PAPER Signatures of host/symbiont coevolution in COLLOQUIUM nutritional endosymbioses

Alex C. C. Wilson1 and Rebecca P. Duncan

Department of Biology, University of Miami, Coral Gables, FL 33146

Edited by John P. McCutcheon, University of Montana, Missoula, MT, and accepted by the Editorial Board May 5, 2015 (received for review February 2, 2015) The role of in bacterial symbiont genome evolution is these ancient, intimate symbiotic associations. Our purpose here well understood, yet the ways that symbiosis shapes host is to highlight three signatures of genome coevolution across or more particularly, host/symbiont genome coevolution in the available holobiont genomes and to draw attention to work at holobiont is only now being revealed. Here, we identify three the frontier of symbiosis research that elucidates mechanisms of coevolutionary signatures that characterize holobiont genomes. The holobiont regulation and integration. first signature, host/symbiont collaboration, arises when completion of essential pathways requires host/ genome com- Three Signatures: Collaboration, Acquisition, and Constraint plementarity. Metabolic collaboration has evolved numerous times Holobiont genome evolution is characterized by patterns of col- in the pathways of amino acid and vitamin biosynthesis. Here, we laboration, acquisition, and constraint. Coevolution typically fea- highlight collaboration in branched-chain amino acid and pantothe- tures host/endosymbiont collaboration on completion of critical nate (vitamin B5) biosynthesis. The second coevolutionary signature metabolic pathways—a set of pathways that is similar across taxa, is acquisition, referring to the observation that holobiont genomes apparently constrained by eukaryotic host gene repertoire and yet, acquire novel genetic material through various means, including concurrently, holobiont genome evolution is dynamic. Dynamic , lateral gene transfer from that are not features include acquisition of novel genomic material like du- their current obligate symbionts, and full or partial endosymbiont plicate genes, genes acquired by lateral gene transfer that enhance replacement. The third signature, constraint, introduces the idea collaboration, and acquisition of coprimary symbionts or even new

that holobiont genome evolution is constrained by the processes primary symbionts by symbiont replacement. EVOLUTION governing symbiont genome evolution. In addition, we propose that collaboration is constrained by the expression profile of the Collaboration cell lineage from which endosymbiont-containing host cells, called Long before it was possible to elucidate the metabolic repertoire , are derived. In particular, we propose that such of an organism by sequencing its DNA, researchers established differences in cell lineage may explain differences in the metabolic basis of many insect nutritional symbioses experi- patterns of host/endosymbiont metabolic collaboration between mentally by quantifying the growth and fecundity of ma- the sap-feeding suborders and Auchenorrhynca. nipulated in a diversity of ways that included some combination of Finally, we review recent studies at the frontier of symbiosis research endosymbiont removal and diet manipulations (13). Such analyses that are applying functional genomic approaches to characterization established that provision host insects with dietary of the developmental and cellular mechanisms of host/endosymbiont components lacking or at low availability in their diets. Thus, at integration, work that heralds a new era in symbiosis research. the beginning of the genomic revolution it was understood that hosts supply endosymbionts with metabolic precursors that en- symbiosis | insect nutritional | coevolution | amino acid biosynthesis | dosymbionts metabolically transform into host-required dietary vitamin biosynthesis components. Notably, from an organismal perspective, insect nu- tritional symbioses in pregenome times were partnerships between any insects have established persistent, intimate associa- two discrete organisms: the host and the endosymbiont. Endo- Mtions with vertically transmitted microbial symbionts. These symbiont whole-genome sequencing affirmed and continues to ancient, commonly intracellular symbionts are, owing to their affirm the nutritional role played by symbionts, whereas host greatly reduced, gene-poor genomes, uncultivable outside their genome and transcriptome sequencing reveals a complex por- hosts. Although the whole-genome sequences of obligate insect trait of the nature and extent of host/endosymbiont metabolic endosymbionts continue to yield novel insight into genome evo- complementarity. lution following adoption of an obligate intracellular lifestyle (e.g., Currently the most well-documented examples of host/endo- ref. 1), it is the recent partnering of these symbiont genome se- symbiont metabolic collaboration involve the biosynthesis of quences with those of their insect hosts that fundamentally shifts essential amino acids in plant sap-feeding insects (2, 7–9, 14–16). conceptualization of “the organism,” propelling us into a new era In most cases, holobiont metabolism has been reconstructed in symbiosis research. using the whole-genome sequence of the symbiont coupled with The first paired insect host/endosymbiont genome available a host bacteriome transcriptome (the bacteriome is the host was that of the , pisum and its bacterial endosymbiont, aphidicola (2, 3). The second was that of the human body louse, Pediculus humanus humanus and its This paper results from the Arthur M. Sackler Colloquium of the National Academy of primary bacterial endosymbiont Candidatus Riesia pediculicola Sciences, “Symbioses Becoming Permanent: The Origins and Evolutionary Trajectories of (4), quickly followed by the carpenter , Camponotus floridanus Organelles,” held October 15–17, 2014, at the Arnold and Mabel Beckman Center of the National Academies of Sciences and Engineering in Irvine, CA. The complete program (5) and its endosymbiont Candidatus Blochmannia floridanus (6). and video recordings of most presentations are available on the NAS website at www. Other insect holobiont genomes now include the citrus mealybug, nasonline.org/Symbioses. Planococcus citri with its dual endosymbionts (7, 8); the hackberry Author contributions: A.C.C.W. and R.P.D. designed research, performed research, ana- petiole gall psyllid Pachypsylla venusta with its endosymbiont lyzed data, and wrote the paper. Candidatus Carsonella ruddii (9); and the tsetse fly, Glossina The authors declare no conflict of interest. morsitans (10) with its endosymbiont Wigglesworthia glossinidia This article is a PNAS Direct Submission. J.P.M. is a guest editor invited by the Editorial (11, 12). This recent accumulation of holobiont genomes facili- Board. tates elucidation of the patterns that characterize coevolution in 1To whom correspondence should be addressed. Email: [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1423305112 PNAS | August 18, 2015 | vol. 112 | no. 33 | 10255–10261 Downloaded by guest on September 28, 2021 symbiosomal membrane bacteriocyte cell membrane

