Low MSP-1 haplotype diversity in the West Palearctic population of the avian malaria parasite Plasmodium relictum

Block 14 of the msp1 gene was amplified from 27 avian blood samples infected with the P. relictum cyt b lineage GRW11 and 111 samples infected with cyt b lineage SGS1 using the primers and annealing temperatures described in Hellgren et al. [21] and sequenced with Sanger methodology (in-house). Note that the name of the internal primer pair, MSP1_3FN and MSP1_3RN, have been named incorrectly regarding their direction to the outside primers, MSP1_3F and MSP1_3R in the original article [21]. These samples were part of previous studies [28, 29], which identified single haemosporidian infections with one of the two Plasmodium cyt b lineages. All samples were collected under legal permits. These samples originated from 36 passerine species captured at 35 localities in Portugal, Morocco, Russia, and Armenia (Table 2). Blood was sampled by brachial venipuncture with a sterile needle and collected into a heparin-free glass capillary tube. Samples were immediately transferred into 2-ml vials with 96% ethanol and stored until DNA extraction as described previously [21].

New msp1 block 14 sequences (GenBank; Pr10: MZ270634 and Pr11: MZ270635) were aligned to those from Hellgren et al. [21] using Geneious Prime 2019.2.3. (www.​geneious.​com). The phylogenetic relationships among new and previously discovered [21] msp1 haplotypes were evaluated by Bayesian phylogenetic analysis, using MrBayes [30] as implemented in Geneious Prime 2019.2.3. Phylogenies was constructed using 1,100,000 iterations and sampling every 200th tree under a general time reversal + invariable site model (GTR + I) allowing for six different substitution rates. After discarding 10% of the sampled trees as a burn-in, the remaining trees were used to construct a majority consensus tree. The phylogeny contained all known msp1 haplotypes originating from cyt b SGS1 and GRW11 infections. As outgroups (for outgroup sequences see [21]) msp1 haplotypes Pr4 (KJ850282) and Pr5 (KJ850281) originating from the cyt b lineage GRW4 was used. For visualization of the relationship and frequency of obtained haplotypes a haplotype network which was inferred using PopArt version 1.7.2 [31], with the TCS inference method [32] was constructed.

For an additional 23 samples, collected under legal permits, with SGS1 (n = 20) or GRW11 (n = 3) infections confirmed using the molecular screening described in Hellgren et al. [33], the complete msp1 gene was sequenced following the sequence capture protocol described in Huang et al. [18], but using a probe set designed to target P. relictum instead (available upon request) after the presence of SGS1 (n = 20) or GRW11 (n = 3) infection was confirmed using the molecular screening protocol described in Hellgren et al. [33]. Capture probes targeting 1,036 genes from the P. relictum genome [15] was constructed and used in a SureSelectXT Target Enrichment kit (Agilent Technologies) as previously described [18], with a few modifications. The protocol started with 200 ng of sheared DNA from each sample and followed the manufacturer’s protocol for sequence capture, including hybridizing probes with the samples for 24 h. For the PCR, 18 cycles of was used in the post-capture amplification with indexed primers (instead of the recommended 10 to 16 cycles) to ensure sufficient amount of DNA for sequencing. Captured fragments were sequenced on an Illumina MiSeq instrument at the Lund University DNA Sequencing Facility.

Raw sequencing reads was processed by removing adapter sequences and low-quality base calls with a sliding window using Trimmomatic v0.39 [34]. Read quality was assessed with FastQC v0.11.9. The reads were mapped to the P. relictum genome (PlasmoDB.org, release 39 [35]) using Nextgenmap v0.5.5 [36] with default options. The resulting sam files were compressed, sorted and indexed using samtools v1.10 [37]. MarkDuplicates v2.20.8 in Picard tools (Broad Institute, 2019) was used to identify duplicate reads. To ensure that the infections were indeed single infections and matched SGS1 or GRW11 cyt b sequences, the consensus sequence was first checked against the SGS1 and GRW11 reference sequences [9]. Then double infections were identified by visually detecting SNPs in the mapped reads of the cyt b gene. For individuals that showed no evidence of double infections at the cyt b gene, the majority consensus sequences of the msp1 gene were generated for sites with a depth of coverage > 5 as implemented in Geneious Prime 2019.2.3. Discarding double infections left 11 samples, of which five had incomplete coverage; these 11 samples were used for further analysis at different degrees. The full-length consensus sequences were first aligned against each other using the MUSCLE alignment algorithm with eight iterations in Geneious Prime 2019.2.3. This allowed the identification of SNPs outside the msp1 block 14 section (targeted in the larger sample set). Next, the full-length msp1 genes were mapped against the shorter reference sequences of msp1 block 14, found in Hellgren et al. [21], to assess the haplotype identity of the complete msp1 sequences. Those previously identified msp1haplotypes of P. relictum have been named Pr1-Pr9 [21] and will be referred to as such thereafter; novel haplotypes were given numbers > Pr9, numbered in the order in which they were discovered.

Block 14 of the msp1 gene was successfully amplified and sequenced in 11 samples of GRW11 originating from nine avian species and in 64 samples of SGS1 from 25 avian species sampled in Russia, Armenia, Morocco, and Portugal (GenBank accession numbers MZ270578-MZ270624). Haplotype Pr2 was discovered in 72 of these 75 infections. Samples that were not sequenced failed in amplification of the nuclear gene, a common problem when trying to sequence nuclear genes of avian malaria due to low parasitaemia in wild hosts [27]. A single adult common whitethroat (Sylvia communis) infected with GRW11 from Armenia had the Pr3 haplotype, which previously had been found only in vectors caught and sampled in the Eastern Palearctic, i.e., Japan [21]. The remaining two haplotypes were new and both found in SGS1 infections. Pr10 was found in an adult rock-thrush (Monticola saxatilis) from Armenia and Pr11 was found in a juvenile Cetti’s warbler (Cettia cetti) from Portugal (Table 2). The majority of the Pr2 haplotypes linked to both GRW11 and SGS1 infections were found in resident bird species confirming their local transmission in the sampling areas.

The phylogenetic reconstruction and the haplotype network placed the Pr11 haplotype together with other haplotypes transmitted either in the Western Palearctic (Pr2) or in the Eastern Palearctic (Fig. 1). In fact, the Pr11 haplotype only differed by a single synonymous substitution from Pr2, the dominant lineage transmitted in the Western Palearctic. The other new haplotype, Pr10, found in an adult Paleotropical migrant (Monticola saxatilis) was a sister to the SGS1/GRW11 clade and differed by a single non-synonymous substitution from Pr1, which has Afrotropical transmission (Fig. 1).

After discarding samples that showed evidence of double infections at the cyt b gene, 11 samples remained and are referred to in the subsequent analysis. The complete msp1 gene, 4740 bp, was sequenced with a coverage depth > 5 in two GRW11 and three SGS1 infected samples. For the remaining six samples, 0.2–43.4% of the nucleotide positions had a coverage depth < 5 × and were excluded from the analysis (Table 3). All five complete samples were unambiguously assigned to the haplotype Pr2 at the MSP1 block 14 and all consensus sequences where monomorphic across the rest of the gene, showing no genetic variation across these samples (represented by GenBank entry MW385258). However, all of them (GenBank accession numbers MZ270625-MZ270635) differed by a single substitution from the reference sequence (KC969175) on GenBank but were identical to the msp1 gene in the genome of the SGS1 parasite available at PlasmoDB [15, 35]. This discrepancy may point to a potential sequencing error in the reference sequence KC969175 [24].