Functional Genomics of Buchnera and the Ecology of Aphid Hosts

Molecular Ecology (2006) 15, 1251–1261 doi: 10.1111/j.1365-294X.2005.02744.x

BlackwellFunctional Publishing, Ltd. genomics of Buchnera and the ecology of aphid hosts

NANCY A. MORAN and PATRICK H. DEGNAN Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721, USA

Abstract In many animal groups, mutualistic bacterial symbionts play a central role in host ecology, by provisioning rare nutrients and thus enabling specialization on restricted diets. Among such symbionts, genomic studies are most advanced for Buchnera, the obligate symbiont of aphids, which feed on phloem sap. The contents of the highly reduced Buchnera genomes have verified its role in aphid nutrition. Comparisons of Buchnera gene sets indicate ongoing, irreversible gene losses that are expected to affect aphid nutritional needs. Furthermore, almost all regulatory genes have been eliminated, raising the question of whether and how gene expression responds to environmental change. Microarray studies on genome-wide expression indicate that Buchnera has evolved some constitutive changes in gene expression: homologues of heat stress genes have elevated transcript levels in Buchnera (relative to other bacteria) even in the absence of stress. Additionally, the microarray results indicate that responses to heat stress and to amino acid availability are both few and modest. Observed responses are consistent with control by the few ancestral regulators retained in the genome. Initial studies on the role of host genes in mediating the symbiosis reveal distinctive expression patterns in host cells harbouring Buchnera. In the near future, a complete genome of pea aphid will accelerate progress in understanding the functional integration of aphid and Buchnera genomes. Although information for other insect symbioses is relatively limited, studies on symbionts of carpenter ants and tsetse flies indicate many similarities to Buchnera. Keywords: bacteria, bacteriocyte, insects, nutrition, phloem-feeding, symbiosis Received 7 June 2005; revision accepted 10 August 2005

symbionts, especially for insect symbioses (e.g. Douglas Historical background & Prosser 1992; Sasaki & Ishikawa 1995; Febvay et al. 1999). Many invertebrates possess endosymbiotic bacteria that The use of molecular phylogenetic data has given robust are vertically transmitted from mother to progeny and confirmation to Buchner’s general view that associations that live in specialized organs. In his overview of animal between insects and their primary symbionts extend into symbioses, Buchner (1965) referred to these as ‘primary’ the deep evolutionary past, dating to the origins of major symbionts and proposed that their main function was groups such as aphids, psyllids, leafhoppers, carpenter ants, to provision nutrients that were limiting in the diet of their and others. Repeatedly, such studies have revealed con- hosts. In insects, such nutritional limitations are expected gruence between phylogenies of insect host and primary for a number of groups that have evolved specialized feed- symbiont, corroborating that symbionts were important ing habits, such as use of plant sap or vertebrate blood as players in the diversification of insect groups that are the sole food. A variety of experimental studies, involving dominant elements in modern terrestrial ecosystems. Thus, removal of native symbionts and/or manipulation of symbiosis was a prerequisite for the colonization of many diets have provided support for nutrient provisioning by specialized ecological niches and for the evolution of groups such as the sap-feeding and blood-feeding insects. In recent years, the emergence of genomics has expanded Correspondence: Nancy A. Moran, Fax: 520-621-3709; our understanding of symbiotic associations of insects and E-mail: [email protected] bacteria. Because bacterial genomes are compact and rich

1252 N. A. MORAN and P . H . DEGNAN in information regarding metabolic capabilities, the sequen- selective retention of particular biosynthetic genes demon- cing of complete genomes is relatively cost-effective. By strated a striking complementarity between host and providing a complete catalogue of genes and pathways symbiont metabolic capabilities. present in organisms, the sequences themselves give a direct Bacterial genomes are tightly packed with genes, indi- means of assessing functional contributions of symbionts cating that superfluous DNA is rapidly eliminated during to hosts. Genome sequences provide the foundation for the evolution of bacterial lineages (Mira et al. 2001; Lerat et al. analyses of the transcriptome or proteome and provide 2005). In symbiont genomes (and those of small genome the potential for assessing how symbiont functions are pathogens), the loss of DNA is extreme, implying strongly regulated in the context of the host environment. that genes that are retained with intact reading frames are The focus of this review is the best-studied insect– functionally significant to the organism. Other indicators symbiont system, that of Buchnera aphidicola and its are consistent with this inference: for example, different aphid hosts. Three genome sequences and several studies genes of a particular biosynthetic pathway are typically all of genome-wide gene expression have been published. We lost or all retained, even when the genes are positioned in summarize these results and relate them to some earlier, different parts of the chromosome. pre-genomics observations on aphid feeding and ecology. Remarkably, the Buchnera gene set is a subset of that of Together, these results suggest a role of Buchnera in both E. coli and closely related bacteria. This fact suggests that, the expansion of aphid ranges and in constraints on their at least in the case of Buchnera, becoming symbiotic did not geographical and host-plant distributions. Additionally, depend on acquiring novel genes. However, the Buchnera we compare results from Buchnera genomes to those from genome is highly derived and highly reduced, making it two other ancient bacteriocyte-associated insect symbionts impossible to determine whether some genes enabling for which genome sequences are available. host invasion were involved in early stages of the symbiosis. Subsequent co-adaptation with the host genome may have rendered symbiont invasion genes obsolete. Some symbionts Gene content and genome sizes of symbionts showing younger associations with insects, exemplified The first studies of symbiont gene content related to by Sodalis of tsetse flies, primary symbionts of Sitophilus function in insect hosts were those of Baumann and weevils, and secondary symbionts of aphids, possess Type coworkers, who cloned and sequenced fragments of the III Secretion Systems (Dale et al. 2001, 2002, 2005; Moran Buchnera chromosome from the aphid host Schizaphis et al. in press), similar to those used by pathogenic bacteria graminum (Baumann et al. 1995). They documented the to invade host cells. The underlying genes are known to presence of pathways for several essential amino acids, be horizontally transferred in many cases, and their acqui- confirming and extending previous evidence for a role of sition could be a step in the original evolution of an intra- Buchnera in provisioning hosts with these nutrients. Most cellular lifestyle. In Buchnera and other specialized symbionts, aphids feed exclusively on phloem sap, which contains early utilization of such a system might be obscured by limited amounts of essential amino acids; that is, the amino subsequent evolution of host mechanisms for ensuring acids used in proteins that animals cannot synthesize infection of bacteriocytes (the cells housing the symbionts) themselves. These pathways are very well studied in and by ongoing reduction of the symbiont genome. bacteria, particularly in the model system of Escherichia coli. The extent to which genome contents of insect symbionts The close relationship between Buchnera and E. coli, and the can be extended to infer metabolic capabilities is illustrated clearly defined homology of their genes, was a major by the detailed analysis of Zientz et al. (2004), which advantage in interpreting the new sequence data from elucidates the metabolic implications of the genomes of Buchnera. Buchnera, Blochmannia and Wigglesworthia. By the time the first Buchnera genome was published for the pea aphid Acyrthosiphon pisum (Shigenobu et al. 2000), Polyploidy it was known that Buchnera possessed genes for biosynthesis of most essential amino acids and that several of Many symbionts, including Buchnera, have very large these genes (for synthesis of tryptophan, isoleucine, valine, cell size: for example, the roughly spherical Buchnera cells leucine, and histidine) were transcribed (Baumann et al. 1995, are 3 µ in diameter and thus about 15× the volume of 1999). Complete genome sequences are the only means of an E. coli cell. In Buchnera at least, this increase in volume confidently showing what genes are absent from a genome, is accompanied by amplification of the genome. In and the first Buchnera genome sequence was perhaps most Acyrthosiphon pisum, individual Buchnera cells contain from notable for revealing which genes had been eliminated. 50 to over 100 chromosome copies, with the degree of These included most of those required for the synthesis of amplification varying with life cycle stage (Komaki & the nonessential amino acids, i.e. the amino acids that Ishikawa 1999, 2000). The functional significance of the animal, such as the aphid hosts, can make themselves. This polyploidy is not clear: it may allow higher productivity

