Generalist Malaria Parasites and Host Imprinting: Unveiling Transcriptional Memory

Abstract

Generalist parasites must adapt to diverse host environments to ensure their survival and transmission. These adaptations can involve fixed genetic responses, transcriptional plasticity, or epigenetic mechanisms. The avian malaria parasite Plasmodium homocircumflexum offers an ideal model for studying transcriptional variation across hosts. We experimentally inoculated P. homocircumflexum into different bird species, bypassing the vector, to assess whether gene expression remains stable across hosts, resets in response to new environments, or reflects epigenetic inheritance. We tested two alternative hypotheses: (i) universal gene expression profile (“one key fits all”), where parasite expression remains consistent across hosts. Our outcomes revealed that gene expression differed significantly depending on the host species and time postinfection, rejecting this hypothesis. (ii) Transcriptional plasticity, where gene expression is determined by the recipient host. Contrary to this hypothesis, we observed that gene expression was primarily influenced by the donor at 8 d postinfection (dpi), whereas gene expression was more aligned with the recipient host at 16 dpi. We also explored two mechanisms to explain these patterns: (i) epigenetic inheritance, whereby early transcription reflects the donor environment but adjusts over time, and (ii) genetic differentiation selecting for specific haplotypes. Our data support mechanism (i): 2,647 differentially expressed genes (DEGs) were associated with the donor at 8 dpi, while only 271 DEGs were linked to the recipient at 16 dpi. Single Nucleotide Polymorphism analyses revealed low genetic differentiation, rejecting mechanism (ii). These findings suggest that P. homocircumflexum undergoes a shift from donor-dependent to recipient-dependent gene expression, likely driven by epigenetic regulation and transcriptional plasticity.

Introduction

The survival of generalist parasites depends on their ability to adapt to new environments (Prati et al. 2022). When conducting host shifts, the parasite get exposed to varying immune responses, body temperatures, metabolism, and nutrient availability, all affecting survival and replication (Agosta et al. 2010; Gupta et al. 2020). Vector-transmitted parasites such as Plasmodium spp. must navigate in both vertebrate and invertebrate hosts, with the vertebrate immune system being a major challenge (Mandala et al. 2021; Matos et al. 2023). Just considering the vertebrate immune responses, they might further vary across species (Palinauskas et al. 2020; Galinski 2022) and geographic regions (Clark et al. 2020; Fecchio et al. 2021), requiring parasites to be highly flexible to overcome these challenges. One way to achieve this flexibility is by relying on transcriptional variation (Turnbull et al. 2022), where gene expression shifts in response to environmental changes, often regulated by epigenetic mechanisms such as DNA methylation (Serrano-Durán et al. 2022) and histone modifications (Connacher et al. 2022). These mechanisms enable adaptation without altering the genome (Hollin et al. 2023), helping parasites to evade host immunity and adjust to host metabolism within a single infection.

A key question is whether generalist parasites adapt through fixed genetic responses or flexible adjustments such as epigenetic modifications and real-time gene expression shifts. One possibility is a “universal key” strategy, where parasites maintain a static gene expression profile that enables infection across different hosts with minimal adjustment, relying on conserved gene responses (Schneider et al. 2023; Naidoo and Oliver 2024). Alternatively, selection-driven variation may favor specific haplotypes with expression profiles better suited to certain hosts (Ellis et al. 2023). Another strategy involves dynamic responses, in which gene expression adapts to different host immune defenses and metabolism (Guha et al. 2021). This plasticity may arise from epigenetic mechanisms or gene expression patterns inherited asexually from the parent population (Govindaraju and Rajavelu 2024), enabling the replication of successful adaptive responses. These inherited profiles may shift when environmental changes demand adaptation, optimizing parasite survival across hosts (Llorà-Batlle et al. 2019).

Plasmodium falciparum exemplifies transcriptional variation, with gene expression changes regulated through complex epigenetic mechanisms (Adejoh et al. 2023). Genes involved in antigenic variation and immune evasion exhibit clonal variation, meaning genetically identical parasites express different gene sets (Wyss et al. 2024). This allows P. falciparum to maintain a diverse population, some better suited to withstand host changes (Schneider et al. 2023). Similarly, avian generalist Plasmodium species may employ analogous strategies, maintaining genetic and epigenetic diversity to adapt rapidly as they shift between avian hosts. However, this remains untested empirically.

