing cysteines 2 and 3; types δ and γ are missing cysteine number 5 [data not shown]). Some cysteine residues are also missing from P. falciparum EBL-1 and P. reichenowi EBP F2 domain (PrEBP F2). Additional cysteine residues are found in certain var types as described (Smith et al. 2000b ).
Interestingly, one of the extra cysteines, which is present 14 residues upstream of cysteine 7 in the EBA-175 F2 domain (Adams et al. 1992 ), is perfectly conserved throughout all var types (cysteine 6a) (Smith et al. 2000b ) and in all F2 domains of the DBL-EBPs from our study as well as in the JESEBL F1 domain (data not shown). The second extra cysteine present in the EBA-175 F2 domain is situated in a region not shared with the var DBL domains. This cysteine is also found in all F2 domains considered in the present study and JESEBL F1 (data not shown). Other amino acid residues are conserved (5 tryptophan, 2 arginine, 1 aspartic acid and 1 glycine as well as other charged and hydrophobic residues).
Phylogenetic trees were constructed with all DBL domains by the MP and NJ methods. Both methods gave similar topology, although identification of clusters was rendered difficult on the NJ tree because of the number of taxa used (data not shown). The PvDBP and its homologues in P. knowlesi and P. cynomolgi, all having a single DBL domain, clustered together (bootstrap values 100% in both methods). All the F1 domains clustered with the PvDBP clade, apart from the F2 domains that grouped with the DBL domains found in PfEMP-1 (bootstrap: MP 88%, NJ 96%—data not shown) (fig. 7 ). With both methods, the var DBL domains grouped according to their type as observed earlier (α with β, δ with γ) (Smith et al. 2000b ); α and δ also formed their own separate clade. DBL domains of the ε type were unresolved. As observed for the C-cys domain, EBA-175 and PrEBP clustered together (bootstrap 99%) for both domains F1 and F2. The phylogenetic relationships for both duplicated domains (F1 and F2) of PEBL, BAEBL, JESEBL, and EBL-1 were either unresolved or were included in clusters but with very low bootstrap values (indicating nonsignificant support).
Discussion
Patterns of Invasion and the Adhesion Molecules of Plasmodium
The malarial merozoite invades the erythrocytes through a complicated multistep process mediated by the molecular interactions between the parasite and host cell erythrocyte. A progressive pattern is evident in the invasion process, with surface contact leading to apical reorientation, junction formation, and release of the apical organelles. Although many proteins associated with the surface and the apical organelles are identified, major questions remain unanswered about the function and the sequence of interactions of most of these molecules. Understanding the origin and evolution of these molecules will provide insights into their function and help devise ways to effectively block invasion.
The invasion paradigm assumes that the initial interactions are relatively nonspecific and reversible until the process reaches a decisive step to enter a host cell (i.e., junction formation), which is dependent on the interaction between one or more specific sets of parasite ligands and specific erythrocyte receptors. This model is based on observations with P. vivax and P. knowlesi, where the presence of a single receptor, the Duffy blood-group antigen, is clearly required for these parasites to invade human erythrocytes (Miller et al. 1975 , 1976 ). When the Duffy blood-group antigen is absent, these parasites cannot invade human erythrocytes. The absence of P. vivax malaria from most of Africa, where the Duffy blood-group negativity is near fixation, clearly supports the model. The DBP mediates the step of invasion in P. vivax, whereas in P. falciparum the homologous EBA-175 is presumed to play the same role by the recognition of a sialic acid–dependent epitope on glycophorin A. However, in the case of P. falciparum, alternative pathways are evident, and many parasite clones can invade human erythrocytes when this principal receptor is absent (Mitchell et al. 1986 ; Perkins and Holt 1988 ; Dolan et al. 1990 , 1994 ). Ligands similar to EBA-175 are presumed to mediate P. falciparum invasion via the alternative pathways to glycophorin A. In other Plasmodium species, additional molecules are identified that play important roles in determining host cell specificity. Such molecules include a family of erythrocyte-binding proteins homologous to the P. vivax reticulocyte-binding proteins (Galinski et al. 1992 ), which includes the P. yoelii high–molecular weight rhoptry proteins (p235) encoded by the y235 multigene family (Freeman et al. 1980 ; Keen et al. 1990 ; Ogun and Holder 1996 ; Preiser et al. 1999 ) and the P. falciparum NBP-1 (Rayner et al. 2001 ). Even though these molecules can play a significant role in determining erythrocyte specificity in some species and strains, how they function and when they act in the invasion process is not certain.
