The ecology, evolution, and physiology of Cardinium: a widespread heritable endosymbiont of invertebrates

Abstract

Candidatus Cardinium hertigii (Cardinium) are maternally transmitted obligate intracellular bacteria found in a wide range of invertebrate hosts, including arthropods and nematodes. Infection with Cardinium has substantial consequences for host biology, with many strains manipulating host reproduction to favor symbiont transmission by (i) feminizing male hosts, (ii) altering host sex allocation, (iii) inducing parthenogenesis, or (iv) causing cytoplasmic incompatibility. Other Cardinium strains can confer benefits to their host or alter host behavior. Cardinium-modified host phenotypes can result in selective sweeps of cytological elements through host populations and potentially reinforce host speciation. Cardinium has potential for applications in controlling arthropod pest species and arthropod-vectored disease transmission, although much remains to be explored regarding Cardinium physiology and host interactions. In this review, we provide an overview of Cardinium evolution and host distribution. We describe the various host phenotypes associated with Cardinium and how biological and environmental factors influence these symbioses. We also provide an overview of Cardinium metabolism, physiology, and potential mechanisms for interactions with hosts based on recent studies using genomics and transcriptomics. Finally, we discuss new methodologies and directions for Cardinium research, including improving our understanding of Cardinium physiology, response to environmental stress, and potential for controlling arthropod pest populations.

In this review, the authors synthesize nearly 30 years of research on the widespread endosymbiont Cardinium and its impressive effects on invertebrate biology and evolution.

Introduction

Bacterial endosymbionts are commonly associated with terrestrial arthropods, as well as other invertebrate lineages. Many of these bacteria rely on their host for survival (i.e. obligate host-association) and use host reproduction to spread from female hosts to offspring (i.e. maternal transmission). Symbionts have evolved many mechanisms for ensuring stable transmission, including provisioning essential nutrients to hosts with nutritionally limited diets (Moran et al. 2008), protecting their host from important natural enemies or abiotic stress factors (Corbin et al. 2017, Oliver and Martinez 2014), and/or manipulating host reproduction in ways that favor infected females and improve symbiont transmission, sometimes at the expense of male hosts or uninfected females (Doremus and Hunter 2020, Werren et al. 2008). Symbionts that influence host reproduction are particularly widespread and have garnered considerable research attention given their roles in shaping arthropod evolution and speciation (Brucker and Bordenstein 2012, Gebiola et al. 2016a, Leclercq et al. 2016) and their potential for controlling arthropod populations (Gong et al. 2020, Hoffmann et al. 2011, Zheng et al. 2019).

Several bacterial lineages have evolved the ability to control and influence host reproduction. The overall phenotypes generated by these “reproductive manipulators” are very similar; however, in some cases the underlying mechanisms are assumed and have been in part shown to be completely different (Ferree et al. 2008, Harumoto et al. 2018, Harumoto and Lemaitre 2018, Penz et al. 2012, Pollmann et al. 2022). There are currently eight lineages of bacteria known to engage in reproductive manipulation, with the Alphaproteobacterium Wolbachia being the most widespread and most extensively researched [for recent reviews on Wolbachia see (Ross et al. 2019, Kaur et al. 2021, Hochstrasser 2023)]. In this review, we focus on another widely distributed endosymbiont that frequently engages in reproductive manipulation, Candidatus Cardinium hertigii (hereafter “Cardinium”). Cardinium is a common heritable endosymbiont found in an estimated 6%–13% of arthropod species, including arachnids, insects, and nonmarine crustaceans, as well as other invertebrates such as plant–parasitic nematodes (Duron et al. 2008a, Nakamura et al. 2009, Schön et al. 2019, Tarlachkov et al. 2023, Weinert et al. 2015, Zchori-Fein and Perlman 2004). Cardinium can have variable effects on hosts, but this symbiont often modifies host reproduction through several mechanisms (Doremus and Hunter 2020). These mechanisms include feminizing genetic males to develop as phenotypic females (Groot and Breeuwer 2006, Weeks et al. 2003), induction of asexual reproduction via parthenogenesis (PI), (Giorgini et al. 2009, Matalon et al. 2007, Zchori-Fein et al. 2001, 2004), and cytoplasmic incompatibility (CI) (Gebiola et al. 2016b, Gotoh et al. 2007, Hunter et al. 2003, Nakamura et al. 2012, Nguyen et al. 2017, Zhang et al. 2012). Beyond reproductive manipulation, Cardinium strains can alter other aspects of host biology, including host sex allocation to favor female offspring (Katlav et al. 2022a, b), oviposition (Kenyon and Hunter 2007), and feeding behavior (Ying et al. 2021). Cardinium may also provide both general fitness benefits (e.g. increased longevity and/or fecundity; Wang et al. 2008, Xie et al. 2016), and more conditional benefits (e.g. increased host resistance to natural enemies; Giorgini et al. 2023). The mechanisms underlying these phenotypes are largely uncharacterized and understanding how Cardinium induces these phenotypes remains a priority.

Initial naming conventions and classification

The earliest reports of Gram-negative Cytophaga-like bacterial symbionts, later named Cardinium, began to emerge in the 1970s in insects (Chang and Musgrave 1972), nematodes (Shepherd et al. 1973, Walsh et al. 1983), and later in arachnids (Kurtti et al. 1996). These reports described an irregularly circular or rod-shaped Gram-negative bacterial endosymbiont belonging to the phylum Bacteroidota (previously Bacteroidetes and Cytophaga–Flexibacter–Bacteroides, or CFB group) with distinctive microfilament-like structures attached to the inner membrane (Fig. 1; Kurtti et al. 1996, Nakamura et al. 2009, Zchori-Fein et al. 2004). Initial naming conventions for Cardinium were according to its taxonomic grouping (e.g. Cytophaga-like; Hunter et al. 2003, Weeks et al. 2003), their hosts (e.g. Encarsia bacterium; Zchori-Fein et al. 2001), or some combination thereof [e.g. CFB-BP (Brevipalpus phoenicis); Weeks and Breeuwer 2003]. In 2004, the name “Candidatus Cardinium hertigii” was proposed (Zchori-Fein et al. 2004), following naming conventions for uncultured bacteria (Murray and Stackebrandt 1995). “Cardinium,” from the Latin word cardo, refers to the column-flanked center of Roman towns, which resemble the unique intracellular structures used to morphologically distinguish Cardinium from other bacterial symbionts (Fig. 1; Zchori-Fein et al. 2004). Separately, Candidatus Paenicardinium endonii was proposed for Cardinium-like symbionts found in nematodes (Noel and Atibalentja 2006). Ca. Paenicardinium and the Cardinium-like symbiont of biting midges were later collapsed into the Ca. Cardinium hertigii classification (Nakamura et al. 2009). Following the previous naming convention used for the endosymbiont Wolbachia pipientis (Werren et al. 2008), a uniform naming convention was proposed in the description of Cardinium hertigii cEper1 (Penz et al. 2012). In this naming convention, Cardinium strains are denoted as cHost1, with “c” referring to Cardinium, “Host” referring to the original host that strain was identified in, and the number distinguishing unique strains co-infecting the same host system (e.g. cEhis1 for Cardinium in Encarsia hispida, strain 1).

Figure 1.

Electron micrograph of Cardinium in a follicle cell (surrounding the oocyte) in Encarsia hispida. One cell shows the parallel array of microfilament-like structures that are distinctive of Cardinium. The structures are hypothesized to be the intracellular components of the phage-derived type 6 secretion system T6SS^iv (Bock et al. 2017). Figure reproduced from Zchori-Fein et al. (2004).

Currently, the literature differentiates Cardinium into at least three phylogenetic clades. Due to the broad host range and high levels of divergence between some Cardinium strains, it has been suggested that Cardinium clades should be considered distinct species rather than be collapsed under the same name (Siozios et al. 2019). Clade A has the widest range of hosts, including Cardinium of both mandibulate and chelicerate arthropods (Nakamura et al. 2009). Cardinium of nematodes make up the proposed clade B (Brown et al. 2018, Guo et al. 2022, Nakamura et al. 2009, Noel and Atibalentja 2006), with strains found in the nematode Pratylenchus penetrans potentially forming a separate clade (Tarlachkov et al. 2023). Clade C is composed specifically of Cardinium strains associated with Culicoides biting midges (Nakamura et al. 2009, Siozios et al. 2019). This phylogenetic structure is labile, with recently proposed phylogenies suggesting as many as seven possible clades accounting for newly identified Cardinium strains (Tarlachkov et al. 2023). For example, a clade including Cardinium of Opiliones (Chang et al. 2010, Stouthamer et al. 2019) as well as clades of Cardinium of freshwater copepods (Edlund et al. 2012) and ostracods (Çelen et al. 2019, Schön et al. 2019) may emerge through further work. Putative Cardinium-like symbionts have also been detected in the muscular foot of freshwater mollusks (Mioduchowska et al. 2020) and may represent an additional clade (Tarlachkov et al. 2023). Only limited instances of possible marine Cardinium have been reported in the literature to date (e.g. a bacterium found in isopods with 98% identity to Cardinium 16S rRNA genes; Wenzel et al. 2018).

We constructed a phylogenetic tree using 138 single-copy genes from the following set of genomes with <200 contigs: (i) assemblies classified as Cardinium deposited to GenBank, (ii) genomes assigned to the Amoebophilaceae family (which includes Cardinium) in the Genome Taxonomy Database (GTDB) (Parks et al. 2022), which cluster with Cardinium or Candidatus Amoebophilus asiaticus (the closest relative to Cardinium and a member of Amoebophilaceae), and (iii) three free-living organisms in the order Cytophagales to serve as an outgroup. The whole-genome tree supports distinct clusters for Cardinium and Amoebophilus (Fig. 2). Further, only two unclassified Amoebophilaceae genomes, one metagenome-assembled genome (MAG) from a soil sample (GCA_035299565.1) and one MAG from a commercial tap water sample (GCA_019744995.1), clustered with Cardinium, but their placement outside previously characterized clades of Cardinium suggests a more distant relationship to known Cardinium genomes. Without additional validation and host information, the taxonomic status of these genomes is unclear. Cardinium genomes form four main clades (Fig. 2), including a large clade A consisting of Cardinium hosted by mites, a spider, and various insects, clade B consisting of Cardinium hosted by nematodes in the genus Heterodera, and with Cardinium cPpe hosted by the nematode P. penetrans forming its own clade F, congruent with what was proposed by the most recent phylogenies constructed using 16S rRNA or gyrB genes (Tarlachkov et al. 2023). The whole-genome tree is distinct from recent single-gene phylogenies due to the inclusion of a Cardinium genome from the cranefly Tipula unca, which was absent from prior trees. Here, it was placed into clade C with Cardinium hosted by Culicoides punctatus (midge), which previously consisted of only midge-hosted Cardinium. Due to the lack of assembled Cardinium genomes from the full range of hosts, the resolution of phylogenetic trees based on whole genomes is still limited, and other clades proposed to date cannot be assessed.

