Phylogeny of the class Actinobacteria revisited in the light of complete genomes. The orders

12 Pages·2014·0.37 MB·English

InternationalJournalofSystematicandEvolutionaryMicrobiology(2014),64,3821–3832 DOI10.1099/ijs.0.063966-0
Phylogeny of the class Actinobacteria revisited in
the light of complete genomes. The orders
‘Frankiales’ and Micrococcales should be split into
coherent entities: proposal of Frankiales ord. nov.,
Geodermatophilales ord. nov., Acidothermales ord.
nov. and Nakamurellales ord. nov.
Arnab Sen,1 Vincent Daubin,2 Danis Abrouk,3 Isaac Gifford,4
Alison M. Berry4 and Philippe Normand3
Correspondence 1NBU Bioinformatics Facility, Department ofBotany, University ofNorth Bengal, Siliguri, 734013,
PhilippeNormand India
[email protected]. 2Biome´trie etBiologie Evolutive, Centre Nationalde la RechercheScientifique UMR 5558,
Universite´ LyonI,Universite´ Lyon, Villeurbanne, France
3EcologieMicrobienne,CentreNationaldelaRechercheScientifiqueUMR5557,Universite´ LyonI,
Universite´ Lyon,Villeurbanne, France
4DepartmentofPlantSciences,UniversityofCalifornia,OneShieldsAvenue,Davis,CA95616,USA
The phylogeny oftheclass Actinobacteria remainscontroversial, essentiallybecause it isvery
sensitivetothechoiceofdatasetandphylogeneticmethods.Weusedatestproposedrecently,
based on complete genome data,which chooses amongcandidate species phylogenies based
onthenumberoflateralgenetransfers(LGT)neededtoexplainthediversityofhistoriesamong
genetreesforasetofgenomes.Weused100completelysequencedgenomesrepresenting35
familiesand 17ordersofthe classActinobacteria andevaluated eight different hypotheses for
theirphylogeny,includingonebasedonaconcatenateof54conservedproteinspresentinsingle
copy inall thesegenomes, trees based on 16Sand23S rRNA genesequences ortheir
concatenation, anda treebased on the concatenation ofMLSA genes (encoding AtpI,GyrA,
FtsZ,SecAandDnaK).WeusedPruniertoinferthenumberofLGTin579proteins(differentfrom
those usedto buildthe concatenated tree) present inatleast 70species,using the different
hypothetical species trees as references. The best tree, with thelowest number oflateral
transfers, wastheone based on the concatenation of54proteins. Inthat tree, theorders
Bifidobacteriales,Coriobacteriales,‘Corynebacteriales’,‘Micromonosporales’,
‘Propionibacteriales’,‘Pseudonocardiales’,Streptomycetales and‘Streptosporangiales’were
recoveredwhiletheorders‘Frankiales’andMicrococcaleswerenot.Itisthusproposedthatthe
order‘Frankiales’,whichhasaneffectivelybutnotvalidlypublishedname,besplitintoFrankiales
ord. nov.(type familyFrankiaceae),Geodermatophilales ord.nov. (Geodermatophilaceae),
Acidothermales ord.nov. (Acidothermaceae) andNakamurellalesord. nov.(Nakamurellaceae).
The order Micrococcalesshould alsobesplit intoMicrococcales (genera Kocuria,Rothia,
Micrococcus, Arthrobacter, Tropheryma, Microbacterium, Leifsonia andClavibacter),
Cellulomonales (Beutenbergia, Cellulomonas,Xylanimonas, Jonesia andSanguibacter) and
Brachybacteriales(Brachybacterium)buttheformalproposalforthiswillhavetowaituntilmore
genomes become available fora significant proportion ofstrains inthisorder.
Abbreviations: DPG, diphosphatidylglycerol; LGT, lateral gene transfer; LL-DAP, LL-diaminopimelic acid; MLSA, multi-locus sequence analysis; MN,
multicopy non-universal;MU, multicopyuniversal;PE,phosphatidylethanolamine; PG, phosphatidylglycerol;PI, phosphatidylinositol;PIM, phosphatidyl-
inositolmannoside;UN,unicopynon-universal;UU,universalunicopy.
Threesupplementarytablesandfivesupplementaryfiguresareavailablewiththeonlineversionofthispaper.
063966G2014CNRS PrintedinGreatBritain 3821
A.Senandothers
INTRODUCTION filaments with the exception of what are called the
deeper branches, namely the members of the order
Ray fungi, called Strahlenpilze by Lieske (1921), or
Bifidobacteriales (Woese, 1987). This treatment was
actinobacteriaastheyarenowdesignated,werehistorically
retained in many successive editions of Bergey’s Manual
considered as ‘intermediary’ between fungi and bacteria
and the class name was published in 1997 (Stackebrandt
with a difficult taxonomy (Krassilnikov, 1941). The name
et al., 1997) although invalidly because no type order had
itself comes from their colony morphology on agar that
been proposed at the time (Euze´by & Tindall, 2001).
exhibits radial growth, a characteristic shared with fungi.
They also share with fungi their mycelial shape and musty However, 16S rRNA genes and the resulting phylogenetic
smell, and with bacteria physiological traits such as a cell- reconstructions cannot be considered the golden standard
wall structure based on peptidoglycan, in common with of bacterial taxonomy for several reasons. First, there are
membersofthephylumFirmicutesasrevealedbytheGram instancesofspecieswhosegenomescontainmorethanone
stain.Thefirsttaxonomicaltreatmentofactinobacteriawas copyofthe16SrRNAgenedifferingbyasmuchas6%,as
that of Buchanan (1917) who had proposed to name an in Thermomonospora chromogena (Yap et al., 1999), which
order Actinomycetales, although with features of debatable wasinterpretedasevidenceoflateraltransfer.Secondly,the
validity(‘Moldlike,nottypicallywaterforms,norwiththe 16S rRNA gene has also been shown to be plasmid-borne
sheath impregnated with iron, true branching often in Bacillus megaterium (Kunnimalaiyaan et al., 2001),
evident...’). A few taxonomical works have been based on furthersupportingtheideaitcouldbetransferredlaterally.
aperceived evolutionary trendfromsimpleformsto more Finally, the 16S rRNA gene is a single marker that thus
complexfeaturessuchashyphaeandsporangia(Kluyver& contravenes one basic tenet of biology, that any analysis
van Niel, 1936). However many authors have argued shouldbeassessedforreproducibility.Thislimitationledto
that the search for a reliable taxonomy was vain given multi-locus sequence analysis (MLSA) developed initially
the numerous instances of loss of a function or of a for the genus Neisseria (Maiden et al., 1998), where
morphologicalfeaturethatwouldresultinfaultyposition- typically five conserved genes or more are sequenced. This
ing, especially in classifications essentially based on a approach is now frequently used, mostly to characterize
dichotomous scheme. specieswithinagenus. Ithasledtothere-characterization
of the genera Nocardia (McTaggart et al., 2010) and
Chemical criteria were initially too few to provide a solid
Streptomyces(Doroghazi&Buckley,2010),withsignificant
taxonomy,yettheyhaveaccumulated,providingeventually
deviations from the topology yielded by 16S rRNA gene
a solid set of data that could compensate for a few lost
sequences, indicative of widespread recombination.