Symbiont with host Bacteriocyte Bacteriome Host insect symbiosomal membrane with symbionts with bacteriocytes with bacteriome

Fig. 1. Organization of endosymbionts in insect hosts. Typically enveloped by a host membrane (green) within the cytoplasm of specialized host bacteriocyte cells, insect nutritional endosymbionts (purple) are maternally transmitted and integrated into host development by currently unknown mechanisms. In- dividual bacteriocyte cells house thousands to tens of thousands of symbionts. Many bacteriocyte cells aggregate to form an organ called the bacteriome. Bacteriomes are typically paired and relatively large.

organ composed of host bacteriocyte cells that house the en- encoding pantothenate synthetase (E.C. 6.3.2.1) that functions dosymbionts) (Fig. 1) and a partial host genome. In the case of only in pantothenate biosynthesis (Fig. 2) (18). Thus, we venture the pea aphid, metabolic reconstruction leveraged the com- that examination of the host genomes and transcriptomes of plete genome of both symbiont and host (2, 14). The many these four endosymbionts will demonstrate that pantothenate similarities and differences in the collaborative amino acid bio- biosynthesis is indeed functional in these holobionts as a result of synthesis of plant sap-feeding holobionts have recently been host/endosymbiont metabolic collaboration. reviewed by Sloan et al. (9) and Hansen and Moran (17). It is clear that metabolic collaboration is a fundamental sig- Therefore, here we illustrate the collaborative signature of host/ nature of host/symbiont genome coevolution. However, the ex- symbiont genome coevolution by drawing attention to one set of tent to which patterns of metabolic collaboration in holobionts amino acid biosynthesis pathways, those for biosynthesis of the result from convergence and not shared ancestry remains to be branched-chain amino acids. Biosynthesis of the branched-chain determined. Distinguishing between traits shared by conver- amino acids, isoleucine, leucine, and valine, involves host/endo- gence and those shared by common ancestry requires charac- symbiont metabolic collaboration in both A. pisum (14, 15, 16) and ter state information for a broad range of closely related taxa P. citri (7, 8). The endosymbionts of both insects have lost the and unfortunately such data are not currently available. That transaminase encoded by ilvE, an enzyme (E.C. 2.6.1.42) that said, on the basis of the data that are available, we venture that mediates the terminal reaction in the biosynthesis of all three metabolic collaboration has evolved independently in multi- branched-chain amino acids. However, the genomes of both host ple metabolic pathways in multiple insect lineages. For ex- insects encode branched-chain amino acid transaminase (BCA). ample, the human body louse, P. humanus and the pea aphid, Thus, in both symbioses, completion of branched-chain amino A. pisum belong to different insect orders and yet both require acid biosynthesis has been hypothesized to occur through host/ host/endosymbiont collaboration for maintenance of functional endosymbiont metabolic collaboration (7, 14). Recently, Russell pantothenate biosynthesis pathways. As described above, A. pisum et al. (16), partially validated the model of BCA complementation (18) and likely P. humanus depend on host supply of β-alanine for of leucine biosynthesis in A. pisum by demonstrating that release symbiont-mediated biosynthesis of pantothenate (Fig. 2). Further, of leucine from isolated Buchnera cells increased in the pres- pantothenate biosynthesis probably requires additional host ence of host cellular fractions containing BCA. Biosynthesis of complementation in P. humanus for synthesis of the metabolic the branched-chain amino acids serves as just one example of intermediate 2-oxoisovalerate, a metabolite synthesized in A. pisum host/endosymbiont metabolic collaboration. Other examples by the endosymbiont Buchnera. Intriguingly, the difference in include phenylalanine (9, 14, 17), methionine (8, 9, 14), and 2-oxoisovalerate biosynthesis in A. pisum and P. humanus likely arginine (8, 9) biosynthesis. Another example of host/endosymbiont metabolic collabora- tion comes from examination of vitamin biosynthesis pathways. L-aspartate β-alanine pantoate Recent work in A. pisum extends evidence of host/endosymbiont metabolic collaboration to include pantothenate (vitamin B5) panD biosynthesis (18). Pantothenate biosynthesis requires a supply of EC 4.1.1.11 β-alanine. Buchnera APS has lost panD (encoding E.C. 4.1.1.11) panC and with it, the ability to synthesize β-alanine. In contrast, A. pisum symbiont gene EC 6.3.2.1 A. pisum, like most insects, retains two pathways for synthesis of A. pisum symbiont missing gene pantothenate β-alanine. Of these two pathways, the expression of the uracil P. citri symbiont gene (vitamin B5) degradation pathway is up-regulated in bacteriocytes (the cells P. citri symbiont missing gene that house the endosymbionts) and the expression of the alternate H. vitripennis symbiont gene pathway is down-regulated. Price and Wilson (18) propose that H. vitripennis symbiont missing gene β-alanine biosynthesis by A. pisum in bacteriocytes serves to pro- P. h. humanus symbiont gene vision Buchnera and facilitate maintenance of a functional panto- P. h. humanus symbiont missing gene thenate biosynthesis pathway. Given the common patterns of G. morsitans symbiont gene collaboration in amino acid biosynthesis across holobionts (reviewed G. morsitans symbiont missing gene in refs. 9 and 17), it is notable that the endosymbionts of P. citri (8), Fig. 2. Patterns of gene retention and loss in the pantothenate biosynthesis Homalodisca vitripennis (19), P. humanus (4), and G. morsitans (12), pathway from insect obligate nutritional symbionts. Symbionts from all five like Buchnera APS, have lost panD and therefore the ability to hosts have lost the ability to synthesize β-alanine, whereas all retain panC, synthesize β-alanine. At the same time their role in pantothenate encoding pantothenate synthetase (E.C. 6.3.2.1), a gene that functions only biosynthesis is strongly supported by retention of panC, a gene in pantothenate biosynthesis.

