Rickettsial evolution in the light of comparative genomics

Transcription

1 Biol Rev (2010), pp doi: /j X x Rickettsial evolution in the light of comparative genomics Vicky Merhej and Didier Raoult Unit for Research on Emergent and Tropical Infectious Diseases (URMITE), CNRS-IRD UMR 6236 IFR48, Faculty of Medicine, University of the Mediterranean, Marseilles, France (Received 31 August 2009; revised 02 July 2010; accepted 12 July 2010) ABSTRACT Rickettsia are best known as strictly intracellular vector-borne bacteria that cause mild to severe diseases in humans and other animals Recent advances in molecular tools and biological experiments have unveiled a wide diversity of Rickettsia spp that include species with a broad host range and some species that act as endosymbiotic associates Molecular phylogenies of Rickettsia spp contain some ambiguities, such as the position of R canadensis and relationships within the spotted fever group In the modern era of genomics, with an ever-increasing number of sequenced genomes, there is enhanced interest in the use of whole-genome sequences to understand pathogenesis and assess evolutionary relationships among rickettsial species Rickettsia have small genomes (11 15 Mb) as a result of reductive evolution These genomes contain split genes, gene remnants and pseudogenes that, owing to the colinearity of some rickettsial genomes, may represent different steps of the genome degradation process Genomics reveal extreme genome reduction and massive gene loss in highly vertebrate-pathogenic Rickettsia compared to less virulent or endosymbiotic species Information gleaned from rickettsial genomics challenges traditional concepts of pathogenesis that focused primarily on the acquisition of virulence factors Another intriguing phenomenon about the reduced rickettsial genomes concerns the large fraction of non-coding DNA and possible functionality of these non-coding sequences, because of the high conservation of these regions Despite genome streamlining, Rickettsia spp contain gene families, selfish DNA, repeat palindromic elements and genes encoding eukaryotic-like motifs These features participate in sequence and functional diversity and may play a crucial role in adaptation to the host cell and pathogenesis Genome analyses have identified a large fraction of mobile genetic elements, including plasmids, suggesting the possibility of lateral gene transfer in these intracellular bacteria Phylogenetic analyses have identified several candidates for horizontal gene acquisition among Rickettsia spp including tra, pat2, and genes encoding for the type IV secretion system and ATP/ADP translocase that may have been acquired from bacteria living in amoebae Gene loss, gene duplication, DNA repeats and lateral gene transfer all have shaped rickettsial genome evolution A comprehensive analysis of the entire genome, including genes andnon-codingdna,willhelptounlockthemysteries of rickettsial evolution and pathogenesis Key words: Rickettsia, diversity, gene loss, pathogenicity, non-coding DNA, pseudogenes, DNA repeats, lateral gene transfer CONTENTS I Introduction 2 II Rickettsia-host relationships 3 (1) Pathogenic Rickettsia in blood-feeding vectors 3 (2) Diverse hosts and relationships 3 (3) Rickettsia-host association 5 III Phylogeny and taxonomy 5 (1) Phenotypic and ecological criteria 5 (2) Molecular data useful for intragenic classification 10 (3) Use of genomics in rickettsial phylogeny and taxonomy 10 * Address for correspondence: Didier Raoult: didierraoult@gmailcom

2 2 Vicky Merhej and Didier Raoult IV Reductive evolution of rickettsiae genome 11 (1) Gene loss and pathogenicity 11 (2) Spacers are not neutral 11 (3) Genome degradation process 15 V Genome expansion 16 (1) Gene families 16 (2) DNA repeats 17 (3) Mobilome and lateral gene transfer 18 VI Future prospects 20 VII Conclusions 20 VIII Acknowledgments 21 IX References 21 I INTRODUCTION The term Rickettsia has long been used to decribe generically small intracellular bacteria that could not be cultivated in axenic media and were not otherwise identified As a consequence, the genus Rickettsia included many fastidious bacteria belonging to different phylogenetic groups The advent of molecular methods has deeply modified the definition of Rickettsia and has allowed new taxonomic and phylogenetic inferences Phylogenetic studies based on the comparison of 16S rrna gene sequences have shown that several of the bacteria classified in the order Rickettsiales, like Coxiella burnetii, Rickettsiella grylli and Eperythrozoon spp do not belong to the alpha-subclass of the phylum Proteobacteria (Neimark & Kocan, 1997; Roux et al, 1997a; Roux & Raoult, 1995b; Stothard, Clark & Fuerst, 1994; Weisburg et al, 1989) Rickettsia spp are best known as human pathogens vectored mainly by hematophagous arthropods Rickettsial diseases are important causes of illness and death worldwide (Raoult et al, 1998) Rickettsial diseases of humans include epidemic typhus (caused by Rickettsia prowazekii), Rocky Mountain spotted fever (Rickettsia rickettsii), Mediterranean spotted fever (Rickettsia conorii), murine typhus (Rickettsia typhi) (Raoult& Roux, 1997) and at least 10 other rickettsial diseases, many of which were only recognized in the last 20 years (Rickettsia africae, Rickettsia honei, Rickettsia aeschlimannii, Rickettsia helvetica, Rickettsia parkeri, Rickettsia slovaca) (Fournieret al, 2000; Paddock et al, 2004; Raoult et al, 1997) Because of the many diseases caused by Rickettsia spp, research efforts have concentrated on vertebrate pathogenic bacteria, and have largely ignored the symbiotic or intracellular rickettsiae associated with non-human eukaryotic cells Nevertheless, the use of molecular methods has shown that Rickettsia spp are a larger and more diverse group of bacteria than previously recognized Thirty-one rickettsial species have been recognized and named to date in the National Center for Biotechnology Information (NCBI) at the National Institutes of Health Sequences for 10 candidatus species, and more than 50 other non-characterized Rickettsia spp have been deposited, waiting for further identification and description A large number of these species have not been associated with pathogenic effects on humans or animals Moreover, different species have been isolated from a wide variety of hosts other than blood-sucking arthropods, including amoebae (Fritsche et al, 1999), leeches (Kikuchi et al, 2002; Kikuchi & Fukatsu, 2005), aphids (Chen, Campbell & Purcell, 1996), leafhoppers (Davis et al, 1998), springtails (Fukatsu & Shimada, 1999), and bruchid beetles (Czarnetzki & Tebbe, 2004; Fukatsu & Nikoh, 2000) The biology of rickettsiae has been poorly investigated because of the inherent difficulty in working with these intracellular bacteria Genetic manipulation of rickettsiae has been reported (Baldridge et al, 2005; Driskell et al, 2009; Qin et al, 2004; Rachek et al, 1998), however it remains laborious and time consuming and is now complemented by the genome sequences of several Rickettsia spp To understand better the molecular mechanisms underlying their evolution and pathogenicity, several genome-sequencing efforts have targeted these bacteria According to the genome project database of the NCBI ( genomes/lprokscgi), 13 genomes of rickettsial species have been sequenced completely: Rickettsia prowazekii str Madrid E, Rickettsia typhi str Wilmington, Rickettsia conorii str Malish 7, Rickettsia rickettsii str Iowa, Rickettsia rickettsii str Sheila Smith, Rickettsia felis URRWXCal2, Rickettsia massiliae MTU5, Rickettsia canadensis str McKiel, Rickettsia bellii RML369-C, Rickettsia bellii OSU , Rickettsia akari str Hartford, Rickettsia africae ESF-5, and Rickettsia peacockii str Rustic Analyses of another six genomes are in progress or at various stages of completion including Rickettsia slovaca, Rickettsia sibirica 246, Rickettsia prowazekii RP22, Rickettsia raoultii, Rickettsia parkeri and the Rickettsia endosymbiont of Ixodes scapularis The availability of complete sequences of different species belonging to a single genus enables comparative genomics to identify differences and commonalities among them Comparative genomic approaches allow better reconstruction and understandingof bacterialevolution, and provide insights into physiology and pathogenesis Genomic data reveal marked similarities among Rickettsia spp, particularly with respect to genome reduction as compared to free-living phylogenetically related bacteria Comparisons of genomic content reveal that gene loss has been an important and ongoing process in evolution in all lineages Rickettsial genomes exhibit significant intragenus variations in size (11 15 Mb) and gene content (about genes) This probably reflects the large diversity of their hosts and infection strategies Here, we review the