functions.Nowadays,todescribeaspecies,itisrequiredto
provide a measure of the DNA G+C content that is a The most recent taxonomical treatment of the phylum
characteristic of the genome; an analysis of the phospho- Actinobacteria (Ludwig et al., 2012a, b), based essentially
lipidsthatarecharacteristicofthemembrane;thediamino on the 16S rRNA gene phylogeny, has considerably modi-
acids that are a characteristic of the cell wall; or the fied all levels of their taxonomy. There are now six classes
quinones that are a characteristic of the respiratory chain. comprising five basal ones (Acidimicrobiia, Coriobacteriia,
None of these elements taken individually is sufficient to Nitriliruptoria, Rubrobacteria and Thermoleophilia) each
identify a microbe; however, taken as a whole they can having only one or two orders, and the main class
yield a solid taxonomic basis in combination with Actinobacteriathatcomprises15orders(Actinomycetaleswith
morphology and growth characteristics. Sokal & Sneath 1 family-5 genera, ‘Actinopolysporales’ 1-1, Bifidobacteriales
(1963) used a combination of such characters as the basis 1-7,‘Catenulisporales’2-2,‘Corynebacteriales’6-13,‘Frankiales’
of numerical taxonomical treatments. These approaches 6-11, ‘Glycomycetales’ 1-3, ‘Jiangellales’ 1-2, ‘Kineosporales’
have permitted the classification of many actinobacterial 1-3,Micrococcales15-92,‘Micromonosporales’1-23,‘Propioni-
lineages starting in 1967 (Goodfellow, 1967), in particular bacteriales’2-18,‘Pseudonocardiales’1-22,Streptomycetales1-3
the genera Streptomyces (Williams et al., 1983) and and‘Streptosporangiales’3-22).
Mycobacterium (Tsukamura et al., 1969).
Complete genomes were initially obtained to gain a
The advent of molecular phylogenetic tools, 16S rRNA comprehensive understanding of the functioning of cells.
gene cataloguing (Stackebrandt & Woese, 1981) and later The first bacterial genome was published in 1995
16S rRNA gene sequencing crowned the search for a (Fleischmann et al., 1995) and the rapid increase since
molecular clock that needed to be present in all lineages, then of sequencing capabilities has resulted in the present
retainedthesamefunctionandtickedatcomparablespeed (December 2013) tally of 4536 published finished bacte-
in all these lineages. These techniques resulted in the rialgenomes(http://www.ncbi.nlm.nih.gov/genome/?term=
actinobacteria emerging as a coherent subdivision in the bacteria), of which 16% or about 731 are classified in the
phylum ‘Gram-Positive bacteria’, that also contained the class Actinobacteria. The first study that aimed to usethese
low G+C firmicutes with the genera Bacillus and completegenomestorevisitbacterialphylogenyresultedin
Clostridium (Woese, 1987). This confirmation buttressed marked changes relative to the 16S rRNA gene-based
the class Actinobacteria as coherent, with short genetic phylogeny,especiallyintheclassActinobacteriawheremany
distances, a generally aerobic metabolism and branched orderswereshowntobeunsupportedbyaconcatenateof31
3822 InternationalJournalofSystematicandEvolutionaryMicrobiology64
PhylogenyofActinobacteriabasedoncompletegenomes
universal proteins. Gao and Gupta (2012) used a similar genomes were clustered into one file on which reciprocal BLAST was
approach with a concatenate of 35 proteins to re-examine performed with BLASTP of blastall version 2.2.25, with Block
SubstitutionMatrix(BLOSUM62)andwherecostandgapextension
actinobacterial phylogeny, adding a study of indels to
values [G & E] were set to 11 and 1, respectively with an e-value of
confirm the main lines of that topology (Wu et al., 2009),
1e204. The number of database sequences to show one-line
an approach also followed specifically for the class descriptions [-v] was kept at 10000 and the number of database
Actinobacteria that showed the orders ‘Frankiales’ and sequencestoshowalignments[-b]wassetat10000.
Micrococcales were in need of reassessment (Verma et al.,
The output of BLAST was subjected to clustering of homologous
2013).Beyondsubsamplingsuchasthebootstrapapproach
sequences using the SiLiX software (http://lbbe.univ-lyon1.fr/SiLiX;
(Felsenstein, 1985), it is nevertheless what the Prunier
Mieleetal.,2011)withtheminimumpercentidentitytoacceptBLAST
approachisdesignedtodo,usingalargenumberofproteins hits for building families [–ident] set at 35, the minimum per cent
toassessthereliabilityofnodesoftopologiesobtainedwith overlaptoacceptBLASThitsforbuildingfamilies[–overlap]setat80,
otherproteins(Abbyetal.,2010).Itwasdecidedtousethis theminimumlengthtoacceptpartialsequencesinfamilies[-l]setat
approach on 100 actinobacterial genomes representing the 100andtheminimumpercentoverlaptoacceptpartialsequencesin
families[-m]setat50.
mainactinobacteriallineages.
In this process several gene families were generated, which were
classified into four groups: 1) UU families (unicopy universal
METHODS families, containing genes present in one copy in each genome and
present in all the 100 genomes); 2) UN families (unicopy non-
Selectionofgenomes.Aminoacidsequencesofallprotein-coding universalfamilies,sameasabovebutnotpresentinallgenomes);3)
genes as well as DNA sequences of 16S and 23S rRNA genes were MUfamilies(multicopyuniversalfamilies,presentinallgenomesbut
downloaded from the JGI-IMG (http://img.jgi.doe.gov/cgi-bin/w/ as more than one copy), and 4) MN families (multicopy non-
main.cgi) and Genoscope (http://www.genoscope.cns.fr/spip) data- universal families, more than one copy and not present in all
bases. One hundred actinobacterial genomes were chosen for the genomes). An in-house python (this study) program was developed
study and care was taken to include most of the orders of the class whichshort-listedthegenefamiliesintovariousgroups.
Actinobacteria.Alistofallgenomesusedinthepresentstudyandthe
Therewereultimately54UUfamiliesandaround5000UNfamilies
resulting statistics are presented in Table S1 (available in the online
recovered. Since the number of UN families was fairly large, the
SupplementaryMaterial).
numberoffamilieswasdeterminedbytakingonlythosefamiliesthat
were present in more than 70 genomes but not in all 100 studied
Species biotopes. Actinobacteria can thrive under various
genomes. This permitted the selection of 579 UN families. The 54
environmental conditions and ecological niches. Based on the
gene UU families were concatenated with Seaview version 4.3.1 and
predominant lifestyle for a given taxon, the selected actinobacterial
aligned with software CLUSTAL Omega version 1.1.0 (Sievers et al.,
genera were assigned to seven different biotopes. These are water,
2011) with default options. A tree was reconstructed with RAxML
plant, thermal, mammal, soil, arthropods and extremophiles (other
usingthisalignedfilewiththeparametersdescribedabovebutwitha
than thermal). A letter code was assigned for each biotope that was
substitution model [m] of PROTCAT_GAMMAIWAG -f a (Fig. 1).
laterusedinthephylogenetictrees.