10256 | www.pnas.org/cgi/doi/10.1073/pnas.1423305112 Wilson and Duncan Downloaded by guest on September 28, 2021 PAPER

results from selection to maintain amino acid biosynthesis pathways the evolutionary history of amino acid transporter genes in plant COLLOQUIUM in Buchnera, the symbiont of plant sap-feeder A. pisum, and loss sap-feeding insects is dynamic with respect to both duplication of these same pathways in Riesia, the symbiont of blood-feeder events and the recruitment of duplicated genes to the host/sym- P. humanus (4). This example aside, definitive determination of biont interface (31). The dynamic evolution of amino acid trans- what host/symbiont collaboration patterns result from conver- porters in these insects, including some very recent duplications in gence vs. descent unavoidably requires generation and analysis of (32), demonstrates that despite millions of years of host/ an extensive dataset that does not yet exist, a dataset that may be endosymbiont coevolution, host genomes are in flux (Fig. 3). some time in coming. Functional genomics work across sap-feeding insects is necessary Common patterns of collaborative maintenance of functional to determine the extent to which these changes in amino acid metabolism continue to be discovered across systems (e.g., refs. transporter evolution contribute to host/endosymbiont integration. 7–9, 14, 17) and yet, common patterns are not always found, even Lateral transfer of bacterial genes to host genomes has the when genomic potential for the evolution of collaboration exists. potential to relieve coevolutionary constraints in a marvelous For example, all sap-feeding insects depend on endosymbiont way. Such lateral gene transfer in the context of a nutritional supply of the branched-chain amino acids that are collaboratively endosymbiosis was first reported in A. pisum (33, 34). Whole- synthesized in A. pisum (14, 15, 16) and P. citri (7, 8) but not genome analysis of A. pisum identified the transfer of 12 genes or collaboratively synthesized in sap-feeding auchenorrhyncans (20, gene fragments from bacteria to the insect genome (34). Of 21) or the psyllid P. venusta (9). We propose below in the context those 12, 7 were highly expressed in bacteriocyte cells but none of the third signature that such differences in patterns of col- were implicated in amino acid or vitamin biosynthesis. Recent laboration across taxa are governed by cellular and develop- work by Husnik et al. (8) in the mealybug P. citri and Sloan et al. mental constraints. (9) in the hackberry petiole gall psyllid P. venusta found that me- tabolism central to symbiotic function depends on host-encoded Acquisition genes of bacterial origin that were acquired laterally from mul- The second signature of host/symbiont genome coevolution is tiple bacterial lineages, together revealing an important addi- acquisition; holobiont coevolution in nutritional insect symbioses tional mechanism facilitating holobiont integration (Fig. 3). Given spans hundreds of millions of years (22–25) and yet recent work the newness of the discovery of lateral gene transfer events that in many holobionts demonstrates that their genomes are dynamic. facilitate host/endosymbiont metabolic collaboration, the impact

Three mechanisms have been documented that contribute to a and extent of lateral gene transfer on host/endosymbiont inte- EVOLUTION genome-level change in insect nutritional symbioses. These in- gration remain to be fully appreciated. clude gene duplication, lateral gene transfer, and partial or full Partial or full endosymbiont replacement by previously facul- symbiont replacement. tative insect symbionts is a third mode by which holobionts gain Gene duplication events facilitate refinement of existing func- new genetic material. Despite the fact that symbiont genome tion, the evolution of new spatial and temporal gene expression evolution is characterized by genome degradation, many endo- patterns, and even the evolution of new gene functions (26–28). symbionts have coevolved with their insect hosts since ancient Recent work in insect nutritional endosymbionts has focused on times (35). When bacteria become obligate symbionts, important the evolution of nutrient amino acid transporters in the genomes population genetic parameters immediately change. For example, of insects that feed on plant sap (29–32). Those studies find that bacterial symbionts, unlike their free-living relatives, experience

Symbiont classification Holometabolous insects β-proteobacteria γ-proteobacteria

Pediculus humanus G,T Genomic resources G BCA Riesia pediculicola G Genome T Transcriptome

G,T Pachypsylla venusta Genome characters slimfast argH rsmJ Carsonella rudii G CG7888 cm ydcJ Host metabolic complementation CG1139 ribC Host amino acid transporter expansion Planococcus citri G,T cysK ddlB bioADB G AAT slimfast rlmI LGT to host genome (metabolism gene) dapF mltD ribAD Tremblaya princeps CGL CG7888 G CBL lysA amiD Moranella endobia murABCDEF LGT to host genome (non-metabolism gene) BCA Acyrthosiphon pisum G,T slimfast ldcA RlpA4 G CG16700 Buchnera aphidicola CG3424 CG8785

Fig. 3. Insects with sequenced host and symbiont genomes. Shown is the phylogeny of representative insect hosts based on Misof et al. (66), Cryan and Urban (67), and Dahan et al. (32). Insect hosts and their respective symbionts appear at the tips of the branches, with symbiont classification represented by text color. Superscripts “G” (genome) and “T” (transcriptome) denote the genomic resources available for each insect and symbiont. Four genomic characters represented by different symbols (Inset) are mapped onto branches indicating the most parsimonious explanation for their distribution among holosymbionts. “Host metabolic complementation” refers to presence of endogenous host genes complementing missing symbiont genes, with insect gene names listed below the character symbol. “Host amino acid transporter expansion” refers to the evolution of lineage-specific duplications in amino acid transporters, with melanogaster orthologs of expanded transporter listed beside the symbol. The D. melanogaster gene slimfast is listed on all three sap-feeding insect lineages because its ortholog expanded independently in each of these insects (9, 31). “LGT to host genome (metabolism gene)” refers to host genome acquisition of bacterial genes that complement missing symbiont genes with roles in metabolism; the names of host- encoded bacterial genes are listed beside the symbol. “LGT to host genome (nonmetabolism gene)” refers to host genome acquisition of bacterial genes that complement missing symbiont genes with nonmetabolic roles.