3 Rickettsial evolution in the light of comparative genomics 3 insights that genomic and computational approaches have provided into the diversity, taxonomy and phylogeny of Rickettsia spp, the mechanisms that have led to the reduction in their genome sizes, and the evolutionary implications of these reductions We also highlight features specific to rickettsial genomes including gene duplication, genome rearrangement, split genes, selfish DNA, and lateral gene transfers II RICKETTSIA-HOST RELATIONSHIPS (1) Pathogenic Rickettsia in blood-feeding vectors Rickettsia spp are best known as arthropod-vectored pathogens of vertebrate hosts (Raoult & Roux, 1997) They have hematophagous arthropod vectors, such as lice, ticks, fleas, and mites and present a range of biological characteristics and transmission strategies Louse-borne rickettsiae are pathogenic to the louse R prowazekii is ingested by the louse with an infected blood meal and multiplies in the epithelial cells of the anterior gut Because of over-replication, infected epithelial cells enlarge and eventually burst, releasing the rickettsiae into the gut lumen and hence eventually the faeces Infected lice often turn red, from the release of ingested blood into the hemocoel Infection with R prowazekii may lead to the death of the louse (Houhamdi et al, 2002) Likewise lice infected by R conorii (Houhamdi et al, 2003) and R typhi (Houhamdi & Raoult, 2006) are killed by bacterial multiplication The transmission of these rickettsiae to vertebrates occurs through infected insect faeces deposited at the bite site During feeding, the louse can transmit the rickettsia to a single host, or may transfer the rickettsia through horizontal transmission to other lice Most Rickettsia spp seem to have no detrimental effect on their arthropod hosts They multiply in most organs and fluids of their host, particularly in the salivary glands and ovaries, which enables horizontal transmission during feeding and transmission from adult to offspring (vertical transovarial transmission, TOT) Rickettsia such as R felis have been observed in the salivary glands of cat fleas in ultrastructural studies (Macaluso et al, 2008), hence saliva produced during an insect bite seems to be a transmission route to susceptible vertebrate hosts TOT has been demonstrated for R typhi within Xenopsylla cheopis (Farhangazad & Traub, 1985), for R akari (Philip & Hughes, 1948), for R felis within Ctenocephalides felis (Wedincamp & Foil, 2002), R slovaca (Rehacek, 1984), R sibirica (Rudakov et al, 1999), R africae (Kelly & Mason, 1991), R helvetica (Burgdorfer et al, 1979), R parkeri (Goddard, 2003), R massiliae (Matsumoto et al, 2005) and R conorii (Blanc & Caminopetros, 1932) TOT of rickettsiae in ticks seems to ensure rickettsial survival and perpetuation appears to rely upon inheritance Ticks serve as reservoir hosts; maintenance of rickettsia within ticks favours infection spread and the geographic distribution of tick-borne rickettsial disease is similar to that of the tick host (Parola & Raoult, 2001) While TOT of rickettsiae is known to occur in the mite Liposcelis sanguineus, its role in maintenance of rickettsiae is not clear and further experiments are needed to investigate the possible role of the mite as a reservoir host of R akari (Eremeeva & Paddock, 2007) Rickettsia-host relationships, particularly regarding tick fitness and in relation to pathogenicity, are not fully understood It has been proposed that vertebrate pathogenic rickettsia transmitted by ticks arose from pre-existing beneficial associations between ticks and their bacterial partners (Darby et al, 2007) This hypothesis has been refuted for R prowazekii, R typhi and R rickettsii Indeed,TOTofR rickettsii appears to be pathogenic to the tick Dermacentor andersoni, reducing its survival and reproductive capacity, and implying a major role for horizontal transmission in its maintenance (Burgdorfer, 1963; Niebylski, Peacock & Schwan, 1999) In other species, the maintenance of rickettsiae via TOT may reach 100% and have no effect on the reproductive fitness and viability of the tick host, eg R massiliae in Rhipicephalus sanguineus (Matsumoto et al, 2005) andr slovacaindermacentor marginatus (Rehacek, 1984) Finally, some Rickettsia spp in blood-feeders, such as Rickettsia peacockii in the tick Dermacentor andersoni, undergo TOT in their arthropod hosts in the absence of transmission to vertebrates (Baldridge et al, 2004) It is not yet possible to predict if a rickettsia maintained through TOT in an arthropod will be a vertebrate pathogen Further experimental studies are needed to understand the role played by TOT in the evolution of pathogenicity in Rickettsia species (2) Diverse hosts and relationships Recent molecular surveys showed that Rickettsia spp are associated with an extremely diverse host range including vertebrates, arthropods, annelids, amoebae and plants (Perlman, Hunter & Zchori-Fein, 2006; Weinert et al, 2009b) (Fig 1) Indeed, some Rickettsia spp are symbionts, with an intimate, although not necessarily beneficial, relationship with a wide range of hosts These species can be considered as facultative endosymbiotic associates or secondary symbionts of invertebrates, that is, symbionts that are not obligate for host survival and reproduction (Perlman et al, 2006) Recent studies have revealed that some Rickettsia species are localized in specific host cells such as bacteriocytes in the parthenogenetic book louse Liposcelis bostrychophila (Perotti et al, 2006) and in the pea aphid Acyrthosiphon pisum (Sakurai et al, 2005), suggesting a necessary and beneficial association Indeed, removal of these Rickettsia spp stops egg production and reproduction in the book louse (Perotti et al, 2006) The closely related Wolbachia spp, in the order Rickettsiales, are endosymbionts of insects and nematodes, and can transfer extensive genetic material to the genome of their hosts (Fenn & Blaxter, 2006; Hotopp et al, 2007; Kondo et al, 2002; Nikoh et al, 2008) Likewise, Rickettsia spp may play an important role in the evolution of their hosts by gene transfer Some Rickettsia spp cause intriguing reproductive disorders in their invertebrate hosts, like male-killing and parthenogenesis Rickettsia spp have been associated with male embryonic lethality in the beetles Adalia bipunctata, Adalia decempunctata (von der Schulenburg et al, 2001; Werren et al, 1994) and