When first identified, the chimeric structure of MAEBL suggested that it might be a molecule involved in alternative invasion pathways because it is homologous to the ebl gene products, except for its amino cysteine-rich erythrocyte-binding domain. The similarity of the ligand domain to AMA-1, and not to the consensus DBL ligand domain of the ebl, suggested that MAEBL could bind alternative receptors while still interacting at its carboxyl end with a putative membrane-associated signaling apparatus. All Plasmodiummaebl studied so far, including the major human malaria parasites, are single-copy genes that have the same consensus 5-exon gene structure with three cysteine domains. Because of this high degree of conservation of maebl, its singularity in the genome, and its ancient origin, it seems unlikely that MAEBL originated and has been conserved as an alternative ligand distinct from the ebl products. Further, our phylogenetic analysis indicates that whereas maebl evolved in a highly conserved manner as a single-copy gene in each species, ebl evolved in a less well conserved manner, duplicating and diversifying in the Plasmodium genome. MAEBL, as well as other transmembrane molecules on the surface of invasive merozoites, may play an important early role in the initial interactions with the host cell by interacting with a ubiquitous or a common host cell surface molecular structure as its receptor. The commonality of its receptor would not necessitate the need for adaptive radiation in the family because the parasite adapts to a new host; in this regard, MAEBL and AMA-1 may be similar. In contrast, the diversity of ebl products, among and within Plasmodium species, reflects a need to interact with (i.e., recognize) unique host cell receptors at a critical step in the invasion process. The demonstrated ability to duplicate and diversify enables the malaria parasite to adapt to new host receptors and develop a repertoire of adhesion molecules for this critical step of the invasion process. Therefore, in some Plasmodium species, like P. falciparum, this has led to the evolution of functionally redundant ligands that can facilitate alternative pathways of invasion.
Similarities Between MAEBL and AMA-1
The conservation of the distinct tandem copies of the MAEBL cysteine-rich domains suggests that both these ligand domains are functionally important and that MAEBL has a similar important function among all Plasmodium species. Phylogenetic analysis of the putative MAEBL and AMA-1 ligand domains (domains 1/2) indicated independent ancient origins for these molecules. Multiple sequence alignments of each of the MAEBL domains M1 and M2 showed marked similarity to portions of the combined AMA-1 domains 1 and 2 (fig. 4 ). Phylogenetic analysis, including protein sequences of these domains from a variety of Plasmodium species and one for the T. gondii AMA-1, showed that all AMA-1 taxa clustered apart from MAEBL and that each of the MAEBL domains M1 and M2 grouped separately from each other (fig. 5 ). This indicated that both AMA-1 and MAEBL are ancient molecules predating Plasmodium speciation. Similarly, the duplication event that created the M1 and M2 domains originated before the split among Plasmodium species. When comparing the grouping of the Plasmodium species, our data were consistent with a previous phylogenetic analysis based on ribosomal genes (Escalante and Ayala 1994 ).