Figure 2.

Phylogenetic tree of Cardinium and Amoebophilus genomes. Assemblies with <200 contigs classified as Cardinium or Amoebophilus (in bold) or clustering with either genus (not bolded) within the Amoebophilaceae family on the GTDB (Parks et al. 2022) were placed into a phylogenetic tree rooted with three free-living members of the Cytophagales order. The tree was constructed using the “Bacterial Genome Tree” tool on the Bacterial and Viral Bioinformatics Resource Center (BV-BRC) website with 138 single-copy genes, 5 duplications, and 5 deletions allowed (Olson et al. 2023). Briefly, amino acid and nucleotide sequences were aligned via MUSCLE (Edgar 2004) and the Codon_align function in BioPython (Cock et al. 2009), respectively, as input for tree generation with 100 rounds of rapid bootstrapping via RAxML (Stamatakis 2014, Stamatakis et al. 2008). A branch length scale bar is given on the bottom of the figure showing the estimated number of substitutions per site as an average of amino acid and nucleotide substitutions. Bootstrap confidence values above 80 are shown above each node. Proposed Cardinium clades supported by the tree are indicated by a black bar and clade ID (A, B, C, and F) alongside the genomes included in that clade. Colors of Cardinium strains correspond with their host group. Accession numbers for non-Cardinium genomes are given after each genome ID. Refer to Table 1 for accession numbers and other information regarding Cardinium genomes included in this figure.

Two major contributions may enhance the resolution of the Cardinium phylogeny: (i) advancements in endosymbiont DNA extractions from minute hosts (Stouthamer et al. 2018), which will be particularly beneficial for hosts, such as mites and microscopic crustaceans, and (ii) the development of a multilocus sequence typing system (Stouthamer et al. 2019), which allows for efficient elucidation of Cardinium phylogeny using evolutionarily informative genes. Future research in the areas of Cardinium genomics and phylogeny should consider these tools for higher-quality genome assemblies and consistent approaches for determining evolutionary relationships.

Cardinium distribution and evolution

Distribution of Cardinium

Cardinium is an obligate endosymbiont globally distributed in a wide range of invertebrate animal hosts (Table S1), with terrestrial arthropods and nematodes the most commonly reported hosts (Nakamura et al. 2009, Tarlachkov et al. 2023). While an early estimate of Cardinium frequency was 6%–7% of arthropod species (Weeks et al. 2003), a more recent estimate suggests ~13% of arthropods are infected with Cardinium, including 60% of chelicerates and 8% of hexapods (Weinert et al. 2015). A recent microbiome survey identified Cardinium sequences from freshwater mollusks, marking the first instance of Cardinium found in hosts other than arthropods and nematodes (Mioduchowska et al. 2020). Much of what is known about Cardinium host range comes from field surveys of invertebrates using 16S rRNA gene-based microbial community profiling (Table S1), so it is likely that current estimates of host range and prevalence are conservative. Published estimates also do not include surveys of other host phyla such as Mollusca and Nematoda.

Evolutionary dynamics of Cardinium

The sister group of Cardinium is Candidatus Amoebophilus asiaticus, a symbiont of amoebae (Horn et al. 2001, Schmitz-Esser et al. 2010). It is possible that early Cardinium were also symbionts of amoebae or other protists in an aquatic environment, a lifestyle which could have led to the movement of the symbiont into diverse animal hosts (Tarlachkov et al. 2023). As a maternally transmitted symbiont, the phylogenetic relationships between Cardinium strains can mirror those of its hosts in some mites, spiders, and Opiliones (Stouthamer et al. 2019). In some cases, these parallel phylogenetic patterns have been used to elucidate evolutionary hypotheses surrounding host divergence events (Kopecky et al. 2013). The phylogenies of Cardinium and its hosts are, however, not always congruent, and there is evidence of host switching events in some Cardinium lineages. The host diversity in clade A, for example, shows closely related pairs of Cardinium taxa found in parasitoid wasps (insects that lay their eggs in and consume their host) and their herbivorous insect hosts (Stouthamer et al. 2019, Tarlachkov et al. 2023, Zchori-Fein and Perlman 2004). It is unclear how often host-switching events occur or if they are evidence of more frequently occurring horizontal transmission. Some proposed mechanisms of horizontal transmission allowing for the observed host-switching include predation (Tarlachkov et al. 2023) and plant-mediated transmission (Chrostek et al. 2017, Gonella et al. 2015, Tarlachkov et al. 2023). If these alternative routes of transmission are indeed found to be prevalent, it is important that future screening measures consider possible differences between transient Cardinium infections (e.g. short-term presence in the gut) and long-term Cardinium symbioses. If possible, such studies on Cardinium prevalence should follow up 16S rRNA-based identifications with additional methods of confirmation such as fluorescence in situ hybridization, electron microscopy, or Cardinium-specific polyermase chain reactions (PCR) or quantitative PCR (qPCR) primers (e.g. gyrB, sufB, EF-G, or groEL) to achieve higher confidence of identification [see Stouthamer et al. (2019) for suggested primers].

Cardinium strains cause different host phenotypes, including several forms of reproductive manipulation (discussed below). Strains inducing similar effects often do not form monophyletic groups, suggesting that if the common ancestor of Cardinium had factors for multiple reproductive manipulations, these manipulative factors were subsequently lost or inactivated as Cardinium diversified and adapted to different hosts (Stouthamer et al. 2019). Alternatively, factors responsible for these phenotypes may have arose through convergent evolution and/or horizontal gene transfer between distantly related symbiont lineages (Stouthamer et al. 2019). While intracellular endosymbionts like Cardinium do not encounter other bacterial species as frequently as extracellular symbionts, instances of Cardinium cooccurring with other intracellular symbionts like Wolbachia in the same host have been observed in several insect, arachnid, and nematode groups (e.g. Brown et al. 2018, Konecka and Olszanowski 2019, Ros et al. 2012, Ros and Breeuwer 2009, White et al. 2009, Xie et al. 2016, Zhang et al. 2012, Zhu et al. 2012). These symbiont co-infections could potentially provide an opportunity for intracellular heritable endosymbionts to exchange genetic material via horizontal gene transfer, which may have played an important role in shaping the accessory functions in Cardinium genomes (Brown et al. 2018, Siozios et al. 2019). There is also the possibility of interkingdom horizontal gene transfer events between intracellular symbionts and their eukaryotic hosts (Hotopp et al. 2007, Leclercq et al. 2016). However, no such occurrences have been reported in Cardinium symbioses to date. Host biology may also have some influence on phenotype induction, as symbiont strains transinfected between species can cause different host phenotypes (Li et al. 2020a). Whether Cardinium had the ability to manipulate eukaryotic reproduction ancestrally or if Cardinium strains acquired their manipulative capabilities through convergent evolution or horizontal gene transfer remains an open question. Resolving this will require further mechanistic characterizations in multiple host species to identify factors responsible for these phenotypes, followed by analyses into the distribution of manipulative factors among Cardinium strains.

Phenotypes associated with Cardinium infection

Cardinium can cause a variety of phenotypes in its hosts (Fig. 3). Some phenotypes manifest via modifications to host reproduction. These phenotypes benefit Cardinium by skewing host reproduction to favor the number or fitness of infected females that transmit the symbiont and may impose a cost to uninfected females or male hosts. This control of host reproduction by symbionts like Cardinium is referred to as reproductive manipulation. Additionally, some Cardinium confer conditionally dependent benefits to their hosts, increasing host fitness and their own fitness alike under particular environments or abiotic conditions. The reproductive manipulation cytoplasmic incompatibility (CI) is commonly associated with Cardinium. CI is a lethal manipulation of infected males that kills uninfected offspring. Other reproductive manipulator phenotypes bias reproduction to favor infected females which transmit Cardinium. To attribute a phenotype to heritable symbionts like Cardinium, a set of criteria have been established and followed in the research community (Stouthamer et al. 1999). The criteria for confirming Cardinium-induced phenotypes are: (i) The phenotype is present only when Cardinium is present and is absent in hosts where the symbiont has been removed. Additionally, for putative parthenogenesis-inducing symbionts, males are produced by uninfected hosts. (ii) Cardinium is the only heritable symbiont present in the host with the phenotype. Differential offspring production or offspring sex ratios between infected and uninfected hosts are suggestive of CI or parthenogenesis and feminization, respectively (Hunter et al. 2003, Zchori-Fein et al. 2004). In this section, we discuss the major phenotypes attributed to Cardinium infection, including how Cardinium induces these phenotypes and the biological and environmental factors that influence their efficacy (Fig. 3).

Figure 3.

Phenotypes associated with Cardinium infection. Known forms of reproductive manipulation induced by Cardinium include CI, PI, and feminization. These phenotypes facilitate Cardinium transmission by providing a fitness advantage to infected females at the expense of uninfected females (CI) or by skewing reproduction to favor female progeny (PI, feminization). CI is a two-step process that drives the infection through a population: first, Cardinium sabotages male hosts causing offspring death when that male mates with uninfected females. Second, Cardinium prevents this modification from killing infected offspring. PI Cardinium cause female hosts to asexually produce other females, while feminizing Cardinium cause genetic males to develop as functional females. Feminizing Cardinium have best been described in haplodiploid Brevipalpus mites, which asexually produce haploid male offspring; Cardinium feminizes these haploid males, resulting in an asexual population of haploid females (Weeks et al. 2001, Groot and Breeuwer 2006). Other possible phenotypes associated with Cardinium infection include thermotolerance, nutritional provisioning (in cosymbiosis), behavior modifications, and protection from natural enemies and pathogens.