4 4 Vicky Merhej and Didier Raoult SFG TG R rickettsii gr R massiliae gr R helvetica gr R akari gr R prowazekii gr R rickettsii spp Astrakhan fever rickettsia str A-167 R conorii str Malish 7 R africae ESF-5 R parkeristrmaculatum20 R sibirica subsp mongolotimonae R sibirica subsp sibirica R slovaca N A 13 B R honei R japonica R heilongjiangensis R hulinensis R peacockii Israeli tick typhus Candidatus R barbariae Rickettsia endosymbiont of Dermacentor hunteri R massiliae Rickettsia sp Bar 29 R rhipicephali R aeschlimanii R montanensis R raoultii Candidatus R kulagi # Rickettsia sp DnS # R helvetica C9P9 R asiatica # R tamurae R monacensis Rickettsia sp IRS * # R akari str Hartford R australis Philip 1950 R felis URRWXCal2 R prowazekii str Madrid E R typhi str Wilmington R canadensis gr R bellii gr R canadensis R bellii spp Rickettsia symbiont of Acyrthosiphon pisum # Rickettsia symbiont of of Bemisia tabaci # Rickettsia symbiont of bombyliid bee fly species # Rickettsia symbiont of chrysopid species lacewing # Rickettsia symbiont of Noctuid moth species # Rickettsia symbiont of Elaterid beetle species # Other Rickettsiae Rickettsia endosymbiont of Nuclearia pattersoni Torix tagoi Torix tukubana Hemiclepsis marginata Deronectes platynotus Cerobasis guestfalica Lutzomyia apach and Rickettsia limoniae Orientia tsutsugamushi Neorickettsia sennetsu Ehrlichia sp Anaplasma sp Wolbachia sp Fig 1 Schematic cladogram of recognized Rickettsia spp Groups are as determined by Roux & Raoult (1999) The classification of additional Rickettsia spp is taken from phylogenetic trees generated from 16S rrna, glta, ompa, and ompb sequences Sequences extracted from GeneBank werealigned withthe multisequence alignment programmuscle (Edgar, 2004) Phylogenetic relationships were inferred using the PHYLIP (Felsenstein,1993)package with the Neighbour Joining method (Saitou & Nei, 1987) Only rickettsial nodes supported by high bootstrap values (>75%) were taken into consideration The spotted fever group (SFG) corresponds to the R rickettsii group (gr), R massiliae gr, R helvetica gr and the R akari gr The typhus group (TG) is represented by the R prowazekii gr The symbols, #, *, and indicate high bootstrap values obtained using 16S rrna, glta, ompa, and ompb, respectively The schematic cladogram shows the positions of Orientia, Neorickettsia, Ehrlichia, and Anaplasma (also in Rickettsiales) as determined from 16S rrna sequences Rickettsia spp with known vertebrate pathogenic effects are indicated in red Arrows indicate host class: ticks in blue, lice in black, mites in green, fleas in red, flies (Neoptera, Hemiptera, Aleyrodiformes) in fushia, bee flies (Neoptera, Diptera) in yellow, Neochrysocharis sp (Neoptera, Hymenoptera) in light blue, booklice (Neoptera, Psocoptera) in orange, aphids (Neoptera, Hemiptera, Aphidiformes) in grey, lacewings (Neoptera, Neuropterida) in dark purple, moths (Neoptera, Lepidoptera) in pink, beetles (Neoptera, Coleoptera) in light green, Nuclearia sp (Eukaryota, Nucleariidae) in light purple, segmented worms (Glossiphoniidae) in gold

5 Rickettsial evolution in the light of comparative genomics 5 Brachys tessellates (Lawson et al, 2001) The male-killing phenotype was confirmed by antibiotic treatment: the number of males that successfully hatched increased (Lawson et al, 2001) Rickettsia spp have been associated with parthenogenesis in the book louse Liposcelis bostrychophila (Yusuf & Turner, 2004) and in the tick Amblyomma rotundatum (Labruna et al, 2004b) However, it is difficult to determine the involvement of these rickettsia in the parthenogenetic phenotype, for many reasons First, their invertebrate hosts harbour multiple symbionts such as Wolbachia spp and without the removal of all but one, it is impossible to determine which symbiont is causing the phenotype (Hurst & Jiggins, 2000) Secondly, the removal of symbionts via rifampicin treatment results in a major reduction in egg hatch rate, as well as offspring production and survival (Yusuf & Turner, 2004), but viable males are not produced after antibiotic treatment Recently, strong evidence for an association between Rickettsia sp Nfor and parthenogenetic reproduction was reported In parthenogenetic populations of the parasitoid Neochrysocharis formosa (Hymenoptera: Eulophidae), over 995% of individuals are females and all appear infected with Rickettsia sp Nfor with no other symbionts (Hagimori et al, 2006) Antibiotic treatment results in production of (uninfected) male offspring, suggesting that the symbiont is responsible for parthenogenetic production of females Many of the rickettsia detected in invertebrates are not or not yet associated with a host phenotype Molecular surveys coupled with biological experiments can help to explore the diversity of Rickettsia-host relationships (3) Rickettsia-host association Specific associations between Rickettsia spp and arthropod vectors have been reported although these observations are based on limited data Rickettsia from the spotted fever group (SFG) have been associated with ticks, fleas and mites and rickettsia from the typhus group (TG) have been associated with ticks, fleas and lice (Table 1) Recent molecular experiments have revealed the great diversity of rickettsial hosts and challenged previously accepted specific associations Besides its currently known biological vector, the cat flea, Ctenocephalides felis, R felis was detected using polymerase chain reaction (PCR) in other species of fleas such as Xenopsylla cheopis and Archeopsylla erinacei from rodents and hedgehogs, and from Anomiopsyllus spp (Bitam et al, 2006; Stevenson et al, 2005), as well as in ticks (Ishikura et al, 2003; Oliveira et al, 2008), mites (Choi et al, 2007) and in a tsetse fly (D Raoult, unpublished data) These molecular data suggest that R felis may be maintained by a variety of arthropod hosts (Reif & Macaluso, 2009) R prowazekii has been reported in acarids from flying squirrels in the USA (Bozeman et al, 1975) and ticks in Africa and Mexico (Medina-Sanchez et al, 2005; Reiss-Gutfreund, 1966) (Table 1) Lice can be infected experimentally with R typhi (Houhamdi et al, 2003) Likewise, the body louse might be able to acquire, maintain and transmit R conorii and R rickettsii (Houhamdi & Raoult, 2006) Many questions regarding the specificity of associations between Rickettsia spp and vector species remain unresolved R helvetica seems to be associated with the tick genus Ixodes: it was first isolated from I ricinus in Switzerland (Beati et al, 1993; Burgdorfer et al, 1979), and subsequently was found in several Ixodes spp in Japan (I ovatus, I persulcatus and I monospinosus) (Fournier et al, 2002) By contrast, R rickettsii is found in several tick genera (Table 1) Between these extremes there are certain Rickettsia spp that are associated with several species within a single genus, such as R africae and R slovaca with various Amblyomma spp and Dermacentor spp, respectively (Parola & Raoult, 2001) (Table 1) It is difficult to know if the specificity of the associations arises from the Rickettsia-tick relationship or from the availability of vertebrate hosts Related Rickettsia spp tend to share related hosts, which suggests ancestral infection followed by Rickettsia-host coevolution However, the host phylogeny does not correspond with the rickettsial phylogeny (Fig 2) suggesting that many host shifts occured between taxonomically distant hosts (Weinert et al, 2009b) The use of molecular methods to detect Rickettsia spp in arthropods and other animals led to the discovery of previously unrecognized rickettsial diversity in invertebrate hosts (Fritsche et al, 1999; Fukatsu & Shimada, 1999; Kikuchi et al, 2002; Weinert et al, 2009b; Werren et al, 1994) These new data aid understanding of rickettsial evolution: Rickettsia spp associated with flies, leeches, protists and freshwater environments form phylogenetic groups distinct from the SFG and TG groups (Perlman et al, 2006; Weinert et al, 2009b) The R bellii group (Fig 1), first found in ticks, contains R bellii and its close relatives that can infect insects and cause malekilling and parthenogenesis phenotypes and plant pathogens inducing rickettsia (Chen et al, 1996; Davis et al, 1998) The common ancestor of Rickettsiales was presumably free-living It has been suggested that the transition to an intracellular lifestyle occurred million years ago In the most common scenario Rickettsia is presumed to be primarily arthropod-associated then some species switched to infect other eukaryotes such as protists and leeches approximately 150 million years ago (Weinert et al, 2009b) Alternative scenarios for the ancestral state are also possiblethe genome of R bellii includes many genes that are related to those of amoebal symbionts (Ogata et al, 2006), perhaps arising from an ancient exchange of genes between an ancestor of R bellii, infecting amoebas, and other amoebal hosts These findings suggest that the first host cell of Rickettsia was a protist Further genomic analyses and eukaryotic host sampling may resolve such questions III PHYLOGENY AND TAXONOMY (1) Phenotypic and ecological criteria Traditional taxonomic methods use morphological, ecological, epidemiological and clinical characteristics to describe bacterial species The genus Rickettsia has historically been divided into three groups based on immunological crossreactivity and vector species: the spotted fever group