FUNCTIONAL GENOMICS OF BUCHNERA 1253 per cell and it is presumably linked to the regulation of cing efforts were detailed enough to detect them. Thus, the symbiont numbers within bacteriocytes, since it indicates extent of amplification of functionally relevant (intact and a change in the coordination between chromosome replication transcribed) copies is not easily determined. This obscures and cell division. the meaning of studies of trpEG copy number in which numbers are determined by hybridization to a probe or by quantitative PCR, without extensive cloning (e.g. Birkle Plasmid location of nutrient biosynthetic genes et al. 2002, 2004). A remarkable early discovery was the finding that trpEG, The number of copies of the plasmid-borne genes rela- the genes encoding the enzyme for the rate-limiting step of tive to chromosomal ones tends to be constant within a tryptophan biosynthesis, were encoded as tandem repeats species but can differ considerably between species, rang- on a plasmid and present in multiple copies relative to ing from less than 1 (e.g. Plague et al. 2003) to 24 (Thao et al. chromosomal genes in the Buchnera of many aphid species 1998). Some intraspecific variation has been identified as (Lai et al. 1994). Soon after, the genes for leucine biosynthesis well. Furthermore, the movement of trpEG and leuABCD to were shown to occur on a different type of plasmid (Bracho a plasmid location occurred only once or very few times in et al. 1995). The two plasmids show different replication evolution (Sabater-Muñoz et al. 2004), implying that the mechanisms and originated independently. The plasmid selection for plasmid location occurred in a common ances- bearing the trpEG genes consist of a tandemly repeated tor of a very large radiation of aphids that includes most unit of about 3.6 kb, with each unit containing an origin of modern species and a wide variety of feeding habits. replication containing binding sites for DnaA protein, Persistence of the plasmid arrangement does not indicate similar to that on the chromosome of Buchnera and other that it remains universally advantageous, but only that related bacteria; the plasmid bearing the leu genes has its most rearrangements are not tolerated in Buchnera genomes. own replication proteins and is related to incompatibility This is suggested by the fact that Buchnera chromosomes plasmids in Salmonella and other bacteria. The amino acid are extremely static, showing a lower incidence of re- biosynthetic genes occur on plasmids only in some Buchnera arrangements relative to sequence divergence, than any other lineages. Plasmid-borne copies were acquired from the bacterial group (Tamas et al. 2002; van Ham et al. 2003). ancestral Buchnera chromosome and not from external Thus, once the genes were transferred to the plasmid, they sources (as for some plasmid-borne genes in bacteria) may persist there even if the original benefit of the plasmid (Rouhbakhsh et al. 1996). location is removed. Overproduction and export of newly synthesized amino acids is a process distinctive of symbionts. Thus, Buchnera Loss of genes and pathways is highly unusual among bacteria in needing to overproduce tryptophan and leucine and is also unique in the Although gene order and orientation have been stable for amplification of the underlying genes and their location on more than 100 million years (Myr) of Buchnera evolution, plasmids. Presumably the two features are related: the comparison of the three available genome sequences unusual genomic arrangements most likely evolved as a reveals ongoing deletions and inactivation of ancestral means of increasing or regulating the expression of genes genes in each lineage (Tamas et al. 2002; van Ham et al. underlying these biosynthetic processes, to enable better 2003). Furthermore, some of the observed gene losses are provisioning of the end products to the hosts (Lai et al. expected to impact host ecology. For example, Buchnera of 1994). Schizaphis graminum possesses degenerate copies of the The initial interpretations of the plasmid position of genes underlying incorporation of inorganic sulphur; these biosynthetic genes has been complicated by several the coding regions for several of the required enzymes are additional findings. First, as mentioned, the Buchnera chro- disrupted by stop codons and deletions, suggesting recent mosome itself is present as many copies per cell, so plasmid- inactivation (Tamas et al. 2002). Because grasses, the host borne genes could be present in either more or fewer copies plants of S. graminum, contain high amounts of sulphur- relative to single copy chromosomal genes. Both situations containing organic compounds, they are likely to ingest have been observed (Thao et al. 1998; Plague et al. 2003). homocysteine or related sulphur-containing intermediates, Second, some of the genes on the plasmid are inactivated, enabling the production of methionine and cysteine without sometimes highly truncated and sometimes only carrying incorporation of inorganic sulphur. Similar losses have small frame shifts and differing only slightly from intact occurred in two of the sequenced genomes; e.g. Buchnera copies (Lai et al. 1996; Wernegreen & Moran 2000). The latter of Baizongia lacks the pathway for arginine and pantothenate are discernable only through cloning and individual biosynthesis, whereas it retains genes underlying biotin sequencing of many copies from individual insects. In fact, biosynthesis that are eliminated from genomes of the trpEG pseudogenes, coexisting with intact copies, have Buchnera of both S. graminum and Acyrthosiphon pisum (van been found in every species in which cloning and sequen- Ham et al. 2003).