Functional similarities exist for the MAEBL M1 and M2 domains and AMA-1 domains 1/2. The P. yoelii MAEBL was identified as an erythrocyte-binding protein in an in vitro cytoadherence assay (Kappe et al. 1998 ). This same type of assay demonstrated that P. yoelii AMA-1 domains 1/2 function as an erythrocyte-binding domain (Fraser et al. 2001 ). Interestingly, Plasmodium and T. gondii were more reliably aligned in this putative ligand domain (contiguous 1 and 2) compared with the rest of the AMA-1 molecule (Donahue et al. 2000 ; Hehl et al. 2000 ), suggesting that this may be a ligand domain in Toxoplasma also. Studies in both Plasmodium and Toxoplasma have shown that anti–AMA-1 antibodies can block parasite invasion. Antibodies against T. gondii AMA-1 domains 1 and 2 significantly reduced host cell invasion in vitro (Hehl et al. 2000 ). Similar but more significant results were obtained in Plasmodium as antibodies against AMA-1 blocked erythrocyte invasion by P. knowlesi (Thomas et al. 1984 ) and immunization with a recombinant AMA-1 prevented infection of a mouse with P. chabaudi (Anders et al. 1998 ). Our phylogenetic analysis infers from their conservation and ancient origins that these independent cysteine-rich domains of MAEBL and AMA-1 play fundamental roles in the parasites' development. The exact nature of their receptors is not known, but in P. yoelii MAEBL and AMA-1 ligand domains have distinct binding preferences for erythrocyte receptors in preliminary studies (P. Michon et al., unpublished data). The nonligand portions of MAEBL and AMA-1 lack any apparent similarity to each other and so would not interact with other parasite surface or cytoplasmic molecules in a similar functional manner. Because both molecules are present on the surface of the invasive merozoite when it initially contacts the erythrocyte, we interpret their binding differences not in terms of alternative function but as a coordinated interaction between ligands. From these data we infer that conservation of the cysteine-rich domains of MAEBL (M1, M2) and AMA-1 (domains 1/2) is related to their inherent functional properties.
DBL Domains of DBL-EBP and PfEMP-1
The DBL domain is present across Plasmodium species expressed as the ligand domain of the merozoite DBL-EBP and in the variant surface protein of P. falciparum PfEMP-1. Multiple alignments of these DBL domains confirmed observations made in previous studies on the conservation of amino acid residues, in particular cysteines (typically 12), present in the ebl family (Adams et al. 1992 ; Okenu et al. 1997 ; Mayer et al. 2001 ; Thompson et al. 2001 ; Triglia et al. 2001 ). The ebl family can be divided into molecules having a single DBL domain, which includes the PvDBP and the related simian malaria parasite proteins plus the putative single-copy ebl of P. yoelii, and molecules having a region II with duplicated DBL domains (F1, F2), such as the P. falciparum and P. reichenowiebl (fig. 1 ). All the main lineages of the DBL domains, including the PvDBP clade, the EBA-175-PrEBP-F1 clade, and the single EBL-1, PEBL, JESEBL, and BAEBL branches, could not be resolved (fig. 7 ) similar to the phylogenic tree for the C-cys region (fig. 6 ). Almost identical polytomy was observed for the F2 domains and associated var DBLs, and such a topology suggests their rapid divergence originating from multiple gene duplication events. In this phylogenetic analysis, five var-type DBLs (α, β, γ, δ, and ε) (Smith et al. 2000b ) were included. In these var domains, only 10 out of the 12 consensus cysteine residues were typically conserved with the DBL-EBP family. Additional cysteines were also observed as described earlier (Smith et al. 2000b ), plus a few other differences in the number of cysteine residues in EBL-1 and P. reichenowi EBP F2 (data not shown). The multiple alignments showed that an additional cysteine residue, which was described at the position 6a in the var DBLs (Smith et al. 2000b ), was also found in all F2 domains examined in this study as well as in the JESEBL F1 domain. The differences in number and position of cysteine residues in the DBL domains imply a possible rearrangement of the disulfide bonds leading to a different three-dimensional conformation. This could reflect an evolutionary divergence in the functional properties associated with each of these domains.