Wealsoreconstructedtreesusinganotherwidelyusedprogram,MEGA
with default parameters (Tamura et al., 2013), except the WAG
Reconstructionofphylogenetictrees.Eightdifferentphylogenetic
substitution model and Gamma Distributed as in the RAxML tree
trees were used as reference to infer lateral gene transfer (LGT): 1)
(Fig. S4); and MrBayes (Bayesian Analysis of Phylogeny) with the
based on the 16S rRNA genes, 2) based on the 23S rRNA genes, 3)
GTRmodelandgamma-distributed ratevariationacrosssitesanda
based on concatenation of 16S and 23S rRNA genes, 4) based on
proportionofinvariablesites(Ronquistetal.,2012)(Fig.S5).
concatenationoffiveMLSAgenes(encodingAtpI,GyrA,FtsZ,SecA
andDnaK)and5)anartificialtreebasedonthepresentclassification To make an MLSA amino acids tree (Margos et al., 2009),
ofBergey’sManual.AllthesetreeswerereconstructedusingRAxML housekeepinggenesindifferentgenefamiliesweresearched:onegene
(Randomized Axelerated Maximum-likelihood; see below). Finally, (encodingmembrane-boundATPsynthase,F1sector,AtpI)wasfound
three trees based on concatenation of 54 proteins universal in inUUfamiliesandfourgenesinMUfamilies(encodingDNAgyrase,
members of the class Actinobacteria were reconstructed, 6) with GyrA, tubulin-like GTP-binding protein, FtsZ ATPase secretory
maximum-likelihood in MEGA software with a distance-based initial preprotein translocase, SecA and chaperone Hsp70 in DNA biosyn-
treeinMEGA,7)withMrBayes,and8)withRAxML. thesis, DnaK). TochooseagenefromeachstrainfromMUfamilies
(genesofthesetypeoffamiliespresentinallgenomesbutasmorethan
Maximum-likelihood phylogenetic analyses were performed with
onecopy),thegeneofAtopobiumparvulum (DSM20469)wastakenas
RAxML 7.2.8 to reconstruct 16S and 23S rRNA gene trees (Figs S1
and S2, respectively) keeping the number of bootstraps [N] at 1000 reference and blasted with the rest of the genes of the same family.
and using the substitution model [m] GTRGAMMAI [General Time Ultimately, five MLSA gene families were used to make a tree in a
mannersimilartothatdescribedforthe‘concatenated’tree(Fig.2).
ReversiblemodelofnucleotidesubstitutionwiththeCmodelofrate
heterogeneity (Yang, 1993)and estimate ofproportion ofinvariable The last tree, which was called ‘artificial’, was reconstructed on the
sites]. basis of 16S rRNA genes and conventional knowledge that
corresponds to the latest taxonomical treatment in Bergey’s Manual
The 16S and 23S rRNA genes were concatenated using Seaview
of Systematic Bacteriology (2nd edition). There were actually some
version 4.3.1. (Gouy et al., 2010). The concatenated sequences were
deviations in the 16S rRNA gene tree (Fig. S1) generated in the
thenalignedwithsoftwarepackageMUSCLE(Edgar,2004a,b).RAxML
present study from Bergey’s taxonomy and those changes were
7.2.8 was then used to reconstruct the tree (henceforth to be
artificiallycorrectedinthe‘artificial’tree.
designated ‘16S–23S’ tree) (Fig. S3) with the above-mentioned
parameters.
Prunier.To identify and quantify putative LGTs among the strains
To make the multi-protein-coding-gene concatenated tree (hence- and their evolutionary role, and to determine the robustness of
forth called ‘concatenated-RAxML’ tree), all the genes of all studied varioustrees,aprogramcalled‘Prunier’(Abbyetal.,2010)wasused.
http://ijs.sgmjournals.org 3823
A.Senandothers
Frankia Datisca[P]
100 100 Frankia Eul1c[P]
100 100FFrraannkkiaia a ClnNi[3P[]P] (‘Frankiales’-I)
100 Frankia Ccl3[P]
Frankia EAN1pec[P]
Stackebrandtia nassauensis[S] (‘Glycomycetales’)
100 Actinoplanes missouriensis[S]
100 100Micromonospora aurantiaca[S] (‘Micromonosporales’)
96 Salinispora tropica[W]
Verrucosispora maris[W]
Pseudonocardia dioxanivorans[W]
100 Actinosynnema mirum[S]
100100 100 SAamccyhcaorloaptooplyssisp omrae deirteytrhraraneeai[[SS]] (‘Pseudonocardiales’)
Saccharomonospora viridis[M]
100 Segniliparus rotundus[M]
Mycobacterium massiliense[M]
100 100Mycobacterium gilvum[S]
100 100 100 100MMycyocboabcatcetreiurimum va innbtraaacleelnluiil[aSre][M]
100 Mycobacterium tuberculosis[M]
Mycobacterium leprae[M]
75 100 NoAcamrdyicao flaicriccionciccau[sM s]ubflavus[S]
100 100 100100100R1h0o0dRoRhcohodocodccouccsoc ecoqccuucisu[ Mse r]oyptharcoupso[Slis][S] (‘Corynebacteriales’)
Rhodoccoccus sp.[S]
58 100 Corynebacterium jeikeium[M]
100 Corynebacterium resistens[M]
100 100 CCooryrnyneebbaaccteteriruiumm a vuarirmiaubciloes[Pu]m[M]
76 83 100 100CoCroynryenbeabcatecrteiurmiu mg luutlcaemriacnusm[M[M]]
Corynebacterium diphtheriae[M]
100 Tsukamurella paurometabola[A]
Gordonia polyisoprenivorans[W]
100100 MoGdeeosdtoeNbrmaackatateomrp umhriaelurllisan uomsbu[sElct]iupraursti[tEa[]W] ((‘‘FFrraannkkiiaalleess’’--IIII)I)
Blastococcus saxobsidens[E]
100 10K0ocurRiao rthhiiazo dpehniltao[cSa]riosa[M]
100 Rothia mucilaginosa[M]
Micrococcus luteus[W]
84 100Arthrobacter aurescens[S]
100 99 100 AArrtthhrroobbaacctteerr cphhleonroanpthherneonliivcoursa[nSs][S] (Micrococcales-I)
Arthrobacter arilaitensis[S]
93 ‘Tropheryma whipplei’[M] I
100 100 100 MicrCoblaavcibtearciutemr mteiscthaicgeaunmen[Psi]s[P]
100 100 CelluBloemutoennabse rfglaiavi Lgceeainvfseaor[nSnai]ae [xSy]li[P]
101000 100 XylanimonaJso cneelsluialo dseilnytiticriafi[cPa]ns[M] (Micrococcales-II)
Sanguibacter keddieii[M]
Bifidobacterium asteroides[A]
100 100 Gardnerella vaginalis[M]
100 Bifidobacterium animalis[M]
100 100 Bifidobacterium dentium[M] (Bifidobacteriales)
100 99 100Bifidobacterium longum[M]
100 Bifidobacterium breve[M]
100 52 95 AMrcoabniolubnaccutesr icuBumritf iihsdaioie[bMma]cotleytriicuumm b[Mifi]dum[M] (Actinomycetales)
Brachybacterium faecium[M] (Micrococcales-III)
54 100 lKntirnaesopcooracncKugysituo rmcaod ccioacltuvosule msrea[Snds]e[nEt]arius[W] ((M‘Kiicnreoocsopcocraialeless-’I)V)
100 Nocardioides sp.[S]
100 100 100 KribbellaM fliacvriodlau[nSa]tus phosphovorus[W] (‘Propionibacteriales’)
100 Propionibacterium acnes[M]
Propionibacterium freudenreichii[S]
Catenulispora acidiphila[S] (‘Catenulisporales’)
Streptomyces cattleya[S]
100 100 100 Streptomyces griseus[S]
91 100 10809SStrterepptotommyycceess c socealibcioelio[Pr[]S] (‘Streptomycetales’)
100 Streptomyces avermitilis[S]
‘Streptomyces bingchenggensis’[S]
Kitasatospora setae[S]
Acidothermus cellulolyticus[T] (‘Frankiales’-IV)
91 100 100NNooccaardrdioioppssisis d aalsbsao[Anv]illei[M]
100 100 TThheerrmmoobbiisfipdoa rfau sbcisap[To]ra[T] (Streptosporangiales)
100 Streptosporangium roseum[S]
Thermomonospora curvata[T] Acidimicrobium ferrooxidans[T] (Acidimicrobiales) II
100 100 Conexibacter woesei[S] – (Solirubrobacterales) VI
Rubrobacter xylanophilus[T] – (Rubrobacterales) V
100 100 Olsenella uli[M]
Atopobium parvulum[M]
0.3 100 100 Cryptobacterium curtum[M] (‘Coriobacteriales’) III
100 Eggerthella lenta[M]
Slackia heliotrinireducens[M]
Biotopes of various strains of actinobacteria
Plant Water Soil Thermal Mammal Extreme Arthropods
[P] [W] [S] [T] [M] [E] [A]
3824 InternationalJournalofSystematicandEvolutionaryMicrobiology64
PhylogenyofActinobacteriabasedoncompletegenomes
Fig.1.Phylogenetictreereconstructedwith54UUgenefamiliesandtheRAxMLprogramforall100actinobacterialstrains.