Compared to mammalian Plasmodium species such as Plasmodium falciparum, avian Plasmodium species offer a unique model for examining the strategies of generalist versus specialist parasites. Plasmodium homocircumflexum is particularly relevant as it infects multiple host species that are phylogenetically diverse (Palinauskas et al. 2015). By analyzing gene expression changes in P. homocircumflexum across different hosts and over time, we can better understand how parasites navigate the balance between adaptation and specialization. The extent to which transcriptomic changes in generalist parasites stem from immediate host conditions versus retained adaptations from previous hosts remains unclear. It is likely that both factors contribute, as evidenced by Plasmodium relictum and P. homocircumflexum, which infect a variety of bird species (Bensch et al. 2009). These parasites may rely on multiple transcriptional variation sources to respond effectively to new immune challenges and physiological conditions (Manzoli et al. 2021). Identifying the specific sources utilized by P. homocircumflexum could enhance our understanding of how generalist parasites optimize their survival across different hosts. To investigate transcriptional variation, we conducted a crosswise infection experiment where we directly inoculated P. homocircumflexum into different host species after the parasites have had time to develop and adjust within either host. The method allowed us to bypass recombination in the vector, ensuring that the parasites initiating the new infection were directly exposed to either the same or a different host environment without time to “reset” their transcriptional profiles. This approach enabled us to assess how the parasite responds to the recipient host and how transcriptional variation unfolds across asexual generations. We considered two alternative hypotheses regarding the regulation of gene expression in P. homocircumflexum: (i) a “one key fits all” strategy, where gene expression remains consistent regardless of host environment, and (ii) transcriptomic plasticity, where gene expression reflects the recipient host environment, as clonal offspring adapt directly to new conditions. We also explored two possible mechanisms that could shape the observed transcriptional profiles: (i) epigenetic inheritance, where gene expression reflects a donor-derived program retained through asexual replication, and (b) selection on specific genotypes, which would result in expression divergence associated with genetic differentiation.

As the infection progresses, we hypothesize that P. homocircumflexum may shift gene expression through selection acting on epigenetically diverse asexually reproducing populations or through plastic responses triggered by host immune signals. By examining these scenarios, our study aims to determine whether gene expression in a generalist parasite is primarily shaped by adaptive responses to the recipient host environment or by retained adaptations from previous hosts. This research will deepen our understanding of the adaptability of generalist parasites and the evolutionary pressures that shape host–parasite interactions.

Results

Parasitemia

Parasitemia from all the individuals changed over the course of the infection (Fig. 1) achieving maximum values around 8 d post infection (dpi) in most of the individuals.

Fig. 1.

Parasitemia levels for each experimental group, showing individual trajectories. Panels indicate combinations of donor and recipient species: a) canaries receiving infected blood from canaries, b) canaries receiving infected blood from siskins, c) siskins receiving infected blood from canaries, and d) siskins receiving infected blood from siskins. Each line represents an individual. Dashed lines indicate days of sampling for each group (8, 12, and 16 dpi).

Principal Component Analysis

The principal component analysis (PCA) results obtained from normalized data show distinct clustering patterns for donors and recipients over time (Fig. 2). At 8 dpi, there was a clear separation based on the donor (Fig. 2a), indicating significant differences in their data profiles. However, this separation was not observed when considering the type of recipient (Fig. 2d). At 12 dpi, the distinction between the groups became less pronounced, suggesting a convergence of their data profiles (Fig. 2b–e). At 16 dpi, the PCA plots have shifted to form a tight clustering of samples in the recipient birds of siskins, regardless of the donor (Fig. 2f). In contrast, grouping by donor resulted in more dispersed data points with no clear clustering (Fig. 2).

Fig. 2.

Biplot from a PCA illustrating the variance in the data for donors and recipients at 8, 12, and 16 dpi. Shapes represent different time points: triangles for 8 dpi, squares for 12 dpi, and circles for 16 dpi. The first column shows PCA plots of donors (top row) and recipients (bottom row) at 8 dpi, the second column at 12 dpi, and the third column at 16 dpi. Each plot indicates the percentage of variance explained by the first two principal components (PC1 and PC2). Ellipses represent the 95% confidence intervals for each group.