6 6 Vicky Merhej and Didier Raoult Table 1 Transmission and diseases caused in humans for validated rickettsiae with known pathogenic effects on vertebrates and rickettsiae of possible or undetermined pathogenicity Pathogenic rickettsiae R prowazekii - Pediculus humanus - Orchopeas howardii - Amblyomma cajennense R typhi - Xenopsylla cheopis - Ctenocephalides felis - Leptopsylla segnis Vector Disease References R conorii subsp conorii - Rhipicephalus sanguineus - Haemaphysalis leachii R conorii subsp israelensis - Rhipicephalus sanguineus - Amblyomma maculatum Epidemic typhus, Brill-zinsser disease Endemic murine typhus Andersson & Andersson (2000); Andersson et al (1998); Green Fishbein & Gleiberman (1990); Guedes et al (2005); Medina-Sanchez et al (2005); Perine et al (1992) Azad (1990); Dumler & Walker (2000); Gikas et al (2002); Raoult & Roux (1997); Walker et al (1989) Mediterranean spotted fever Raoult & Roux (1997) Israel tick typhus Zhu et al (2005a) R conorii subsp caspia - Rhipicephalus pumilio Astrakham fever Zhu et al (2005a) - Rhipicephalus sanguineus R conorii subsp indica - Rhipicephalus sanguineus Indian tick typhus Zhu et al (2005a) R sibirica subsp sibirica R sibirica subsp Mongolitimonae Astrakhan fever Rickettsia - Dermacentor nutallii - Dermacentor marginatus - Dermacentor silvaum - Dermacentor pictus - Dermacentor sinicus - Dermacentor auratus - Haemaphysalis concinna - Hyalomma asiaticum - Hyalomma truncatum Siberian tick typhus Lymphangitis-associated rickettsioses Mediannikov Parola & Raoult (2007) Fournier & Raoult (2004); Parola et al (2001); Yu et al (1993) - Rhipicephalus pumilio Astrakhan fever Eremeeva et al (1994) R heilongjiangensis - Dermacentor silvarum - Haemophysalis japonica douglasi - Haemophysalis concinna R australis - Ixodes holocyclus -Ixodestasmani - Ixodes cornuatus R japonica - Haemaphysalis flava - Haemaphysalis longicornis - Dermacentor taiwanensis - Ixodes ovatus R honei - Aponomma hydrosauri - Ixodes granulatus R marmionii - Haemaphysalis novaeguinae - Ixodes holocytus R slovaca - Dermacentor marginatus - Dermacentor reticulatus R aeschlimannii - Hyalomma m marginatum - Hyalomma mrufipes - Rhipicephalus appendiculatus - Haemaphysalis punctata Far-eastern tick-borne rickettsiosis Fournier et al (2003); Mediannikov et al (2006); Mediannikov et al (2009) Queensland tick typhus Campbell & Domrow (1974); Graves et al (1993); Roberts (1960) Japanese or Oriental spotted fever Fournier et al (2002); Uchida et al (1989); Uchida et al (1992); Uchiyama & Uchida (1989) Flinders Island spotted fever Graves & Stenos (2003); Stenos et al (1998); Unsworth et al (2007); Whitworth et al (2003) Australian spotted fever Stenos Graves & Unsworth (2005); Unsworth et al (2007) Tick-borne lymphadenopathy (Tibola) and Dermacentor-borne-necrosiserythema-lymphadenopathy (DEBONEL) Ibarra et al (2005); Lakos (1999); Oteo et al (2004); Raoult et al (1997, 2002); Sekeyova et al (1998) Tick-transmitted disease Beati et al (1995, 1997)

7 Rickettsial evolution in the light of comparative genomics 7 Table 1 (Cont) Vector Disease References R felis - Ctenocephalides felis - Ctenocephalides canis - Archeopsylla erinacei - Pulex irritans - Xenopsylla cheopis - Anomiopsyllus nudata R akari - Allodermanyssus sanguineus - Liponyssoides sanguineus R australis - Ixodes holocyclus -Ixodestasmani - Ixodes cornuatus R parkeri - Amblyomma maculatum (triste) - Amblyomma americanum R massiliae - Rhipicephalus sanguineus - Rhipicephalus turanicus - Rhipicephalus mushamae - Rhipicephalus lunulatus - Rhipicephalus sulcatus R rickettsii - Dermacentor andersoni - Dermacentor variabilis - Rhipicephalus sanguineus - Amblyomma cajennense - Amblyomma aureolatum R africae - Amblyomma hebraeum - Amblyomma variegatum - Rhipicephalus (Boophilus) decoloratus R raoultii - Dermacentor nutallii - Dermacentor silvarum - Dermacentor reticulatus - Dermacentor marginatus - Dermacentor niveus Rickettsiae of possible or undetermined pathogenicity R bellii - Dermacentor variabilis - Dermacentor andersoni - Dermacentor occidentalis - Dermacentor albopictus - Haemophysalis lepopalutris - Ornithodoros concanensis - Argas cooleyi - Amblyomma cooperi - Amblyomma aureolatum - Amblyomma dubitatum - Amblyomma humerale - Amblyomma rotundatum - Amblyomma oblongoguttatum - Amblyomma scalpturatum - Amblyomma ovale - Ixodes loricatus - Haemaphysalis juxtakochi R canadensis - Haemaphysalis leporispalustris - Dermacentor andersoni - Dermacentor variabilis - Amblyomma americanum Flea-borne spotted fever Azad et al (1997); Bitam et al (2006); Eremeeva et al (2008); Higgins et al (1996); Stevenson et al (2005); Venzal et al (2006) Rickettsial pox Eremeeva & Paddock (2007); Paddock et al (2006) Queensland tick typhus Graves et al (1993); McBride et al (2007) Tick-transmitted disease skin lesions and lymphadenitis Spotted fever Rocky Mountain spotted fever, American spotted fever, or tick typhus African tick bite fever Goddard (2003); Paddock et al (2004); Parker et al (1939); Venzal et al (2004) Bacellar et al (1995); Beati et al (1996); Beati & Raoult (1993); Rydkina et al (1999) Burgdorfer (1988); Childs & Paddock (2007); Demma et al (2005); Guedes et al (2005); Pinter & Labruna (2006) Kelly et al (1996); Parola et al (1999); Portillo et al (2007); Raoult et al (2001) R slovaca-like infection Mediannikov et al (2008); Shpynov et al (2001) ND Gage et al (1994); Horta et al (2006); Labruna et al (2004b, c, 2007a, 2007b); Philip et al (1983); Pinter & Labruna (2006) Rocky Mountain spotted fever (RMSF)-like disease in USA (?) Suspected cause of acute cerebral vasculitis in Ohio Bozeman et al (1970); Burgdorfer & Brinton (1970); McKiel Bell & Lackman (1967)