1254 N. A. MORAN and P . H . DEGNAN

Such inactivations are the likely explanation for strain- most or all of the essential amino acids implies that losses specific dietary requirements in aphids. A review of aphid of particular pathways are deleterious in the longer run nutritional physiology (Srivastava 1987) summarizes and generally lead to lineage extinction. earlier studies of dietary requirements for amino acids and concludes that aphids ‘appear to constitute a rare group Transfer of bacterial genes to host genome? where different insect populations within one species differ in their requirement for essential amino acids.’ For Full genome sequences of animals and plants have example, certain A. pisum biotypes (representing individual produced recent revelations regarding the large extent of clonal isolates with distinctive host plant use) have dietary gene transfer from ancestral organelles to the nuclear requirements for particular essential amino acids, includ- genome (Martin 2003). Buchnera resembles organelles in ing arginine, phenylalanine, tryptophan and valine, based many ways, including ancient derivation from a free-living on studies of aphids with intact symbionts (Srivastava bacterium, obligate residence within the host cytoplasm, et al. 1985; Srivastava 1987). The ability of sulphate to and large-scale genome reduction, raising the possibility spare the requirement for sulphur-containing amino acids that Buchnera has also transferred genes to the host was another variable in dietary requirements of A. pisum genome. But, in contrast to mitochondria and chloroplasts, genotypes (Srivastava 1987). Additionally, Aphis fabae it is not apparent that Buchnera must import gene clones grown on defined diets differ in their amino acid products from its host for the processes underlying its own requirements, with individual clones (with symbionts replication; the gene inventories from the completed intact) requiring histidine, methionine, threonine, or valine genomes indicate that it retains its own polymerases for optimal growth (Wilkinson & Douglas 2003). These and other basic cellular machinery. In fact, aside from observations strongly suggest that Buchnera inhabiting the retention of nutrition-provisioning genes, the genome particular aphid lineages differ in biosynthetic capabilities reduction in Buchnera is similar to that observed in obligate due to mutational inactivation of particular pathways. pathogens with small genomes, such as mycoplasmas and Because Buchnera genomes show no cases of gene uptake, rickettsiae, organisms that are horizontally transferred such loss of function is almost certainly irreversible, at least among hosts and that are generally deleterious. All of these once a pathway is inactivated by more than one mutation. organisms are parasites in the sense of requiring host An implication is that Buchnera acts as a potential constraint metabolic products, but the genome reduction of the on the long-term ecological diversification of aphid hosts. pathogens implies that extensive gene loss can occur For example, loss of the pathway for sulphur fixation without dependence on products of genes transferred implies that S. graminum is now destined to remain on grasses, from symbionts to the host genome. (Selection on hosts or other hosts with sufficient sources of fixed sulphur, for would not maintain genes for the benefit of pathogens.) its entire evolutionary future (Tamas et al. 2002). The extent One reason that Buchnera and similar mutualistic symbi- of such inactivations as a basis for polymorphism within onts of animals might differ from organelles in the extent a species is not clear from the genomic sequences. For of gene transfer to the host germ line is that the obstacles to S. graminum and A. pisum, the Buchnera genomes were such transfer are more formidable. Whereas mitochondria sequenced for a single aphid genotype. For Buchnera of and chloroplasts originated in single-celled ancestors, Baizongia pistaciae, a population sample was used, and the enabling close proximity between their genomes and host resulting genomic sequence revealed five polymorphic sites nuclear genomes, Buchnera is confined to non-germ-line that may represent snapshots of recent gene inactivations. cells except for brief stages in the life cycle following egg Thus, several lines of evidence indicate that Buchnera infection. Despite these caveats, gene transfer cannot be genome evolution results in the ongoing loss of capabilities excluded as a possibility. Indeed, transfer to the host for provisioning hosts and never in the acquisition of new nucleus of a chromosomal segment of a Wolbachia sym- capabilities. It may appear puzzling that symbionts lose biont appears to have occurred in a beetle (Kondo et al. 2002), capabilities that might benefit their mutualistic hosts. indicating that such events are possible. The sequencing of Because plants, and plant parts, vary in phloem content the Acyrthosiphon pisum genome, expected within the near of amino acids (Sandström & Moran 1999; Koyama et al. future, will largely answer this question, at least for the 2004), aphid colonies living on a particular plant may case of Buchnera. experience local relaxation of selection for retention of the corresponding Buchnera pathway. Furthermore, the muta- Gene expression tional target resulting in inactivation of the pathway can be quite large, especially for the longer pathways requiring Reduced bacterial genomes typically possess very more enzymes, increasing the likelihood of loss during a few regulatory genes, and this generalization extends to particular period of relaxed selection. The fact that Buchnera Buchnera, which retains only about five of the more than of major aphid lineages retain capacity for biosynthesis of 200 regulatory genes found in E. coli. Additionally, the

Fig. 1 Response of transcriptome of Buchnera of Schizaphis graminum to severe heat stress, from Wilcox et al. (2003). (A) The average fold- change induced in Buchnera (dark grey) is much less than that induced by heat stress in the set of homologous genes in Escherichia coli (light grey); only a few Buchnera genes show significant increases. Variance in gene expression values is significantly lower in Buchnera than in E. coli (**P < 0.00001). (B) The normalized distribution of transcript abundance under non-stress (dark grey) and heat stress (light grey) for Buchnera genes. Most of the homologues of heat shock genes show high expression even under non-stress conditions (shaded tail indicates > 90th percentile; see list in box). Only a few of these show significant increases in response to heat stress (gene names in bold in second box).

