Sequence information, chromosomal locations and transcriptional directions of genes in the 3D7 genome were obtained from the Plasmodium Genome Resource 36. Sequences from the Hb3 and Dd2 sequencing projects were retrieved from the Microbial Sequencing Center, Broad Institute 37. Sequences from the Ghanaian isolate and the It/FCR3 strain were downloaded from The Plasmodium genome project, Welcome Trust Sanger Institute 38. The coverage of Hb3, Dd2 and the Ghanaian isolate were 8.07×, 7.13× and 8× respectively, whereas the coverage of the It strain was estimated to be 3.84× by averaging the sizes of Hb3, Dd2 and Ghanaian isolate genomes and comparing the number of reads sequenced for the four strains.

Sequence reads were aligned to the n-, o-, pfmc-2tm and q-genes using BLASTN without low complexity filtering. The identity cutoff was set to 95% with a minimum accepted length of an overlap of ≥ 36 bp. The number of bps aligning to the genes was compared to the length of each gene, which yielded an estimated coverage for each gene in each of the parasite lines. This gene specific coverage was subsequently compared to the total coverage for the strains and a copy number estimate was calculated.

A graphical output of all genes in the subtelomeric block 4–5 for all 14 chromosomes was generated (Figure 1). The boundaries of the subtelomeric ends were defined based on the whole genome synteny mapping of P. falciparum with rodent malaria parasites (P. berghei, P. chabaudi and P. yoelii) 18. Subtelomeric gene-families are categorized into 18 groups (Additional File 1) and are displayed in different colors. Grouping of the subtelomeric genes was based on information from literature, the OrthoMCL Database 39 and/or protein features (possession of PEXEL/VTS domain and transmembrane regions) acquired from the Plasmodium database 36 where protein domains were predicted using HMM against the Pfam database, version 17.

Nucleic acids (gDNA and RNA) were extracted using either the Easy-DNA™ (Invitrogen) or the RNeasy^® (Qiagen) kits according to the recommendations of the suppliers. Total RNA was isolated from 3D7AH1, FCR3 and 7G8 at 8 to 28 hours post invasion with four-hour intervals for two consecutive parasite cycles. To ensure DNA-free RNA, the isolated RNA was treated with TURBO DNA-free™ DNAse (Ambion).

Standard polymerase chain reaction (PCR) was used for the amplification of n-, o-, pfmc-2tm-, and q-genes of the SDs. Primers were designed based on the published 3D7 sequences: n-gene: forward 5'-TTT TTT TCA AGT AAG AGA TGC-3', reverse 5'-CCA CAA CCA CAC AAG AAG-3'; o-gene: forward 5'-CAA TAA ATA TAG CAA GTC G-3', reverse 5'-TAA ATC ATG TTC TGT GTG-3'; pfmc-2tm: forward 5'-ATC ATA CCA TAA TGG AGG-3', reverse 5'-ACC TAT TTT CAT GTC AGG-3' and q-gene: forward 5'-TGA AAA TAC CAA AGT ACC-3', reverse 5'-ATT GTA ATC CTT TAG CTC-3'. Amplification products were cloned into Topo vectors (TOPO TA cloning kit, Invitrogen) before transformation into TOP10 competent E.coli. DNA from at least four bacterial clones was sequenced for each target using M13 forward and reverse primers. ClustalW multiple alignments were performed thereafter, using BioEdit software version 7.0.5 (Tom hall, Ibis Therapeutics, Carlsbad, CA).