The colour code indicates the biotopes of the sequenced genome. Numbers at nodes are the bootstrap results for 1000
subsamplingsexpressedasapercentage.Bar,0.3substitutionspersitedistance.
Prunieressentiallyreconcilesagenetreewithareference(species)tree othertheordersMicrococcales,Bifidobacteriales,Actinomyce-
and searches for a maximum statistical agreement forest (MSAF) tales, ‘Propionibacteriales’, ‘Catenulisporales’, Streptomyce-
betweenthem.The‘fast’PrunierusedinthepresentstudyusesLR-
tales, ‘Streptosporangiales’ and families Dermabacteraceae,
ELW edge support values (Strimmer & Rambaut, 2002) to identify
Kineosporiaceae, Dermacoccaceae, Intrasporangiaceae and
discrepancies between gene and species trees. Prunier infers LGT
Acidothermaceae. The genus Acidothermus was found to
eventswhenitfindssignificanttopologicalconflictsbetweenthetwo
trees.ThefollowingsettingswereusedforPrunier:boot.thresh.con- groupcloselytothermophilesintheorder‘Streptosporangi-
flict590 (i.e. support value threshold for topological conflict); ales’,supportedbyagood(91%)bootstrapresult(Fig.1).
fwd.depth52 (i.e. maximal depth at which Prunier looks forward
AllordersdescribedinthelatestversionofBergey’sManual
to find a significant LGT when the current LGT is not significant).
Multi_root.bool [5true] (this indicates Prunier should compute an (Ludwig et al., 2012b), were recovered with the exceptions
LGT scenario for each possible root and was used because the real of the orders ‘Frankiales’ and Micrococcales.
rootsofthetreeswerenotcertain).
In the order ‘Corynebacteriales’, the genera Corynebac-
The ‘boot.thresh.conflict’ denotes the lowest LR-ELW edge support terium,RhodococcusandMycobacteriumwereallrecovered.
value for a given node in the gene tree used for recognizing its
In the order Bifidobacteriales, the genus Gardnerella was
topologicalconflictspeciestree,whereasthe‘fwd.depth’standsforthe
not outside the genus Bifidobacterium but within it. All
maximal depth at which the Prunier (fast) looks forward to find a
other genera with more than one genome studied were
significantLGTwhenthecurrenteventisnotsignificant.Thedepth
value was set at 2, which means that if significantly supported recovered as coherent.
disagreementcannotberemovedwithoneLGTevent,Prunierwillgo
two steps (one at a time) ‘forward’ to determine whether the next
Topology of the MLSA tree
LGT in the list will remove a significant conflict. If Prunier finds a
bettersolutionwithadepthvalueof2or3,itprovidesthatsolution
ThetopologyoftheMLSAtree(Fig.2)issimilartothatof
intheoutput.ForfurtherdetailsaboutPrunierpleaserefertohttp://
the concatenated tree with the order Coriobacteriales
pbil.univ-lyon1.fr/software/prunier.
clustered with the genera Conexibacter, Rubrobacter and
Acidimicrobium, then the family Frankiaceae at the root.
The position of the genus Acidothermus is different, being
RESULTS
here close to the root.
Eight trees were generated using different genes or gene
The two sets seen with the concatenated tree (I:
concatenates with contrasted topologies, two of which will
‘Glycomycetales’, ‘Micromonosporales’, ‘Pseudonocardiales’,
be discussed in detail.
‘Corynebacteriales’ Nakamurellaceae, Geodermatophilaceae;
II:Micrococcales,Bifidobacteriales,Actinomycetales,‘Propioni-
Topology of the concatenated-RAxML tree bacteriales’, ‘Catenulisporales’, Streptomycetales, ‘Streptospor-
angiales’,Dermabacteraceae,Kineosporiaceae,Dermacoccaceae,
AllthetreesgeneratedwererootedfollowingGao&Gupta
Intrasporangiaceae) were recovered in the MLSA tree except
(2012) and Embley & Stackebrandt (1994) with the order
for the family Acidothermaceae that was at the root
Coriobacteriales including Atopobium parvulum (DSM
immediatelyaftertheFrankiaceaebranch.
20469) that has an obligately anaerobic metabolism, as
doallmembersoftheclade(Fig.1).Thebranchleadingto All genera with more than one genome analysed (Arthro-
Conexibacter woesei and Rubrobacter xylanophilus, both of bacter, Bifidobacterium, Corynebacterium, Frankia, Mycobac-
which are aerobic, emerges next. terium, Nocardiopsis, Propionibacterium, Rhodococcus, Rothia,
Streptomyces) were recovered as coherent with the excep-
ThenextbranchtoemergeisthatofthegenusFrankiawith
tion of the genus Bifidobacterium that contained the genus
cluster 2 at the root, followed by cluster 4, cluster 3 and
Gardnerella.
then clusters 1a and 1b as most derived. The other genera
traditionally grouped with the genus Frankia within the
‘Frankiales’-Geodermatophilus,Modestobacter,Blastococcus, Quantification of LGT
Acidothermus and Nakamurella were not anywhere close
Revision of phylogeny of organisms, especially bacteria
to the genus Frankia. Bootstraps were high that posi-
using complete genome sequences reveals a degree of
tionedthesefamiliesingroupsawayfromthegenusFrankia
incongruence due to biological or methodological differ-
(Figs1&2).