8 8 Vicky Merhej and Didier Raoult Table 1 (Cont) R helvetica - Ixodes ricinus - Ixodes ovatus -Ixodespersulcatus - Ixodes monospinosus R montanensis (formerly R montana) Vector Disease References - Dermacentor variabilis - Dermacentor andersoni Tick-transmitted disease Headache and myalgias, and less frequently with rash or an eschar Perimyocarditis (?) Cardiac valve pathology (?), sarcoidiosis (?) ND Bell et al (1963) Beati et al (1993); Burgdorfer et al (1979); Fournier et al (2002, 2004a); Nilsson et al (2002); Nilsson Lindquist & Pahlson (1999); Parola et al (1998); Walker Valbuena & Olano (2003) R asiatica - Ixodes ovatus ND Fujita et al (2006); Ishikura et al (2002) R tamurae - Amblyomma testudinarium Suspected cause of acute illeness in Laos R amblyommii - Amblyomma americanum - Amblyomma cajennense - Amblyomma coelebs - Amblyomma longirostre Suspected cause of Texas tick fever, lone star fever, bullis fever Fournier et al (2006a); Phongmany et al (2006) Anigstein & Anigstein (1975); Labruna et al (2004a, b); Livesey & Pollard (1943); Woodland McDowell & Richards (1943) R rhipicephali - Rhipicephalus sanguineus ND Hayes & Burgdorfer (1979); Labruna et al (2007b) R monacensis - Ixodes ricinus ND Jado et al (2007); Simser et al (2002) Candidatus R - Rhipicephalus turanicus ND Mura et al (2008) barbariae Candidatus R - Dermacentor andersoni ND Niebylski et al (1997) paecockii Candidatus R kotlanii - Haemaphysalis concina ND Sreter-Lancz et al (2006) Candidatus R kulagini - Rhipicephalus sanguineus ND Raoult et al (2005) Candidatus R uilenbergi - Amblyomma thollini ND Matsumoto et al (2007) Candidatus R andeanae Candidatus R tarasevichiae - Amblyomma maculatum ND Blair et al (2004); Jiang et al (2005) - Ixodes boliviensis -Ixodespersulcatus ND Shpynov et al (2003) Candidatus R gravesii - Amblyomma triguttatum ND Owen et al (2006b) Candidatus R -Ixodespersulcatus ND Owen et al (2006a) antechini Candidatus R principis - Haemaphysalis japonica ND Mediannikov et al (2006) - Haemaphysalis sulcata Candidatus R rara - Haemaphysalis concinna ND Raoult et al (2005) For vector, ticks are in blue, lice in black, fleas in red, and mites in green Only human diseases are shown ND not determined; Hyalomma mforhyalomma marginatum R sibirica subsp sibirica observed by light microscopy in Haemaphysalis concinna in the pre-molecular biology era not confirmed by molecular microbiology (SFG), the typhus group (TG), and the scrub typhus group (STG) SFG rickettsiae, mainly associated with ticks, have an optimal growth temperature of 32 C, a guanosine plus cytosine (G+C) content between 32% and 33%, can polymerize actin and thereby enter the nuclei of host cells (Heinzen et al, 1993; Teysseire, Boudier & Raoult, 1995; Teysseire, Chiche-Portiche & Raoult, 1992), and cause spotted fever in humans Typhus group (TG) rickettsiae are associated with body lice (R prowazekii) orfleas(r typhi), have an optimal growth temperature of 35 C, a G+C content of 29%, are only found in the cytoplasm of host cells (Heinzen et al, 1993; Teysseire et al, 1992), and cause typhus in humans Use of these ecological and phenotypic criteria to describe Rickettsia species corresponds to the definition of species in eukaryotes where genetic isolation is associated with species emergence (Mayr, 1942) and, indeed, genetic isolation of Rickettsia spp in arthropods has resulted in the formation of new species However, the above classification is perhaps simplistic, and some Rickettsia spp do not fit well within this grouping For example, SFG rickettsiae are defined as living in ticks, but exceptions include R akari (transmitted by mites) and R felis (transmitted by cat and dog fleas) In cladograms such as Fig 1 R bellii clusters separately from the TG and SFG rickettsiae It has been proposed

9 Rickettsial evolution in the light of comparative genomics 9 R rickettsii R conorii Rhipicephalus sanguineus Rhipicephalus spp 99 R sibirica R africae Hyalomma truncatum Hyalomma spp R parkeri R slovaca R marmioni R honei R montanensis Dermacentor andersoni Dermacentor spp Dermacentor marginatus Haemaphysalis leachi Haemaphysalis spp Haemaphysalis punctata R heilongjansis R rhipicephali R amblyommii Amblyomma americanum Amblyomma spp Amblyomma maculatum R massiliae R helvetica R asiatica R tamurae Ixodes ricinus Ixodes persulcatus Ixodes spp Ixodes tasmani R australis R akari Ctenocephalides felis Ctenocephalides canis R felis R bellii R canadensis R typhi Rickettsia symb Nuclearia pattersoni Orientia tsutsugamushi Rickettsia symb Hemiclepsis marginata Ricekttsia symb Torix tukubana Neorickettsia sennetsu Wolbachia spp Rickettsia sp PAR R prowazekii Ehrlichia spp Anaplasma spp Pulex irritans Leptopsylla segnis Xenopsylla cheopis Pediculus humanus capitis Pediculus humanus corporis Pediculus humanus humanus Acyrthosiphon pisum Nuclearia pattersoni Glossiphoniidae spp Fig 2 Evolution of Rickettsia spp and their hosts Neighbour-joining phylogenies of Rickettsia spp and their hosts obtained using 16S rrna and 18S rrna gene sequences, respectively The analyses include only the hosts (and corresponding Rickettsia spp) for which 18S rrna data are available on NCBI Only high bootstrap values (>75%) are shown on the branch nodes Branch lengths are proportional to sequence divergence and can be measured relative to the scale bars shown (002 nucleotide substitutions per site) The colours of the arrows correspond to the different host genus that this species diverged prior to the division between the SFG and the TG, forming a separate group misleadingly termed the ancestral group (Stothard et al, 1994) even though many rickettsial clades had already emerged (Fig 1) R canadensis has some similarities with the TG, including G+C content and serological cross-reactivity, and other features that are shared with the SFG, including growth in both the cytoplasm and nucleus of the host cell, and ticks as an