organization of transcription units is highly disrupted be GroEL protein (orthologue of HSP60), a chaperonin that by the loss of genes and by rearrangements (Moran & is universally distributed and that functions in the folding Mira 2001). Finally, the distinctive biology of symbionts, of new and destabilized proteins (Ohtaka et al. 1992). This which live in a relatively constant environment and which protein comprises about 10% of Buchnera protein under overproduce some metabolic products for hosts, is expected non-stress conditions (Baumann et al. 1996), a level equivalent to impose distinctive selection pressures on gene regulation. to that in E. coli under the most elevated state, as induced In most bacteria, such as the well-characterized E. coli, genes by heat or other stresses that cause proteins to destabilize. are transcribed in discrete transcription units (operons) Furthermore, full genome microarrays have revealed that with a high level of independence of expression, choreo- most of the 20 homologues of heat shock genes retained in graphed by a large set of regulatory elements, including the Buchnera genome are overexpressed constitutively both proteins and RNAs. In contrast, Buchnera does not compared to their levels in other bacteria (Moran et al. 2003; show clearly defined transcriptional terminators, and the Wilcox et al. 2003; Fig. 1). Together, these gene products are binding sites for the generalized sigma-70 promoter, involved in recycling and refolding proteins that have required for initiation of transcription for most genes, are become degraded. The overexpression of this set of genes, not readily distinguished. which are dispersed on the Buchnera chromosome, indicates Although the inability to culture Buchnera makes genetic a general compensation for problems with protein stability manipulations impossible, analysis of gene expression (Wilcox et al. 2003). Because Buchnera shows accumulation patterns is possible, using RNA extracted from aphids and of mutations effecting amino acid changes throughout the symbionts. The full genome sequences for Buchnera have genome, due to relaxed selection or ineffective selection allowed the examination of global gene expression levels resulting from small genetic population size, this constitu- using gene microarrays. Such studies, summarized below, tive overexpression is best explained as a compensatory can reveal the extent to which Buchnera responds to ecolo- adaptation to mutational destabilization of proteins gically realistic environmental change. (Moran 1996). Experiments in E. coli and other organisms demonstrate that overexpression of GroEL and other chaperonins can mask mutations that would otherwise be Changes in constitutive expression — heat shock deleterious (e.g. Rutherford & Lindquist 1998; Fares et al. 2002). genes An early puzzle in Buchnera biology was the extremely Reduction in regulatory mechanisms controlling high levels of a protein that was initially named transcription ‘symbionin’ and hypothesized to play a central role in the aphid–bacterial interaction (Ishikawa 1982). Later, once As noted above, an observation applying to all highly sequencing was available, symbionin was determined to reduced bacterial genomes and to the Buchnera genomes

For considerations of host nutritional ecology, the expression of Buchnera’s amino acid biosynthetic genes is of particular interest. Amino acid profiles in phloem differ among plant species and also among growth stages of individual plants (e.g. Sandström & Moran 1999; Koyama et al. 2004), implying fluctuations in the need for different biosynthetic contributions by Buchnera. In E. coli and related bacteria, these pathways have a total of 16 different specific transcriptional regulators. Buchnera of S. graminum retains only one of these, metR, which is a regulator of metE, underlying methionine synthesis; in Buchnera of Acyrthosiphon- pisum, metR is inactivated through several frame shifts and a large deletion. Experiments manipulating amino acid levels in the plants and aphids showed that transcript levels of these genes lack responses that are statistically significant and substantial (greater than twofold), with the striking exception of metE (Fig. 2; Moran et al. 2005). The metE transcript was significantly up-regulated with the addition of several nonessential amino acids, which may Fig. 2 Relative transcript levels in Buchnera of Schizaphis graminum be equivalent for Buchnera physiology due to interconver- in response to manipulation of individual amino acids in aphid hosts. Values are from microarrays (Moran et al. 2005). The genes sions in aphid tissues (Moran et al. 2005). Experiments that are negatively regulated in Escherichia coli as a result of adding using reverse transcriptase quantitative PCR confirmed the end product amino acid are shown in bold type. The only that metE expression is most sensitive to levels of homo- substantial response is limited to metE which shows higher cysteine, demonstrated to be a co-activator of metR in transcript levels in response to added arginine. *Similar response E. coli. Furthermore, the response was completely lacking is seen to glutamic acid and glutamine. in Buchnera of A. pisum, in which metR is degraded. Thus, Buchnera has lost almost all mechanisms for regulating transcript levels of genes underlying amino acid in particular, is the extensive loss of regulatory genes. biosynthesis in response to amino acid pools. Moreover, it Inspection of the genome sequence does not reveal any has not invented new ones during the course of its evolu- obvious new regulatory mechanisms, but these would tion in aphid hosts, at least none with substantial effects likely not be recognizable and might be expected in view of on transcript levels. In Buchnera of S. graminum, the only the fact that Buchnera has co-evolved with hosts for over strong response to amino acid fluctuations is dependent on 100 Myr. the single retained regulator (metR), and even that system Full genome microarrays for Buchnera, consisting of is missing from both of the other sequenced Buchnera individually amplified double-stranded DNA sequences, genomes (van Ham et al. 2003). allow direct evaluation of global transcriptional responses The few regulatory genes that Buchnera retains may to environmental challenges. To date, these studies have control shifts in gene expression corresponding to differ- been carried out for two kinds of changes that routinely ent stages of its life cycle within hosts. A period of rapid affect naturally occurring aphids and their Buchnera: heat division and distinctive gene expression appears to follow stress and fluctuations in the composition of dietary amino colonization of a new host embryo (Wilkinson et al. 2003), acids. Both studies indicate that transcriptional responses and later growth of the Buchnera population closely tracks are few and consist only of ancestral responses that have growth of the host (Baumann & Baumann 1994). These been retained. observations indicate that distinct phases do occur in the Responses to heat shock indicate that, in addition to Buchnera life cycle, and these are expected to correspond to overexpression under non-stress conditions, Buchnera distinct phases of gene expression. Homologues of himA of Schizaphis graminum does retain a modest heat shock (encoding Integration Host Factor) are intact in all three response involving a global heat shock regulator and four Buchnera genomes; in other bacteria, this gene is involved promoters affecting expression of six genes (Fig. 1, Wilcox in detecting changes in the cellular environment and et al. 2003). The regulatory mechanisms appear to be the switching gene expression accordingly. In Buchnera, himA same as in E. coli, and the consensus heat shock promoter may be involved in the switch between life cycle stages, is the same. As mentioned, most of the heat shock gene such as replicative stages and vegetative stages, during homologues have lost heat shock promoters and are which cells are metabolizing and producing nutrients for overexpressed constitutively. hosts but not dividing.