Host reproductive manipulation via CI

The phenotype most widely associated with Cardinium infection is CI, with CI Cardinium strains observed in parasitoid wasps (Gebiola et al. 2016b, Hunter et al. 2003), planthoppers (Nakamura et al. 2012), thrips (Nguyen et al. 2017), and mites (Gotoh et al. 2007, Ros and Breeuwer 2009, Xie et al. 2016, Zhu et al. 2012). CI is a two-step lethal manipulation composed of a modification step in males and a rescue step in eggs (Fig. 4). During modification, the symbiont first sabotages male sperm or other paternal products to kill their offspring during embryogenesis. Next, the maternally transmitted symbiont “rescues” infected embryos by nullifying the toxicity of modified sperm and restoring the viability of rescued embryos; uninfected females’ embryos are not rescued and die early in development. Together, the modification and rescue steps of CI grant a fitness advantage to infected females since they produce viable infected offspring with both infected and uninfected mates. This drives symbiont infections to high frequencies in most host populations and promotes stable transmission of the symbiont (Harris et al. 2010, Hoffmann and Turelli 1997). CI can also greatly influence host evolution by contributing to host speciation, with symbionts acting as reproductive barriers (Brucker and Bordenstein 2012, Gebiola et al. 2016a) and causing selective sweeps in host populations as CI symbionts drive themselves through a population (Hurst and Jiggins 2005).

Figure 4.

Hypothetical mechanism of Cardinium-induced CI and rescue based on cytology and fluorescence microscopy of Encarsia suzannae testes and embryos. Cardinium-induced CI in E. suzannae is initiated during spermatogenesis, possibly by altering male chromatin or DNA-associated proteins, or by loading sperm with a toxin that is released upon fertilization. The second scenario is illustrated here. Either scenario causes asynchronous chromosomal segregation during mitosis in early embryogenesis that kills uninfected offspring. To rescue CI, Cardinium may reverse the alterations made to male chromatin or similarly alter female chromatin to restore mitotic synchrony or produce an antitoxin to prevent the sperm-delivered toxin from killing the embryo.

Due to their ability to drive themselves and associated cytoplasmic elements through a population, CI symbionts also have promising applications for controlling the transmission of arthropod-vectored pathogens if the drive provided by CI symbionts can be paired with other desirable traits like host pathogen resistance (Gong et al. 2020, Hoffmann et al. 2011, Moreira et al. 2009, Zheng et al. 2019). This is exemplified by the successful suppression of Dengue transmission in Australia following release of CI Wolbachia-infected Aedes aegypti: Wolbachia alone increases A. aegypti resistance to dengue infection along with inducing CI to drive the infection and pathogen resistance through natural populations (Hoffmann et al. 2011, Moreira et al. 2009, Ryan et al. 2020). It is currently unknown whether Cardinium can also confer host resistance to vector-borne diseases, but as Cardinium research advances, novel host impacts may be revealed. The evolutionary ramifications of symbiont-induced CI, as well as the potential application of CI symbionts in the control of pest populations and vector-borne diseases, have spurred sustained interest in the cellular and genetic mechanisms underlying CI.

Recent work on the cellular localization of two CI Cardinium strains in male reproductive tissues of Encarsia parasitoid wasps has provided some insight into how Cardinium induces CI in its hosts (Doremus et al. 2022, 2020). In Encarsia suzannae, the Cardinium strain cEper1 densely infects most developing sperm cells during pupation, the life stage critical to CI modification in this host. Cardinium cells are removed from developing sperm during spermiogenesis, the final stage of sperm development associated with large-scale morphological and nuclear restructuring of sperm cells (Ferree et al. 2019). The loss of cEper1 coincides with the onset of sperm maturity and the end of the CI modification window, suggesting that cEper1 modifies developing sperm cells internally to induce CI prior to its removal during spermiogenesis (Doremus et al. 2020). This localization pattern is similar to that of some CI Wolbachia strains, including wMel and wRi, which infect Drosophila (Clark et al. 2003, 2002) as well as CI Wolbachia that infect the wasp Nasonia vitripennis (Clark et al. 2008). Both Cardinium and Wolbachia infect developing sperm cells and seemingly modify these cells prior to symbiont removal in spermiogenesis.

A second CI Cardinium strain, cEina3, shows a strikingly different infection density and localization pattern in the male reproductive system of the wasp Encarsia partenopea (Doremus et al. 2022). Despite apparently few cells infecting the male reproductive system, cEina3 causes a more consistently lethal CI phenotype than cEper1. This low-density cEina3 infection is also associated with a different infection pattern: instead of infecting sperm cells, most cEina3 cells infect somatic cells at the base of the testis and around the seminal vesicle, where mature sperm cells are stored. This unique localization pattern suggests that cEina3 uses a different modification method and/or set of CI effectors to indirectly modify either developing sperm cells within the testes or fully mature sperm within the seminal vesicle (Doremus et al. 2022). It is possible that this second type of localization pattern contributes to the more lethal CI phenotype associated with cEina3, although variation in CI effectors and their expression, as well as host factors could also contribute to the observed differences in CI strength. Regardless of the underlying mechanism, CI-inducing symbionts can display a variety of localization patterns in the tissues of closely related hosts. Characterization of the localization patterns of other CI Cardinium strains could help determine the extent of this variation, whether localization correlates with CI lethality and CI effector identity, and which host tissues are relevant to CI induction.

The cytological defects arising from Cardinium CI were first characterized for the cEper1 strain in E. suzannae. Following fertilization, CI-affected embryos display several mitotic defects including abnormal chromatin condensation and chromatin segregation that can produce chromatin bridges between daughter nuclei. These mitotic defects persist through multiple cleavage divisions, with the accumulation of these defects ultimately leading to late embryonic mortality (Gebiola et al. 2017a). The cytological defects caused by cEper1 CI are similar to those caused by Wolbachia CI (Callaini et al. 1997, Tram and Sullivan 2002), although Wolbachia CI usually kills the embryo earlier in development than Cardinium CI (Callaini et al. 1996, Tram et al. 2006). These similarities between Cardinium and Wolbachia CI at the cellular level seem to be an instance of convergent evolution, as Cardinium genomes do not encode the Wolbachia CI factors (Penz et al. 2012). While genomic and transcriptomic analyses of cEper1 have not yet revealed obvious candidate Cardinium CI factors (Mann et al. 2017, Penz et al. 2012), these studies, along with a new understanding of Cardinium localization in male tissues during CI modification, form a basis to potentially identify cEper1 proteins involved in CI. Identifying these CI factors and their host targets will provide insights into the origin of Cardinium CI, and reveal where the CI mechanisms of Cardinium and Wolbachia converge.

Parthenogenesis and feminization

Cardinium is commonly associated with two phenotypes that result in female-biased sex ratios: parthenogenesis induction (PI) and feminization. PI enables infected females to asexually produce more infected females, eventually resulting in an entirely female asexual host population (Fig. 5). In insects, Cardinium-induced PI has only been demonstrated in hosts exhibiting haplodiploidy, in which diploid females can produce haploid male offspring asexually while females develop from fertilized diploid embryos (Matalon et al. 2007, Provencher et al. 2005, Zchori-Fein et al. 2001, 2004). In haplodiploid hosts, PI Cardinium cause diploid females to asexually produce diploid daughters. More recently, Cardinium has also been associated with PI in nonmarine ostracods, whose ploidy status is unclear (Schön and Martens 2020).

Figure 5.

Cardinium feminization and parthenogenesis both manipulate the asexual production of haplodiploid hosts, turning male offspring into females. In haplodiploid hosts, unfertilized haploid eggs develop as males, while fertilized diploid eggs develop as females. The feminization model for Cardinium is based on Brevipalpus mites, where haploid unfertilized male eggs develop as functional phenotypic females (Weeks et al. 2001, Groot and Breeuwer 2006). This results in populations of haploid individuals that reproduce asexually. Parthenogenesis causes the asexual production of female offspring via a two-step process in E. hispida. Cardinium first inhibits offspring ploidy reduction by disrupting the anaphase 1 step of meiosis, then feminizes diploid eggs to develop as females (Giorgini et al. 2007, 2009).

Cardinium has been associated with parthenogenesis in several lineages of hosts, including parasitoid wasps (Matalon et al. 2007, Zchori-Fein et al. 2001, 2004), armored scale insects (Provencher et al. 2005), and nonmarine ostracods (Schön and Martens 2020). Unfortunately, symbiont-induced PI is difficult to confirm in many cases, as parthenogenetic hosts often lose the ability to reproduce sexually and PI-associated symbionts can become essential for reproduction, limiting the potential for experimental manipulation of the symbiont (Fricke and Lindsey 2024a, Stouthamer et al. 2010, Zchori-Fein et al. 2001). In most instances, the role of Cardinium in host parthenogenesis has not been confirmed and is instead correlated with the presence of Cardinium in asexual, but not sexual species (Provencher et al. 2005, Schön and Martens 2020, Zchori-Fein et al. 2001). This is the case for the parthenogenetic wasp Encarsia tabicivora (formerly classified as Encarsia pergandiella; Gebiola et al. 2017b), in which Cardinium is associated with parthenogenesis but is also essential for successful wasp reproduction (Zchori-Fein et al. 2001). The role of Cardinium as the causal agent of PI has been confirmed by symbiosis disruption via antibiotics or heat in only two parasitoid wasp species, Plagiomerus diaspidis (Gordh and Lacey 1976, Matalon et al. 2007, Zchori-Fein and Perlman 2004) and Encarsia hispida (Giorgini 2001, Giorgini et al. 2009, Zchori-Fein et al. 2004). In both cases, these wasps began asexually producing male offspring following the loss of Cardinium, as would be expected in a haplodiploid host (Giorgini 2001, Giorgini et al. 2009, Gordh and Lacey 1976, Zchori-Fein et al. 2004).

Cardinium PI has been best characterized in haplodiploid Encarsia wasps. In E. hispida and E. tabacivora, restoration of diploidy in unfertilized eggs is caused by the central fusion of sister chromosomes either during meiosis I or immediately following it (Doremus and Hunter 2020, Giorgini et al. 2007). This contrasts with Wolbachia PI, which restores diploidy via gamete duplication following the completion of meiosis (Giorgini et al. 2007). These differences in the timing and manner of diploidy restoration suggest that Cardinium and Wolbachia may have independently evolved PI, as is the case for CI. The antibiotic treatment of E. hispida curiously results in the production of diploid males, indicating that diploidization alone is not sufficient for PI and that Cardinium must additionally feminize diploid hosts to induce parthenogenesis (Giorgini et al. 2009). These results suggest that Cardinium uses a two-step mechanism to induce PI, composed of initial diploidization followed by feminization later in development. If PI is a two-step process, feeding adult female E. hispida antibiotics may have disrupted the feminization but not diploidization steps of PI in their offspring leading to the production of diploid males. This work highlights the importance of characterizing offspring ploidy following symbiosis disruption in suspected PI systems, which can indicate whether diploidy restoration is sufficient for inducing female development or if a second feminizing step is required. A two-step PI mechanism involving host feminization via manipulation of the sex determination system has also been proposed for some PI Wolbachia strains (Ma and Schwander 2017). Recent work identifying Wolbachia PI factors supports a two-factor mechanism for some Wolbachia strains, with one of the two factors acting as a potential mimic of transformer, a key insect sex determination factor (Fricke and Lindsey 2024b, Li et al. 2024a). Whether Cardinium PI strains use a similar mechanism to feminize hosts remains to be seen.