ences(Boussau&Daubin,2010).Thereforeitisimportant
Thenemergesadichotomywiththeorders‘Glycomycetales’, to verify any phylogenetic tree generated. One relatively
‘Micromonosporales’, ‘Pseudonocardiales’, ‘Corynebacteriales’ reliable method of verification is the detection of LGT
and the families Nakamurellaceae and Geodermatophilaceae based on the search for a maximum statistical agreement
formerly in the ‘Frankiales’ on the one hand and on the forestbetweenagenetreeandareferencetree(Abbyetal.,
http://ijs.sgmjournals.org 3825
A.Senandothers
10045626F2raFFnrarkaniank kiDaia aa tElinsAci NaA[1CPp]Ne1cF4[rPaan][Pki]a Ccl3[P] (‘Frankiales’)
100 FFrraannkkiiaa CENul31[cP[P]]
100AcidothermusK cietallCsualaottoleysntpiucoluirsasp [sToe]rtaa eac[Sid]iphila[S] ((‘‘FCraatnekniualleissp’)orales’)
100 10010S100tr0e1p0S6t0o0trmeSSSpyttttcrrroeeeempppstt tyoococmmametyytsylcc ecgeeyerssasis [ caSesovcu]eeasrlb[imScieo]itili[olPisr][[SS]] (‘Streptomycetales’)
100 100 ‘Streptomyces bNinogccahredniogidgeesn ssips’. [[SS]]
100 100 100 100 Microlunatus PprhPoorposipophniooibnviaobcratucestre[iuWrmiu]m fr eaucdneensr[eMic]hii[S] (‘Propionibacteriales’)
Kribbella flavida[S]
100 100 IntrasKpyotroacnogciucums csaeldvuemnt[aSri]us[W] (Micrococcales)
Kineococcus radiotolerans[E] (‘Kineosporiales’)
Bifidobacterium asteroides[A]
100 Bifidobacterium animalis[M]
100 60 Gardnerella vaginalis[M]
82 Bifidobacterium dentium[M] (Bifidobacteriales)
40 100 2140 1B00ifidBBobiiffaiiddcootebbraaiuccmttee rrbiiuuifmmid ulbomrne[gvMeu][mM[]M]
32 83 921007040 71100 BXSeyluaatnneginmubioebnraJgacoistane cecr esaklivelaued ldrodnesiaenileiyiit[[trSMiicf]i]ac[aMPn]osAb[Mriclu]anncoubsa ccuterrtiiusimi[M h]aemolyticum[M] (Acidimicrobiales)
Cellulomonas flavigena[S]
100 100 100100 100 LeMifiscoKCronolabicavau ixcbryitaaelic [rrPithue]imrz om tpeichshitlaai‘gcT[Saeronu]pemnh[sePirs]y[mP]a whipplei’[M] (Micrococcales)
100 100 Rothia mucilaginosa[M]
100 Rothia dentocariosa[M]
Arthrobacter arilaitensis[S]
33 61 100 100100 AABrrttrhharrcoohbbyaabcca1tteec0rrt0 ecprhhiuleAomnrro atfhnparethohcebrineuaoncmlitive[cMoru_r]saa[nuSsr]e[Ss]cens[MS]icrococcus luteus[W]
10080 101000TheTrhmeormm1oo0nb0oisspNNpoooorccraaaa b rrcddiusiiooprvppoassrStaiiass[t[Tr Tdae]]lpabstaos[soApn]ovriallenig[Miu]m roseum[S] (Streptosporangiales) I
Thermobifida fusca[T]
100 10400MVSeiacrlrirnouimcsoposonirsoaps toprrooapr amic aaau[rWirsa[n]Wtia]ca[W] (‘Micromonosporales’)
100 Actinoplanes missouriensis[S]
Stackebrandtia nassauensis[M] (‘Glycomycetales’)
90 10060 MBodlGaesNestootaodckbeoaarmcmccutuaertsero lmslpaaha xmirloiunubsulst sioipd[bEaesr]ntcistua[rE[uW]s[]E] (‘Frankiales’)
83 100100107M00yMcMo1ycb0yoMa0cbcoytaMcbecoarytibceuctramoeicubr tmimaeucrm aittusue smibrniieult irlremeacnpcu svrelaeaolle[nsuM[iblMsa]a[rM]ael[]eMn]ii[S] Mycobacterium gilvum[S]
Segniliparus rotundus[M]
100 100 77 10010046 100 CCCoCorroyoyrnrnyyeennbbeeaabbccaattcecetrtreieiuurrimimuum m du i lgpaclhueutrrthiaamenmrusiia[ccMeou[s]mMu[]mM[]M] (‘Corynebacteriales’)
86 79 100 100 CCoroCyrnyoenrybenabecabtceatreciurtiemurmi ur emjes iivskateeriiunamsb[iM[lMe][]P]
69 88 80 NocarAdmiaGy fcoaTorrsdcliucionkinciacioamac [pucMorue]lsyll iass uopbpafrulearnvoiumvsoe[rStaa]nbso[Wla[A]]
78 90100 1R10h00o0RdohcoRRodhhcoooccdduooosc ccceoouqccsucc ieuu[Mrssy ]tsohppro.a[pScou]lsis[S[S]]
48 Saccharopolyspora_erythraea[S]
98 100 AScactinchoasryonmneomnoas mpoirruam v[iSrid]is[M] (‘Pseudonocardiales’)
Amycolatopsis mediterranei[S]
100 Pseudonocardia dioxaniv1o0ra0nsA[Wcid]imicrobium ferro1o0x0idans[TC]ryptobAOatcoltspeeornibuemilulam c uu plria[tuMrvm]u[lMum][M] ((A‘Ccoidriiombiaccrotebriiaalleess)’) IIIII
0.2 100 90 100 80ConexibacEtegr gweSorltaehcsekelliiaa[S lhe]enltioa[tMrin]ireducens[M] (Solirubrobacterales) VI
Rubrobacter xylanophilus[T] (Rubrobacterales) V
Biotopes of various strains of actinobacteria
Plant Water Soil Thermal Mammal Extreme Arthropods
[P] [W] [S] [T] [M] [E] [A]
Fig.2.PhylogenetictreereconstructedwithfiveMLSAgenefamiliesandtheRAxMLprogramforall100actinobacterialstrains.
The colour code indicates the biotopes of the sequenced genomes. Numbers at nodes are the bootstrap results for 1000
subsamplingsexpressedasapercentage.Bar,0.2substitutionspersitedistance.
3826 InternationalJournalofSystematicandEvolutionaryMicrobiology64
PhylogenyofActinobacteriabasedoncompletegenomes
2012). As described above, two sets of gene trees were themostreliableone.Concerningthepositionoftheroot,
generated, one set of 54 gene trees with UU families and statistically negligible differences of transfer were found
anothersetof579gene treeswithUMfamilieswithLGTs. among various roots within the trees. It was therefore
Gene trees were then simulated from various reference decided to continue to consider the genus Atopobium in
trees by using the subtree pruning and regrafting opera- the Coriobacteriales clade as root for interpretation of the
tions as described by Abby et al. (2010). Precisely, eight trees.
different reference trees were used, the concatenated-
RAxML tree, the concatenated-MEGA tree, the concate-
nated-Bayesian tree, the MLSA tree, the 23S tree, the 16S– DISCUSSION
23S tree, the 16S tree and the tree based on conventional
knowledge [an artificial tree that corresponds to the latest Actinobacteria constitute a major group of bacteria,
exceeded only by the phyla Proteobacteria and Firmicutes
taxonomical treatment in Bergey’s Manual of Systematic
Bacteriology (2nd edition)]. inthenumberofavailable completegenomes (http://www.
genomesonline.org/cgi-bin/GOLD/phylogenetic_distribution.
ThehighestnumberoftransfersinvariousUUgenefamily cgi; Pagani et al., 2012). The taxonomic treatment of the
trees ranged from 69 to 375 whereas in MU gene family actinobacteria has fluctuated considerably due to technical
trees, it ranged from 19 to 39 (for details see Table 1). developments, the most recent being the phylogenetic
analysis based on complete genomes (Wu et al., 2009),
It was found that the concatenated tree (RAxML) had the
which has yielded a topology markedly different from that
lowest number of transfers (69 and 1130) across all the
obtained previously with the 16S rRNA gene. There is
rootsinbothtypesoftreefollowedbytheBayesiantree(71
thus a major need to assess the reliability of the different
and 1136), the MLSA tree (125 and 1490), 23S (123 and
trees. The Prunier method (Abby et al., 2010) can provide
1554), 16S–23S (150 and 1870), artificial (346 and 3633)
andfinallythe16Stree(375and4039).Thereforeitmaybe such an independent measure, and we have thus applied
inferred that among the eight trees analysed, the con- it to actinobacterial phylogeny.
catenated-sequence trees are the most reliable ones.
