Vive la différence: Exploiting the differences between rodent and human malarias

Abstract

Experimental research into malaria biology and pathogenesis has historically focused on two model systems, in vitro culture of the human parasite Plasmodium falciparum and in vivo infections of laboratory animals using rodent parasites. While there is clear value in having a manipulatable animal model for studying malaria, there have occasionally been controversies around how representative the rodent model is of the human disease, and therefore significant emphasis has been placed on the similarities between the two biological systems. By focusing on basic nuclear functions, we wish to highlight that identifying key differences in the parasites and their interactions with their mammalian hosts can be equally informative and provide remarkable insights into the biology and evolution of these important infectious organisms.

Contributions of the rodent malaria models

Malaria remains a major threat throughout the developing world, in particular in Sub-Sahara Africa where it is responsible for over 400,000 deaths annually [1]. The human disease is caused by five species of parasites of the genus Plasmodium that infect circulating red blood cells (RBCs), causing illness ranging in severity from high fevers and flu-like symptoms to death [2]. Only one of these species, P. falciparum, is routinely cultured in vitro, and thus it has become the primary experimental model for the human malarias. This system is widely used and has contributed substantially to our current knowledge of the cell and molecular biology of these organisms. With the advent of the CRISPR/Cas9 (see Glossary) system for genome editing [3, 4], discoveries based on genetic manipulations have become even more readily achievable. While cultured parasites are therefore enormously useful for many basic studies, the primary deficiency of this system is the inability to directly study host-parasite interactions. All malaria parasites display relatively stringent host-species requirements, therefore in vivo experiments with parasites that infect people have been limited to experimental human infections, chimpanzees and in some cases New World monkeys, all at considerable expense and with significant ethical limitations. The primate parasite P. knowlesi can infect both humans and Old World monkeys and its use as an experimental system is rapidly improving [5], however for routine experimental investigations, there are no readily available in vivo models, with the exception of highly specialized systems in which human tissue is engrafted into immunodeficient mice [6, 7].

This deficiency has been addressed by the development and use of rodent infectious parasites in laboratory mice and rats. The lion’s share of this work has utilized the parasites P. berghei, P. yoelii and P. chabaudi originally isolated from the African thicket or tree rat [8, 9]. Studies using these parasites have been extensively applied to deciphering the mammalian immune response to malaria [10] employing the ever-growing catalogue of transgenic mice that are deficient in various immune parameters. Thus, much of our current knowledge of malaria immunology is based on the mouse model [11]. Similarly, studies of the liver stage of the parasite’s lifecycle [12] and our understanding of parasite development in the mosquito have benefited greatly from work with rodent parasites [13, 14]. In addition, the drug development pipeline routinely applies studies of rodent parasites to help determine bioavailability and efficacy of new antimalarial compounds [15].

Despite these advantages, there are some deficiencies in how well the rodent model represents human malaria, and some controversies have occasionally arisen regarding possible over-interpretation of conclusions derived from rodent malaria infections [16, 17]. In general, these concerns have focused on aspects of pathogenesis, in particular the relevance of the rodent model for understanding severe disease, including cerebral malaria and placental malaria [18]. These discussions generally have focused on the similarities between rodent and human infections, with the emphasis placed on how well the rodent models represent the human disease. While such discussions are clearly necessary and important, we believe it is equally important not to lose sight of what can be gained by focusing on the differences between the rodent and human parasites. Here we focus on nuclear functions that are typically highly conserved between closely related species but have proven to be remarkably divergent when comparing parasites that infect rodents and humans. Given the parallels between their lifecycles and their extensive genome homology, distinct discrepancies in their biology likely reflect important evolutionary and physiological differences that can provide key insights into the selective pressures that have shaped these parasites and enable them to survive and transmit between hosts.

Insights from the parasite’s nucleus

Both rodent and human malaria parasites have very similar life cycles, including sexual recombination and transmission via Anopheles mosquitoes and an obligatory initial infection of the liver prior to release into the blood stream and propagation within the RBCs of their mammalian hosts. The sequences of their genomes display extensive synteny [19], and transcriptional regulatory elements have been shown to function cross-species [20], thus demonstrating their very close evolutionary relationship. Given the highly conserved nature of nuclear functions (DNA replication and repair, transcription, nuclear division, etc.) and how closely related these parasites appear to be, one would expect very few differences in basic nuclear pathways. Nonetheless, several remarkable and surprising differences have been described in basic nuclear capabilities, providing insights into several aspects of the parasites’ lifestyles.