Differential Expression Analyses

During the course of infection, P. homocircumflexum underwent transcriptional changes as it adapted to the host environment (Figs. 3 and 4). At 8 and 12 dpi, parasites originating from the same donor showed more homogeneous gene expression patterns, regardless of the species they infected. This observation suggests that early in the infection, parasites relied on preestablished gene expression profiles rather than immediately adapting to the new host environment. This pattern was reflected in the overall expression trends (Fig. 3) and the number of differentially expressed genes (DEGs), which were found to be higher when grouped by donor compared to recipient (Fig. 4). During this phase, parasites may prioritize strategies for immune evasion, red blood cell invasion, and replication while maintaining a transcriptional signature inherited from their original host.

Fig. 3.

Percentage of DEGs between donors and recipients at three time points: 8, 12, and 16 dpi. For each time point, two adjacent bars represent the percentages for donors and recipients, respectively, allowing direct comparison.

Fig. 4.

Minus average plots showing differential gene expression analysis at 8, 12, and 16 dpi. The three plots on the top (a–c) show the results when analyzing the data by taking into account the donor species. The three plots on the bottom (d–f) show the results when analyzing the data by taking into account the recipient species. The x axis represents the log₂ fold change, and the y axis shows the −log₁₀–adjusted P–value. Points above the significance threshold indicate significantly upregulated or downregulated genes, while the remaining points are nonsignificant. Highlighted gene names correspond to the top 20 DEGs (FDR 2). The total number of DEGs is indicated for each plot.

However, by 16 dpi, the influence of the recipient species emerged as the primary factor shaping parasite gene expression (Fig. 3). At this stage, transcriptomic profiles were more distinct between host species, suggesting an active response to new physiological and immunological pressures. DEG patterns also shifted, with a higher percentage of DEGs associated with the recipient rather than the donor (Fig. 4). The most significant changes occurred between 8 and 16 dpi. At 8 dpi, 2,647 DEGs were detected when grouping by donors (Fig. 4a), while only 60 DEGs were observed when considering recipients (Fig. 4d). By 16 dpi, this trend reversed, with 271 DEGs when grouping by recipients (Fig. 4f) and only 19 DEGs by donors (Fig. 4c). These findings suggest that as infection progressed, parasites shifted from a generalist gene expression strategy to a more specialized adaptation tailored to the recipient host, likely involving metabolic adjustments and mechanisms for immune evasion. This highlights the plasticity of P. homocircumflexum in modulating its transcriptome to optimize survival across different host environments.

Variant Calling

A distribution of the samples according to their Single Nucleotide Polymorphism (SNP) was obtained for day 8 dpi (Fig. 5). As indicated in the Materials and Methods, similar analysis could not be performed for 12 and 16 dpi. The results indicate that samples do not cluster clearly when classified according to either recipient (Fig. 5a) or donor (Fig. 5b) species. The distribution of the samples showed no clear pattern when considering either the recipient (Fig. 5a) or donor (Fig. 5b) species. Additionally, two Fixation Index (FST) analyses were conducted. The first analysis yielded an FST value of 0.09 (95% CI 0.02 to 0.15) when considering donor species, while the second analysis resulted in an FST of 0.03 (95% CI 0.01 to 0.14) when considering recipient species. These FST values further support a lack of clear genetic differentiation among the samples based on the species classification.

Fig. 5.

PCA based on SNPs in expressed RNA in samples from 8 dpi classified based on two different criteria. a) The distribution of samples when classified according to the recipient species. b) The distribution of the same samples but classified based on the blood donor.

Discussion

Our results provide insights into P. homocircumflexum infection in two avian host species, showing how generalist parasites adapt to phylogenetically distinct hosts. Parasitemia peaked at 8 dpi, with RNA expression shifting from donor-dependent at 8 dpi to recipient-dependent by 16 dpi. Regarding the scenarios, (i) the “one key fits all” strategy could be ruled out, as gene expression shifted according to the host environment, and (ii) transcriptomic plasticity could be partially supported, as expression in later stages reflected the recipient host environment, suggesting adaptation. Hence, our findings reject both extremes, revealing an intermediate profile with a shift from donor- to recipient-aligned expression. This pattern is consistent with the idea that gene expression is plastic but may initially reflect donor-derived regulation. We explored two mechanisms that could underlie this pattern: (i) epigenetic inheritance and (ii) selection-driven genetic divergence. While the first remains speculative, it aligns well with our data; the second is not supported by our genetic control analyses. Therefore, the observed pattern is consistent with epigenetic inheritance (i), whereby transcriptional profiles are retained from the donor via asexual replication and gradually adjust to the recipient host environment. While we did not directly measure epigenetic marks, this mechanism offers a plausible explanation for the early donor-driven expression. However, selection-driven differences (ii) are not supported, as our analysis of SNP variation served as a control to assess whether expression profiles could be explained by underlying genetic divergence. The low FST values and lack of clustering in PCA indicate that the observed transcriptional differences are unlikely to result from selection on divergent haplotypes. These findings underscore the plasticity and adaptability of generalist parasite plasticity in diverse host environments. Although we did not find evidence for selection acting on genetic variants (haplotypes), it is also possible that natural selection acts directly on gene expression levels (Nourmohammad et al. 2017; Cope et al. 2025). This form of selection could be considered an extension of the second mechanism, acting not on sequence variation but on regulatory variation. Below, we discuss these results in detail.