The overall picture of gene expression is that it is rela- Other up-regulated genes included homologues of genes tively static, compared to free-living bacteria. Additionally, encoding transporters and lysozymes that act to degrade transcription units are less clearly defined. For example, bacterial cell walls. These findings may be relevant to heat shock increases transcription of heat shock genes as outstanding mysteries regarding Buchnera function: how well as downstream sequences, even when the downstream small molecules are moved between host and symbiont, as genes are on the complementary strand (Wilcox et al. required for the nutritional exchange which has been 2003). demonstrated in physiological studies, and how symbiont numbers are regulated. The Buchnera genome lacks most homologues of genes encoding transporters, including Population biology of nutrient-provisioning genes ones used by related bacteria to move amino acids across Analyses of Buchnera genomes and gene sequences are cell walls. Gene expression profiles obtained for bacteriocytes consistent with the occurrence of elevated levels of genetic raises the possibility that host transporters interact with drift (Moran 1996; van Ham et al. 2003; Rispe et al. 2004). In bacterial membranes so as to effect movement of amino particular, sequence analyses consistently indicate rapid acids or other metabolites. sequence evolution. However, polymorphism in Buchnera Recent studies have begun to shed light on the gene genes within host species is extremely low, as is expression underlying specification of the bacteriocytes polymorphism in aphid nuclear genes. Low intraspecific during early development (Braendle et al. 2003). These polymorphism reflects the typical population structure of data indicate that the Buchnera inoculum is a major feature aphids, which are clonal during most of their life cycle, of early embryogenesis. Bacteriocytes are determined and highly mobile and subject to large-scale bottlenecks within colonized at an early stage, and no further host cells are a geographical region (Funk et al. 2001; Abbot & Moran recruited to this role later. 2002). These developmental specializations presumably reflect Shifts in selection on the pathways underlying amino a long history of co-evolution. The most plausible view is acid production are suggested by the few studies compar- that aphids acquired these symbionts only once, early in ing closely related species. For example, Diuraphis species their evolution, enabling them to exploit a novel ecological that differ in the extent of damage inflicted on host grasses niche, phloem-feeding on vascular plants. Such a lifestyle also differ in the presence of pseudogenes for tryptophan would be untenable without symbionts, due to nutritional biosynthesis (Wernegreen & Moran 2000). Also, Buchnera inadequacies. Co-adaptation on the part of the aphids of the genus Uroleucon appear to have undergone relaxation subsequently rendered them dependent on Buchnera for of selection on some amino acid biosynthetic pathways, reasons other than the nutritional dependence that was the with elevated sequence evolution and the appearance of original basis for the association. In modern aphids, early pseudogenes in some lineages (Wernegreen et al. 2000). development is disrupted if Buchnera is removed using These comparisons of closely related species provide a antibiotics, and even nutritional supplementation does not picture of the initial stages of gene and pathway loss, leading remove the requirement for Buchnera during embryonic to major differences in content of provisioning abilities development (Buchner 1965; Douglas 1998). Thus, the between divergent lineages, as revealed by the complete hosts require the symbionts, even as they lose capabilities genome sequences (Tamas et al. 2002; van Ham et al. 2003). to provide nutrients and to regulate production of these nutrients in response to host needs. Mutations in the aphid genome can also affect essential processes, but these can be Genomics on the host side restored by sex and recombination, whereas the depend- The functioning of an intimate symbiont such as Buchnera ence on a mutation-prone, asexual genome may constitute cannot be understood without addressing, at the molecular a major limitation on aphid evolution. Possibly secondary level, its integration within hosts. This aspect has been symbionts sometimes replace nutritional functions of relatively neglected until recently, due to the lack of Buchnera; however, known contributions of secondary information on the aphid genome. However, one recent symbionts are not nutrient related (Montllor et al. 2002; study has addressed coordination of aphid gene expression Oliver et al. 2003). with that of Buchnera. Nakabachi et al. (2005) characterized The planned sequencing of the A. pisum genome will transcripts that were differentially expressed in Acyrthosiphon vastly accelerate understanding of how host and symbiont pisum bacteriocytes. The findings reinforce the view of genomes function together. A particularly unknown arena the bacteriocyte as a factory for amino acid metabolism, is that of the host immune system which must be modu- requiring the coordination of host and symbiont contributions. lated to allow the proliferation of bacteria within the host Aphid genes with high transcript levels in bacteriocytes body. Additionally, the regulation of symbiont numbers relative to other tissues included those underlying processes must be controlled in some manner by the host; as yet, the of amino acid transamination, catabolism and transport. mechanisms are unknown.