Polymorphisms in RNA polymerase II

One of the most highly conserved proteins amongst eukaryotic organisms is the catalytic subunit of RNA polymerase II, a protein call Rpb1. This protein is required for the synthesis of messenger RNA. Rpb1 displays a conserved domain structure that in most eukaryotic organisms includes a C-terminal domain (CTD) consisting of a series of heptad repeats of variable number [21]. The number of repeats is distinct within broad evolutionary lineages, for example all mammals have precisely 52 repeats of the amino acid sequence YSPTSPY/K [22]. The lack of variation in this repetitive sequence is one of the hallmarks of Rpb1, thus it was surprising to discover that primate and rodent malaria parasites display differing numbers of heptad repeats within their respective CTDs, with the primate parasites possessing a significant increase in the number of heptads [23]. This was the first observation of variability of heptad number within a single genus and suggested a possible deviation in the function of this domain of the protein, something that would have been difficult to predict given how closely related these organisms are.

Investigation into the possible functional significance of the CTD expansion in the primate parasites led to interest in the histone methyltransferase Su(var)3–9, Enhancer of Zeste, Trithorax 2 (PfSET2) and its cognate demethylase Jumonji C1 (PfJmjC1). Orthologues of SET2 in higher eukaryotes are known to bind to the CTD of Rpb1 during mRNA transcription, thus being recruited to sites of active gene expression where it marks the associated chromatin by methylating lysine 36 of histone H3 (H3K36me3) [24]. Remarkably, orthologues of both SET2 and JmjC1 are missing in the genomes of the rodent parasites, providing another, perhaps related, example of unexpected variability of core nuclear functions between the rodent and human parasites (Figure 1A) [25]. PfSET2 was then independently identified as an important component of the machinery involved in the epigenetic regulation of the var gene family involved in antigenic variation in P. falciparum [26]. Rodent parasites also undergo antigenic variation, but the mechanism appears to differ significantly, with changes in expression of antigen encoding genes linked to the transition from acute to chronic infection rather than to antibody mediated immunity [27]. These observations led to the hypothesis that PfSET2 might be recruited to var genes by RNA pol II, perhaps during the transcription of noncoding RNAs, where it deposits the H3K36me3 mark necessary for proper gene regulation. This model was partially confirmed through experiments designed to disrupt recruitment of PfSET2 by RNA pol II, which cause marked changes in var gene expression [25]. Thus, observations of surprising differences between RNA pol II in rodent and human parasites led to novel mechanistic insights into the process of antigenic variation.

Figure 1: Phylogenetic trees displaying the loss in the rodent parasite lineage of genes encoding proteins involved in key nuclear functions.

(A) Genes encoding orthologs of phosphoethanolamine-N-methyltransferase (PMT), gametocyte development 1 protein (GDV1), Su(var)3–9, Enhancer of Zeste, Trithorax 2 (SET2) and Jumonji C1 (PfJmjC1) have been lost from the rodent parasites but are identifiable and syntenic in the parasites that infect primates and birds. (B) Genes encoding components of canonical nonhomologous end joining (cNHEJ) as represented by Ku are missing in both primate and rodent parasites, but a Ku-like protein is observed in avian parasites. Components of the translesion polymerase complex, including Rev1, pol ζ, a SNF2 helicase and a ring finger ligase, are missing from the rodent parasite genomes. Adapted from [43], Copyright © American Society for Microbiology, mBio, 11, e03272–19.

An extra regulator in the sexual differentiation pathway

While the symptoms of malaria result from the asexual replication of parasites within the red blood cells of the host, transmission through the mosquito vector is dependent on differentiation into the sexual forms, a process called gametocytogenesis. Sexual differentiation requires expression of a conserved, master-regulator called AP2G, a transcription factor thought to initiate the gene expression cascade that ultimately leads to the formation of male and female gametocytes [28, 29]. Expression of AP2G has been shown to be induced by changes in environmental conditions, and AP2G is conserved in parasites that infect both rodents and humans [30]. However, induction of gametocytogenesis in the rodent and primate lineages displays some key differences, in particular the absence of an additional regulatory protein called Gametocyte Development 1 (GDV1) from rodent parasites (Figure 1A) [31]. GDV1 appears to be necessary for commitment to sexual differentiation in P. falciparum and the gene encoding this protein is controlled by expression of an antisense noncoding RNA [32]. While present in the genomes of primate and bird parasites, the entire locus is missing from the syntenic position in the genomes of rodent parasites, and no clear alternative gene has been identified. This implies the loss of a level of control of sexual differentiation in rodent parasites, but the consequences of this loss are not yet understood.