Although the infection dynamics of some parasites are well understood, classifying a parasite as a generalist or specialist depends on the density of its reservoir host (Fecchio et al. 2021). In this context, the biodiversity and abundance of species within an ecosystem may influence whether a parasite lineage is more or less generalist in its host range (de Angeli Dutra et al. 2021; Doussang et al. 2021; Prati et al. 2022). Indeed, previous studies have determined that the same parasite species can adapt to a more specialist role in environments with a limited number of potential host species, compared to scenarios where many potential hosts are available (de Angeli Dutra et al. 2021; Dharmarajan et al. 2021). The molecular mechanisms and adaptations that a generalist parasite population can undergo during infection and reproduction in different host species have been insufficiently analyzed. Our results provide new insights into this issue. Despite exhibiting similar levels of parasitemia, we revealed that the source of infection significantly influences the RNA expression patterns of the parasite. In essence, the donor species seems to dictate parasite expression in the days following infection. However, after 16 dpi, expression patterns appear to be primarily determined by the infected species (i.e. the recipient). According to the existing scientific literature, factors contributing to these changes in RNA expression patterns may include genetic variations such as SNPs, epigenetic modifications, or direct plasticity (Lorà-Batlle et al. 2019).

Single nucleotide polymorphisms in human malaria have been pivotal in identifying genetic factors that influence susceptibility to infection and drug resistance (Gill and Sharma 2022; Dinzouna-Boutamba et al. 2023). Over the last decade, SNP studies have been extensively used to uncover host–pathogen interactions and guide the development of targeted therapies and vaccines (e.g. Maïga-Ascofaré et al. 2010). In this study, we have used established SNP analysis workflows to determine whether the differences found during peak infection (8 dpi) could be due to natural selection within the parasite population. Our findings indicate that genetic selection during peak infection is 9% based on donor species and 3% based on recipient species. While these percentages suggest a small degree of selection at this stage of the infection cycle, we cannot definitively conclude that this process accounts for the observed differences. Additional methods (e.g. discriminant PCA or phylogenetics) could complement this analysis, but the limitations of our RNA-seq-derived SNP dataset restrict their applicability.

Epigenetic mechanisms, such as histone modifications and DNA methylation, have been recognized as significant drivers of transcriptional variation in parasites. These mechanisms enable heritable and reversible changes in chromatin state, which are crucial for the adaptive capacity of these organisms (Holliday 2006; Handel et al. 2010). This flexibility is particularly important for generalist parasites, as it facilitates their adaptation to changing environments during infection (Duraisingh and Skillman 2018; Hollin et al. 2023). In mammalian malaria systems, including P. falciparum, epigenetic factors have been shown to drive early gene expression changes (Adejoh et al. 2023). It is plausible that P. homocircumflexum employs similar mechanisms inherited from the donor population to regulate gene expression at 8 dpi. These inherited epigenetic responses may provide a strategic advantage during the early stages of infection, allowing for rapid adaptation to the new host environment. As the infection progresses, the host immune system could exert selective pressure (Abdrabou et al. 2021), resulting in recipient-dependent patterns of gene expression. This shift underscores the parasite's dynamic adaptability in response to host-derived signals and immune challenges, highlighting the complexity of generalist parasite–host interactions. To test this hypothesis more directly, RNA expression data from the donor individuals at the moment of blood transfer would be required, enabling comparison with their corresponding derived 8 dpi populations. Although, these specific samples were not collected for transcriptomic analysis, such analyses could clarify whether transcriptional patterns reflect inherited epigenetic regulation or other adaptive processes.