10 10 Vicky Merhej and Didier Raoult arthropod reservoir It was placed in the ancestral group with R bellii, although this grouping is now questioned on the basis of genetic (Fournier et al, 2003) and genomic criteria (Merhej, Karkouri & Raoult, 2009a) Because of the few remarkable phenotypic characters expressed by these bacteria, traditional classification methods used in bacteriology are hard to apply to Rickettsia spp (2) Molecular data useful for intragenic classification The use of DNA-DNA hybridization for description of novel species (Grimont, 1988; Wayne et al, 1987) on the basis of a cutoff of 70% DNA-DNA relatedness was established for Enterobacteria spp (Wayne et al, 1987), but has low discriminatory ability for Rickettsia spp: R rickettsii, R conorii, R sibirica,andr montanensis are considered by this technique to be the same species The advent of molecular methods allowed more accurate identification of Rickettsia spp, improving evolutionary studies of these bacteria The 16S rrna gene sequence of the STG species, Rickettsia tsutsugamushi, was found to be distinct enough to warrant transfer into the genus Orientia (Tamura et al, 1995) Nevertheless, 16S rrna sequencing was insufficient to distinguish different Rickettsia spp (996% sequence identity between R conorii and R massiliae), precluding phylogenetic inferences using this method Recent advances in the taxonomy and phylogeny of Rickettsia spp have mainly relied on the analysis of multiple molecular sequences With the use of molecular methods, the number of representatives of the genus Rickettsia has increased dramatically over the last 20 years with 31 currently recognized species, and many as-yet uncharacterized strains, subspecies, and uncultured species ( wwwncbinlmnihgov/taxonomy/browser/wwwtaxcgi? id=780) Phylogenetic relationships have been inferred from individual gene sequences including citrate synthetase (glta) (Roux et al, 1997b) (Fig 1), and genes encoding for the surface cell antigen (sca) family:ompa (Fournier,Roux& Raoult, 1998), ompb (Roux & Raoult, 2000), sca4 (Sekeyova, Roux & Raoult, 2001), sca1 (Ngwamidiba et al, 2006), and sca2 (Blancet al, 2005a; Ngwamidiba et al, 2005) However, a robust and accurate phylogeny of rickettsia has not been achieved using single-gene phylogenies, and ambiguities remain in the phylogenetic position of R canadensis and species within the SFG As progressively more genes have been sequenced, multi-genic approaches (Vitorino et al, 2007; Weinert, Welch & Jiggins, 2009a) have been used to investigate species relationships as well as to develop taxonomic strategies (Fournier et al, 2003; Maiden et al, 1998; Roux & Raoult, 1999; Stackebrandt et al, 2002; Tautz et al, 2003) The first attempt to reclassify rickettsial species using a multigenic approach, combining several genes including glta andsca family members, allowed the establishment of a polyphasic taxonomy of Rickettsia spp compatible with other characteristics such as DNA-DNA homology and G+C content (Roux & Raoult, 1999) This allowed phylogenetic relationships among members of the genus Rickettsia to be estimated, using a sort of consensus between the different individual gene phylogenies Species within the SFG can be subdivided into four subgroups in accordance with their phenotypic profiles: the R rickettsii, R massiliae, R helvetica, and R akari groups (Fig 1) Recently, phylogenetic analysis of the prf plasmid-encoded genes of R felis indicated an affinity between the clade of R felis/r akari and the so-called ancestral group (Gillespie et al, 2007) Thus, it has been proposed to create a transitional group, containing R felis and R akari, between R prowazekii and R typhi on one hand and the rickettsia transmitted by ticks (R conorii, R rickettsii, and R sibirica) on the other (Gillespie et al, 2007) However, species of this transitional group, together with the tick-transmitted R australis, belong to the R akari group as defined by Roux & Raoult (1999) (3) Use of genomics in rickettsial phylogeny and taxonomy Availability of complete genome sequences has made a profound change in our understanding of rickettsial phylogeny and taxonomy Trees produced from the concatenation of 704, 668 and 635 core gene proteins (Blanc et al, 2007b; Fournier et al, 2008; Merhej et al, 2009a) for 7, 8 and 11 available genomes of Rickettsia spp, respectively, show high bootstrap support These analyses delineated the two groups TG and SFG, and the SFG subgroups (Roux & Raoult, 1999) and showed that R belli and R canadensis lie outside the SFG and TG (Merhej et al, 2009a) Phylogenomic analysis revealed that R felis/r akari share dozens of similar genes with R canadensis and R bellii genomes (Gillespieet al, 2008), according to those authors supporting the creation of the transitional group although this remains a controversial While adopting the new designation, phylogenetic analysis on the basis of some concatenated genes did not support the creation of a monophyletic transitional group, on the contrary it grouped R prowazekii with R akari (Weinert et al, 2009b) Genome comparison between R prowazekii and R conorii showed that overall intergenic spacers were significantly more variable than genes (P < 001) (Ogata et al, 2001) It has been demonstrated that non-coding sequences are superior to genes for the phylogenetic and genotypic classification of botanical species (Demesure, Sodzi & Petit, 1995; Yang et al, 2002) To date, the most widely used intergenic spacer in bacteria has been the 16S 23S rdna spacer (Hassan et al, 2003; Hill et al, 2002; Maiwald, Lepp & Relman, 2003; Roux & Raoult, 1995a; Stamm, Bergen & Walker, 2002) However, the 16S 23S rdna spacer is not usable in Rickettsia species as these genes are not contiguous (Andersson et al, 1995, 1999; Ogata et al, 2001) Genotyping studies based on the presence or absence of single-nucleotide polymorphisms (SNP) (Karpathy, Dasch & Eremeeva, 2007; Zhu et al, 2008) and variable nucleotide tandem repeats (VNTR) (Eremeeva et al, 2006) in intergenic spacers have been applied to Rickettsia spp Moreover, multi-spacer typing (MST) has been developed selecting the most variable noncoding regions between R prowazekii and R conorii The MST analyses combining three intergenic spacers (dksa-xerc,