Similarly, a protein involved in lipid metabolism called phosphoethanolamine-N-methyltransferase (PMT) has been shown to influence gametocyte formation and development in P. falciparum [33], suggesting that aspects of lipid acquisition and metabolism could play important roles in sexual differentiation. The potential importance of lipid metabolism for sexual commitment in P. falciparum was reinforced by studies showing that lipid metabolite lysophosphatidylcholine can strongly suppress sexual commitment rates [34]. While PMT appears to be important for gametocytogenesis in P. falciparum, the gene encoding this protein has also been lost in rodent parasites (Figure 1A), again implying that key differences exist in the regulatory cascade that determines how and when sexual differentiation occurs in the different parasite lineages. Understanding the selective pressures inherent to the different host-parasite relationships found in rodent vs primate hosts are likely to provide key insights into important aspects of conditions that influence malaria transmission.

Loss of translesion polymerases for DNA repair and recombination

The genomes of malaria parasites display extreme sequence diversity between members of the multicopy gene families that encode proteins localized to the surface of infected red blood cells [35]. These gene families can contain from tens to hundreds of copies per genome, with each individual gene having a unique sequence. This high degree of sequence diversity enables the parasites to avoid cross-reactive antibody responses to the surface proteins, thus by changing which gene is expressed, parasites can avoid adaptive immunity and perpetuate chronic infections [36]. This vast sequence variability is thought to largely derive from frequent recombination events between family members, occurring during both asexual replication and the sexual cycle [37–39]. The genomes of malaria parasites do not encode key components of the canonical nonhomologous end joining (cNHEJ) DNA repair pathway, therefore DNA double strand breaks must be repaired by homologous recombination (HR) or an inefficient alternative NHEJ pathway [40, 41]. The loss of cNHEJ has been proposed to facilitate more frequent recombination between variant antigen encoding genes, thus driving the diversification process and providing a selective advantage for the absence of a major DNA repair pathway [37]. Interestingly, the genomes of avian parasites appear to have retained a single, highly divergent component of cNHEJ (Figure 1B), however in the absence of the rest of the pathway, the function of this protein is unknown (see Box 1).

Box 1: A mysterious role for Ku in bird malaria parasites.

Since the initial assembly and publication of the full P. falciparum genome sequence, it was noted that all components of the cNHEJ pathway are missing. There are no identifiable homologs of key components of the pathway including Ku70/Ku80, DNA ligase 4, XRCC4 or Artemis. This pattern held true as additional genome sequences of parasites of rodents, primates and bats were sequenced and annotated. However, it was noted recently that Plasmodium species that infect birds, P gallinaceum and P relictum, harbor syntenic genes that contain a Ku70/Ku80 beta-barrel domain (see Figure 1B). The sequence of this putative Ku protein is quite divergent from Ku proteins of higher eukaryotes and lacks other conserved domains present in Ku proteins in model organisms. No homologs of Ligase 4, XRCC4 or Artemis can be identified suggesting that despite the presence of this Ku-like protein, cNHEJ is likely not functional. Trypanosomes similarly harbour homologs of Ku 70 and Ku 80 but are known to not have a functional cNHEJ pathway [53]. In these organisms the Ku proteins instead are required for telomere maintenance [54, 55], and Ku knockouts display no change in sensitivity to DNA damaging agents [54]. In contrast, Toxoplasma gondii, an apicomplexan parasite related to Plasmodium, has robust cNHEJ and knock outs of the Ku proteins result in a substantial increase in sensitivity to DNA damaging agents, as expected [56]. The presence of Ku-like proteins in avian malaria parasites therefore represents an interesting puzzle that has yet to be explored.

In P. falciparum, the sequence diversity for any multicopy gene family observed within an individual parasite’s genome can be further extended through comparisons between the genomes of different parasite isolates [42]. Essentially, no two parasite isolates look alike, a property that enables parasites to re-infect hosts that have previously harboured an infection. Remarkably however, examination of full genome sequence assemblies from several independent isolates of the rodent parasite P. chabaudi found that the repertoires of multicopy gene families were highly similar, with many syntenic genes encoding proteins of identical sequence [43, 44]. This degree of sequence identity is never observed when different isolates of P. falciparum are compared, indicating that recombination between members of the multicopy gene families is somehow constrained in rodent parasites. Closer examination of the DNA repair pathways found that genes encoding all four components of the translesion (TLS) polymerases are readily identifiable in primate and bird parasites but entirely missing from the genomes of all rodent parasites examined (Figure 1B) [43]. The TLS polymerases facilitate recombination between non-identical sequences and are likely required for the recombination events that drive diversification of the multicopy gene families. Their loss in the rodent parasite lineage therefore provides a probable explanation for the lack of sequence diversity between parasite isolates. In addition, by reducing or eliminating recombination between family members, individual genes could evolve independently, resulting in proteins with alternative functions or subcellular localizations, as has been reported in rodent parasites [45]. Thus the highly conserved, basic molecular mechanisms responsible for the maintenance of genome integrity have diverged significantly in malaria parasites, first through the loss of cNHEJ in all Plasmodium species, and second through the loss of TLS polymerases in the rodent parasite lineage.