Wilson and Duncan PNAS | August 18, 2015 | vol. 112 | no. 33 | 10257 Downloaded by guest on September 28, 2021 relaxed selection and greatly reduced population size, resulting By convergence, not by descent, metabolic pathways are in- in elevated genetic drift (35–37). As a consequence of elevated tegrated so that completion of a collaborative pathway requires genetic drift and small population size, mildly deleterious mu- contribution of gene products from both symbiont and host. We tations become fixed at a higher rate than they would in a large described examples of collaborative biosynthesis of amino acids population. Further, because of an absence of sex and re- and vitamins in the first section; here we argue that patterns of combination, any symbiont population can carry a mutation load collaborative biosynthesis have evolved independently in multi- only equal to or higher than that of its least-loaded lines; this ple insect lineages because they are constrained by host gene feature of evolution in small, nonrecombining populations is repertoire. When endosymbiont genomes encode genes that are know as Muller’s ratchet (38). Over evolutionary time, Muller’s functionally redundant with host genes (or coresident symbiont ratchet results in endosymbionts having highly reduced and de- genes), endosymbionts experience relaxed selection to maintain graded genomes. Despite the degenerative nature of endosym- their functionally redundant genes. Given the process of gene biont genome evolution, there is strong selection for symbionts loss in symbiont genomes, endosymbionts are more likely to lose to retain genes essential to performing their metabolic duties, a gene if it is complemented by the host genome. Thus, it is not except when another symbiont with the same metabolic capacity surprising that within Sternorrhyncha, the biosynthesis of the infects the host. Mechanistically, then, the acquisition of facul- branched-chain amino acids, as well as phenylalanine and me- tative symbionts both relaxes selective pressures on the main- thionine, requires host complementation at the same steps, steps tenance of essential genes in the primary symbiont and provides mediated by enzymes broadly conserved across the Metazoa. opportunity for the replacement or complementation of ances- Although shared ancestry precludes our ability to confidently tral endosymbiont function or functions that are lost as a result distinguish between convergence and homology in shared ster- of relaxed selection. In such cases, facultative symbionts become norrhynchan metabolic complementation, other lines of evidence obligate, thereby facilitating maintenance of functions that have support host gene repertoire as a constraining factor in symbiont become eroded over millions of years by Muller’s ratchet (38). gene loss. For example, BCA independently enabled the evolution Full symbiont replacement and partial symbiont replacement— of metabolic collaboration in different pathways and in different i.e., the occurrence of coprimary symbionts—are common features symbiotic contexts. In Sternorrhyncha, BCA mediates the terminal of holobiont evolution. Examples of full symbiont replacement in step in branched-chain amino acid biosynthesis, complementing sap-feeding nutritional symbioses include aphids with a second- the missing symbiont gene, ilvE (14, 16). On the basis of the Kyoto arily acquired fungal symbiont (39) and the mealybug Encyclopedia of Genes and Genomes (KEGG) annotations, we Rastrococcus, in which the ancestral, betaproteobacterial symbiont propose that in the divergent blood feeders P. humanus and Tremblaya was replaced by a Bacteroidetes (40). Further, in the G. morsitans, BCA also complements the missing symbiont ilvE, this time mediating conversion of valine to pantoate in pan- Auchenorrhyncha, which typically have two primary symbionts, thothenate biosynthesis (4, 12, 44). These examples demonstrate one symbiont (Sulcia) has been maintained through evolution that at least three times, symbionts have independently lost the whereas its coprimary symbiont has been replaced many times same gene (ilvE) whose function was complemented by a con- such that different auchenorrhynchan lineages have alphaproteo- served host gene, indicating that symbiont gene loss is indeed bacterial, betaproteobacterial, gammaproteobacterial, or yeast-like constrained by host gene repertoire. Additional evidence for a symbionts (20, 21, 41) (Fig. 3). Likewise, there are many examples role of host gene content in shaping symbiont gene content is of coprimary symbionts in aphids, mealybugs, and auchenor- – illustrated by symbiont gene losses when hosts have acquired rhynchans that partition metabolic roles (19 21) and sometimes functionally equivalent genes from bacteria through lateral gene even provide genes that complement missing genes from the transfer (8, 9). Thus, as lateral gene transfer to the host genome genomes of their cosymbiont (7, 42, 43). Symbiont replacement expands host gene repertoire, it additionally relaxes selective can even arise when an ancestral symbiont diverges into two in- constraints on symbiont gene content, providing further oppor- terdependent lineages (1). The replacement or complementation tunities for host/endosymbiont metabolic collaboration. of a primary endosymbiont by a more recently acquired (or de- “ ” Apart from host gene repertoire, another factor likely con- rived) symbiont has the potential to functionally reset genes straining holobiont genome evolution is the gene expression pro- and pathways that have been eroded by mutation accumulation file of the cells that give rise to bacteriomes. This constraint is over evolutionary time. evident in symbiont retention of essential metabolic genes that are functionally redundant with host-encoded genes. For example, Constraint returning again to branched-chain amino acid biosynthesis, P. citri Emerging as the third signature of coevolution in insect nutritional and A. pisum both evolved metabolic collaboration in branched- endosymbioses, constraint influences the evolution of metabolic chain amino acid biosynthesis, where the terminal step is carried collaboration, the acquisition of bacterial genes by host genomes, out by the insect gene BCA, but branched-chain amino acid bio- and symbiont replacement. The role of constraint in metabolic synthesis is intact in symbionts of P. venusta, other psyllids, and all collaboration becomes clear in the observation that by conver- auchenorrhynchan sap feeders studied to date. This difference in gence and not by descent, the same insect genes in different collaboration is not driven by host gene content, as BCA is holobionts are engaged in host/symbiont metabolic collabora- encoded in the genome of P. venusta (9) and transcribed in the tion, complementing the same symbiont gene losses. Opportu- cicada Diceroprocta semicincta (31). The presence of BCA in in- nities to evolve collaborative biosynthesis are constrained by the sects whose symbionts have intact branched-chain amino acid gene content of host genomes. Additionally, we propose that the biosynthesis demonstrates that host gene content by itself is not coevolutionary potential of an endosymbiosis is constrained by sufficient to enable the evolution of metabolic collaboration. The the cell type that gave rise to the bacteriome in each taxon. evolution of metabolic collaboration requires expression of both Although all nucleated cell types contain the complete host ge- complementary host metabolism genes and host and/or symbiont nome, all cell types do not express all genes. Therefore, differ- transporters to facilitate movement of metabolites and/or enzymes ences in basal expression of the cell lineage that gives rise to the between host and symbiont. We raise the possibility that the dif- bacteriome in each taxon will constrain patterns of holobiont co- ferences we describe above in branched-chain amino acid bio- evolution. In fact, as we discuss below, we propose that cell line- synthesis complementation arise from differences in host gene age differences may explain differences in collaborative branched- expression profile in the cells giving rise to the bacteriomes in chain amino acid biosynthesis among the Sternorrhyncha and sternorrhynchans, psyllids, and auchenorrhynchans. We argue Auchenorrhynca (Fig. 3). this because host/symbiont complementation appears to be a