Copy numbers relative to the 3D7AH1 parasite of the n-gene, PFA0675w, PFA0685c, PFA0690w and PFA0700c were determined for FCR3, 7G8, UAM25, HB3, Dd2, TM180 and TM284. Primers specific for the n-gene (5'-AGG GCA ATT GAT TTT AGC AGG TAT-3' and 5'-CAA AAC TAC TGA ATG CTA TAA ATG AAG GA-3'), PFA0675w (5'-TAT AAG ACC AAC TCT TTT CAT TTG TCT TTA C-3' and 5'-AAA ATC CTG TTG TAT GTA CGA TTA GCA T-3'), PFA0685c (5'-AAT ATA TAA CAA GTC GAG CAC TAA CGG A-3' and 5'-TCC TCT TAT TTG TGG ATT TTT ATT TCC-3'), PFA0690w (5'-ACC AAG AGC CTT GTG AAA CGA-3' and 5'-TTT CTT CCT TCT TCA GTT TTT TTG TG-3'), PFA0700c (5'-AGG AGA TTA CTA GCC GAA CCA CAC-3' and 5'-TTT ATG GGT TTT CAA TAT ATG TGA TTT GT-3') and the endogenous control gene PF10_0084 (5'-ACA ACG AAG CAA CAG GAG GTA GAT-3' and 5'-AGT CCA TCA ATA TAG CTC TTG GAA CAT A-3') were all designed using Primer Express 2.0 (Applied Biosystems) towards perfectly conserved stretches of the genes. Approximately 1 ng of DNA was used as template in quadruplicate amplification reactions in MicroAmp 96 well plates in 20 μl containing SYBR Green master mix and 300 nm of each primer. Amplifications were carried out in an ABI sequence detector 7500 (Applied Biosystems) for 40 cycles (95°C for 15 seconds and 60°C for 1 min). PCR-efficiencies of all primer-pairs were evaluated on dilution series of 3D7AH1 genomic DNA and found to be sufficiently close to obviate the need for any correction factor. Results were analysed using the ΔΔCt method (User bulletin 2, Applied Biosystems) based on the tested assumption that the target genes are amplified with the same efficiency as the endogenous control.

Total RNA was reversibly transcribed with SuperScript III Rnase H reverse transcriptase (Invitrogen), random hexamers and oligo(dT)_12–18 (300 ng/μl and 25 ng/μl respectively, both from Invitrogen) for two hours at 50°C. For each cDNA synthesis reaction, a control reaction without reverse transcriptase was performed with identical amounts of template. For qPCR-based determination of n-gene transcription the same primers were used as listed above except for the endogenous control, where seryl-tRNA synthetase was employed. The primers were: 5'-TAT CAT CTC AAC AGG TAT CTA CAT CTC CTA-3' and 5'-TTT GAG AGT TAC ATG TGG TAT CAT CTT TT-3'. The amplification reactions were conducted as described above, with the only difference that 2 ng of template was used. Transcription levels were achieved by dividing the x¯Ctn-gene MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGaciGaaiaabeqaaeqabiWaaaGcbaGafmiEaGNbaebadaWgaaWcbaacbiGae83qamKae8hDaqNae8hiaaIae8NBa4Mae8xla0Iae83zaCMae8xzauMae8NBa4Mae8xzaugabeaaaaa@386C@ with the x¯Ctseryl-tRNAsynthetase MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGaciGaaiaabeqaaeqabiWaaaGcbaGafmiEaGNbaebadaWgaaWcbaacbiGae83qamKae8hDaqNae8hiaaIae83CamNae8xzauMae8NCaiNae8xEaKNae8hBaWMae8xla0Iae8hDaqNae8NuaiLae8Nta4Kae8xqaeKae8hiaaIae83CamNae8xEaKNae8NBa4Mae8hDaqNae8hAaGMae8xzauMae8hDaqNae8xyaeMae83CamNae8xzaugabeaaaaa@4BF5@ for each strain and time point. The standard deviation of the quotient was calculated according to the User Bulletin 2, Applied Biosystems. Results were visualized as log₂ transformed values and plotted using SigmaPlot 9.0 (Systat Software Inc.).

FISH targeting the n-gene was conducted according to previously described methodology 35. The fluorescein labeled (Fluorescein-High Prime, Roche Applied Science) n-gene probe was generated from 3D7AH1 gDNA using the primers 5'-TTT TTT TCA AGT AAG AGA TGC-3' and 5'-CCA CAA CCA CAC AAG AAG-3'.

Comparative analysis of the P. falciparum genome with rodent plasmodium species has disclosed synteny breaks at the boundaries of the subtelomeric compartments 18. Here, we have analysed the subtelomeric gene content of the 3D7 genome by grouping the genes into families as shown in Figure 1. Eight homologous regions were found, all sharing the same genomic organization being located on seven chromosomes (Chromosomes 1, 2, 3, 6, 7, 10 and 11). This duplicated DNA segment (named SD1) was found to contain six genes: rif, pfmc-2tm, a var pseudogene and three hypothetical genes (n-, o- and q-gene) (Figure 2A). The breakpoints of these segmental duplicons vary slightly, with the 5' break point being either within or downstream with respect to the rif gene and the 3' break point being either upstream or downstream of the var pseudogene. The most extended duplicated loci (approximately 32 kb in size) are both located on chromosome 6, but on opposite chromosomal ends. Although the rif genes are not identical in-between the SD1, homologous rif copies can be found within all SD1 (Figure 2B). Most of the genes within SD1 encode PEXEL-containing export proteins, with the exception of the q-gene and the var pseudogenes (Additional File 2). SD1-fragments harbouring only two or three of the SD1- genes (o-gene, pfmc-2tm, q-gene) were also found in the 3D7 genome (Additional File 1).