Different lifestyles, different selective pressures

The examples mentioned above all describe distinct differences between rodent and human parasites in fundamental aspects of nuclear function, specifically in gene regulation and DNA repair. The changes in RNA pol II, SET2/JmJC1 and the translesion polymerases can all be linked to immune evasion and antigenic variation, demonstrating the powerful role this selective pressure has played on the evolution of these parasites. However, given the similarities between the primate and rodent immune systems, it is not immediately clear what differences could have led to such substantial changes to these highly conserved pathways. One model proposes that differences in lifespans and fecundity could underlie these differences. Specifically, in long-lived hosts that have relatively low reproductive rates like humans, the majority of the potential hosts within any given population are likely to have previously harboured an infection and thus have an extensive repertoire of antibodies that recognize variant antigens (Figure 2A). Such semi-immune individuals are thought to represent a major reservoir for parasite transmission [46], and successful reinfection of such hosts requires extensive variability between parasites. This selective pressure should favour frequent recombination between genes encoding variant antigens as well as tightly controlled transcriptional regulation of antigen encoding genes, properties influenced by the extended CTD of RNA pol II, SET2/JmJC1 and the translesion polymerases. In contrast, a high reproductive rate and relatively short lifespan of a rodent host (more specifically, the host of the common ancestor of extant rodent parasite species) would result in host populations dominated by individuals that are often immunologically naïve, thus greatly reducing the pressure to evade pre-existing immunity (Figure 2B). The observed differences in basic nuclear functions therefore have provided surprising insights into the host-parasites interactions that influence rates of parasite transmission, the length and chronicity of infection, the likelihood of re-infection and the acquisition of immunity.

Figure 2: Hypothetical model for how different infection dynamics could provide alternative selection pressures on parasites that infect primates versus rodents.

(A) Primates tend to be long-lived and have relatively few offspring, thus transmission is dominated by reinfections of semi-immune hosts (red arrows) rather than infection of immunologically naïve, young offspring (yellow arrows). The need to avoid pre-existing immunity could provide strong selection pressure for extensive antigen diversification and for tightly regulated antigenic variation. (B) In contrast, rodents are generally shorter lived and have greater numbers of offspring, thus providing a larger pool of immunologically naïve hosts that dominate transmission (red arrows) compared to previously infected adults (yellow arrows). This would reduce the selective pressure for evading pre-existing immunity.

Similarly, differences in host lifestyle could also have selected for the differences in how rodent and primate parasites control sexual differentiation. For example, humans inhabit geographical regions that can display widely variable transmission dynamics, ranging from stable, continuous transmission to seasonal transmission typified by alternating dry/rainy seasons to unpredictable transmission found in areas with unstable weather patterns. Such variable access to the mosquito vector could have provided the selective pressure for precise regulation of sexual differentiation, potentially enabling them to coordinate gametocyte production to periods when mosquitoes are abundant. How such environmental signals are translated to the parasites is unknown, but the observed links to lipid metabolism perhaps offers clues. If the ancient host of the common ancestor of rodent parasites lived in a more stable transmission environment, such environmental responsiveness would be less necessary. The loss of the gametocyte regulator GDV1 and the metabolic enzyme PMT in the rodent parasite lineage and their retention in parasites that infect both birds and primates make such speculation intriguing.

Concluding Remarks

The insights provided through comparisons with rodent parasites, in particular by identifying unexpected differences, can serve to generate new hypotheses that might not otherwise have been evident. Beyond the nuclear functions described above, comparisons of other aspects of parasite biology have yielded similarly important insights [47], thereby reinforcing the value of extensive investigations into a variety of different parasites systems. Extending such comparative analyses beyond primate and rodent parasites and into additional species of Plasmodium, for example parasites of farmed animals [48–50] and those that infect birds and reptiles [51, 52], is likely to be similarly fruitful (see Outstanding Questions). For example, annotation of whole genome sequences of two bird parasite species revealed the surprising existence of atypical orthologues of Ku, a protein generally involved in cNHEJ and missing in all previously examined lineages of Plasmodium (Figure 1B). The function of these Ku-like proteins and why they have been retained specifically in bird parasites is not understood (see Box 1). Similarly, closer examination of rodent parasites might reveal alternative mechanisms for DNA repair that partially compensate for the loss of TLS polymerases. Comparative studies therefore represent a valuable source of information that is only just beginning to be tapped.