11 Rickettsial evolution in the light of comparative genomics 11 mppa-purc, and rpme-trnafmet) allowed identification of rickettsial species at the strain level (Fournier et al, 2004b, 2006b; Fournier & Raoult, 2007; Zhu et al, 2005b) Altogether, current findings emphasize the importance of genome analysis in evolutionary studies and taxonomy IV REDUCTIVE EVOLUTION OF RICKETTSIAE GENOME (1) Gene loss and pathogenicity Rickettsia spp have reduced genomes that vary in size from 11 Mb for the TG, Mb for the SFG, and 15 Mb for R bellii accounting for protein-coding genes (Table 2) Gene loss is thought to be a feature of the evolution of intracellular pathogenic bacteria (Andersson & Andersson, 1999b; Andersson & Kurland, 1998; Blanc et al, 2007b;Darbyet al, 2007; Merhej et al, 2009b; Moran, 1996) R prowazekii is much more virulent than R conorii, however genome sequencing of R prowazekii has not found any genes directly identifiable as virulence determinants (Ogata et al, 2001) In fact, the genome of R prowazekii [834 open reading frames (ORFs)] represents a subset of R conorii (1374 ORFs) and it possesses almost no genes that are not present in R conorii, with the exception of four TG-specific genes that also are present in R typhi (RP624, RP338, RP164 and RP174) and absent from, or split in, other rickettsiae (Gillespie et al, 2008; Ammerman et al, 2009) Further genetic experiments are needed to elucidate the role of these genes in the pathogenicity of murine and epidemic typhus Genome analysis revealed the presence of ricka inr conorii and its absence in R prowazekii (Ogata et al, 2001) The RickA protein plays a role in actin-based motility (Gouin et al, 2004) and actin comet tail formation has been shown to be an important virulence determinant in Listeria spp and Shigella spp (Frischknecht & Way, 2001; Pollard & Borisy, 2003) However, ricka is absent from R prowazekii while it is commonly found in the genomes of other Rickettsia spp without identified pathogenicity (Balraj et al, 2008a, b; Balraj,Renesto & Raoult, 2009) Sca2 has been associated recently with adherence (Cardwell & Martinez, 2009) and actin-based motility (Kleba et al, 2010) Interestingly, R prowazekii does not appear to express a functional Sca2 protein and lacks actin-based motility, whereas the expression of a more developed Sca2 protein in R typhi is consistent with its limited actin-based motility and the short actin tails observed in this bacteria However, Sca2 protein and actin-based motility are a common feature of most SFG rickettsiae although some members of this group have never been associated with human disease (Kleba et al, 2010) Genome comparison of the virulent R rickettsii strain Sheila Smith with the avirulent R rickettsii strain Iowa revealed disruptionin rompa(sca0) in R rickettsii strain Iowa and defects in the processing of the autotransporter and surface antigen rompb (sca5) The rickettsial outer membrane proteins, rompa and rompb play important roles in adhesion and invasion, respectively, of mammalian cells (Chan et al, 2009; Li & Walker, 1998; Uchiyama, Kawano & Kusuhara, 2006) This disruption and defective processing most likely contributes to the avirulence of R rickettsii Iowa (Ellison et al, 2008) Genetic manipulation of R prowazekii showed that mutation of the gene encoding for a homolog of phospholipase D (PLD) (Renesto et al, 2003) attenuates the virulence of R prowazekii in a guinea pig model (Driskell et al, 2009) Moreover, animals immunized with the mutant strain were protected against subsequent challenge with the virulent Breinl strain, suggesting that this transformant could serve as a nonreversible, attenuated vaccine strain Interestingly, the expression of PLD genes in Salmonella enteric serovar Typhimurium allowed the bacteria to escape from the phagosome (Whitworth et al, 2005) These studies suggest a possible role for RickA, Sca2, rompa, rompb and PLD in rickettsial virulence and underline how the complementary use of genomic analysis with genetic manipulation can illuminate rickettsial biology and pathogenesis Vertebrate-pathogenic Rickettsia spp lack pathogenicity islands (clusters of genes that encode virulence traits) that are present in many bacterial pathogens (Hacker & Kaper, 2000) It has been suggested that genes encoding proteins necessary for host recognition, invasion and pathogenicity may occur in rickettsial plasmids; however, the presence of plasmids in Rickettsia species (R felis, R massiliae, R africae, R parkeri and R peacockii:blanc et al, 2007a; Felsheim, Kurtti & Munderloh, 2009; Fournier et al, 2009; Ogata et al, 2005a; Paddock et al, 2004) shows no correlation with virulence Rickettsial genomic analysis reveals that a shift to pathogenicity does not necessarily require acquisition of novel genes In some cases, loss of gene function may be implicated in the emergence of virulence because it may provide a selective advantage, as exemplified by the succession of genetic events that contribute to virulence in Shigella spp and Yersinia spp (Maurelli et al, 1998; Sun, Hinnebusch & Darby, 2008) When Shigella spp evolved from E coli to become pathogens, they shed genes via deletion of a large genomic segment The formation of these black holes, ie deletions of genes that are detrimental to lifestyle, provides an evolutionary pathway that enables a pathogen to enhance virulence (Darby et al, 2007) This finding prompted a massive comparative genomic analysis of the occurrence of gene loss in intracellular bacteria, including pathogenic bacteria (Merhej et al, 2009b) Interestingly, a recent study comparing R africae and R rickettsii pointed out the loss of essential genes such as regulatory genes in R rickettsii as a possible factor involved in the development of pathogenicity (Fournier et al, 2009) (2) Spacers are not neutral The trend towards a small genome size in rickettsiae did not prevent the accumulation of non-coding DNA in some Rickettsia spp, leading to a high percentage of non-coding DNA (Table 2) compared to that in other bacteria sequenced to date (6 14% of the genome) (Rogozin et al, 2002) Genome analysis of various rickettsial genomes shows that the genome reduction process described for alpha-proteobacteria (Boussau et al, 2004) has occurred independently in different

12 12 Vicky Merhej and Didier Raoult Table 2 Genome features of sequenced Rickettsia spp Genome Refseq Genome size (Mb) G+C (%) % coding Genes Protein coding RPE and RR Genes encoding eukaryotic-like motifs References Rickettsia prowazekii str Madrid E Rickettsia typhi str Wilmington Rickettsia canadensis McKiel NC Ank (RP226, RP714, RP716) 1 TPR (RP599) NC Ank (RT0218) 1 TPR (RT0587) NC ND 1 Ank (A1E 02000) 2TPR(A1E 02670, A1E 02675) Rickettsia akari Hartford NC ND 5 Ank (A1C 01340, A1C 01680, A1C 04205, A1C 04380, A1C 04475) 1TPR(A1C 00275) Rickettsia sibirica ND ND Rickettsia rickettsii str NC ND 8 Ank (A1G 00065, A1G 00070, Sheila Smith A1G 00075, A1G 01255, A1G 01260, A1G 02955, A1G 02960, A1G 04850) 1TPR(A1G 00150) Rickettsia rickettsii Iowa NC ND 3 TPR (RrIowa 0034, RrIowa 0849, RrIowa 1596) Rickettsia conorii str Malish 7 NC Ank (RC0308, RC0502, RC0700, RC0859, RC0860, RC0877, RC0955) 4 TPR (RC0914, RC0957, RC1366, RC1367) Rickettsia slovaca 127 ND ND Rickettsia africae ESF-5 NC ND 8 Ank (RAF ORF0213, RAF ORF0264, RAF ORF0286, RAF ORF0482, RAF ORF0702, RAF ORF0782, RAF ORF0795, RAF ORF0860) 2TPR(RAF ORF0829, RAF ORF0861) Rickettsia peacockii str Rustic Rickettsia massiliae MTU5 NC ND 2 Ank (RPR 00905, RPR 03380) 1 TPR (RPR 07730) NC Ank (rma ORF0010, rma ORF0284, rma ORF0304, rma ORF0505, rma ORF0506, rma ORF0522, rma ORF0768, rma ORF0873, rma ORF0874, rma ORF0884, rma ORF0957, rma ORF0958, rma ORF1264) 4TPR(rma ORF0022, rma ORF0711, rma ORF0921, rma ORF1352) Andersson et al (1998) McLeod et al (2004) Ellison et al (2008) Ellison et al (2008) Ogata et al (2001) Fournier et al (2009) Felsheim et al (2009) Blanc et al (2007a)

13 Rickettsial evolution in the light of comparative genomics 13 Table 2 (Cont) Genome Refseq Genome size (Mb) G+C (%) % coding Genes Protein coding RPE and RR Genes encoding eukaryotic-like motifs References Rickettsia felis URRWXCal2 Rickettsia bellii RML369-C Rickettsia bellii OSU Rickettsia massiliae MTU5, plasmid prma Rickettsia peacockii str Rustic plasmid prpr NC NC Ank (RF 0011, RF 0063, RF 0266, RF 0314, RF 0381, RF 0422, RF 0580, RF 0591, RF 0782, RF 0783, RF 0922, RF 0923, RF 0939, RF 0950, RF 0987, RF 1081, RF 1087, RF 1099, RF 1306) 8TPR(RF 0026, RF 0260, RF 0310, RF 0366, RF 0421, RF 0444, RF 1175, RF 1394) 27 Ank (RBE 0150, RBE 0151, RBE 0220, RBE 0231, RBE 0261, RBE 0317, RBE 0319, RBE 0347, RBE 0357, RBE 0489, RBE 0555, RBE 0585, RBE 0586, RBE 0589, RBE 0601, RBE 0623, RBE 0801, RBE 0902, RBE 0921, RBE 0984, RBE 0985, RBE 0997, RBE 1025, RBE 1110, RBE 1215, RBE 1220, RBE 1411) 9TPR(RBE 0036, RBE 0125, RBE 0424, RBE 0554, RBE 0634, RBE 0677, RBE 0791, RBE 0850, RBE 1180) NC ND 26 Ank (A1I 01140, A1I 01165, A1I 01775, A1I 02250, A1I 02375, A1I 02825, A1I 03170, A1I 03345, A1I 03360, A1I 03425, A1I 03670, A1I 03730, A1I 03735, A1I 04060, A1I 04455, A1I 05135, A1I 05995, A1I 06035, A1I 06200, A1I 06210, A1I 06495, A1I 06575, A1I 06640, A1I 07130, A1I 07140, A1I 07855) 10 TPR (A1I 03535, A1I 00190, A1I 01380, A1I 02670, A1I 03165, A1I 03560, A1I 04380, A1I 05610, A1I 07295, A1I 07300) Ogata et al (2005a) Ogata et al (2006) NC Kb ND ND Blanc et al (2007a) NC Kb ND ND