Clustering together of the var DBL types is in accordance with the results from another study (Smith et al. 2000b ) (fig. 7 ). The phylogenetic relatedness of the DBL correlates with known ligand-receptor interactions attributed to some of these DBL domains, indicative of a pattern of adaptive radiation, and is consistent with a sequential pattern of repeated duplication and diversification. For example, DBL α types bind to complement receptor 1 and heparan sulfate and are responsible for erythrocyte rosetting (Rowe et al. 1997 ; Chen et al. 1998 ). DBL β types participate in binding to ICAM-1 (Smith et al. 2000a ), thought to be responsible for the sequestration of erythrocytes infected with trophozoites and schizont in the brain. DBL γ is the ligand domain for chondroitin sulfate, a receptor for human placental infections by P. falciparum (Buffet et al. 1999 ; Reeder et al. 1999 ), leading to malarial complications associated with pregnancy.
DBL domains present as single copies formed a markedly separate clade, the PvDBP clade (fig. 7 ). All F1 domains in the DBL-EBPs having a duplicated DBL appeared more closely related to the single DBLs of the PvDBP clade than to F2 domains present in the same molecules. Interestingly, only a single ebl is identified so far in the nearly completed P. yoelii genome (5× coverage, May 30, 2001 release), which has a five exon structure and encodes a single F1-type DBL domain most similar to the P. vivaxebl (data not shown). EBA-175 was tightly clustered with the P. reichenowi EBP for both F1 and F2 domains as shown by the very short branch lengths connecting them, suggesting that these two genes are true orthologues. The F2 domains clustered together and with the DBL domains of the var genes. The latter finding is in accordance with our previous observation on the conservation of cysteine 6a in F2 domains and var DBLs.
These data present a relatively obvious dichotomy of the DBL domains and a possible pattern of evolution. The P. falciparumebl have two distinct DBL domains; the F1 domain is common to all ebl genes and is not in var, and the F2 domain is found only in the P. falciparumebl (and phylogentically closely related species) and is the progenitor of the var gene DBL. The DBL of PfEMP-1 was derived by duplication and then sequential diversification of the F2 domain of an ancestral ebl gene. Initial gene duplication and diversification led to variant binding phenotypes. Secondary gene duplications led to multiple gene variants within each group, representing antigenically distinct variants for each binding phenotype. Further duplication and diversification gave rise to an extensive repertoire of tandem DBL in a continuing process, as exemplified by the multiple var-encoded DBL with similar receptor interactions or binding phenotypes. Expansion of the DBL domain coding sequence in the P. falciparum genome presumably occurred by mechanisms common among eukaryotic organisms, which would include repeated intra chromosomal tandem duplications, deletions or local rearrangements (or both), and followed by proliferation through interchromosomal rearrangements and duplications (Ruvkun and Hobert 1998 ; Achaz et al. 2001 ). The two types of processes driving proliferation and diversity within var genes are selection for binding to different classes of receptor and immune selective pressure. Although proliferation of the DBL domain does not appear to be common among Plasmodium, other similar amplifications of antigenically variant molecules have occurred (Cheng et al. 1998 ; Al-Khedery et al. 1999 ; Kyes et al. 1999 ; Del Portillo et al. 2001 ). The subtelomeric location of the var genes, and the other multicopy genes on the chromosome ends, seems conducive for gene duplication as well as intergenic recombination (Rubio et al. 1996 ; Fischer et al. 1997 ; Hernandezrivas et al. 1997 ; Gardner et al. 1998 ; Bowman et al. 1999 ; Del Portillo et al. 2001 ).
Comparisons Based on the C-cys Regions
The MAEBL is a chimeric molecule with its C-cys region having identity to the DBL-EBP family (fig. 1 ). This region, also referred to as region VI, has approximately 100 amino acids and contains eight conserved cysteine residues, except for EBL-1, which has only four cysteines. The amino acid conservation among all the paralogues suggests an important functional role for this domain, although it does not have erythrocyte-binding activity, nor does it appear to contribute to the binding activity of the DBL domains for P. vivax DBP (Chitnis and Miller 1994 ) or P. falciparum EBA-175 (Sim et al. 1994 ).