Outstanding Questions.

• What selective pressures led to an additional level of regulation of sexual differentiation in non-rodent malaria parasites, as exemplified by the transcription factor GDV1?

• Why does the liver stage of infection take human parasites approximately a week to complete while rodent parasites complete this stage of the lifecycle in 2 days?

• Why do the gametocytes of rodent parasites mature in 27–30 hours while the human parasite P. falciparum requires 10–12 days?

• Does infecting a nucleated red blood cell (as bird and reptile parasites do) result in differences in interactions with the host cell?

• How does domestication of host animals affect parasite transmission, virulence and the acquisition of immunity?

Highlights.

Experimental malaria research is largely focused on two systems, in vitro culture of the human parasite Plasmodium falciparum and in vivo study of rodent parasites reared in laboratory mice. Both systems have contributed enormously to our understanding of this important parasitic disease.

The rodent malaria model has emphasized its advantages for deciphering aspects of pathogenesis and the mammalian immune response to infection. However, concerns about important differences, in particular regarding the severe complications of cerebral malaria, have raised questions about the utility of the rodent malarias as models for human disease.

Recent work has demonstrated that by focusing on the differences between primate and rodent malarias can yield powerful insights into basic biology, the interactions of parasites with their mammalian hosts and the evolution of the Plasmodium lineage.

Glossary

Antigenic Variation

The systematic alteration over time of the surface antigens of infectious organisms that are exposed to the host immune system.

AP2G

A transcription factor of malaria parasites linked to sexual differentiation. It is a member of the Apetala 2/Ethylene Response Factor (AP2/ERF) family of transcription factors found in plants and apicomplexan parasites.

Canonical Nonhomologous End Joining (cNHEJ)

A major pathway for the repair of DNA double strand breaks. This pathway is generally thought to be more error prone than the other dominant pathway, homologous recombination (HR). The cNHEJ pathway appears to be missing in malaria parasites.

Cerebral malaria

A frequently lethal complication of infection by the malaria parasite Plasmodium falciparum. This syndrome arises from adherence of parasite infected red blood cells within the post-capillary venules of the brain, causing inflammation, brain swelling and obstructed blood flow. This can lead to severe symptoms, including seizures and coma.

CRISPR/Cas9

Clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9. A system for genome editing derived from a bacterial defence system against bacteriophages. When used for genome editing, it enables the creation of targeted DNA double strand breaks.

Fecundity

The ability to produce an abundance of offspring

Gametocytogenesis

The production of sexually differentiated forms, called gametocytes, from asexually replicating forms of malaria parasites. Male and female gametocytes mature into gametes upon arrival in the midgut of a blood feeding mosquito.

Gametocyte Development 1 protein (GDV1)

A transcription factor shown to be required for sexual differentiation in P. falciparum.

H3K36me3

Trimethylation of histone H3 at the lysine in the 36^th position. This is an epigenetic mark implicated in the regulation of transcriptional activity through changes in chromatin structure.

Homologous recombination (HR)

A major pathway for the repair of DNA double strand breaks. This pathway is generally thought to be less error prone than the other dominant pathway, nonhomologous end joining (NHEJ), due to its utilization of a template for repair.

Jumonji C1 (JmjC1)

A lysine demethylase that removes methyl groups from histones. This enzyme activity is implicated in chromatin assembly and epigenetic mechanisms of transcriptional regulation.

Phosphoethanolamine-N-methyltransferase (PMT)

A methyltransferase involved in lipid metabolism, specifically the synthesis of N-methylethanolamine phosphate and N-adenosylhomocysteine from S-adenosyl-L-methione and ethanolamine phosphate.

SET2

The histone methyltransferase Su(var)3–9, Enhancer of Zeste, Trithorax 2. This methyltransferase is thought to deposit methyl groups onto the lysine residue at position 36 of the histone H3. This enzyme activity is implicated in chromatin assembly and epigenetic mechanisms of transcriptional regulation.

Synteny

The conservation of chromosomal position of orthologous genes when comparing two or more species.

Translesion (TLS) polymerases

DNA polymerases that have the ability to synthesize DNA across a damaged template. They have been implicated in aiding homologous recombination between non-identical sequences.