The concatenated-RAxML tree with 1130 lateral transfers
We also sought to detect the most reliable root. Though was the one with the lowest number of predicted LGTs,
conventionally the order Coriobacteriales including the and thus may be regarded as the most reliable. This tree
genus Atopobium is considered as the root, the default represents 54 gene families with 17642 amino acids in
parameters were used where Prunier proposes a scenario total. The major difference that emerged compared with
foreverypossiblerootofthespeciestree(Abbyetal.,2010) previous treatments in Bergey’s Manual (Ludwig et al.,
and the root where least transfer took place would be the 2012a, b) is in the topology of the order ‘Frankiales’. This
actualroot.Theresultoftransferinvariousrootsisshown order was previously described as artificial, comprising
inTablesS2andS3(forUUgenefamilytreesandMUgene families with considerable differences in morphology,
family trees, respectively). It was found that the con- physiology or genome features (Barabote et al., 2009;
catenated-RAxML tree had the minimum transfer (69–73 Gtari et al., 2012; Normand et al., 2007). The use of
and 1140–1265) across all the roots in both type of trees, complete genomes thus permits resolution of that troub-
followed by the Bayesian tree (72–75 and 1132–1243), the ling aspect by splitting the paraphyletic order, yielding
concatenated-MEGA tree (103–122 and 1261–1376), the at least five orders, each with a single family: Frankiales
MLSA tree (125–130 and 1501–1622), 23S (124–131 and ord. nov. (family Frankiaceae), Geodermatophilales
1583–1682), 16S-23S (151–163 and 1897–1982), artificial ord. nov. (Geodermatophilaceae), Nakamurellales ord.
(347–367and3665–3793)andfinallythe16Stree(380–395 nov. (Nakamurellaceae) and Acidothermales ord. nov.
and 4106–4329). Therefore it may be inferred that among (Acidothermaceae) as well as the order ‘Kineosporiales’
the eight trees analysed, the concatenated-RAxML tree is alreadyproposedin2012(Normand&Benson,2012a).
Table1.Characteristics of theeight genegroups analysed
Characteristic* 16S 23S 16S-23S Concatenated ConcatenatedMEGA MLSA Bayes Artificial
RAxML
LGT(max)from579MU 39 26 27 22 22 19 22 27
LGT(sum)from579MU 4039 1554 1870 1130 1238 1490 1136 3633
LGTfrom54UU 375 123 150 69 103 125 71 346
Meanno.ofpositions(AA/NA) 1525 3069 4594 17642 17642 2624 17642 n.a.
Lengthofline-ups(AA/NA) 3066 3487 6553 25589 25589 2630 25589 n.a.
*LGT,lateralgenetransfer;AA,aminoacids;NA,nucleicacids;n.a.,notapplicable.
http://ijs.sgmjournals.org 3827
A.Senandothers
An order for the ‘Frankiales’/Frankiaceae is the easiest to environment may have contributed to exchange of genes
defend, given that it is positioned at the root of the whole in their shared biotopes.
class Actinobacteria in the RAxML-based tree (Fig. 1).
Theorder‘Glycomycetales’isasmallordercomprisingonlyone
Additionally, the genus Frankia occupies a unique
family,threegeneraandonlyoneavailablegenomeatpresent.
ecological niche in the symbiotic root tissues of plants
Itisthushardtodiscerngeneralcharacteristicsofthegroup.
and a unique physiological function, that of biological
Conversely,theorder‘Micromonosporales’isalargeorderwith
nitrogen-fixation.MembersofthegenusFrankiacontaina
23generaandseveralavailablegenomes,isolatedmostlyfrom
cluster of nif genes, plasmid-borne in at least one strain
soilandoceanicwaters.Stackebrandtianassauensis,whichhas
(Simonet et al., 1986), but not in all other lineages
theonlycharacterizedgenomeoftheorder‘Glycomycetales’,is
(Normand et al., 2007), the phylogeny of which positions
asoilmicrobe,non-motileandproducessubstrateandaerial
the genus Frankia away from all other lineages (Normand
myceliabutnospores.Itisaerobic,withafattyacidpattern
& Bousquet, 1989; Normand et al., 1992). All strains
dominated by saturated branched-chain acids, anteiso-C
analysed also have a rare sugar, 2-O-methyl-D-mannose 17:0
(26.8%), iso-C (8.7%), iso-C (8.7%) and iso-C
(Mortetal.,1983).Manymorestrainscanbehypothesized 15:0 16:0 17:0
(9.0%). The identified polar lipids are phosphatidylglycerol
to exist in nodules but are so far only known by direct
(PG) and diphosphatidylglycerol (DPG), and the predom-
characterizationinvariousspecies(Nazaretetal.,1989).It
inant menaquinones are MK-10(H4), MK-10(H6), MK-
is thus proposed to create the order Frankiales ord. nov.
11(H4) and MK-11(H6) (Munk et al., 2009). Members of
with a single family, Frankiaceae. This order may be
theorder ‘Micromonosporales’ aresoil and marinemicrobes,
augmented in the coming years to include additional
withseverallineagesrecentlyisolatedfromavarietyofplant
micro-organisms identifiedfrom soil, in a manner, similar
tissues(Carroetal.,2013).Theyaremotilewithsubstrateand
to the finding of 16S rRNA gene sequences recovered
scantaerialhyphae,carryspores,andhaveafattyacidpattern
from the rhizosphere of Alnus (Normand and Chapelon,
dominated by saturated branched-chain acids, iso-C ,
1997). 15:0
anteiso-C andanteiso-C .Theidentifiedpolarlipidis
15:0 17:0
The family Geodermatophilaceae forms a coherent cluster phosphatidylethanolamine(PE),andpredominantmenaqui-
with the orders ‘Corynebacteriales’, ‘Micromonosporales’ nones are all types of the MK-9 and MK-10 series. There is
and‘Pseudonocardiales’, away from the family Frankiaceae. thereforelittleincommonbetweenthetwoorders.However,
Members of the order Geodermatophilales ord. nov. also theapparentphylogeneticproximitybetweenthetwoorders
occupy unique ecological niches, such as stone walls and mayneedtowaitforfurtherchemicalandgenomicdatabefore
desert soils, and possess unusual resistance to oxidative thetwoordersarefusedandanincertaesedisstatusintheclass
stress (Gtari et al., 2012). They share generic genomic Actinobacteriaisproposed.
featuressuchasmultiplecopiesofthetrwCgene(Chouaia
The order ‘Catenulisporales’, which contains only two
et al., 2012; Normand et al., 2012). The family
genera (Actinospica with two species and Catenulispora
Geodermatophilaceae is evolving rapidly with several novel
withfivespecies)issystematicallyassociatedwiththeorder
species proposed recently (Montero-Calasanz et al., 2013)
Streptomycetales (three genera, 561 species), in the 16S
with isolates from desert soils and plant rhizospheres
rRNA gene tree, the MLSA tree and the concatenated tree.