10258 | www.pnas.org/cgi/doi/10.1073/pnas.1423305112 Wilson and Duncan Downloaded by guest on September 28, 2021 PAPER

ubiquitous signature of holobiont genome evolution. Thus, if level in Buchnera. Taken together, these results reopen the door COLLOQUIUM we do not find complementation when host gene content makes for recognizing Buchnera as more than a symbiotic puppet under complementation possible, the most parsimonious explanation host control. becomes that we do not find complementation because genes The second recent study that addressed the question of met- necessary for the evolution of complementation are not ex- abolic control in the A. pisum/Buchnera symbiosis pursued the pressed in symbiont-hosting tissues. That BCA is enriched in suggestion by Moran et al. (46) that host control of nonessential P. citri (8) and A. pisum (15) bacteriomes, but is not enriched in amino acid transport to Buchnera could be a mechanism by the bacteriomes of P. venusta (9), is consistent with the notion which aphids regulate endosymbiont metabolism. By functionally that gene expression in symbiont-hosting tissues is an important characterizing A. pisum amino acid transporters via heterologous factor constraining holobiont genome evolution. expression in Xenopus laevis oocytes, Price et al. (52) identified In the same way that the accumulation of whole symbiont an amino acid transporter, ApGLNT1, that transports glutamine genomes has solidified understanding of the signatures of bac- from hemolymph into bacteriocyte cells. ApGLNT1 is a high- terial genome evolution following adoption of a symbiotic life- affinity glutamine transporter whose transport of glutamine is style, the accumulation of holobiont genomes will continue to inhibited by arginine, an end-product metabolite synthesized by refine understanding of the signatures of host/symbiont genome Buchnera. Remarkably, the localization of ApGLNT1 in the coevolution. Metabolic collaboration, the acquisition of novel bacteriocyte plasma membrane coupled with its transport capa- genomic material, and host genomic constraints are emergent bilities strongly suggests ApGLNT1 functions as an important features of host/symbiont genome coevolution. Although few in regulator of amino acid biosynthesis in the holobiont. The gen- number, the holobiont genomes currently available are facili- erality of symbiotic regulation by substrate feedback inhibition in tating advances at the frontier of symbiosis research, a research other insect nutritional endosymbiosis requires investigation. frontier focused on elucidating the cellular and developmental Questions of developmental integration are fundamentally mechanisms of host/endosymbiont integration. important to advancing understanding of holobiont evolution generally and are particularly important for elucidating the role Elucidating the Cellular and Developmental Mechanism of of constraint in host/symbiont coevolution. Nutritional endo- Host/Endosymbiont Integration symbiont transmission has been studied best in A. pisum. Like all Compelling questions at the frontier of insect obligate nutritional aphids, A. pisum is cyclically parthenogenetic; thus, a single endosymbiosis research concern mechanisms of holobiont meta- A. pisum lineage () has the capacity to reproduce both bolic and developmental integration (45). With respect to metabolic sexually and asexually during different times of the year. When EVOLUTION integration, current research focuses on questions of regulation and aphids reproduce sexually, they are oviparous; when they re- control of nutrient biosynthesis. Whereas with respect to develop- produce asexually, they are viviparous. The ease of maintaining a mental integration, current research focuses on the cellular locali- continuous supply of asexual embryos vs. the difficulty of main- zation of novel bacteriocyte-expressed proteins. Notably, many taining a steady supply of sexually produced embryos means that studies at the frontier are characterized by application of tools that most recent work has focused on Buchnera transmission during facilitate elucidation of the mechanisms of holobiont integration. (53). That said, during both sexual and Addressing questions about holobiont metabolic and devel- asexual reproduction in A. pisum it is clear that Buchnera do not opmental integration requires appreciation of two features typical localize to the germarium at any point in development and that of most insect obligate nutritional endosymbioses. First, endo- each developing oocyte (in the case of sexual reproduction) or symbionts are vertically transmitted and have been for long evo- embryo (in the case of asexual reproduction) receive Buchnera lutionary time periods (35). Strict means that from a single maternal bacteriocyte cell (13, 53–55). The most the passage of symbionts is predictable and integral to insect de- recent model of endosymbiont transmission, which was built velopment. However, little is known currently about the cell and from examination of transmission electron micrograph images of developmental pathways responsible for nutritional endosymbiont Buchnera transmission in asexual A. pisum, demonstrates that transmission, and the developmental origin of bacteriocyte cells in Buchnera transmission occurs by a process of exo/endocytosis all taxa is unknown. The second feature typical of most obligate (53). Individual Buchnera are exocytosed from a single maternal nutritional endosymbioses is the localization of symbionts inside bacteriocyte that is closely associated with the posterior end of an host membranes, inside host bacteriocyte cells, a pattern of con- embryo at stage 7 of embryogenesis. During exocytosis, Buchnera tainment that creates many metabolically active compartments lose their host-derived symbiosomal membrane and migrate across (Fig. 1). With respect to membranes, recall that membranes are the “transmission center” as naked Buchnera cells. When they barriers to the free movement of molecules. Thus, membrane reach the passage between the enlarged follicle cells, they are barriers have the potential to regulate metabolic inputs from host endocytosed by the multinucleate syncytium within the blasto- to symbiont and metabolic outputs from endosymbiont to host. derm. During endocytosis, the naked Buchnera regain their Two recent studies have addressed the question of metabolic symbiosomal membrane and move into the central syncytium control in the A. pisum/Buchnera symbiosis. Where Buchnera’s where they await cellularization before germband formation weak transcriptional responses to changing nutritional demand (55). It is important to note that although Buchnera are closely (e.g., refs. 46–49) has resulted in the suggestion that control of associated with the germline during development, they never holobiont metabolism lies with the host (46, 50), recent work by invade the germline. This distinction is significant when consid- Hansen and Degnan (51) renews the possibility that Buchnera ering transmission of the endosymbiont Wolbachia, which infects plays an important role in regulating holobiont function. Using the germline of its host (56). Although the cell biology of Wol- label-free mass-spectrometry quantification, Hansen and Degnan bachia host interactions is particularly well characterized and has (51) found differential protein expression from Buchnera in em- benefitted from a lot of careful and elegant work in the model bryonic vs. maternal bacteriomes. These differences did not result insect Drosophila (e.g., refs. 57–60), we caution at this point from differences in Buchnera gene expression. Although they against extrapolating mechanisms of Wolbachia transmission to foundnoevidencefordifferentialBuchnera gene expression, they the transmission of bacteriome-restricted nutritional symbionts, did find over 100 small RNAs and hundreds of untranslated re- because, to the best of our knowledge, bacteriome-associated gions that were significantly expressed by Buchnera, some of which nutritional symbionts do not infect and are not associated with matched differentially expressed proteins. Based on their find- germline stem cells (13, 55, 61). That said, because in aphid ings, Hansen and Degnan (51) proposed that small RNAs me- development Buchnera are surrounded by actin filaments (55), diate regulation of gene expression at the posttranscriptional the discovery that Wolbachia makes use of the host microtubule