Direct transcriptional variation refers to transient, nonheritable changes in gene expression that are triggered by environmental stressors (Llorà-Batlle et al. 2019). In the context of host–parasite interactions, these mechanisms can be influenced by the host immune response, which evolves dynamically over the course of an infection. Around 12 dpi, the host immune system transitions from innate to adaptive immunity (Sorci 2013; Delhaye et al. 2018), marking a critical turning point in the interaction between the host and the parasite. In our study, we observed a distinct shift in RNA expression patterns between 8 and 16 dpi. Early in the infection (8 dpi), parasite gene expression appeared to be dependent on the donor population. However, by 16 dpi, it was predominantly shaped by the recipient species. This temporal change aligns with the timing of immune system modulation, as the influence of the donor diminishes and the adaptive immune signals from the recipient take precedence. Our findings support the idea that parasite transcriptional responses are not only reactive but also adaptive to the host immune environment. This highlights the role of immune system dynamics in shaping gene expression. These results provide valuable insights into how immune responses drive transcriptional variation, furthering our understanding of the complex interplay between host and parasite.

A critical next step is to investigate the host immune response over time, with a particular focus on distinct immune pathways or gene groups that are activated at different stages of infection. RNA-seq analyses could be instrumental in uncovering pathways associated with innate immunity during the early stages and adaptive immunity in the later stages of infection, thereby providing valuable insights into host–parasite dynamics. Furthermore, exploring cytokine responses, T-cell activation pathways, and other immune-related processes would help clarify the role of host immunity in shaping parasite gene expression. Such analyses could elucidate whether specific pathways are consistently targeted by P. homocircumflexum across hosts, which would improve our understanding of host-specific versus generalist responses.

Our experimental design, which involved blood inoculations, ensured that parasite populations began from the same starting point. This controlled setup allowed us to detect significant differences in RNA profiles influenced by donor species at 8 dpi and by recipient species at 16 dpi. The donor effect observed during early stages highlights the importance of parasite synchronization with the donor host. These results suggest that gene expression patterns may differ in natural infections transmitted by vectors due to the additional epigenetic regulation imposed by vector stages. Follow-up studies that include vector–host–parasite interactions would provide valuable insights into RNA expression dynamics and the influence of epigenetic regulation during transmission. Moreover, comparisons with human malaria experiments, where vector-stage regulation has been well documented (Yu et al. 2022), could further refine our understanding of these processes in avian systems.

In conclusion, our study highlights the dynamic transcriptional plasticity of P. homocircumflexum during infection in two avian host species, emphasizing how generalist parasites respond to distinct host environments. We observed that parasitemia peaked at 8 dpi, with RNA expression patterns initially influenced by the donor species before shifting to profiles driven by the recipient by 16 dpi. This transition may coincide with the progression of the host immune response from innate to adaptive immunity, suggesting that host-derived signals could influence parasite transcriptional responses. However, since we did not directly assess host immune markers, further studies are needed to confirm this relationship. Our SNP analyses revealed limited genetic selection, indicating that the observed transcriptional shifts are unlikely driven by genetic changes but may instead result from epigenetic mechanisms such as histone modifications or DNA methylation. These mechanisms may provide the parasite with transcriptional flexibility necessary to rapidly adjust to host-specific environments. Nonetheless, whether these changes are adaptive in the context of transmission remains unclear. In particular, the relationship between within-host transcriptional shifts and transmission success warrants further investigation. If the genes undergoing regulation are involved in gametocyte production, this could indicate an adaptive response facilitating transmission. Future research incorporating gametocyte-stage expression data and vector-mediated infections will be crucial to fully understand the interplay between host immunity, epigenetic regulation, and parasite adaptation.

Materials and Methods

Experimental Animals and Parasite Acquisition

The research was conducted in 2019 at the Nature Research Centre in Vilnius, Lithuania, using juvenile domestic canaries (Serinus canaria domestica) and Eurasian siskins (Spinus spinus) from commercial suppliers. All procedures complied with European and Lithuanian regulations on animal research, with ethical approval from the State Food and Veterinary Service of Lithuania (approval no. 2018/05/03-G2-84).

Birds were housed individually in cages, allowing social interaction, in a vector-free room with a stable 21 ± 1 °C temperature and a natural light/dark cycle. Food and water were provided ad libitum. Before experiments, Polymerase Chain Reaction (PCR) and microscopy confirmed all birds were malaria-free. The parasite used was P. (Giovannolaia) homocircumflexum, lineage COLL4 (GenBank KC884250), originally isolated in 2010 from a wild red-backed shrike (Lanius collurio) captured at the Biological Station of the Zoological Institute of the Russian Academy of Sciences on the Curonian Spit, Baltic Sea. The strain was cryopreserved following Dimitrov et al. (2010) and later used for experiments.