Feminization differs from PI by causing infected hosts that harbor the chromosomal configuration for male development to develop as fully functional, phenotypic females (Fig. 5). Unlike populations affected by PI, in which males are no longer necessary for females to reproduce, males are usually still required for reproduction in feminized populations. A curious exception is in Brevipalpus mites feminized by Cardinium, wherein genetic haploid male mites instead develop into haploid females, resulting in an entirely haploid female population that reproduces asexually (Groot and Breeuwer 2006, Weeks et al. 2001). Exposure to antibiotics results in an increased production of male offspring, confirming the role of Cardinium in causing female development (Groot and Breeuwer 2006). The feminization of haploid females results in a scenario functionally similar to PI, as infected females can asexually produce more feminized females (Groot and Breeuwer 2006, Weeks et al. 2001). If Cardinium PI is caused by two independent diploidization and feminization steps, it is possible that the Cardinium strains in Brevipalpus mites have retained the ability to feminize hosts but lost the ability to restore diploidy. Alternatively, feminizing Cardinium in Brevipalpus may use an independent method of feminization that is not analogous to the process used by PI Cardinium strains. Identification of PI factors involved in diploidization and feminization will help resolve whether these Cardinium use the same processes to induce asexual female reproduction.

While most sex-ratio distorting Cardinium are associated with asexual reproduction, there may be other methods by which Cardinium can bias host sex ratio to favor females. One such alternative method involves female-biased offspring sex allocation in the haplodiploid citrus thrips Pezothrips kellyanus, which is co-infected with CI-inducing strains of Cardinium and Wolbachia. Egg fertilization in P. kellyanus is egg size-dependent, with larger eggs preferentially fertilized to produce female offspring (Katlav et al. 2021). This results in individual females producing broods with either strongly female-biased or male-biased sex ratios. While uninfected females can display female-biased sex allocation, those hosting Cardinium are even more likely to produce female-biased broods, suggesting that symbiont infection biases sex allocation to favor female offspring (Katlav et al. 2022a). This sex allocation effect may in part be caused by a general increase in fitness, as Cardinium-infected females are larger than uninfected ones or those co-infected with Wolbachia. These larger Cardinium-infected females in turn produce larger eggs than uninfected or co-infected females, which are then preferentially fertilized (Katlav et al. 2022a). Cardinium seemingly affects maternal resource investment in eggs, although research on the metabolic contribution of this Cardinium may reveal more information on this unique phenotype. Similarly, egg size-based selective fertilization has been described in other haplodiploid arthropods, although it remains to be seen if symbionts in other systems also influence sex allocation (Macke et al. 2010).

Other host phenotypes associated with Cardinium infection

Cardinium is also associated with a range of host phenotypes beyond reproductive manipulation. These include modifying host oviposition behavior of Encarsia wasps (Kenyon and Hunter 2007) and mating preferences of P. kellyanus (Tourani et al. 2024), the latter of which may improve the fitness of Cardinium-infected females through the avoidance of lethally incompatible males that have Wolbachia. Cardinium can also provide general beneficial effects such as increased host fecundity and higher survival rates (Katlav et al. 2022a, Wang et al. 2008, Xie et al. 2016). How Cardinium increases host fitness is not clear, but it may involve some degree of metabolic provisioning (Newton and Rice 2020). Cardinium can also be costly for the host, reducing fecundity and survival by using host nutritional resources or, in the case of some Mediterranean (MED) biotype Bemisia whiteflies, altering feeding behavior (Shan and Liu 2021, Ying et al. 2021). Cardinium may also provide more conditional benefits, although most have only been observed in whitefly hosts. Such benefits include potentially augmenting host resistance to parasitoids in MED Bemisia whiteflies (Giorgini et al. 2023), altering interactions between whitefly hosts and their host plants (Liu et al. 2023), or improving MED Bemisia whitefly thermotolerance (Yang et al. 2021). MED Bemisia whiteflies collected from warmer regions of China also show higher Cardinium infection rates, supporting a link between higher temperatures and Cardinium infection in these populations (Li et al. 2023a). Further characterization of the impact of Cardinium infection on host thermotolerance is warranted given the increasing severity of global climate change and its implications for arthropod biology.

In addition to modifying host abiotic stress tolerance, heritable endosymbionts can augment host immunity via providing protection from natural enemies and pathogens (Oliver and Martinez 2014). While some symbionts protect their host by producing specific toxins targeting parasites or pathogens (Degnan and Moran 2008, Hamilton et al. 2016), other symbionts, most notably Wolbachia, may provide more general protection by competing with pathogens or modulating the host immune response (Rances et al. 2012, Caragata et al. 2013, Bhattacharya et al. 2017). The role of Cardinium in host immunity presents a clear gap in our understanding of Cardinium biology, particularly given the potential for endosymbionts to help control arthropod-vectored disease transmission exemplified by the World Mosquito Program (O’Neill 2018). A study comparing the responses of a lepidopteran cell line to Cardinium infection found that Cardinium elicited an upregulated immune response, suggesting some degree of interplay between Cardinium and insect immunity (Nakamura et al. 2011). Whether this upregulated immune response in an in vitro cell line translates to in vivo effects in natural hosts has not been thoroughly explored. In addition to potentially upregulating host immune responses to parasites and pathogens (Kambris et al. 2010), these interactions between Cardinium and the host immune system could also play roles in regulating these symbioses and may impede the establishment of new symbioses (Weinert et al. 2015). One hindrance to studying the defensive benefits of Cardinium has been the lack of known natural enemies or pathogens of the more extensively studied Cardinium hosts (e.g. Encarsia wasps and Brevipalpus mites). Yet, Cardinium is also present in certain agricultural and medical pests like planthoppers, whiteflies, and biting midges, providing a promising opportunity to study how Cardinium interacts with anthropogenically relevant insect-vectored pathogens (Gilbertson et al. 2015, Mellor et al. 2000, Zhou et al. 2008).

While Cardinium-associated phenotypes have been explored in a number of different arthropod hosts, almost nothing is known of the phenotypic consequences of Cardinium infection in nematodes, aquatic crustaceans, or spiders, the latter of which appear particularly prone to hosting Cardinium (Duron et al. 2008b). In P. penetrans, a plant–parasitic nematode co-infected with Cardinium and Wolbachia, Wolbachia was found to be associated with a female-biased sex ratio, although it is unclear whether Cardinium is a reproductive manipulator in this host as well (Brown et al. 2018, Wasala et al. 2019). Cardinium and Wolbachia may also participate in metabolic complementation for the biosynthesis of methionine or fatty acids in P. penetrans (Brown et al. 2018). In nonmarine ostracods, Cardinium infection is associated with parthenogenetic populations (Schön et al. 2019, Schön and Martens 2020). As this association was only recently identified, Cardinium has not yet been confirmed as the causal agent of parthenogenesis. It is likely that we currently underestimate both the prevalence of Cardinium in nematode and aquatic systems and its phenotypic impact on these hosts. The presence of manipulative symbionts in aquatic systems is particularly interesting, given that the terrestrial models for reproductive manipulation may not hold true in marine environments, where broadcast spawning is a prevalent reproductive strategy (Kustra and Carrier 2022). New approaches may be required for identifying reproductive manipulation phenotypes in such host groups (Kustra and Carrier 2022).

As with many obligate host-associated symbionts, most of what is known about Cardinium comes from genomic information and studies on infected insect cultures since direct culturing of the bacterium is currently not possible. However, successful efforts to maintain Cardinium in arthropod cell lines provide a promising avenue for exploring phenotypic effects on the cellular level. The initial isolation of Cardinium in an Ixodes tick cell line (Kurtti et al. 1996) was followed by successful propagation of Cardinium in mosquito cell lines (AeAI-2 and C7-10) and in a lepidopteran cell line (HZ-AM1) (Morimoto et al. 2006). The Ixodes-derived Cardinium were also propagated in silkworm cells (Bm-aff3) (Nakamura et al. 2011). Successful propagation of Cardinium derived from host lineages besides Ixodes ticks has not been reported, to date.

Factors affecting Cardinium symbioses and phenotypes

Numerous biotic and abiotic factors influence the stability of heritable symbioses and the penetrance of symbiont-induced phenotypes, such as host factors, symbiont titer, co-infecting symbionts, or external environmental stressors. These effects may also be mediated by genotype differences among symbiont and host lineages, although the influence of host genotype on Cardinium-induced phenotypes has not been widely explored.

Host genotype interactions with Cardinium symbioses have been most extensively studied in the MED biotype Bemisia tabaci whiteflies, with the fitness benefits of Cardinium infection, including increased fecundity, decreased developmental time, and increased longevity, varying between different whitefly lineages (Li et al. 2023b). The influence of host lineage also extends to Cardinium-conferred thermotolerance, with only certain whitefly lineages benefiting from Cardinium infection following heat exposure (Yang et al. 2021). Experimental transinfection of Cardinium between host species also suggests that host genotype in part determines the phenotypic effects of Cardinium. A CI-inducing Cardinium strain was successfully transferred between two planthopper species. While this Cardinium strain caused CI in its native host, it did not induce CI in the novel planthopper (Li et al. 2020b).

Other host factors, like host age, can also modify the manifestation of Cardinium-induced phenotypes. Cardinium CI lethality decreases as male adults age in the planthopper Sogatella furcifera and the carmine spider mite Tetranychus cinnabarinus (Nakamura et al. 2012, Noda et al. 2001, Xie et al. 2010). Yet, in the wasp E. suzannae, adult male age does not affect Cardinium CI; instead, male pupal developmental time is the most important factor for determining Cardinium CI strength in this wasp (Doremus et al. 2019, Perlman et al. 2014). This is likely due to the timing of sperm production and CI male modification in E. suzannae, both of which are completed primarily during pupation (Doremus et al. 2020). Host age increases feminization efficacy in Cardinium-infected Brevipalpus mites, with the production of male offspring rapidly decreasing as adult females age. In this case, Cardinium titer may increase as females age, resulting in a greater proportion of eggs undergoing successful feminization in older females (Groot and Breeuwer 2006). Many of these differences observed between Cardinium symbioses likely arise from idiosyncrasies of host biology.