Wilson and Duncan PNAS | August 18, 2015 | vol. 112 | no. 33 | 10259 Downloaded by guest on September 28, 2021 network and its associated proteins to move around Drosophila attention to advances that have been made in understanding oocytes (57) suggests an exciting line of investigation for those host/symbiont metabolic and developmental integration and also interested in addressing the question of how during development to draw attention to the importance of postgenome approaches nutritional endosymbionts migrate from bacteriocytes to oocytes to hypothesis testing and exploration. Going forward, the field is (or in the peculiar case of asexual aphids, to embryos). especially ripe for the engagement of cell and developmental Two recent studies have investigated the cellular localization biologists passionate about understanding the role of symbiosis of novel bacteriocyte-expressed proteins. In the first, Shigenobu in evolution. and Stern (62) took a de novo approach to functional charac- terization of A. pisum orphan proteins, and in doing so they Conclusion identified novel secreted proteins that are overrepresented in A new era in symbiosis research arrived in 2010 with publication bacteriocyte transcriptomes. The novel secreted proteins fell into two groups that they named bacteriocyte-specific cysteine-rich of the A. pisum genome. Since that time genomics have trans- (BCR) proteins and secreted proteins (SP). Interestingly, all 7 formed our understanding of host/endosymbiont relationships. BCR genes and half of the 6 SP genes show identical patterns of Once understood as mutualisms between biologically autonomous in situ hybridization, being first expressed in stage 7 asexual entities, obligate symbiotic relationships are more accurately embryos in the syncytium, a location and developmental stage composed of interdependent parts of an intimately coevolving coincident with Buchnera infection of the blastula (55). Expres- whole, the holobiont. Despite the current limited availability of sion of all 10 of those genes was maintained in bacteriocyte cells paired genomes, these genomes support identification of signa- throughout aphid development and in adults. The role these tures of holobiont genome evolution. These signatures include novel secreted proteins play in mediating and maintaining hol- host/endosymbiont metabolic collaboration, acquisition of novel obiont function is yet to be elucidated. However, this work by genomic material, and evolutionary constraint. As more holobiont Shigenobu and Stern (62) highlights an approach that can use- genomes are sequenced, we anticipate that these signatures will fully be applied to studying developmental integration, an area of investigation that offers both promise and opportunity. continue to be supported and that other as yet unidentified sig- The second study that investigated the cellular localization of natures will likely emerge. New paired genomes will additionally novel bacteriocyte-expressed proteins investigated the protein enhance the shift of symbiosis research toward adoption of func- localization of RlpA4 (rare lipoprotein A paralog number 4), an tional genomics approaches for enhanced elucidation of the cellular A. pisum gene acquired by lateral gene transfer from an unknown and developmental mechanisms of host/endosymbiont integration. bacterial host (33, 34). Nakabachi et al. (63) found that the Thus, paired host/symbiont genomes provide the foundation for protein product of A. pisum genome-encoded RlpA4 is targeted identification of the fundamental principles of holobiont evolu- to Buchnera cells in maternal bacteriocytes, a result that has tion, advances in knowledge that are shaping our paradigm of important implications when considering the distinctions tradi- what it means to be an organism. tionally drawn between organelles and endosymbionts. For ex- ample, it is commonly argued that an important transition in ACKNOWLEDGMENTS. We acknowledge members of the A.C.C.W. laboratory organelle evolution has been the evolution of mechanisms that past and present, in particular Phillip F. Kushlan, and several undergraduate facilitate targeting of host genome-encoded proteins back to or- students for their time mining paired genomes for evidence of complemen- tarity. Hsiao-Ling Lu and Athula Wikramanayake helped with useful discus- ganelles (64, 65). Thus, the recent discovery that host genome- sions of insect development. We thank John McCutcheon, W. Ford Doolittle, encoded RlpA4 is targeted to Buchnera cells blurs the distinction and Patrick Keeling for the opportunity to participate in the Arthur M. Sackler between organelles and endosymbionts and further raises the Colloquia, “Symbioses Becoming Permanent: The Origin and Evolutionary possibility that more host-encoded genes are similarly targeted to Trajectories of Organelles.” We are grateful for the input of two anonymous reviewers; their feedback strengthened our synthesis. This work was supported endosymbionts in aphids and other holobionts. by National Science Foundation Awards IOS-1121847 and IOS-1354154 (to In highlighting recent work at the frontier of insect obligate A.C.C.W.) and DEB-1406631 (to A.C.C.W. and R.P.D.) and Graduate Research nutritional endosymbiosis research our goal has been to draw Fellowship DG1E-0951782 (to R.P.D.).