The C-cys domain with its conserved features among a diverse range of paralogues has an ancient origin in the evolutionary history of Plasmodium species. The C-cys showed significant separation of the clade comprising the PvDBP and other ebl homologues found in simian parasites from that of all other ebl and maebl (fig. 6 ). Duplication of this domain in various structurally similar ebl paralogues of P. falciparum and the more distant paralogue, maebl, suggests that this highly conserved domain has an important functional role yet to be determined. The reason for the lack of resolution using the C-cys domain for other taxa, branching at the central node, is not clear. It could be explained by the small number of characters examined (108 amino acids), although the high number of parsimony informative characters (80) does not lend support to such a conclusion and suggests that the observed polytomy may be indicative of a genuinely rapid divergence event. The apparent lack of multiple ebl in some species (P. yoelii, P. vivax) supports a more recent divergence event for the ebl in P. falciparum. Conversely, the lack of nucleotide identity and dispersal of the P. falciparumebl on separate chromosomes is consistent with a long-standing divergence of the ebl genes with further amino acid divergence restricted by functional constraints. Nevertheless, the separate grouping of maebl from all the ebl indicates that maebl evolved as an intact gene in Plasmodium and is not the result of a recent recombination between ama-1 and an ebl.
Implications of Gene Duplication and Diversification
An extensive redundancy characterizes the P. falciparum genome (Gardner et al. 1998 ; Bowman et al. 1999 ) and is indicative of an adaptive character allowing a diversity of phenotypes by a single organism. In terms of parasitic invasion of erythrocytes, molecular diversity facilitates multiple alternative pathways of invasion. Alternative pathways enable parasites to infect a range of polymorphic erythrocytes within a host population and possibly maintain an infection when inhibitory antibodies block a critical invasion pathway. Thus, there are clear advantages for a parasite to develop alternative pathways of invasion. Evolutionarily divergent Plasmodium species, P. falciparum and P. yoelii, exhibit the ability to invade via alternative receptors using two distinct mechanisms.
Numerous studies have provided evidence for alternative invasion pathways for P. falciparum (reviewed in Preiser et al. 2000 ). Nevertheless, direct evidence for homologous ligands in P. falciparum functional for alternative pathways of invasion was only supported by the recent characterization of BAEBL, which recognizes a sialic acid–dependent epitope distinct from the epitope on glycophorin A recognized by EBA-175 (Mayer et al. 2001 ; Thompson et al. 2001 ). Similar to the PfEMP-1 DBL that share receptor-binding phenotypes and cluster together in clades, the BAEBL and EBA-175 domains were grouped together (fig. 7 ). Although their level of clustering had only weak support within this data set, the correctness of their phylogenetic relatedness is supported by the similarity of their receptor binding type (i.e., both bind sialic acid–dependent epitopes). The roles of the other identified ebl products (PEBL, JESEBL, EBL-1) during invasion are not determined. It is possible that some of these other ebl products are not functional genes, at least for the erythrocytic stages. No product was identified for PEBL (a.k.a. EBA-165) in blood-stage parasites, a result attributed to frameshifts in the open reading frame (Triglia et al. 2001 ). Surprisingly, transcription of the pebl, like that of the other ebl, occurs in a developmentally regulated, stage-specific manner during the erythrocytic growth cycle of P. falciparum (Blair et al. 2002a ). Such regulated transcriptional control is very unusual for vertebrate pseudogenes (Mighell et al. 2000 ). Nevertheless, pfRH3 is another apparent transcriptionally active pseudogene reported for P. falciparum; it is a member of the multigene family related to the P. vivaxrbp and P. yoeliiy235 genes (Taylor et al. 2001 ).