The influence of Cardinium titer in determining phenotype expression also varies depending on the symbiosis. In the carmine spider mite, T. cinnabarinus, Cardinium CI lethality is tightly associated with symbiont titer, with a reduction in titer resulting in weaker CI (Xie et al. 2010). However, in E. suzannae and the mite Tetranychus urticae, reduction in Cardinium titer does not always correspond with reduced CI (Doremus et al. 2019, Xie et al. 2016). As discussed above, in the wasp E. partenopea, a low titer Cardinium strain is capable of inducing uniformly lethal CI, further indicating that some Cardinium strains can consistently impose phenotypes upon the host despite their low density (Doremus et al. 2022). Why titer affects some Cardinium symbioses more than others remains an open question, but likely involves differences in Cardinium-encoded factors, gene expression, localization, and/or host susceptibility to manipulation.

Heritable symbionts can share their host with other co-infecting symbiont lineages. These symbiont co-infections offer an opportunity for host-restricted symbiont lineages to interact with one another, often with consequences that extend to the host. Cardinium is frequently found co-infecting hosts with other manipulative symbionts like Wolbachia (Hubert et al. 2021b, Nakamura et al. 2012, Nguyen et al. 2017, White et al. 2011). The outcomes of these co-infections seem to be variable, ranging from additive (Nakamura et al. 2012, Nguyen et al. 2017), to neutral (White et al. 2011), or weakening effects on symbiont-induced phenotypes (Hubert et al. 2021b, Ros and Breeuwer 2009, Zhang et al. 2012, Zhao et al. 2018) and likely depend on the symbiont strains and host involved. In several cases, co-infecting Cardinium strains are asymptomatic, at least with regard to reproductive manipulation (Doremus et al. 2022, White et al. 2011). In these cases, the asymptomatic Cardinium strain co-infects hosts along with a CI-inducing symbiont (Doremus et al. 2022, White et al. 2009). How these asymptomatic Cardinium strains spread is unclear; because of their occurrence in coinfections, they may act as heritable hitchhikers passively spreading via the CI and/or selective advantages provided by their neighbor. These strains may also provide conditional benefits that facilitate their spread or may be former reproductive manipulators that subsequently lost their manipulative capabilities. The propensity of Cardinium for co-infection, along with the variable outcomes of these interactions, make it an intriguing model for studying how symbionts respond to one another, as well as the evolutionary and ecological ramifications of co-infections.

The external environment can exert a strong influence on heritable symbioses. Heritable symbionts are notorious for their susceptibility to environmental stress, particularly thermal stress, which can reduce symbiont titer, transmission rate, and phenotypic expression (Corbin et al. 2017). Yet, a survey of global symbiont infections found a positive association between Cardinium infection and temperature, particularly in mandibulate arthropods like true bugs, suggesting that Cardinium may better tolerate warmer temperatures than symbionts like Wolbachia (Charlesworth et al. 2019). More localized surveys of Cardinium infections in specific host populations have yielded similar results. A study of Cardinium prevalence in Chinese populations of the spider mite T. cinnabarinus only found Cardinium in host populations in milder warm and wet regions compared to more arid regions prone to daily extreme temperature fluctuations (Liu et al. 2006). A similar infection pattern was found for Cardinium infecting MED B. tabaci whiteflies in China (Li et al. 2023a) and in Culicoides biting midges in Israel (Morag et al. 2012), suggesting that these infection patterns may be comparable across Cardinium symbioses. Together these surveys suggest that extreme fluctuations in temperature may be more important than mean temperatures for limiting Cardinium prevalence in nature. These surveys also highlight humidity as an additional factor that may be important for Cardinium infection stability. Whether Cardinium itself is susceptible to changes in humidity (seemingly unlikely because of its intracellular habitat), or if low humidity indirectly inhibits Cardinium spread by compounding its fitness costs or altering the host intracellular environment is not clear. In Culicoides midges, an indirect effect of humidity may be more likely as these midges rely on moist environments for their development (Morag et al. 2012).

While infection surveys can provide some insight into the climatic forces shaping Cardinium symbioses, there are few studies directly testing the effect of environmental stressors on Cardinium symbioses. In some cases, like in MED B. tabaci whiteflies, Cardinium seem to respond positively to warmer temperatures, with Cardinium titer (Yang et al. 2023a) and whitefly survival and fecundity increasing following warm temperature exposure (Yang et al. 2021). Yet, in the whitefly parasitoid E. suzannae, similarly warm temperatures negatively affect the symbiosis by reducing Cardinium titer, transmission rate, and CI strength (Doremus et al. 2019). In citrus thrips, exposure to warm temperatures reduced the female-biasing effect of Cardinium but did not affect Cardinium CI or titer (Katlav et al. 2022b). These three studies show that substantial variation in response to temperature stress exists among Cardinium strains, with the caveats that these different observations could be the result of different methodological approaches (i.e. severity of temperature exposures), host biology, or variation in Cardinium thermotolerance. Additional research testing the effect of temperature and humidity on Cardinium symbioses is essential for understanding how Cardinium symbioses more generally respond to stressful environmental conditions.

Genomic, transcriptomic, and proteomic insights into Cardinium physiology and ecology

The increasing accessibility of genome-scale sequencing and analysis has contributed to an increase in published Cardinium genomes. Despite the diversity of Cardinium symbioses, genetic information is only available for limited host groups. Even more sparse is information on Cardinium function at the mRNA and protein level, with only two published Cardinium gene transcription datasets (Mann et al. 2017, Gardner et al. 2018). Functional data are essential for deducing how Cardinium responds to changes in its cellular environment across host development or environmental conditions at the gene, mRNA, and protein levels. Further, molecular data from Cardinium systems are very rarely connected to host phenotype, highlighting a need for integrating molecular data with the phenotypic characterization of Cardinium strains to allow for comparative analyses of host–microbe interactions (e.g. reproductive manipulation effectors). In the following sections, we describe our current knowledge of Cardinium physiology and its interactions with hosts based on available genomic, transcriptomic, and proteomic data. We will also compare this knowledge to current insights from Wolbachia, a distantly related but functionally similar symbiont, to provide a relevant frame of reference for Cardinium physiology and ecology as a bacterial endosymbiont.

Structure of Cardinium genomes

There are 21 publicly available Cardinium genomes (Table 1), 7 of which are closed: cEper1 from the parasitoid wasp E. suzannae, cHgTN10 from the nematode Heterodera glycines, cSfur from the planthopper S. furcifera, cOegib-Wal from the spider Oedothorax gibbosus, DF from the mite Dermatophagoides farinae, icPhiSpin1 from the beetle Philonthus spinipes, and idTipUnca1 from the cranefly T. unca. The remaining 14 genomes are contig or scaffold-level genome assemblies, containing anywhere from 5 to 307 contigs. Of the 21 available Cardinium genomes, only 2 have been linked to a reproductive manipulation phenotype. Most Cardinium genome assemblies are fragmented, likely due to challenges during assembly from insufficient sequencing coverage of the symbiont genome due to large amounts of host DNA (Stouthamer et al. 2018) and repetitive DNA regions in the symbiont genome. Similar to genomes of other obligate intracellular endosymbionts, Cardinium genomes have reduced size (between 0.9 and 1.5 Mbp), GC content (between 33.5% and 39.0%), and functional potential (encoding around 1000 predicted genes) (McCutcheon and Moran 2012, Moran 2003, Moran et al. 2008, Moran and Bennett 2014, Moran and Plague 2004). Similarly, Ca. Amoebophilus asiaticus, an obligate intracellular symbiont of free-living amoebae and closest known relative to Cardinium, has a 1.9 Mbp genome with around 1500 genes and 35% GC content (Schmitz-Esser et al. 2010).

Table 1.

Publicly available Cardinium genomes and associated phenotypic and assembly data. Abbreviations: cytoplasmic incompatibility (CI), no data/unknown (n.d.), and Darwin Tree of Life Project (DTLP).

Strain name	Host	Reproductive manipulation	% GC content	Number of contigs	Plasmids	Genome assembly, plasmid size (kb)	GenBank accession	References
cBcalN1	Mite—B. californicus	n.d.	36.5	307	n.d.	1046	GCA_022810505.1
cBcalN2	Mite—B. californicus	n.d.	36.5	275	n.d.	1049	GCA_022810445.1
cByotB1	Mite—Brevipalpus yothersi	n.d.	36.5	216	n.d.	1086	GCA_022763005.1
cByotN1	Mite—B. yothersi	n.d.	36.5	292	n.d.	1145	GCA_022810525.1
TP	Mite—Tyrophagus putrescentiae	n.d.	39	33	n.d.	914.8	GCA_025215055.1	Xiong et al. (2023)
JH06282024_1	Mite—T. putrescentiae	n.d.	39	28	n.d.	1052	GCA_036541165.1
JH06282024_2	Mite—T. putrescentiae	n.d.	38.5	59	n.d.	1090	GCA_036541185.1
JH06282023_2	Mite—T. putrescentiae	n.d.	39	55	n.d.	1052	GCA_030441515.1
cDfar	Mite—D. farinae	n.d.	37	5	n.d.	1482	GCA_007559345.1	Erban et al. (2020)
DF	Mite—D. farinae	n.d.	38	1	2	1419, 33.8, and 125.3	GCA_025268635.1	Xiong et al. (2023)
cOegib-Wal	Spider—O. gibbosus	n.d.	36.5	1	0	1137	GCA_936981045.1	Halter et al. (2023)
cBtQ1	Whitefly—B. tabaci	n.d.	36	11	1	1012 and 52.0	GCA_000689375.1	Santos-Garcia et al. (2014)
CanCar	Whitefly—B. tabaci	n.d.	36	50	n.d.	996.8	GCA_004300865.1
cSfur	Planthopper—S. furcifera	CI	39.2	1	0	1103	GCA_003351905.1	Zeng et al. (2018)
cEper1	Wasp—E. suzannae	CI	36.5	1	1	944.9 and 57.8	GCA_000304455.1	Penz et al. (2012)
icPhiSpin1	Beetle—P. spinipes	n.d.	39	1	0	1093	GCA_964030745.1	DTLP
idTipUnca1	Cranefly—T. unca	n.d.	36.5	1	0	1374	GCA_964020025.1	DTLP
cCpun	Biting midge—C. punctatus	n.d.	33.5	63	0	1137	GCA_004354815.1	Siozios et al. (2019)
cHgTN10	Nematode—H. glycines	n.d.	38	1	0	1193	GCA_003176915.1	Showmaker et al. (2018)
cPpe	Nematode—P. penetrans	n.d.	35.5	45	0	1358	GCA_003788695.1	Brown et al. (2018)
cHhum	Nematode—Heterodera humuli	n.d.	38.5	166	0	1056	GCA_028766965.1	Tarlachkov et al. (2023)