(Normand &Benson,2012b).Itisthusproposedtocreate
At the same time, the order ‘Catenulisporales’ is character-
theorderGeodermatophilalesord.nov.withasinglefamily,
ized by the presence of LL-diaminopimelic acid (LL-DAP),
Geodermatophilaceae.
the predominance of iso-C and anteiso-C fatty acids,
16 17
The remaining two families, Nakamurellaceae and MK9-(H ), -(H ) and -(H ) isoprenoids, PG, DPG,
6 8 4
Acidothermaceae, are some distance from the family phosphatidylinositol (PI) and phosphatidylinositol man-
Frankiaceae in the concatenated-sequence trees with the nosides (PIM) as predominant phospholipids, aerial
best Prunier score, and thus could be taken out of the hyphae and chains of cylindrical arthrospores (Donadio
current order ‘Frankiales’. However, the presence of a et al., 2012) while the order Streptomycetales is character-
single genome in each of the two cases, contrary to izedbythepresence ofLL-DAP andmeso-DAP, MK9-(H6)
Frankialesord.nov.andGeodermatophilalesord.nov.with and -(H ) isoprenoids, DPG, PE, PI and PIM as
8
several genomes published, prevents defining generic predominant phospholipids, aerial hyphae and chains of
features for the moment. Nevertheless, the two genomes arthrospores (Ka¨mpfer, 2012a, b). On balance, the inclu-
have long branches and no specific association with any sion of ‘Catenulisporales’ within the order Streptomycetales
other lineagesanditisthusproposedtocreate neworders is thus to be contemplated.
forthem:Nakamurellalesord.nov.andAcidothermalesord.
The order Micrococcales is not recovered intact in the
nov. in the class Actinobacteria.
concatenated tree, rather it is found with three anomalous
The cluster comprising Acidothermales ord. nov. and the groups: order Bifidobacteriales, order Actinomycetales and
order ‘Streptosporangiales’ is quite solid with 91% boot- order ‘Kineosporiales’. Since the physiology and ecology of
strap. In this cluster there are a majority of genomes that the anaerobic bifidobacteriales are quite distinct from
belong to thermophiles, a fact that given the strong those of the micrococcales beyond a preponderance of
biochemical pressures present in a high-temperature mammal-inhabitingstrains,reunitingtheorderMicrococcales
3828 InternationalJournalofSystematicandEvolutionaryMicrobiology64
PhylogenyofActinobacteriabasedoncompletegenomes
with the other three orders should be avoided. Instead, Prunierdata.ItmaythereforebeconcludedthattheRAxML
thisordershouldbesplitintoseveralmonophyleticorders: treeismostaccurateamongalltreesstudiedatleastfromthe
onethatwouldkeepthenameMicrococcaleswiththegenera Pruniertest.
Micrococcus, Kocuria, Rothia, Arthrobacter, Tropheryma,
The status of the families Cryptosporangiaceae and
Microbacterium, Clavibacter and Leifsonia; one that could
Sporichthyaceaeintheabsenceofgenomedataisuncertain
be designated Cellulomonales and would contain the
withintheclassActinobacteriaandwillhavetowait.There
genera Beutenbergia, Cellulomonas, Xylanimonas, Jonesia
is little doubt that the rate of acquisition of genome
and Sanguibacter; and finally the orders Brachybacteriales
sequences from isolates and of metagenomes from various
(Brachybacterium) and Dermacoccales (Kytococcus and
biotopes will continue unabated in the coming years, and
Intrasporangium); these latter two orders being difficult to
given the diminishing costs of sequence acquisition and
defendatthistime,becauseonlyoneortwogenomeshave
ease of use of computer analytical tools, all described
beenanalysedsofar.
species will soon have a genome sequence available in the
Actinobacteria have colonized several ecological niches databases. This should result in a generalization of
beyond the soil, where they are known to be enriched. approaches such as the Prunier (Abby et al., 2010) to
Amongthe100genomesanalysedhere,lessthanathird(28) buttress phylogeny and identify the details of genes’
areconsideredassoildwellers,whileanotherthird(33)have evolution. Actinobacteria have been postulated to have
been isolated from mammalian hosts. There is certainly a been the originators of innovations such as synthesis of
biasinthesedataduetotheeconomicandsocialimportance cholesterol (Lamb et al., 1998), the ability to recycle
of disease-causing microbes, but this can modify the proteins through a proteasome (De Mot, 2007), and the
perspective on soil as the emblematic biotope of actino- presence of suspended metabolism structures such as
bacteria. Soil actinobacteria can nevertheless be credited as exospores, which together have led to the suggestion that
possessing relatively large genomes, slow growth rates, rich actinobacteria were both more ancient than generally
secondarymetabolismandhighratesofgeneexchanges. considered and possible ancestors of archaea and eukar-
yotes (Cavalier-Smith, 2002).If thepresent studydoes not
In the concatenated-RAxML tree, it was found that
touch on these aspects, it nevertheless supports a position
many of the species sharing the same biotope were
for the phylum Actinobacteria as ancient, diversified and
grouped. For instance, all Frankia species which were
prone to frequent gene exchange.
considered plant-related were together. In case of the
order ‘Micromonosporales’, two water-inhabiting species,
Salinispora tropica and Verrucosispora maris, were phylo- Description of Frankiales ord. nov.
genetically close together with two other soil-inhabiting
Frankiales [Frank.i.a9les. N.L. masc. n. Frank a German
species,i.e. Actinoplanes missouriensis andMicromonospora
botanist (1839–1900) who studied root symbioses; suff.
aurantiaca. Similarly, in case of the order ‘Pseudono-
-alesendingtodenoteorder;N.L.fem.pl.n.Frankialesthe
cardiales’, out of five species studied, one water-inhabiting
Frankia order].
(Pseudonocardia dioxanivorans) and one mammal-inhab-
iting(Saccharomonosporaviridis)weredistantly positioned Cells are Gram-positive, non-motile, form multi-locular
in two parts of the branch. In the case of the sporangia, most form thick-walled vesicles, fix nitrogen
order‘Corynebacteriales’, out of seven Corynebacterium and establish root nodules on actinorhizal plants. A
species, except Corynebacterium variabile (which is a plant member of the class Actinobacteria. Contains a single
inhabitant), all species were mammal-inhabiting and family, Frankiaceae (Stackebrandt et al., 1997). The type
Corynebacterium variabile clearly appears as a side branch. genus is Frankia. Phylogenetic analyses have been pub-
Likewise, in case of the order Micrococcales I, ‘Tropheryma lished previously (Normand & Bousquet, 1989; Normand
whipplei’, which is a mammal-inhabiting species, appears et al., 1996). The name of the order including families
clearly asadistantspecieswhereasplant-inhabiting Micro- Frankiaceae, Geodermatophilaceae, Nakamurellaceae,
bacterium testaceum, Clavibacter michiganensis and Sporichthyaceae, Acidothermaceae and Cryptosporangiaceae
Leifsoniaxyli wereclustered. Lastly inthecaseoftheorder was effectively published in the latest Bergey’s Manual
Bifidobacteriales, out of six Bifidobacterium species, five (Normand & Benson, 2012b).
were mammal-inhabiting along with Gardnerella vaginalis,
and Bifidobacterium asteroides was arthropod-related. In
Description of Geodermatophilales ord. nov.
the concatenated tree, all the mammal-inhabiting
Bifidobacterium species grouped along with Gardnerella Geodermatophilales(Ge.o.der.ma.to.phil.a9les.N.L.fem.pl.
vaginalis, and clearly positioned at the root of the order. n. Geodermatophilus type genus of the order; suff. -ales
ending to denote order; N.L. fem. n. Geodermatophilales
Fromatechnicalpointofview,itmaybementionedthatout
the Geodermatophilus order).
oftwomulti-geneconcatenatedtrees(concatenated-RAxML
treeandconcatenated-MEGAtree),theconcatenated-RAxML Cells are Gram-positive, aerobic, have motile and non-
tree reconstruction approach has been found to perform motile forms, inhabit stone surfaces and interiors, and are
better in terms of LGT as evident from analysis of the pigmented.AmemberoftheclassActinobacteria.Contains
http://ijs.sgmjournals.org 3829
A.Senandothers
asinglefamily,Geodermatophilaceae(Normand,2006).The Acidothermus cellulolyticus 11B provides insights into its ecophysio-
typegenusisGeodermatophilus;therearetwoothergenera, logicalandevolutionaryadaptations.GenomeRes19,1033–1043.