Conclusions
Invasion of a Plasmodium merozoite into an erythrocyte is a complex process involving a panel of merozoite ligands interacting with different erythrocyte receptors at different junctures during the invasion process. In P. falciparum, alternative pathways of invasion exist, presumably through the expression of alternative ligands, which have redundant functions and recognize different receptors. Identification of MAEBL homologues in the human malaria parasites is a step toward understanding the function of this chimeric molecule. The origins of maebl,ebl, and ama-1 predate speciation of Plasmodium, whereas at least some duplication of ebl and var genes occurred later toward the evolution of P. falciparum as a species. The maebl and other ebls obviously have a common ancestor gene but represent distinct lineages. The maebl is of ancient origin and evolved as a single locus, including its unique chimeric structure, in each Plasmodium species, in conjunction with the ama-1 and the ebl gene families. The ancient origin of the maebl, its highly conserved nature among all Plasmodium species studied, its different time of expression during development, and the distinct localization in merozoites, all suggest that the MAEBL has a distinct role from the DBL-EBP products during the invasion process. Therefore, MAEBL does not have a redundant or alternative role in invasion to ebl products, such as EBA-175 and BAEBL. On the other hand, other P. falciparumebl paralogues, which have occurred by duplication and diversification in this and other Plasmodium species, potentially provide multiple functionally equivalent ligands to EBA-175 for alternative invasion pathways. EBA-175 appears to act as critical merozoite ligand, recognizing a sialic acid–dependent epitope on glycophorin A, whereas the similar ebl products with DBL domains, like BAEBL, appear to be functionally equivalent ligands recognizing an alternative sialic acid–dependent receptor.
Supplementary Material
Accession numbers: Plasmodium knowlesimaebl AY042082; Plasmodium vivaxmaebl AY042083; Plasmodium falciparummaebl AY042082.
Pekka Pamilo, Reviewing Editor
Present address: Department of Zoology, University of Oxford, UK
Abbreviations: AMA-1, apical membrane antigen-1; DBP, Duffy antigen-binding protein; DBL, Duffy-binding–like; EBA-175, erythrocyte-binding antigen-175; ebl; erythrocyte-binding–like; EBP, erythrocyte-binding protein; MPTs, most parsimonious trees.
Keywords: malaria phylogenetic relationships Plasmodium MAEBL erythrocyte-binding protein apical membrane antigen-1
Address for correspondence and reprints: John H. Adams, Department of Biological Sciences, University of Notre Dame, Notre Dame, Indiana 46556. jadams3@nd.edu
This work was supported by the National Institutes of Health (R29/R01 AI33656). J.H.A. is a Burroughs Wellcome Fund New Investigator in Molecular Parasitology. J.R.S. is supported by a Wellcome Trust Biodiversity Fellowship (Grant No 050808/Z/97/Z).
We wish to thank the scientists and funding agencies comprising the international Malaria Genome Project for making sequence data from the genome of P. falciparum (3D7) public before the publication of the completed sequence. The Sanger Centre (UK) provided the sequences for chromosomes 1, 3–9, and 13, with financial support from the Wellcome Trust. A consortium composed of The Institute for Genome Research, along with the Naval Medical Research Center (USA), sequenced chromosomes 2, 10, 11, and 14, with support from NIAID/NIH, the Burroughs Wellcome Fund, and the Department of Defense. The Stanford Genome Technology Center (USA) sequenced chromosome 12, with support from the Burroughs Wellcome Fund. The Plasmodium Genome Database is a collaborative effort of investigators at the University of Pennsylvania (USA) and Monash University (Melbourne, Australia), supported by the Burroughs Wellcome Fund.