slightly over half of the genes of each genome, are shared between Buchnera, Blochmannia, and Wigglesworthia. Each symbiont also contains genes not found in the other two. Both sequenced Blochmannia genomes retain a suite of essential amino acid biosynthetic pathways similar to that of Buchnera, as well as the pathways involved in the production of several additional cofactors, and urea degrada- tion (Fig. 3; for a review see Zientz et al. 2004). In contrast, Wigglesworthia has lost nearly all amino acid biosynthetic capabilities but retains the genes involved in producing a number of cofactors including B complex vitamins (e.g. thiamine, biotin, pyridoxine; Fig. 3). Neither Blochmannia strain retains any plasmids, and the single plasmid found in Wigglesworthia does not encode any genes known to be involved in either amino acid or cofactor biosynthesis (rep, speG, traT, ybbK, yggB, ibp; Akman et al. 2002). The nutritional contribution of the latter endosymbiont to Fig. 3 Presence of amino acid and vitamin pathways in the genomes of Buchnera (aphid symbiont), Blochmannia (carpenter ant the tsetse hosts is consistent with the composition of mam- symbiont) and Wigglesworthia (tsetse fly symbiont), based on malian blood, which is rich in protein but poor in such complete genome sequences. Pathway is indicated as present in vitamins (Nogge 1981). Many carpenter ants (Camponotus, Buchnera or Blochmannia if it is intact in at least one of their Colobopsis, and Polyrhachis) are omnivorous; thus, Bloch- respective sequenced genomes. mannia’s contribution of essential amino acids and cofactors is somewhat unexpected. Their metabolic contribution might be significant for temporary phases of larval or pupal development or for seasonal periods during which ants are Comparisons to other bacteriocyte-associated heavily dependent on plant exudates (Degnan et al. 2004; symbionts Wolschin et al. 2004). Building on the success and insights provided by studies The recent completion of the Blochmannia pennsylvanicus of the aphid–Buchnera association, genomic approaches genome has permitted comparison within and between have been applied to several other associations of insects two phylogenetically and compositionally unique sym- with intracellular bacteria. To date, three more full genome biont lineages (Bl. pennsylvanicus–Bl. floridanus and Buchnera sequences have been completed for two of these symbionts, of the three aphid species; Degnan et al. 2005). As in the Blochmannia species, from carpenter ants, and Wigglesworthia three Buchnera genomes, Bl. pennsylvanicus and Bl. floridanus glossinidius, from tsetse flies (Akman et al. 2002; Gil et al. exhibit complete conservation of gene order, differences in 2003; Degnan et al. 2005). Remarkably similar patterns gene content are strictly due to irreversible gene losses, and of genome evolution have occurred in all three symbiont Blochmannia shows extremely high rates of protein evolu- lineages (Buchnera, Blochmannia and Wigglesworthia); these tion. These parallels reinforce conclusions from studies on include drastic reductions in genome size and gene Buchnera regarding the processes involved in genome content, extreme AT bias, and elevated rates of nucleotide erosion of obligate intracellular bacteria. In particular, the substitutions. The repeated emergence of these features dramatic reduction in genome size appears to have occurred in these and other intracellular mutualists and patho- very early in both symbioses, in part as a result of relaxed gens suggests that prolonged intracellular associations selection on metabolic pathways that are redundant in the enforce similar shifts in selective pressures and population host cell. structure. Some differences in genome content probably reflect The precise evolutionary steps underlying reductive biological differences among these three symbiont groups. genome evolution in these symbionts are still unclear; how- Phylogenetic analyses indicate different ages of their ever, growing evidence suggests that an innate deletional obligate host associations of Buchnera, Blochmannia, and bias of bacterial genomes combined with the proliferation Wigglesworthia (150–250 Myr, 30–40 Myr, and 50–100 of repeat elements (e.g. transposons) facilitate the process Myr, respectively; Moran et al. 1993; Akman et al. 2002; (Mira et al. 2001; Moran & Plague 2004). As a result of these Degnan et al. 2004). Also, Wigglesworthia and Blochmannia processes, genomes of both Wigglesworthia and Blochmannia both exist freely in the host cell cytoplasm and retain a rod- have shrunk to less than one-fifth the size of that of E. coli. like morphology (Aksoy 1995; Sauer et al. 2000), in contrast As is typical in small genome bacteria, all three groups to Buchnera, which is large, spherical and surrounded by a retain a core set of metabolic proteins; in fact, 331 genes, or host-derived membrane. Blochmannia and Wigglesworthia

© 2006 Blackwell Publishing Ltd, Molecular Ecology, 15, 1251–1261 FUNCTIONAL GENOMICS OF BUCHNERA 1259 cells have been observed to be quite long (12–30 µm and Baumann L, Baumann P, Thao ML (1999) Detection of the mes- 5–9 µm, respectively), although polyploidy has not yet senger RNA transcribed from genes encoding enzymes of been noted. amino acid biosynthesis in Buchnera aphidicola (endosymbiont of aphids). Current Microbiology, 38, 135–136. The wealth of genomic data for the symbionts has Birkle LM, Minto LB, Douglas AE (2002) Relating genotype and generated interest in transcriptional analyses of the host phenotype for tryptophan synthesis in an aphid-bacterial insects. Transcriptome and full genome analyses are symbiosis. Physiological Entomology, 27, 302–306. underway for the tsetse fly, vector for African trypano- Birkle LM, Minto LB, Walters KFA, Douglas AE (2004) Microbial somes (> 60 000 ESTs for Glossinia moristans; Aksoy et al. genotype and insect fitness in an aphid-bacterial symbiosis. 2005), and limited expression data are becoming available Functional Ecology, 18, 598–604. for carpenter ants (Goodisman et al. 2005). Bracho AM, Martinez-Torres D, Moya A, Latorre A (1995) Discovery and molecular characterization of a plasmid localized in Buchnera sp., bacterial endosymbiont of the aphid Rhopalosiphum Conclusions padi. Journal of Molecular Evolution, 41, 67–73. Braendle C, Miura T, Bickel R et al. (2003) Developmental origin Genomic studies of Buchnera reveal small genomes that and evolution of bacteriocytes in the aphid-Buchnera symbiosis. have lost and continue to lose many capabilities, including PLoS Biology, 1, 70–76. those for provisioning hosts and for regulating gene Buchner P (1965) Endosymbiosis of Animals with Plant Microorganisms. expression. This genetic erosion must occur in the context Interscience, New York. Dale C, Young SA, Haydon DT, Welburn SC (2001) The insect of ongoing co-adaptation by hosts, the nature of which has endosymbiont Sodalis glossinidius utilizes a type III secretion yet to be characterized. One of the most intriguing of the system for cell invasion. Proceedings of the National Academy of many remaining issues regarding the functional genomics Sciences, USA, 98, 1883–1888. of symbiosis is the question of how host gene products Dale C, Jones T, Pontes M (2005) Degenerative evolution and interact with symbionts to compensate for gene losses functional diversification of type-III secretion systems in the and to integrate symbiont functions into the mutualistic insect endosymbiont Sodalis glossinidius. Molecular Biology and system. For example, we do not know the answer to the Evolution, 22, 758–766. Dale C, Plague GR, Wang B, Ochman H, Moran NA (2002) Type question of whether different Buchnera lineages could III secretion systems and the evolution of mutualistic endo- be transferred among divergent species of aphids succes- symbiosis. Proceedings of the National Academy of Sciences, USA, sfully or whether symbionts and hosts show specialized 99, 12397–12402. co-adaptation that would prevent such transfer. In the next Degnan PH, Lazarus AB, Brock C, Wernegreen JJ (2004) few years, we can hope for major advances in this relatively Host–symbiont stability and fast evolutionary rates in an unknown area, due to the planned completion of the genome ant–bacterium association: Cospeciation of Camponotus species of the model host species, Acyrthosiphon pisum. and their endosymbionts, Candidatus Blochmannia. Systematic Biology, 53, 95–110. Degnan PH, Lazarus AB, Wernegreen JJ (2005) Genome sequence of References Blochmannia pennsylvanicus indicates parallel trends in genome evolution among bacterial mutualists of insects. Genome Research, Abbot PA, Moran NA (2002) Extremely low levels of genetic 15, 1023–1033. polymorphism in endosymbionts (Buchnera) of aphids. Douglas AE (1998) Nutritional interaction in insect-microbial Molecular Ecology, 11, 2649–2660. symbioses: aphids and their symbiotic bacteria Buchnera. Annual Akman L, Yamashita A, Watanabe H et al. (2002) Genome Review of Entomology, 43, 17–37. sequence of the endocellular obligate symbiont of tsetse flies, Douglas AE, Prosser WA (1992) Synthesis of the essential amino Wigglesworthia glossinidia. Nature Genetics, 32, 402–407. acid tryptophan in the pea aphid (Acyrthosiphon pisum) symbiosis. Aksoy S (1995) Wigglesworthia gen. nov. and Wigglesworthia glossi- Journal of Insect Physiology, 38, 565–568. nidia sp. nov., taxa consisting of the mycetocyte-associated, Fares MA, Ruiz-Gonzalez MX, Moya A, Elena SF, Barrio E (2002) primary endosymbionts of tsetse flies. International Journal of Endosymbiotic bacteria: GroEL buffers against deleterious Systematic Bacteriology, 45, 848–851. mutations. Nature, 417, 398. Aksoy S, Berriman M, Hall N et al. (2005) A case for a Glossina Febvay G, Rahbe Y, Rynkiewicz M, Guillaud J, Bonnot G (1999) genome project. Trends in Parasitology, 21, 107–111. Fate of dietary sucrose and neosynthesis of amino acids in the Baumann L, Baumann P (1994) Growth-kinetics of the endo- pea aphid, Acyrthosiphon pisum, reared on different diets. Journal symbiont Buchnera aphidicola in the aphid Schizaphis graminum. of Experimental Biology, 202, 2639–2652. Applied and Environmental Microbiology, 60, 3440–3443. Funk DJ, Wernegreen JJ, Moran NA (2001) Intraspecific variation Baumann P, Baumann L, Clark MA (1996) Levels of Buchnera in symbiont genomes: Bottlenecks and the aphid–Buchnera aphidicola chaperonin GroEL during growth of the aphid Schiza- association. Genetics, 157, 477–489. phis graminum. Current Microbiology, 32, 279–285. Gil R, Silva FJ, Zientz E et al. (2003) The genome sequence of Baumann P, Baumann L, Lai CY, Rouhbakhsh D, Moran NA, Blochmannia floridanus: comparative analysis of reduced genomes. Clark MA (1995) Genetics, physiology, and evolutionary Proceedings of the National Academy of Sciences, USA, 100, 9388–9393. relationships of the genus Buchnera: intracellular symbionts Goodisman MAD, Isoe J, Wheeler DE, Wells MA (2005) Evolution of aphids. Annual Review of Microbiology, 49, 55–94. of insect metamorphosis: a microarray–based study of larval