Experimental Procedures

Eight canaries and six siskins were randomly assigned to four experimental groups (Fig. 6). The P. homocircumflexum strain COLL4, obtained from a canary, was used to infect four canaries and three siskins. The same strain, sourced from a siskin, was used to infect four additional canaries and three siskins. The experiments used the parasite’s seventh passage.

Fig. 6.

Experimental design for this study involved using blood from an initial donor for consecutive experimental infections. One canary and one siskin served as blood donors, while a total of six siskins and eight canaries were used as recipients. Colored arrows indicate the origin of the infected blood: orange arrows represent blood from siskin donor, while green arrows represent blood from canary donor. (Drawings obtained with the artificial intelligence of Artbreeder and the final figure designed with Canva).

To initiate infections, cryopreserved blood was introduced into a noninfected canary and siskin, which then served as donors: canary-derived parasites for canaries and siskins and siskin-derived parasites for canaries and siskins. Each bird received 100 µL of a freshly prepared suspension containing infected blood, 3.7% sodium citrate, and 0.9% saline (4:1:5 ratio), with an erythrocytic meront intensity of 0.4% and an estimated dose of 6 × 10⁵ meronts. The mixture was injected into the pectoral muscle following Iezhova et al. (2005). Birds were monitored for 32 dpi, with blood samples collected every 4 d for microscopy and PCR. Additional samples for RNA sequencing were taken at 8, 16, and 24 dpi.

Sampling, Microscopy, and Molecular Analysis

Blood samples were obtained by puncturing the brachial vein. A small drop of blood from each bird was used to prepare two blood smears, while approximately 30 µl was stored in SET buffer (0.05 M Tris, 0.15 M NaCl, 0.5 ethylenediaminetetraacetic acid, pH 8.0) for molecular analysis (Hellgren et al. 2004). Samples for RNA analysis were preserved in TRIzol LS Reagent (Invitrogen, Carlsbad, CA). Blood samples were kept at −20 °C until processed. The blood smears were air-dried, fixed with absolute methanol, and stained with Giemsa (Valkiūnas et al. 2008). The slides were examined under an Olympus BX51 light microscope equipped with an Olympus DP12 (Olympus, Shinjuku City, Japan). Approximately 100 fields were reviewed at a magnification of 1000×. Parasitemia was quantified by counting the number of parasites per 1,000 erythrocytes, or per 10,000 erythrocytes in cases of low infection intensity, following the protocol recommended by Godfrey et al. (1987). Detailed procedures for preparing, staining, and examining blood smears were outlined by Valkiūnas et al. (2008).

DNA extraction was performed using the ammonium acetate protocol (Sambrook et al. 1989). A nested PCR protocol was employed for genetic analysis (Waldenström et al. 2004). Amplification products (1.5 µl) were resolved on a 2% agarose gel. Sequencing followed the protocol by Bensch et al. (2000), targeting the 5′ region with the primer HaemF. Dye terminator cycle sequencing (BigDye) was conducted on an ABI PRISMTM 3100 capillary sequencer (Applied Biosystems, USA). Sequence editing and alignment were carried out using BioEdit (Hall 1999).

RNA Extraction and Sequencing

Blood samples for RNA sequencing were collected at three different time points (8, 16, and 12 dpi). The samples were stored in TRIzol. Total RNA was extracted from 20 μL whole blood using 1000 μL TRIzol LS Reagent (Invitrogen, Carlsbad, CA) and homogenized by vortexing from all seven individuals. Samples were then incubated at room temperature for 5 min before the addition of 200 μl chloroform (Merck KGaA, Darmstadt, Germany). After a further incubation at room temperature for 3 min, the samples were centrifuged at 11,000 rpm for 17 min at 4 °C. The supernatant was then transferred to new tubes and processed using an RNeasy Mini Kit (Qiagen, GmbH, Hilden, Germany). Following the manufacturer's protocol, 1 volume of 70% ethanol was added to the lysate. The total extracted RNA was shipped on dry ice to Novogene Bioinformatics Technology, Hong Kong, for RNA quality control, DNAse treatment, and rRNA reduction and amplification using the SMARTer Ultra Low Kit (Clontech Laboratories, Inc.). Novogene performed library preparation, cDNA synthesis, and paired-end RNA sequencing using the Illumina HiSeq 2000. We quality checked all demultiplexed RNA-seq reads using FastQC (v.0.10.1) (Andrews 2010).