References
Achaz G., P. Netter, E. Coissac,
Adams J. H., D. E. Hudson, M. Torii, G. E. Ward, T. E. Wellems, M. Aikawa, L. H. Miller,
Adams J. H., O. Kaneko, P. L. Blair, D. S. Peterson,
Adams J. H., B. K. Sim, S. A. Dolan, X. Fang, D. C. Kaslow, L. H. Miller,
al-Khedery B., J. W. Barnwell, M. R. Galinski,
Altschul S. F., W. Gish, W. Miller, E. W. Myers, D. J. Lipman,
Anders R. F., P. E. Crewther, S. Edwards, M. Margetts, M. L. Matthew, B. Pollock, D. Pye,
Baruch D. I., B. L. Pasloske, H. B. Singh, X. Bi, X. C. Ma, M. Feldman, T. F. Taraschi, R. J. Howard,
Blair P. L., A. Witney, J. D. Haynes, J. K. Moch, D. J. Carucci, J. H. Adams,
Blair P. L., S. H. I. Kappe, J. E. Maciel, B. Balu, J. H. Adams,
Bowman S., D. Lawson, D. Basham, et al. (25 co-authors)
Buffet P. A., B. Gamain, C. Scheidig, et al. (14 co-authors)
Chen Q., A. Barragan, V. Fernandez, A. Sundstrom, M. Schlichtherle, A. Sahlen, J. Carlson, S. Datta, M. Wahlgren,
Cheng Q., N. Cloonan, K. Fischer, J. Thompson, G. Waine, M. Lanzer, A. Saul,
Chitnis C. E., M. J. Blackman,
Chitnis C. E., L. H. Miller,
Chou Q., M. Russell, D. E. Birch, J. Raymond, W. Bloch,
Cooke B., R. Coppel, M. Wahlgren,
Crewther P. E., J. G. Culvenor, A. Silva, J. A. Cooper, R. F. Anders,
del Portillo H. A., C. Fernandez-Becerra, S. Bowman, et al. (13 co-authors)
Devon R. S., D. J. Porteous, A. J. Brookes,
Dolan S. A., L. H. Miller, T. E. Wellems,
Dolan S. A., J. L. Proctor, D. W. Alling, Y. Okubo, T. E. Wellems, L. H. Miller,
Don R. H., P. T. Cox, B. J. Wainwright, K. Baker, J. S. Mattick,
Donahue C. G., V. B. Carruthers, S. D. Gilk, G. E. Ward,
Escalante A. A., F. J. Ayala,
Felsentein J.,
Fischer K., P. Horrocks, M. Preuss, J. Wiesner, S. Wunsch, A. A. Camargo, M. Lanzer,
Fraser T. S., S. H. Kappe, D. L. Narum, K. M. VanBuskirk, J. H. Adams,
Freeman R. R., A. J. Trejdosiewicz, G. A. Cross,
Galinski M. R., C. C. Medina, P. Ingravallo, J. W. Barnwell,
Gardner M. J., H. Tettelin, D. J. Carucci, et al. (25 co-authors)
Haynes J. D., J. P. Dalton, F. W. Klotz, M. H. McGinniss, T. J. Hadley, D. E. Hudson, L. H. Miller,
Hehl A. B., C. Lekutis, M. E. Grigg, P. J. Bradley, J. F. Dubremetz, E. Ortega-Barria, J. C. Boothroyd,
Hendy M. D., D. Penny,
Hernandez-Rivas R., D. Mattei, Y. Sterkers, D. S. Peterson, T. E. Wellems, A. Scherf,
Hinterberg K., D. Mattei, T. E. Wellems, A. Scherf,
Hodder A. N., P. E. Crewther, M. L. Matthew, G. E. Reid, R. L. Moritz, R. J. Simpson, R. F. Anders,
Kappe S. H., G. P. Curley, A. R. Noe, J. P. Dalton, J. H. Adams,
Kappe S. H. I., A. R. Noe, T. S. Fraser, P. L. Blair, J. H. Adams,