Another feature among some endosymbiont genomes undergoing reduction is an increased abundance of mobile genetic elements (Naito and Pawlowska 2016). This also seems to be true for Cardinium. For example, transposable elements represent 12.4% of coding sequences (CDS) in Cardinium cEper1 and ~30% of the cOegib-Wal genome (Halter et al. 2023, Penz et al. 2012). In contrast, prophages, which are present in Wolbachia genomes, seem to be largely absent in available Cardinium genomes (Brown et al. 2018, Halter et al. 2023, Penz et al. 2012). Plasmids appear to be common, although the true prevalence of plasmids in Cardinium is difficult to ascertain due to the limited number of closed genomes. Of the seven closed genomes, strains cEper1 and DF were found to harbor one and two plasmids, respectively, and one draft genome (cBtQ1) was predicted to have a plasmid. However, plasmid presence needs to be verified for draft assemblies. Cardinium plasmids range in size from 33.8 to 125.3 kb and have a reduced % GC content compared to chromosomal DNA. Commonly encoded features on known Cardinium plasmids include a plasmid partitioning protein (e.g. ParA), a homolog to a conjugation/recombination enzyme (e.g. TraG), many genes common in smaller mobile genetic elements (e.g. transposases), and many genes with unknown functions. While the functions of Cardinium plasmids are currently unknown (hence their status as “cryptic” plasmids), it has been hypothesized that they contain genes important for host interaction (Penz et al. 2012, Santos-Garcia et al. 2014, Xiong et al. 2023).

Despite similarity between Cardinium genomes in overall genomic properties, these genomes are highly variable at the gene and nucleotide levels, with a wide range in the average nucleotide identity from around 70% to >99% between genome pairs (Table S2). For example, our analysis using OrthoVenn3 (Sun et al. 2023) on 11 Cardinium genomes from various hosts (cEper1, cBtQ1, cSfur, cOegib-Wal, DF, TP, icPhiSpin1, idTipUnca1, cCpun, cPpe, and cHgTN10) revealed a core set of 389 gene clusters out of the ~1000 genes encoded by a Cardinium genome (Fig. S1). Conserved functions (inferred via GO terms of orthologous clusters) are mainly those essential for bacterial life, including DNA replication and repair, transcription, translation, cell homeostasis, protein folding and transport, ATP synthesis, and a limited set of biosynthesis pathways such as those for peptidoglycan and liposaccharides.

Cardinium metabolism and physiology

Due to its obligate intracellular lifestyle and overall limited encoded metabolic capacity, it is difficult to infer the core metabolism and energy generation mechanisms of Cardinium. We compiled results from the KEGG GhostKOALA service (Kanehisa et al. 2016) on all available genomes to provide a general overview of Cardinium metabolism for this review (Fig. 6).

Figure 6.

Heatmap comparing the metabolic potential of Cardinium genomes to their sister lineage A. asiaticus and to Wolbachia wPip. Wolbachia wPip represents a second endosymbiont lineage that shares multiple general features with Cardinium, including occupying a similar intracellular niche, commonly occurring in a range of invertebrate hosts, and causing a variety of manipulative and beneficial host phenotypes. A general assessment of the metabolic potential of the genomes was generated using the GhostKOALA service from KEGG (Kanehisa et al. 2016) by combining results from the “module” tab of the KEGG Mapper Reconstruction output for each genome into a heatmap using JColorGrid (Joachimiak et al. 2006). Numbers in parentheses at the end of each pathway name indicate the total number of genes in that pathway, and the accession for each KEGG pathway module is given in brackets. Further, bolded letters above the tree indicate proposed Cardinium clades from Fig. 2 and asterisks indicate genomes that have additional pathways not shown in this heatmap, since only select pathways present in Wolbachia or Amoebophilus but absent in Cardinium genomes were included in the figure. Note that high fragmentation of genome assemblies may cause inaccuracies in assessing presence or absence of pathways so functional capacity might be underestimated. Refer to Table 1 for accession numbers and other information regarding Cardinium genomes included in this figure. GenBank accession numbers for A. asiaticus and Wolbachia wPip are GCA_000020565.1 and GCA_000073005.1.

Cardinium has limited metabolic capacity like other obligate intracellular symbionts, with the highest biosynthetic potential occurring in the CI-causing Cardinium strains cEper1 and cSfur, while the nematode-associated (cPpe, cHhum, and cHgTN10) and mite-associated (cDfar) Cardinium strains have lower biosynthetic potential (Fig. 6).

Even highly conserved pathways of central metabolism are reduced or lost entirely in Cardinium. This suggests that Cardinium relies on many host-derived metabolites and pathway intermediates to make up for its incomplete biosynthetic pathways. Our analysis shows most Cardinium genomes encode a partial glycolysis pathway, where only the reactions involving 3-carbon compounds are retained (Fig. 6). Glycolysis is further reduced in cDfar and cOegib-Wal and is completely absent in nematode-associated Cardinium (cHgTN10, cPpe, and cHhum). This suggests even greater host dependence in these strains. Gluconeogenesis is also incomplete in Cardinium and 6-carbon sugars cannot be formed by most strains. This pathway is further reduced in cDfar, cOegib-Wal, and nematode-associated Cardinium. The oxidation of pyruvate to acetyl-CoA via pyruvate dehydrogenase is conserved (except for cDfar and cPpe). Additionally, the biosynthesis of lipoate, which is an important cofactor for many enzymes (e.g. pyruvate dehydrogenase) (Spalding and Prigge 2010), is present in all Cardinium except cPpe. However, the tricarboxylic acid (TCA) cycle and other central metabolism pathways such as the Entner–Doudoroff pathway and the pentose phosphate pathway appear to be absent in Cardinium (Fig. 6). All subunits of adenosine triphosphate (ATP) synthase are also encoded in nearly all Cardinium genomes based on BLASTp comparisons, so the ability to generate ATP via a proton gradient across the cell membrane has likely been retained. However, no other components of an electron transport chain are present (Penz et al. 2012). Only some pathways that use intermediates of central carbon metabolism are present. For example, most Cardinium are capable of both fatty acid biosynthesis initiation and elongation; however, no Cardinium strain can synthesize nucleotides de novo and there are very few complete biosynthesis pathways for amino acids, even with Cardinium genomes commonly encoding tRNAs for all 20 amino acids (Fig. 6). DNA maintenance also appears to be well-conserved in Cardinium with ~42 DNA repair and recombination genes per genome predicted by GhostKOALA (Kanehisa et al. 2016). This includes RecA, which seems to be conserved in all sequenced genomes, likely enabling homologous recombination (Penz et al. 2012).

Cardinium encodes pathways for core elements of the bacterial cell membrane, retaining the entire pathway for the synthesis of phosphatidylethanolamine (Fig. 6), a core lipid of bacterial cell membranes (Murzyn et al. 2005), and most of the pathway for producing peptidoglycan based on amino acid similarities to known peptidoglycan biosynthesis proteins. Cardinium has also retained its ability to produce some components of lipopolysaccharide (LPS), an outer membrane feature characteristic of Gram-negative bacteria. For example, most of the genes required for the biosynthesis of lipid A, which causes the endotoxic effects of LPS (Gronow and Brade 2001), are conserved across all Cardinium. In addition, the production pathway of dTDP-l-rhamnose, a precursor to rhamnose and prominent component of the O-antigen of LPS (Tsukioka et al. 1997), is also mostly complete in all Cardinium other than those associated with nematodes (cHgTN10, cPpe, and cHhum) (Fig. 6). It has been proposed that the presence of LPS in the Cardinium outer membrane may cause it to induce a stronger immune response in some hosts, thus potentially limiting its host range compared to Wolbachia, which has lost most of the genes required for LPS biosynthesis (Weinert et al. 2015). Future work incorporating expression data will discern the relevance of these cell surface features, which will be important for evaluating this hypothesis.

Cardinium strains cEper1 and cSfur encode a complete biotin synthesis pathway, but this pathway is only partially encoded by cBtQ1, CanCar, and cHhum, and completely absent in A. asiaticus and all other sequenced Cardinium strains (Fig. 6) (Brown et al. 2018, Penz et al. 2012, Santos-Garcia et al. 2014, Siozios et al. 2019, Zeng et al. 2018). The degradation or absence of the biotin synthesis pathway in most Cardinium suggests that supplementation of this nutrient is not needed for the bacterium itself, nor is it a key component of most Cardinium–host symbioses. It is possible the biotin synthesis capabilities of cEper1 and cSfur provide a nutritional benefit to their hosts, as B-vitamins are broadly important for arthropod fitness but are absent in specialized diets like blood and phloem (Serrato-Salas and Gendrin 2023). Therefore, the supplementation of B-vitamins by endosymbiotic bacteria, such as Wolbachia, to their hosts can play an important role in nutritional symbiosis (Ju et al. 2020). A role in biotin provisioning is plausible for cSfur, which is hosted by a sap-feeding planthopper. However, it seems unlikely for cEper1 to provide a large nutritional benefit through biotin provisioning as its host is a parasitoid that presumably acquires biotin from its whitefly diet. Evidence points toward a potential horizontal transfer of the core biotin synthesis gene cluster encoded by cEper1 and cSfur (bioA, bioD, bioC, bioH, bioF, and bioB) between Cardinium and Wolbachia, as these genes occur in the same arrangement in Cardinium and Wolbachia and have greater sequence similarity to Wolbachia-encoded homologs than to genes encoded by other bacteria in the order Cytophagales (phylum: Bacteroidota) (Nikoh et al. 2014, Penz et al. 2012, Zeng et al. 2018). Overall, much more remains to be explored regarding the role, origins, and retention of the biotin synthesis pathway in Cardinium.