and adult gene expression in the ant Camponotus festinatus. and virulence genes. Proceedings of the National Academy of Sci- Evolution, 59, 858–870. ences, USA (in press). Ishikawa H (1982) Host–symbiont interactions in the protein- Nakabachi A, Shigenobu S, Sakazume N et al. (2005) Transcrip- synthesis in the pea aphid, Acyrthosiphon pisum. Insect Biochemistry, tome analysis of the aphid bacteriocyte, the symbiotic host cell 12, 613–622. that harbors an endocellular mutualistic bacterium, Buchnera. Komaki K, Ishikawa H (1999) Intracellular bacterial symbionts of Proceedings of the National Academy of Sciences, USA, 102, aphids possess many genomic copies per bacterium. Journal of 5477–5482. Molecular Evolution, 48, 717–722. Nogge G (1981) Significance of symbionts for the maintenance Komaki K, Ishikawa H (2000) Genomic copy number of intracel- of an optimal nutritional state for successful reproduction in lular bacterial symbionts of aphids varies in response to hematophagous arthropods. Parasitology, 82, 101–104. developmental stage and morph of their host. Insect Biochemistry Ohtaka C, Nakamura H, Ishikawa H (1992) Structures of chaper- and Molecular Biology, 30, 253–258. onins from an intracellular symbiont and their functional Kondo N, Nikoh N, Ijichi N, Shimada M, Fukatsu T (2002) expression in Escherichia coli groE mutants. Journal of Bacteriology, Genome fragment of Wolbachia endosymbiont transferred to 174, 1869–1874. X chromosome of host insect. Proceedings of the National Academy Oliver KM, Russell JA, Moran NA, Hunter MS (2003) Facultative of Sciences, USA, 99, 14280–14285. bacterial symbionts in aphids confer resistance to parasitic Koyama Y, Yao I, Akimoto SI (2004) Aphid galls accumulate high wasps. Proceedings of the National Academy of Sciences, USA, 100, concentrations of amino acids: a support for the nutrition 1803–1807. hypothesis for gall formation. Entomologia Experimentalis et Plague GR, Dale C, Moran NA (2003) Low and homogeneous Applicata, 113, 35–44. copy number of plasmid-borne symbiont genes affecting host Lai CY, Baumann L, Baumann P (1994) Amplification of trpEG: nutrition in Buchnera aphidicola of the aphid Uroleucon ambrosiae. adaptation of Buchnera aphidicola to an endosymbiotic associa- Molecular Ecology, 12, 1095–1100. tion with aphids. Proceedings of the National Academy of Sciences, Rispe C, Delmotte F, van Ham RCHJ, Moya A (2004) Mutational USA, 91, 3819–3823. and selective pressures on codon and amino acid usage in Lai CY, Baumann P, Moran N (1996) The endosymbiont (Buchnera) Buchnera, endosymbiotic bacteria of aphids. Genome Research, of the aphid Diuraphis noxia contains plasmids consisting of 14, 44–53. trpEG and tandem repeats of trpEG pseudogenes. Applied and Rouhbakhsh D, Lai CY, von Dohlen CD et al. (1996) The tryptophan Environmental Microbiology, 62, 332–339. biosynthetic pathway of aphid endosymbionts (Buchnera): Lerat E, Daubin V, Ochman H, Moran NA (2005) Evolutionary genetics and evolution of plasmid-associated trpEG within the origins of genomic repertoires in bacteria. PLoS Biology, 3, Aphididae. Journal of Molecular Evolution, 42, 414–421. e130. Rutherford SL, Lindquist S (1998) Hsp90 as a capacitor for Martin W (2003) Gene transfer from organelles to the nucleus: morphological evolution. Nature, 396, 336–342. frequent and in big chunks. Proceedings of the National Academy Sabater-Muñoz B, van Ham RC, Moya A, Silva FJ, Latorre A (2004) of Sciences, USA, 100, 8612–8614. Evolution of the leucine gene cluster in Buchnera aphidicola: Mira A, Ochman H, Moran NA (2001) Deletional bias and the insights from chromosomal versions of the cluster. Journal of evolution of bacterial genomes. Trends in Genetics, 17, 589–596. Bacteriology, 186, 2646–1654. Montllor CB, Maxmen A, Purcell AH (2002) Facultative bacterial Sandström J, Moran NA (1999) How nutritionally imbalanced is endosymbionts benefit pea aphids Acyrthosiphon pisum under phloem sap for aphids? Entomologia Experimentalis et Applicata, heat stress. Ecological Entomology, 27, 189–195. 91, 203–210. Moran NA (1996) Accelerated evolution and Muller’s ratchet in Sasaki T, Ishikawa H (1995) Production of essential amino-acids endosymbiotic bacteria. Proceedings of the National Academy of from glutamate by mycetocyte symbionts of the pea aphid, Sciences, USA, 93, 2873–2878. Acyrthosiphon pisum. Journal of Insect Physiology, 41, 41–46. Moran NA, Mira A (2001) The process of genome shrinkage in Sauer C, Stackebrandt E, Gadau J, Hölldobler B, Gross R (2000) the obligate symbiont, Buchnera aphidicola. Genome Biology, 2, Systematic relationships and cospeciation of bacterial endo- research0054.1–0054.12. symbionts and their carpenter ant host species: proposal of the Moran NA, Plague GR (2004) Genomic changes following host new taxon Candidatus Blochmannia gen. nov. International Journal restriction in bacteria. Current Opinion in Genetics and Development, of Systematic and Evolutionary Microbiology, 50, 1877–1886. 14, 627–633. Shigenobu S, Watanabe H, Hattori M, Sakaki Y, Ishikawa H (2000) Moran NA, Munson MA, Baumann P, Ishikawa H (1993) A molecular Genome sequence of the endocellular bacterial symbiont of clock in endosymbiotic bacteria is calibrated using the insect aphids Buchnera sp. APS. Nature, 407, 81–86. hosts. Proceedings of the Royal Society of London. Series B, Biological Srivastava PN (1987) Nutritional physiology. In: World Crop Pests, Sciences, 253, 167–171. Vol. 2A, Aphids, Their Biology, Natural Enemies and Control (eds Moran NA, Plague G, Sandström J, Wilcox JL (2003) A genomic Minks AK, Harrewijn P), pp. 99–122. Elsevier, Amsterdam. perspective on nutrient-provisioning by bacterial symbionts Srivastava PN, Gao Y, Levesque J, Auclair JL (1985) Differences in of insects. Proceedings of the National Academy of Sciences, USA, amino acid requirements between 2 biotypes of the pea aphid, (Colloquium Issue), 100, 14543–14548. Acyrthosiphon pisum. Canadian Journal of Zoology, 63, 603–606. Moran NA, Dunbar HE, Wilcox JL (2005) Regulation of transcrip- Tamas I, Klasson L, Näslund K et al. (2002) Fifty million years tion in a reduced bacterial genome: nutrient-provisioning genes of genomic stasis in endosymbiotic bacteria. Science, 296, 2376– of the obligate symbiont, Buchnera aphidicola. Journal of Bacteriology, 2379. 187, 4229–4237. Thao ML, Baumann L, Baumann P, Moran NA (1998) Endo- Moran NA, Degnan PH, Santos SR, Dunbar HE, Ochman H (2005) symbionts (Buchnera) from the aphids Schizaphis graminum and The players in a mutualistic symbiosis: Insects, bacteria, viruses Diuraphis noxia have different copy numbers of the plasmid