De Novo Transcriptome Assembly

As the genome of P. homocircumflexum is not available, alternative tools were employed in order to isolate information belonging exclusively to the P. homocircumflexum lineage pCOLL4. Initially, the canary genome (NCBI RefSeq assembly serCan2020) and a portion of the siskin genome (NCBI RefSeq assembly ASM3478079v1) were employed to eliminate reads belonging to the bird. Subsequently, the P. relictum genome (NCBI RefSeq assembly GCA_900005765.1; Böhme et al. 2018) was utilized to keep reads belonging to the malaria parasite. Finally, the P. homocircumflexum transcriptome previously published (García-Longoria et al. 2020) was employed to obtain a greater number of reads and ensure that the reads used exclusively belonged to the malaria parasite. Analyses with and without the reference genome were performed using the bioinformatics software STAR (Dobin et al. 2013) and Bowtie (Langmead and Salzberg 2012). Standard adjustments were implemented to determine alignment sensitivity. Finally, the de novo assembly of the parasite was conducted using the transcriptome assembler Trinity (v. 2.15.1) (Grabherr et al. 2011), resulting in 14,649 contigs.

Guanine–Cytosine (GC) content varies across different eukaryotic organisms, with where Plasmodium species evolving toward a high Adenine–Thymine (AT) richness compared to its hosts (Galtier et al. 2001; Birdsell 2002; Hershberg and Petrov 2010; Videvall 2018), which provides a valuable tool for transcript separation in host–parasite studies. Consequently, the subsequent transcript was subjected to filtration based on GC content in order to exclude any host contigs and to include sequences with a mean GC content below 23% (n = 12,058).

Differential Gene Expression Level

The expression levels of the parasite genes were quantified using Salmon (v. 1.3.2) (Patro et al. 2017). Salmon is a software program that produces expected read counts for every contig and identifies genes that are expressed between species, along with the name of each contig. The read counts for every contig were stored in a file that was statistically analyzed inside the R statistical environment (v. 4.5.0) (R Core Team 2023). Read counts were normalized using regularized log transformation in order to account for potential variation in sequencing depth and the large differences in the number of parasites present in the blood (parasitemia levels). Regularized log transformation of counts was performed in order to represent the data without any prior knowledge of the sampling design in the PCA and sample distance calculations. This method of presenting counts without bias is preferable to variance stabilizing of counts when the size factors vary greatly between the samples, as is the case in our data. The package ggplot2 (Wickham 2016) was employed for the generation of all graphs.

Statistical Analyses

For the differential expression analyses, we used the R package DeSeq2 (love et al. 2014), which is designed for the specific task of performing this type of analysis. This package (version 1.16.1) was utilized to correct the variance–mean dependence in count data derived from high-throughput sequencing assays and to test for differential expression based on a model that employs the negative binomial distribution. When significant differences in expression were identified, and to circumvent potential issues associated with sequencing depth, gene length, or RNA composition, count data were initially normalized using the DeSeq2 method.

The statistical comparisons are presented in Fig. 6. Firstly, the samples were classified according to the donor. This distinction was made regardless of the species of bird receiving the blood (Fig. 6). Subsequently, the samples were classified based on the recipient of the blood (Fig. 6). In this instance, only the species of bird receiving the blood was considered. This approach enabled the number of DEGs to be determined in relation to either the origin of the blood (from birds of mixed species but with the same donor species) or the recipient (from birds of the same species but with infection originating from different species of donor).

Author Contributions

V.P. and J.A. performed all experimental work through experimental infections with and all the laboratory procedures. O.H. and V.P. conceived the study, designed the overall experiment, and contributed to the conceptual development. O.H., V.P., and A.M. secured funding for the research. L.G.-L. conducted all bioinformatics analyses and statistical assessments, drafted the manuscript, and contributed to the conceptual development and interpretation of results. D.O. provided support and assistance with bioinformatics analyses. All authors reviewed, edited, and approved the final version of the manuscript.

Data Availability

All the data used in this study has been uploaded to the National Center for Biotechnology Information (NCBI) under the accession number PRJNA1250583.