Cardinium has a greatly reduced biosynthetic potential compared to Wolbachia. Relative to Cardinium, Wolbachia strain wPip is enriched in KEGG functional categories for the metabolism of carbohydrates, nucleotides, cofactors, vitamins, lipids, and amino acids. Wolbachia encodes the entire TCA cycle, most of the nonoxidative pentose phosphate pathway, and can synthesize phosphoribosyl pyrophosphate, nucleotides, and glutathione, all of which are mostly absent in Cardinium. Wolbachia also has nearly complete pathways for the biosynthesis of lysine, riboflavin, heme, ubiquinone, and tetrahydrofolate (Fig. 6). The only biosynthetic pathways, which appear to be more complete in Cardinium than Wolbachia wPip are those involved in LPS production and biotin synthesis, the latter of which is complete in only two Cardinium genomes (Fig. 6). Overall, the potential for Cardinium to provision nutrients to its hosts appears to be less than that of Wolbachia; however, additional genomes will provide further context for the full extent of Cardinium metabolism.

Genome-encoded features for host interaction

While lacking in biosynthetic potential, Cardinium genomes are rich in predicted eukaryotic-interacting proteins (e.g. proteins with ankyrin repeats, tetratricopeptide repeats, leucine-rich repeats, and F- and U-box domains; Frank 2019, Martyn et al. 2022), as well as transport proteins and secretion systems. These features may be involved in interactions ranging from host immune evasion to reproductive manipulation and are likely integral to the host-associated lifestyle of Cardinium. However, much is still unclear regarding how Cardinium uses these genes for interactions with its hosts, and, as such, they are important targets for characterizing how Cardinium interfaces with and influences a wide range of host organisms.

Due to its drastically reduced biosynthetic capacity, Cardinium likely relies heavily on its transporters to obtain necessary host-derived metabolites, nucleotides, and other molecules. Most Cardinium genomes encode ~38 transport proteins as predicted by GhostKOALA (∼4% of CDS), but this number is likely an underestimate based on reported numbers from previously published genomes [60 predicted transporters in cEper1 and 80 in cSfur (7.1% and 10% of CDS, respectively) (Penz et al. 2012, Zeng et al. 2018)]. These are predicted to include oligopeptide transporters, dicarboxylate transporters, an ATP/ADP (adenosine diphosphate) antiporter, and other putative nucleotide transporters, among others, suggesting that Cardinium can import many essential compounds from its surrounding host cell (Penz et al. 2012, Zeng et al. 2018).

Cardinium also encodes multiple secretion systems for exporting synthesized effectors and other proteins to host cells. All Cardinium sequenced, thus far have both the sec-dependent secretion pathway for transporting unfolded proteins across the cytoplasmic membrane and homologs of the novel phage-derived type VI secretion system (T6SS^iv) characterized in A. asiaticus (Böck et al. 2017) based on gene annotations and amino acid similarities to known proteins. The intracellular components of the T6SS^iv likely form the prominent columns for which Cardinium is named and which are a key characteristic of its cellular ultrastructure (Fig. 1; Böck et al. 2017). The effector proteins secreted by this secretion system are unknown, but the prominence and high transcription level of the assembled T6SS^iv within Cardinium cells (Mann et al. 2017) and prevalence of this secretion system across the entire lineage suggest it plays an important role in protein secretion and/or interaction with hosts or other microbes. MacSyFinder (Néron et al. 2023) also predicts that some Cardinium strains additionally encode a type I secretion system (T1SS), including CanCar, cBtQ1, cCpun, cDfar, DF, icPhiSpin1, and TP. Further research is required to confirm whether the T1SS is assembled and functional in these Cardinium strains and what its role in cell physiology may be, as many sequenced Cardinium appear to lack T1SS component homologs.

Some Cardinium genomes, including cCpun, cBtQ1, cSfur, cOegib-Wal, icPhiSpin1, and idTipUnca1, encode at least four homologs to genes relating to gliding motility (such as gldK, gldL, gldM, and gldN) (Halter et al. 2023, Santos-Garcia et al. 2014, Siozios et al. 2019, Zeng et al. 2018). The role of these Cardinium genes in gliding motility is debatable, however, since many core genes are missing and Cardinium strains like cPpe that lack this machinery can be found throughout the body of their host (Brown et al. 2018). It has been posited that these four proteins could instead be involved in protein secretion via a role in building the type 9 secretion system (T9SS) common to Bacteroidota (McBride 2019). However, it should be noted that many of the genes required for the T9SS are absent. An alternative hypothesis speculates that the T9SS may be ancestral to Cardinium and may have undergone gradual loss following replacement with the T6SS^iv as the dominant protein secretion system in Cardinium. This idea has yet to be examined (Siozios et al. 2019), and more research is required to identify the significance and evolutionary origins of gliding motility proteins in Cardinium.

Finally, the recent increase in availability of Cardinium genomes and the identification of factors involved in reproductive manipulations caused by other symbionts provides an opportunity for expanded comparative analyses identifying the presence, or lack thereof, of homologous manipulation factors within Cardinium genomes. The initial comparative analysis between CI-inducing cEper1 and CI Wolbachia strains found little homology between the distantly related bacteria (Penz et al. 2012). Since that study, factors involved in Wolbachia CI have been identified, with a more recent comparative analysis confirming that Cardinium genomes do not encode the factors responsible for Wolbachia CI, and thus acquired its CI manipulation independently (Lindsey et al. 2018). Unfortunately, the absence of genomes for feminizing and parthenogenesis-inducing Cardinium precludes a comparative analysis of effector genes, like the recently identified factors implicated in Wolbachia PI, at this time (Fricke and Lindsey 2024b, Li et al. 2024b). Comparative genomics studies comparing phenotypically similar Cardinium symbionts would provide insights into potential Cardinium reproductive manipulation factors.

Cardinium transcriptomics

To date, there are two published studies on transcription levels of Cardinium. The first Cardinium transcriptome was published by Mann et al. (2017) comparing transcription levels between Cardinium cEper1 from male and female adult E. suzannae to identify potential CI candidate genes. A second study, published by Gardner et al. (2018), assembled 468 Cardinium transcripts from the nematode, H. glycines, and provided a brief Gene Ontology (GO) overview. Both studies found housekeeping genes (i.e. genes involved in cell maintenance like DNA maintenance, transcription, and translation) and transporters to be transcribed at the highest overall levels (Mann et al. 2017, Gardner et al. 2018). Below, we summarize the unique features of the strains presented in these two studies.

Most of what is known about Cardinium transcription is based on findings in the cEper1 strain. Mann et al. (2017) found biotin synthase was highly transcribed, along with other genes in the biotin synthesis pathway, which were moderately transcribed. As discussed above, the benefit of biotin for cEper1 or E. suzannae is unclear as E. suzannae should obtain B-vitamins from their larval whitefly diet. Additionally, 55.8% of the 129-predicted transposases within the Cardinium cEper1 genome were transcribed, although it is unclear whether they are active in transposition (Mann et al. 2017). Finally, there were also many highly transcribed hypothetical proteins with no known or predicted function, some of which were classified as candidates for causing CI based on predicted protein domains, annotations, and gene expression levels (Mann et al. 2017). Although Cardinium transcription does not differ greatly based on host sex generally, there were 15 differentially transcribed Cardinium genes between male and female hosts, with three upregulated in cEper1 infecting male hosts and 12 upregulated in cEper1 infecting female hosts (Mann et al. 2017). Five of the Cardinium genes upregulated in female hosts encode ribosomal proteins, one encodes a tRNA ligase, and two encode RNA polymerase subunits, suggesting an overall increase in transcription and translation by cEper1 when infecting female E. suzannae. Only one cEper1 gene upregulated in females is an uncharacterized hypothetical protein, while all three Cardinium genes upregulated in males are uncharacterized, with two being adjacent plasmid-encoded genes. It is possible that more pronounced differences may be found among Cardinium infecting different sexes of other hosts, Encarsia of earlier life stages, or when analyzing tissue/organ-specific gene expression patterns.

Gardner et al. (2018) applied transcriptomics to identify host cell manipulation effectors from the early life stages of the soybean cyst nematode H. glycines and, as a by-product, assembled 468 Cardinium transcripts from infected nematodes. Subsequent analysis of GO terms revealed that the most abundant biological processes assigned to Cardinium transcripts in this host were mainly involved in cell maintenance and growth: translation (14%), transport (10%), DNA metabolic processes (8%), and carboxylic acid metabolic processes (8%) (Gardner et al. 2018). The main molecular functions assigned to Cardinium transcripts included ATP binding (24%), DNA binding (17%), RNA binding (14%), and metal ion binding (11%), suggesting that these features may play key roles in the association of Cardinium with H. glycines (Gardner et al. 2018).

Overall, our knowledge on Cardinium gene expression at the transcript level remains limited, with only two strains represented. Additional published transcriptomic studies of Cardinium hosts in which host gene expression was the focus (Abbà et al. 2022, Li et al. 2020a, Liu et al. 2023, Yang et al. 2021) could be leveraged, along with future transcriptomic and proteomic data, for a more comprehensive overview of regulatory, functional, and host interaction features associated with Cardinium.

Host responses to Cardinium infection

Understanding how a variety of hosts respond to infection by Cardinium is important for developing a general understanding of Cardinium—host interactions. Yet, our knowledge on host transcriptomic and proteomic changes in response to Cardinium largely derives from two hosts: the house dust mite, D. farinae, and MED Bemisia whiteflies (also called B. tabaci Biotype Q).

In a transcriptome study of D. farinae, Cardinium infection was negatively correlated with the expression of several host metabolic pathways, such as glycolysis and the TCA cycle (Hubert et al. 2021a), suggesting that Cardinium infection may decrease rates of D. farinae central metabolism. Hubert et al. (2021a) also hypothesized that, since Cardinium gene transcription was correlated with the transcription levels of mite genes involved in terpenoid backbone biosynthesis, Cardinium may stimulate the production of mite pheromones involved in reproduction or aggregation, which could potentially enhance the transmission of the maternally inherited symbiont. Additionally, the expression of some Cardinium genes was correlated with the transcription of various mite immune system genes, suggesting that Cardinium may influence the immune system of its mite host, perhaps through activation of mite Toll-like receptors or by altering the transcription level of caspases, which induce apoptosis of host cells (Hubert et al. 2021a). An additional study found that D. farinae mite exosomes, which are small mite-secreted organelles implicated in human allergic respiratory disease due to the molecules they carry as cargo, contain Cardinium-derived proteins: 17 out of 72 proteins detected in exosomes were Cardinium proteins (Yang et al. 2023b). This suggests Cardinium may also play a role in the allergenicity of the house dust mite (Yang et al. 2023b).