A previous CGH project from this laboratory revealed a subtelomeric gene segment (PFA0685c, PFA0690w and PFA0695c), located on the right end of chromosome 1 in the 3D7 strain, to be duplicated in a fresh clinical isolate (UAM25) 35 (Figure 3A). Further analysis indicates that this locus shares three of the same paralogous genes as SD1s described above, with the same gene order and orientation but with less sequence homology (55% identity). This SD was named SD2. Compared to the eight SD1, SD2 was found to carry the n-gene as a pseudogene and the q-gene (PFA0675w) was found to harbour RESA-like repeats and a DNAJ domain (PFAM database: PF0026; amino acid 1097–1160), which the q-gene of SD1 does not possess. PSI-BLAST analyses of the genes in the SD2 (converged at iteration 3) showed that the q-gene has orthologous genes in P. vivax and in rodent malaria parasites (P. yoelii, P. chabaudi and P. berghei). However, no orthologous genes could be identified for the other SD2 gene-members.

To elucidate whether the sequence conservation of the SD1 remains across different P. falciparum parasites, we sequenced the n-, o-, pfmc-2tm and q-gene of five parasites originating from different geographical areas: FCR3 (The Gambia), TM180 (Thailand), 7G8 (Brazil), UAS31 and UAS39 (both from Uganda). In addition, sequence information for HB3 (Honduras) and Dd2 (Indochina) 37 and It (Brazil) 38 was retrieved for the analysis. ClustalW multiple alignments revealed that genes within the SD1s are of a high sequence identity (99%), with the exception of a ≈ 23 amino acid hypervariable loop within pfmc-2tm which is predicted to be surface-exposed 1240. Polymorphisms other than those of pfmc-2tm in the eight SD1s of 3D7 were mainly situated within repetitive sequence stretches of the intra- and intergenic regions. Comparisons of sequences to single nucleotide polymorphism (SNPs) data published recently 41 (Additional File 3) revealed four novel non-synonymous SNPs in the n-gene, and four non-synonymous and two synonymous SNPs in the q-gene.

Using the n-gene as a representative member of SD1, the SD1 copy number in different P. falciparum strains relative to the 3D7 parasite was estimated using qPCR. The genomes of HB3 and the clinical isolate (UAM25) were found to contain the same number of SD1 copies as 3D7 (n = 8), whereas Dd2 was found to carry ≤ 4 (Figure 4A). Comparable numbers of pfmc-2tm was previously reported for HB3 relative to 3D7 40, signifying a copy number association between the n-gene and pfmc-2tm.

The results were further confirmed by fluorescent in situ hybridizations (FISH). In addition to a clear pattern of variable copy numbers (Figure 4A) most of the signals were distributed at the rim of the parasite-nuclei where chromosomal ends are known to tether 23, confirming the subtelomeric localization of the SD1s (Figure 4B).

The amplification of SD2 was also verified by qPCR targeting the pseudo n-gene (PFA0690w), as well as the adjacent genes, PFA0675w (pseudo q-gene), PFA0685c (pseudo o-gene) and PFA0700c (Figure 3B). In contrast to SD1, the SD2 in UAM25 did not include PFA0675w (paralogous to the q-gene).

The intraerythrocytic developmental expression of the genes in the SDs was previously studied using microarrays (42 : E-MEXP-128) 4344. Only the n-gene was found significantly transcribed, with maximum expression in the ring stages. In addition, the pseudo n-gene (PFA0690w) of SD2 was found to be expressed, despite of its supposedly truncated ORF, with maximum transcript abundance at 36h post-invasion 44.

In order to investigate the impact of gene dosage on transcription levels, n-gene transcription was investigated for three parasites with varying numbers of SD1s. 3D7AH1, FCR3 and 7G8 parasites were harvested at 4-hour intervals from eight to 28 hours post-invasion and relative mRNA levels were studied by qPCR. The maximum level of transcription of the n-gene was found in ring-stage parasites, which coincides with previous transcription data 4344. A clear transcriptional difference was observed when comparing 3D7AH1 and 7G8, which carry eight and five copies in the genome, respectively, but similar level of transcription was found for 3D7AH1 and FCR3, although the latter carries fewer copies of the n-gene (Figure 5).