containing the leucine biosynthetic genes. Current Microbiology, Wilkinson TL, Fukatsu T, Ishikawa H (2003) Transmission of 36, 238–240. symbiotic bacteria Buchnera to parthenogenetic embryos in the van Ham RC, Kamerbeek J, Palacios C et al. (2003) Reductive aphid Acyrthosiphon pisum (Hemiptera: Aphidoidea). Arthropod genome evolution Buchnera aphidicola. Proceedings of the National Structure and Development, 32, 241–245. Academy of Sciences, USA, 100, 581–586. Wolschin F, Hölldobler B, Gross R, Zientz E (2004) Replication Wernegreen JJ, Moran NA (2000) Decay of mutualistic potential in of the endosymbiotic bacterium Blochmannia floridanus is cor- aphid endosymbionts through silencing of biosynthetic loci: related with the developmental and reproductive stages of its Buchnera of Diuraphis. Proceedings of the Royal Society of London. ant host. Applied and Environmental Microbiology, 70, 4096–4102. Series B, Biological Sciences, 267, 1423–1431. Zientz E, Dandekar T, Gross R (2004) Metabolic interdependence Wernegreen JJ, Richardson AO, Moran NA (2000) Parallel acceler- of obligate intracellular bacteria and their insect hosts. ation of evolutionary rates in symbiont genes underlying Microbiology and Molecular Biology Reviews, 68, 745–770. host nutrition. Molecular Phylogenetics and Evolution, 19, 479– 495. Wilcox JL, Dunbar HE, Wolfinger RD, Moran NA (2003) Con- Nancy Moran and Patrick Degnan both study the genomics of sequences of reductive evolution for gene expression in an bacterial symbiosis in animals. The main study systems are aphids obligate endosymbiont. Molecular Microbiology, 48, 1491–1500. and other sap-feeding insects. These insects contain a variety of Wilkinson TL, Douglas AE (2003) Phloem amino acids and the heritable bacteria, which affect host nutrition, resistance to natural host plant range of the polyphagous aphid, Aphis fabae. Entomo- enemies, and tolerance to thermal stress. logia Experimentalis et Applicata, 106, 103–113.