1. Van Leuven JT, Meister RC, Simon C, McCutcheon JP (2014) Sympatric in a 13. Buchner P (1965) Endosymbiosis of with Plant Microorganisms (Interscience, bacterial endosymbiont results in two genomes with the functionality of one. Cell New York). 158(6):1270–1280. 14. Wilson ACC, et al. (2010) Genomic insight into the amino acid relations of the pea 2. International Aphid Genomics Consortium (2010) Genome sequence of the pea aphid aphid, Acyrthosiphon pisum, with its symbiotic bacterium Buchnera aphidicola. Insect Acyrthosiphon pisum. PLoS Biol 8(2):e1000313. Mol Biol 19(Suppl 2):249–258. 3. Shigenobu S, Watanabe H, Hattori M, Sakaki Y, Ishikawa H (2000) Genome sequence 15. Hansen AK, Moran NA (2011) Aphid genome expression reveals host-symbiont co- of the endocellular bacterial symbiont of aphids Buchnera sp. APS. Nature 407(6800): operation in the production of amino acids. Proc Natl Acad Sci USA 108(7):2849–2854. 81–86. 16. Russell CW, Bouvaine S, Newell PD, Douglas AE (2013) Shared metabolic pathways in a 4. Kirkness EF, et al. (2010) Genome sequences of the human body louse and its primary coevolved insect-bacterial symbiosis. Appl Environ Microbiol 79(19):6117–6123. endosymbiont provide insights into the permanent parasitic lifestyle. Proc Natl Acad 17. Hansen AK, Moran NA (2014) The impact of microbial symbionts on host plant uti- Sci USA 107(27):12168–12173. lization by herbivorous insects. Mol Ecol 23(6):1473–1496. 5. Bonasio R, et al. (2010) Genomic comparison of the Camponotus floridanus and 18. Price DRG, Wilson ACC (2014) A substrate ambiguous enzyme facilitates genome Harpegnathos saltator. Science 329(5995):1068–1071. reduction in an intracellular symbiont. BMC Biol 12(1):110. 6. Gil R, et al. (2003) The genome sequence of Blochmannia floridanus: Comparative 19. Wu D, et al. (2006) Metabolic complementarity and genomics of the dual bacterial analysis of reduced genomes. Proc Natl Acad Sci USA 100(16):9388–9393. symbiosis of sharpshooters. PLoS Biol 4(6):e188. 7. McCutcheon JP, von Dohlen CD (2011) An interdependent metabolic patchwork in 20. McCutcheon JP, McDonald BR, Moran NA (2009) Convergent evolution of metabolic the nested symbiosis of mealybugs. Curr Biol 21(16):1366–1372. roles in bacterial co-symbionts of insects. Proc Natl Acad Sci USA 106(36):15394–15399. 8. Husnik F, et al. (2013) from diverse bacteria to an insect 21. Bennett GM, Moran NA (2013) Small, smaller, smallest: the origins and evolution of genome enables a tripartite nested mealybug symbiosis. Cell 153(7):1567–1578. ancient dual symbioses in a Phloem-feeding insect. Genome Biol Evol 5(9):1675–1688. 9. Sloan DB, et al. (2014) Parallel histories of horizontal gene transfer facilitated extreme 22. Rasnitsyn AP, Quicke DLJ (2002) History of Insects (Kluwer, Dordrecht, The reduction of endosymbiont genomes in sap-feeding insects. Mol Biol Evol 31(4):857–871. Netherlands). 10. Watanabe J, et al.; International Glossina Genome Initiative (2014) Genome sequence 23. Thao ML, et al. (2000) Cospeciation of psyllids and their primary prokaryotic endo- of the tsetse fly (Glossina morsitans): Vector of African trypanosomiasis. Science symbionts. Appl Environ Microbiol 66(7):2898–2905. 344(6182):380–386. 24. Moran NA, Munson MA, Baumann P, Ishikawa H (1993) A molecular clock in endo- 11. Akman L, et al. (2002) Genome sequence of the endocellular obligate symbiont of symbiotic bacteria is calibrated using the insect hosts. Proc R Soc B Biol Sci 253(1337): tsetse flies, Wigglesworthia glossinidia. Nat Genet 32(3):402–407. 167–171. 12. Rio RVM, et al. (2012) Insight into the transmission biology and species-specific func- 25. Moran NA, Tran P, Gerardo NM (2005) Symbiosis and insect diversification: An ancient tional capabilities of tsetse (Diptera: glossinidae) obligate symbiont Wigglesworthia. symbiont of sap-feeding insects from the bacterial phylum Bacteroidetes. Appl En- MBio 3(1):e00240-11. viron Microbiol 71(12):8802–8810.

10260 | www.pnas.org/cgi/doi/10.1073/pnas.1423305112 Wilson and Duncan Downloaded by guest on September 28, 2021 PAPER