Keen J., A. Holder, J. Playfair, M. Lockyer, A. Lewis,
Kyes S. A., J. A. Rowe, N. Kriek, C. I. Newbold,
Mayer D. C. G., O. Kaneko, D. E. Hudson-Taylor, M. E. Reid, L. H. Miller,
Mighell A. J., N. R. Smith, P. A. Robinson, A. F. Markham,
Miller L. H., S. J. Mason, D. F. Clyde, M. H. McGinniss,
Miller L. H., S. J. Mason, J. A. Dvorak, M. H. McGinniss, I. K. Rothman,
Mitchell G. H., T. J. Hadley, M. H. McGinniss, F. W. Klotz, L. H. Miller,
Narum D. L., A. W. Thomas,
Nielsen H., J. Engelbrecht, S. Brunak, G. von Heijne,
Noe A. R., D. J. Fishkind, J. H. Adams,
Oaks S. C. Jr.,, V. S. Mitchell, G. W. Pearson, C. J. Carpenter,
Ogun S. A., A. A. Holder,
Okenu D. M., P. Malhotra, P. V. Lalitha, C. E. Chitnis, V. S. Chauhan,
Okoyeh J. N., C. R. Pillai, C. E. Chitnis,
Orlandi P. A., F. W. Klotz, J. D. Haynes,
Ozwara H., C. H. Kocken, D. J. Conway, J. M. Mwenda, A. W. Thomas,
Perkins M. E., E. H. Holt,
Peterson D. S., L. H. Miller, T. E. Wellems,
Peterson D. S., T. E. Wellems,
Peterson M. G., V. M. Marshall, J. A. Smythe, P. E. Crewther, A. Lew, A. Silva, R. F. Anders, D. J. Kemp,
Preiser P., M. Kaviratne, S. Khan, L. Bannister, W. Jarra,
Preiser P. R., W. Jarra, T. Capiod, G. Snounou,
Prickett M. D., T. R. Smarz, J. H. Adams,
Ranjan A., C. E. Chitnis,
Rayner J. C., E. Vargas-Serrato, C. S. Huber, M. R. Galinski, J. W. Barnwell,
Reeder J. C., A. F. Cowman, K. M. Davern, J. G. Beeson, J. K. Thompson, S. J. Rogerson, G. V. Brown,
Rowe J. A., J. M. Moulds, C. I. Newbold, L. H. Miller,
Rubio J. P., J. K. Thompson, A. F. Cowman,
Ruvkun G., O. Hobert,
Saitou N., M. Nei,
Sanger F., S. Nicklen, A. R. Coulson,
Sim B. K., C. E. Chitnis, K. Wasniowska, T. J. Hadley, L. H. Miller,
Sim B. K. L., T. Toyoshima, J. D. Haynes, M. Aikawa,
Smith J. D., C. E. Chitnis, A. G. Craig, D. J. Roberts, D. E. Hudson-Taylor, D. S. Peterson, R. Pinches, C. I. Newbold, L. H. Miller,
Smith J. D., A. G. Craig, N. Kriek, D. Hudson-Taylor, S. Kyes, T. Fagen, R. Pinches, D. I. Baruch, C. I. Newbold, L. H. Miller,
Smith J. D., G. Subramanian, B. Gamain, D. I. Baruch, L. H. Miller,
Su X. Z., V. M. Heatwole, S. P. Wertheimer, F. Guinet, J. A. Herrfeldt, D. S. Peterson, J. A. Ravetch, T. E. Wellems,
Swofford D. L.,
Taylor H. M., T. Triglia, J. Thompson, M. Sajid, R. Fowler, M. E. Wickham, A. F. Cowman, A. A. Holder,
Thomas A. W., J. A. Deans, G. H. Mitchell, T. Alderson, S. Cohen,
Thompson J. D., D. G. Higgins, T. J. Gibson,
Thompson J. K., T. Triglia, M. B. Reed, A. F. Cowman,
Thornton J. W., R. DeSalle,
Triglia T., J. K. Thompson, A. F. Cowman,
Tusnady G. E., I. Simon,
Verra F., A. L. Hughes,