ModestobacterandBlastococcus.Aphylogeneticanalysishas Boussau, B. & Daubin, V. (2010). Genomes as documents of
beenpublishedpreviously(Normandetal.,1996). evolutionaryhistory.TrendsEcolEvol25,224–232.
Buchanan, R. E. (1917). Studies in the nomenclature and classi-
fication of the Bacteria: II. The primary subdivisions of the
Description of Acidothermales ord. nov.
Schizomycetes.JBacteriol2,155–164.
Acidothermales(Aci.do.therm.a9les.N.L.masc.n.Acidothermus Carro, L., Pujic, P., Trujillo, M. E. & Normand, P. (2013). Micro-
typegenusoftheorder;suff.-alesendingtodenoteorder;N.L. monosporaisanormaloccupantofactinorhizalnodules.JBiosci38,
685–693.
fem.pl.n.AcidothermalestheAcidothermusorder).
Cavalier-Smith, T. (2002). The neomuran origin of archaebacteria,
Cells are Gram-positive, non-motile and thermo-resistant the negibacterial root of the universal tree and bacterial mega-
(55 uC). A member of the class Actinobacteria. Contains a classification.IntJSystEvolMicrobiol52,7–76.
single family, Acidothermaceae (Stackebrandt et al., 1997). Chouaia, B., Crotti, E., Brusetti, L., Daffonchio, D., Essoussi, I.,
A phylogenetic analysis has been published previously Nouioui,I.,Sbissi,I.,Ghodhbane-Gtari,F.,Gtari,M.&otherauthors
(Barabote et al., 2009). (2012). Genome sequence of Blastococcus saxobsidens DD2, a stone-
inhabitingbacterium.JBacteriol194,2752–2753.
De Mot, R. (2007). Actinomycete-like proteasomes in a Gram-
Description of Nakamurellales ord. nov.
negativebacterium.TrendsMicrobiol15,335–338.
Nakamurellales(Na.ka.mu.rel.la9les.N.L.fem.n.Nakamurella Donadio, S., Cavaletti, L. & Monciardini, P. (2012). Order VI
typegenusoftheorder;suff.-alesendingtodenoteorder;N.L. Catenulisporales ord. nov. In Bergey’s Manual of Systematic
fem.pl.n.NakamurellalestheNakamurellaorder). Bacteriology, vol. 5, p. 225. Edited by W. B. Whitman, A. Parte,
M.Goodfellow,P.Ka¨mpfer,H.-J.Busse,M.E.Trujillo,W.Ludwig&
CellsareGram-positive,non-spore-forming,coccus-shaped, K.Suzuki.NewYork:Springer.
non-motile and inhabit soil. A member of the class Doroghazi, J. R. & Buckley, D. H. (2010). Widespread homologous
Actinobacteria. Contains a single family, Nakamurellaceae recombination within and between Streptomyces species. ISME J 4,
(Stackebrandtetal., 1997).Thetypegenus isNakamurella; 1136–1143.
there are two other genera, Humicoccus and Saxeibacter. A Edgar,R.C.(2004a).MUSCLE:multiplesequencealignmentwithhigh
phylogenetic analysis has been published previously (Kim accuracyandhighthroughput.NucleicAcidsRes32,1792–1797.
etal.,2012). Edgar,R.C.(2004b).MUSCLE:amultiplesequencealignmentmethod
withreducedtimeandspacecomplexity.BMCBioinformatics5,113.
Embley,T.M.&Stackebrandt,E.(1994). The molecular phylogeny
ACKNOWLEDGEMENTS
and systematics of the actinomycetes. Annu Rev Microbiol 48, 257–
289.
A.S.acknowledgesthereceiptoftheDBT-CRESTaward.Thanksare
Euze´by, J. P. & Tindall, B. J. (2001). Nomenclatural type of orders:
expressed to the French Research Agency ANR (Sesam, ANR-10-
corrections necessary according to Rules 15 and 21a of the
BLAN-1708)andtotheFrance-Berkeley-Fund(collaborativegrantto
BacteriologicalCode(1990Revision),anddesignationofappropriate
A.M.B.andP.N.).ThisworkwaspartiallyfundedbyaUGCproject
nomenclaturaltypesofclassesandsubclasses.Requestforanopinion.
on Frankia and actinorhizal plants. The authors gratefully acknow-
IntJSystEvolMicrobiol51,725–727.
ledge support from the CNRS/IN2P3 Computing Center (Lyon/
Villeurbanne, France), for providing a significant amount of the Felsenstein,J.(1985).Confidencelimitsonphylogenies:anapproach
computing resources needed for this work. V.D. acknowledges usingthebootstrap.Evolution39,783–791.
support from the ANCESTROME project (ANR-10-BINF-01-01). A Fleischmann,R.D.,Adams,M.D.,White,O.,Clayton,R.A.,Kirkness,
portion of the work was also done at NBU Bioinformatics Facility. E.F.,Kerlavage,A.R.,Bult,C.J.,Tomb,J.F.,Dougherty,B.A.&other
ThesuggestionsandhelpofMaximeBruto,PhDstudent,University authors(1995).Whole-genomerandomsequencingandassemblyof
ofLyon1ininterpretationandanalysisofdataarerecognized.Helpof
HaemophilusinfluenzaeRd.Science269,496–512.
ProfessorJeanEuze´by(EcoleNationaleVe´te´rinairedeToulouse)with
Gao,B.&Gupta,R.S.(2012).Phylogeneticframeworkandmolecular
nomenclatural details is acknowledged. Thanks are also due to
signaturesforthemaincladesofthephylumActinobacteria.Microbiol
Sanghati Bhattacharya, PhDstudent at NorthBengal University, for
MolBiolRev76,66–112.
lendingahandindrawingoutnumerouscolourtrees.
Goodfellow, M. (1967). Numerical taxonomy of some named
bacterialcultures.CanJMicrobiol13,1365–1374.
REFERENCES
Gouy, M., Guindon, S. & Gascuel, O. (2010). SeaView version 4: a
multiplatform graphical user interface for sequence alignment and
Abby, S. S., Tannier, E., Gouy, M. & Daubin, V. (2010). Detecting phylogenetictreebuilding.MolBiolEvol27,221–224.
lateral gene transfers by statistical reconciliation of phylogenetic Gtari,M.,Essoussi,I.,Maaoui,R.,Sghaier,H.,Boujmil,R.,Gury,J.,
forests.BMCBioinformatics11,324. Pujic,P.,Brusetti,L.,Chouaia,B.&otherauthors(2012).Contrasted
Abby,S.S.,Tannier,E.,Gouy,M.&Daubin,V.(2012). Lateral gene resistance of stone-dwelling Geodermatophilaceae species to stresses
transferasasupportforthetreeoflife.ProcNatlAcadSciUSA109, knowntogiverisetoreactiveoxygenspecies.FEMSMicrobiolEcol80,
4962–4967. 566–577.
Barabote, R. D., Xie, G., Leu, D. H., Normand, P., Necsulea, A., Ka¨mpfer, P. (2012a). Family I. Streptomycetaceae Waksman and
Daubin, V., Me´digue, C., Adney, W. S., Xu, X. C. & other authors Henrici 1943, 339AL emend. Rainey, Ward-Rainey and Stackebrandt
(2009). Complete genome of the cellulolytic thermophile 1997,486emend.Kim,Lonsdale,SeongandGoodfellow2003b,113
3830 InternationalJournalofSystematicandEvolutionaryMicrobiology64