26. Näsvall J, Sun L, Roth JR, Andersson DI (2012) Real-time evolution of new genes by 47. Moran NA, Dunbar HE, Wilcox JL (2005) Regulation of transcription in a reduced COLLOQUIUM innovation, amplification, and divergence. Science 338(6105):384–387. bacterial genome: Nutrient-provisioning genes of the obligate symbiont Buchnera 27. Deng C, Cheng CHC, Ye H, He X, Chen L (2010) Evolution of an antifreeze protein by aphidicola. J Bacteriol 187(12):4229–4237. neofunctionalization under escape from adaptive conflict. Proc Natl Acad Sci USA 48. Reymond N, et al. (2006) Different levels of transcriptional regulation due to trophic 107(50):21593–21598. constraints in the reduced genome of Buchnera aphidicola APS. Appl Environ Mi- 28. Force A, et al. (1999) Preservation of duplicate genes by complementary, de- crobiol 72(12):7760–7766. generative mutations. 151(4):1531–1545. 49. Bermingham J, et al. (2009) Impact of host developmental age on the transcriptome 29. Duncan RP, Nathanson L, Wilson ACC (2011) Novel male-biased expression in paralogs of the symbiotic bacterium Buchnera aphidicola in the pea aphid (Acyrthosiphon of the aphid slimfast nutrient amino acid transporter expansion. BMC Evolutionary pisum). Appl Environ Microbiol 75(22):7294–7297. Biology 11:253. 50. Macdonald SJ, Lin GG, Russell CW, Thomas GH, Douglas AE (2012) The central role of – 30. Price DRG, Duncan RP, Shigenobu S, Wilson ACC (2011) Genome expansion and dif- the host cell in symbiotic nitrogen metabolism. Proc Biol Sci 279(1740):2965 2973. ferential expression of amino acid transporters at the aphid/Buchnera symbiotic in- 51. Hansen AK, Degnan PH (2014) Widespread expression of conserved small RNAs in – terface. Mol Biol Evol 28(11):3113–3126. small symbiont genomes. ISME J 8(12):2490 2502. 31. Duncan RP, et al. (2014) Dynamic recruitment of amino acid transporters to the insect/ 52. Price DRG, et al. (2014) Aphid amino acid transporter regulates glutamine supply to – symbiont interface. Mol Ecol 23(6):1608–1623. intracellular bacterial symbionts. Proc Natl Acad Sci USA 111(1):320 325. 32. Dahan RA, Duncan RP, Wilson ACC, Dávalos LM (2015) Amino acid transporter ex- 53. Koga R, Meng XY, Tsuchida T, Fukatsu T (2012) Cellular mechanism for selective pansions associated with the evolution of obligate endosymbiosis in sap-feeding in- vertical transmission of an obligate insect symbiont at the bacteriocyte-embryo in- terface. Proc Natl Acad Sci USA 109(20):E1230–E1237. sects (: sternorrhyncha). BMC Evol Biol 15:52. 54. Wilkinson TL, Fukatsu T, Ishikawa H (2003) Transmission of symbiotic bacteria Buch- 33. Nikoh N, Nakabachi A (2009) Aphids acquired symbiotic genes via lateral gene nera to parthenogenetic embryos in the aphid Acyrthosiphon pisum (Hemiptera: transfer. BMC Biol 7:12. Aphidoidea). Struct Dev 32(2-3):241–245. 34. Nikoh N, et al. (2010) Bacterial genes in the aphid genome: Absence of functional 55. Miura T, et al. (2003) A comparison of parthenogenetic and sexual embryogenesis of gene transfer from Buchnera to its host. PLoS Genet 6(2):e1000827. the pea aphid Acyrthosiphon pisum (Hemiptera: Aphidoidea). J Exp Zoolog B Mol Dev 35. Bennett GM, Moran NA (2015) Heritable symbiosis: The advantages and perils of an Evol 295(1):59–81. evolutionary rabbit hole. Proc Natl Acad Sci USA 112:10169–10176. 56. Serbus LR, Casper-Lindley C, Landmann F, Sullivan W (2008) The genetics and cell 36. Moran NA, McCutcheon JP, Nakabachi A (2008) Genomics and evolution of heritable biology of Wolbachia-host interactions. Annu Rev Genet 42:683–707. bacterial symbionts. Annu Rev Genet 42:165–190. 57. Ferree PM, et al. (2005) Wolbachia utilizes host microtubules and Dynein for anterior 37. McCutcheon JP, Moran NA (2012) Extreme genome reduction in symbiotic bacteria. localization in the Drosophila oocyte. PLoS Pathog 1(2):e14. Nat Rev Microbiol 10(1):13–26. 58. Frydman HM, Li JM, Robson DN, Wieschaus E (2006) Somatic stem cell niche tropism in 38. Muller HJ (1964) The relationship of recombination to mutational advance. Mutat Res Wolbachia. Nature 441(7092):509–512. 106:2–9. 59. Fast EM, et al. (2011) Wolbachia enhance Drosophila stem cell proliferation and 39. Fukatsu T, Ishikawa H (1992) A novel eukaryotic extracellular symbiont in an aphid, target the germline stem cell niche. Science 334(6058):990–992. Astegopteryx styraci (Homoptera, , Hormaphidinae). J Insect Physiol 38(10): 60. Toomey ME, Panaram K, Fast EM, Beatty C, Frydman HM (2013) Evolutionarily con- – 765 773. served Wolbachia-encoded factors control pattern of stem-cell niche tropism in EVOLUTION 40. Gruwell ME, Hardy NB, Gullan PJ, Dittmar K (2010) Evolutionary relationships among Drosophila ovaries and favor infection. Proc Natl Acad Sci USA 110(26):10788–10793. primary endosymbionts of the mealybug subfamily phenacoccinae (hemiptera: Coc- 61. von Dohlen CD, Kohler S, Alsop ST, McManus WR (2001) Mealybug β-proteobacterial – coidea: Pseudococcidae). Appl Environ Microbiol 76(22):7521 7525. endosymbionts contain gamma-proteobacterial symbionts. Nature 412(6845):433–436. 41. Moran NA (2007) Symbiosis as an adaptive process and source of phenotypic com- 62. Shigenobu S, Stern DL (2013) Aphids evolved novel secreted proteins for symbiosis – plexity. Proc Natl Acad Sci USA 104(Suppl 1):8627 8633. with bacterial endosymbiont. Proc Biol Sci 280(1750):20121952. 42. Pérez-Brocal V, et al. (2006) A small microbial genome: The end of a long symbiotic 63. Nakabachi A, Ishida K, Hongoh Y, Ohkuma M, Miyagishima SY (2014) Aphid gene of relationship? Science 314(5797):312–313. bacterial origin encodes a protein transported to an obligate endosymbiont. Curr Biol 43. Lamelas A, et al. (2011) Serratia symbiotica from the aphid Cinara cedri: A missing link 24(14):R640–R641. from facultative to obligate insect endosymbiont. PLoS Genet 7(11):e1002357. 64. Keeling PJ, Archibald JM (2008) Organelle evolution: What’s in a name? Curr Biol 44. Kanehisa M, Goto S (2000) KEGG: Kyoto encyclopedia of genes and genomes. Nucleic 18(8):R345–R347. Acids Res 28(1):27–30. 65. McCutcheon JP, Keeling PJ (2014) Endosymbiosis: Protein targeting further erodes the 45. Shigenobu S, Wilson ACC (2011) Genomic revelations of a mutualism: The pea aphid organelle/symbiont distinction. Curr Biol 24(14):R654–R655. and its obligate bacterial symbiont. Cell Mol Life Sci 68(8):1297–1309. 66. Misof B, et al. (2014) Phylogenomics resolves the timing and pattern of insect evo- 46. Moran NA, Plague GR, Sandström JP, Wilcox JL (2003) A genomic perspective on lution. Science 346(6210):763–767. nutrient provisioning by bacterial symbionts of insects. Proc Natl Acad Sci USA 100(Suppl 67. Cryan JR, Urban JM (2012) Higher-level phylogeny of the insect order Hemiptera: Is 2):14543–14548. Auchenorrhyncha really paraphyletic? Syst Entomol 37(1):7–21.

Wilson and Duncan PNAS | August 18, 2015 | vol. 112 | no. 33 | 10261 Downloaded by guest on September 28, 2021