14 14 Vicky Merhej and Didier Raoult Table 2 (Cont) Genes encoding eukaryotic-like motifs References Protein coding RPE and RR Genome size (Mb) G+C (%) % coding Genes Genome Refseq Ogata et al (2005a) NC Kb Ank (RF p14, RF p42) 3TPR(RF p16, RF p17, RF p18) Rickettsia felis URRWXCal2, plasmid prfdelta NC Kb ND ND Ogata et al (2005a) Rickettsia felis URRWXCal2, plasmid prf NC Kb ND ND Fournier et al (2009) Rickettsia africae ESF-5 plasmid praf Refseq is the NCBI genome reference; G+C, guanosine plus cytosine content; RPE, repeat palindromic elements; RR, rickettsial repeats; Ank, ankyrin repeat; TPR, tetratricopeptide repeat ND not determined rickettsial lineages leading to the existing species assemblage (Blanc et al, 2007b) Consequently, there is substantial variation in the level of genome decay and coding capacity (Table 2) While the genome of R massiliae shows a relatively low coding capacity (69%) (Blanc et al, 2007a), the genomes of R bellii and R felis exhibit higher values (84% and 83%, respectively) close to that of other symbiotic bacteria (81% for Wigglesworthia spp, 80% for Buchnera spp, 97% for Carsonella ruddii) (Akmanet al, 2002; Nakabachi et al, 2006; van Ham et al, 2003) Hence, variations among Rickettsia spp genomes may originate in accelerated genome degradation in species with a low coding capacity (Andersson & Andersson, 1999b) Many non-coding intergenic sequences in Rickettsia spp contain no recognizable ORFs but show similarity to short overlapping ORFs of other species (Amiri, Davids & Andersson, 2003; Frank, Amiri & Andersson, 2002; Ogata et al, 2001) From the high degree of colinearity between R prowazekii and R conorii (Fig 3), Ogata et al (2001) showed that 229 (41%) of the R conorii-specific ORFs have a significant sequence similarity in intergenic regions of R prowazekii genome Thus the intergenic non-coding regions in R prowazekii represent gene remnants that have been so extensively degraded that they are no longer recognizable as genes (Ogata et al, 2001) The accumulation of degraded genes in Rickettsia spp genomes has been accounted for by reductive genome evolution and small effective population sizes (Andersson & Kurland, 1998; Blanc et al, 2007b) It has been suggested that the transition to an intracellular lifestyle results in relaxation of purifying selection across many genes causing rapid accumulation of degraded genes and large deletions that encompass several loci (Andersson & Kurland, 1998; Moran, 1996; Wernegreen & Moran, 1999) The genomes of Rickettsia spp display elevated levels of protein sequence evolution mainly attributable to an increased background mutation rate rather than to modification of selective pressure (Blanc et al, 2007b) The critical factors that determine the retention or elimination of degraded genes remain elusive (Mira, Ochman & Moran, 2001) The reduced threat of genetic parasites in the protected intracellular environment may lower the genomic deletion rate, slowing the elimination of degraded genes (Lawrence, Hendrix & Casjens, 2001) Because of their physical sequestration within distinct hosts, the mechanism of rickettsial speciation has been considered allotropic with genetic isolation of the species Thus, lateral gene transfer was regarded previously as rather exceptional in Rickettsia with a limited contribution to evolution in this genus However, it was proposed recently that most degraded genes in Rickettsia spp originated from lateral acquisition (Fuxelius et al, 2008) In this scenario the presence of pseudogenes can be attributed to failed horizontal gene transfers (Liu et al, 2004) Some of the non-coding DNA regions appear more preserved than if they were experiencing neutral drift Because their degree of conservation is similar to the most conserved genes in Rickettsia spp, non-coding genes may be under evolutionary constraint and have some function (Jukes & Kimura, 1984; Ophir et al, 1999) These regions may include gene terminators, promoters and additional signals, including

15 Rickettsial evolution in the light of comparative genomics 15 Fig 3 Whole genome alignment of Rickettsia spp The figure was generated by the Mauve rearrangement viewer (Darling et al, 2004) It shows a linear representation of the genomes of R bellii RML369-C, R felis URRWXCal2, R conorii str Malish7, R prowazekii Madrid E, and R typhi str Wilmington The size of the horizontal bars corresponds to genome size (Kb) operators and transcriptional signals (Rogozin et al, 2002) Highly conserved intergenic spacers in eukaryotes (Dermitzakis et al, 2003; Kellis et al, 2003) have long been recognized as good candidates for functional regions (Hardison, Oeltjen & Miller, 1997; Pennacchio & Rubin, 2001), and several have been confirmed as gene regulatory sequences (eg Balakirev & Ayala, 2003; Hirotsune et al, 2003; Loots et al, 2000) Evidence suggests that the non-coding DNA regions in Rickettsia spp appear to be more than just genetic waste regions and future genomic studies may explain the frequent occurence of these non-coding spacers relative to the small size of rickettsial genomes (3) Genome degradation process Comparisons between the genomes of R conorii and R prowazekii have allowed recognition of the different steps of genome degradation, ie from split genes to pseudogenes and gene remnants (Ogata et al, 2001) A split gene is a gene fragmented into several pieces of ORFs by internal stop codons It is generated through the processes of mutation, such as the introduction of a termination codon, creation of a frameshift in the gene, or deletion At this stage, the short remaining ORFs can still be recognized because of sequence similarity over the full-length ortholog in another species The stop codons may occur in polycistronic regions that allow the conservation of one of the enzymatic activities, thus the ORFs can be transcribed and translated and may retain some functions Alternatively, the damage may be repaired and the gene restored This is the case for genes with poly(a) tracts in the small genomes with low G+C content typical of intracellular genomes, where polymerase infidelity at poly(a) tracts rescues the functionality of genes with frameshift mutations (Tamas et al, 2008) By contrast, the reductive evolution process may lead to gradual degradation rather than gene recuperation, giving rise to a pseudogene Pseudogenes have been defined as the molecular remains of broken genes that have accumulated debilitating mutations over time that render them incapable of functioning (Balakirev & Ayala, 2003) They are usually identified as several short consecutive ORFs exhibiting sequence similarities to functional genes but interrupted by premature stop codons (Gerstein & Zheng, 2006) The gene degradation process may continue, giving rise to a gene remnant Finally, the sequence corresponding to the gene may be eliminated (Fig 4) When the process of nonfunctionalization concerns a crucial gene within a complex pathway or network, this gradual genedeath sequence of events may eventually trigger a mass gene extinction of the dependent genes according to the domino theory for gene death (Dagan, Blekhman & Graur, 2006; Moran & Mira, 2001) For example, R typhi is missing the entire cytochrome c oxidase pathway (McLeod et al, 2004) It is noteworthy that automatic annotation techniques have failed to recognize some remnants; computational