Cardinium has been linked to a variety of changes in MED Bemisia gene expression, which may contribute to changes in host phenotypes. For example, Cardinium infection can upregulate host genes involved in cellular homeostasis, metamorphosis, autophagy, and the ubiquitin/proteasome system, all of which may be involved in the increased thermotolerance of some Cardinium-infected whiteflies (Yang et al. 2021). Cardinium is also associated with decreased transcription of some whitefly detoxification genes (Liu et al. 2023). This change in detoxification gene transcription may be in part a response to altered host plant transcription, as cotton plants in this study also exhibited a downregulation of defense response genes when the plants were infested with Cardinium-infected whiteflies versus uninfected whiteflies (Liu et al. 2023). It is possible that Cardinium may mediate whitefly–plant interactions by suppressing plant antiherbivory defenses.

The only example of host responses to Cardinium at the protein level also focused on uninfected and Cardinium-infected MED Bemisia whiteflies (Li et al. 2018a). In that study, there were 146 differentially expressed host proteins in response to Cardinium infection. Upregulated proteins may be involved in the whitefly immune response, as they were associated with RNA metabolism and pathways involved in signaling, drug resistance, apoptosis, and the metabolism of glyoxylate and dicarboxylate. The metabolism of retinol, a vitamin crucial for many aspects of host growth, development, and immune function was also upregulated (Li et al. 2024a, Morriss-Kay and Wardt 1999). Downregulated genes were enriched in functions involved in DNA processing and the spliceosome, suggesting that host pre-mRNA splicing may be altered during Cardinium infection (Li et al. 2018a). A later study by the same group found that Cardinium infection induced 23 differentially expressed MED Bemisia microRNAs (miRNAs), which posttranscriptionally regulate genes via prevention of mRNA translation (Bartel 2004, Li et al. 2018b). These Cardinium-responsive miRNAs may be responsible for some outcomes seen via proteomics, as they are predicted to regulate genes involved in female host development, immune response, and energy metabolism, as well as inhibit apoptosis and decrease host resistance to thermal stress and insecticides during Cardinium infection (Li et al. 2018b).

Overall, Cardinium and host gene expression analyses are severely lacking and largely limited to a single host species. New experiments designed to interrogate specific questions (e.g. Cardinium gene expression differences in different host sexes, life stages, organs, and so on) are essential to understand the mechanisms of Cardinium–host interactions. Furthermore, additional datasets and analysis of Cardinium infecting various hosts with confirmed phenotypes (e.g. CI, PI, increased fecundity, and so on) would be an asset to emerging Cardinium research.

Present and future directions in Cardinium research

Substantial progress has been made over the past 30 years in our understanding of the widespread invertebrate endosymbiont, C. hertigii. Research has shown Cardinium to be a symbiont of numerous invertebrate lineages, capable of manipulating host biology in a variety of ways. The sequencing of Cardinium genomes provided additional insights into the biology of this endosymbiont, such as revealing its limited metabolic potential, identifying a novel phage-derived Type VI secretion system, and confirming an independent origin of Cardinium CI. Still, gaps in our understanding of Cardinium remain, leaving many important areas of Cardinium biology open for exploration, such as the general cellular biology of Cardinium, the host–microbe interactions of Cardinium symbioses on molecular and phenotypic scales, and the global dynamics of the broader ecophysiology and evolution of Cardinium (Fig. 7). These areas of research will be critical to address as we start to consider potential applications for Cardinium in agricultural and human health contexts.

Figure 7.

Future directions and open questions in Cardinium research. These directions include researching Cardinium cellular biology and Cardinium–host interactions to better understand how this symbiont modifies host biology. Investigations into Cardinium interactions with co-infecting microbes, particularly pathogens, could provide a basis for incorporating Cardinium into arthropod pest control strategies. Expanded characterizations of Cardinium host range and the influence of environmental factors on Cardinium infection dynamics may help predict how climate shifts impact widespread heritable endosymbionts like Cardinium and their invertebrate hosts.

The increased accessibility of genomic sequencing has been a boon for symbiosis research broadly, generating insights into the molecular biology and evolution of uncultivable bacterial symbionts like Cardinium. For example, several Cardinium genomes have been published over the last decade, providing information on the functional potential of these symbionts. Yet, studies exploring Cardinium gene expression at the transcript and protein levels remain limited. Such analyses are essential for learning how Cardinium survives in and responds to its host environment. Proteomic analyses are especially needed, as they have played a critical role in the identification of candidate factors involved in symbiont physiology and host interaction, including the Wolbachia CI factors (Beckmann and Fallon 2013, Beckmann et al. 2017, LePage et al. 2017). Furthermore, the increasing availability of more advanced sequencing techniques and expression analyses, such as single-cell transcriptomics and proteomics, may further facilitate research into how Cardinium responds to and modifies its native host cell environment.

To date, Cardinium–host interaction research has largely focused on the phenotypic consequences of infection for animal hosts; however, information on the molecular characteristics and products of Cardinium that drive these phenotypes remains limited. For example, an understanding of the molecular mechanisms underlying Cardinium reproductive manipulation is currently lacking. While research on Encarsia wasps has provided new details on Cardinium CI modification and its resulting cytological defects, it remains to be seen what molecular features induce these drastic phenotypes. Gaining a general understanding of how Cardinium manipulates host biology will require mechanistic information from a variety of host systems. The recent advances in the accessibility of genome sequencing and gene expression analyses, coupled with the use of heterologous expression systems in model organisms like Saccharomyces cerevisiae, offer an opportunity to unravel the molecular mechanisms underlying heritable symbioses even in less tractable host systems (Murphy and Beckmann 2024).

Our understanding of Cardinium host effects beyond reproductive manipulation is even more limited. In particular, studies exploring interactions between Cardinium and host pathogens are surprisingly lacking given symbiont-mediated protection and potential for disease mitigation granted by other common symbionts (O’Neill et al. 2018, Ballinger and Perlman 2019, Gong et al. 2023). Research into how Cardinium modifies the vector potential of important crop pests like planthoppers or animal disease vectors, such as biting midges will be integral to understanding the potential of this symbiont as a tool for arthropod-vectored disease control (Mellor et al. 2000, Morag et al. 2012, Li et al. 2020a). Research into the host benefits granted by Cardinium infection will also improve our understanding of Cardinium/host interactions, including how this symbiont spreads through host populations, and the evolutionary dynamics of these symbioses.

Direct molecular interrogation of Cardinium biology has been challenging due in part to the obligate intracellular lifestyle of the bacterium. Developing tools for direct molecular manipulation of Cardinium would open a number of doors for understanding basic aspects of Cardinium biology and interactions with its host, ranging from functionally characterizing candidate factors for host manipulation or growth regulation to identifying novel host interaction features important for maintaining symbiosis. For example, peptide nucleic acids (Pelc et al. 2015) and CRISPR-based technologies have seen increasing use in manipulating formerly intractable bacteria, including Rickettsia spp. which commonly associate with insects (McClure et al. 2017, Fisher and Beare 2023). Chemical mutagenesis is another intriguing option for manipulating Cardinium, as it does not require Cardinium cell cultures and could potentially be applied to Cardinium living in a natural host (McClure et al. 2017). Using chemical mutagenesis in combination with genetics approaches could enable researchers to generate and identify hosts harboring Cardinium mutants with altered phenotypes (McClure et al. 2017, Duarte et al. 2021). Such an approach has been successfully used to identify genes important for regulating Wolbachia proliferation (Duarte et al. 2021). The ability to culture Cardinium cells in different arthropod cell lines (Kurtti et al. 1996, Morimoto et al. 2006, Nakamura et al. 2011) combined with these emerging methods for molecular manipulation of obligate intracellular bacteria is a promising route to elucidate the molecular mechanisms that underlie Cardinium–host interactions.

Just as important as understanding how Cardinium interacts with its host, is the range of hosts with which Cardinium interacts. Recent surveys suggest that the host range of Cardinium is broad, including species across multiple phyla, indicating that Cardinium likely acts as an unseen player in the biology of a substantial number of invertebrate species. For example, identification of Cardinium in entirely new host groups, like mollusks, suggests that current estimates of Cardinium prevalence do not reflect its true global prevalence and host range (Weinert et al. 2015). Clearly, more surveys across host taxa are needed to better understand this symbiont’s host range, particularly in host taxa that have only had limited screening, like nonmarine crustaceans, or taxa belonging to groups commonly harboring Cardinium, like spiders. The increasing ease and usage of high throughput sequencing techniques, like 16S rRNA amplicon sequencing, will enable these advancements, making larger surveys that target a broader range of potential hosts feasible.

The effect of geographic and climate variation on Cardinium infection dynamics should also be explored. Even with their sustained successful spread across invertebrates, some heritable symbionts, like Cardinium, can be highly susceptible to thermal stress. Understanding how these symbionts respond to a climate with increasing and more variable temperatures remains a high priority research topic. Given the fundamental roles that Cardinium and other symbionts play in their host’s biology (e.g. influencing sex allocation, offspring survival, and fecundity), the effects of the environment on heritable symbionts will also have wide ranging consequences for their invertebrate hosts. This could also have anthropogenic implications, as symbiont thermal sensitivity may affect the potential application of symbionts in pest management and control of insect-vectored diseases, highlighting the need to better understand the interactions between climate and heritable symbioses. Surveys of natural Cardinium-infected host populations, particularly across seasonal or spatial gradients, could provide a more accurate understanding of present day Cardinium infection dynamics, while laboratory studies can be designed to better replicate natural thermal conditions of the present day and predicted future climate conditions. Global surveys may also offer a great opportunity to recruit and educate members of the general public and scientific community across the world via citizen science through a program similar to the Wolbachia Project and its associated database (Lemon et al. 2020). Together, these types of studies will generate a clearer understanding of global Cardinium infection dynamics.

While much remains to be uncovered regarding Cardinium evolution, ecology, and molecular physiology, advancements in the fields of bioinformatics and molecular biology offer a wealth of opportunities to learn more about Cardinium. The coming years will likely bring new information on how Cardinium manipulates host biology, as well as how Cardinium and their hosts may respond to a changing climate. This will lay the groundwork for potential applications of Cardinium in areas, such as pest management and arthropod-vector disease control and prevention in the not-so-distant future.

Funding

This work was funded by the National Science Foundation (#2426306 to M.S.H., #2002987 and #2426304 to S.S.-E., and IOS #2426305 and IOS #2003107 to M.K.) and the National Institute for Food and Agriculture-United States Department of Agriculture (2023-67012-39352 to M.R.D. and #2023-67013-39897 to M.S.H.).