1Road map of the phylum ActinobacteriaWolfgang Ludwig, Jean Euzéby, Peter Schumann, Hans-Jürgen Busse, Martha E. Trujillo,Peter Kämpfer and William B. WhitmanThis revised road map and the resulting taxonomic outlineupdate the previous versions of Garrity and Holt (2001) andGarrity et al. (2005) with the description of additional taxaand new phylogenetic analyses. While the road map seeks tobe complete for all taxa validly named prior to 1 January 2008,some taxa described after that date are included.The new phylogenetic trees are strict consensus trees basedon various maximum-likelihood and maximum-parsimony anal-yses and corrected according to results obtained when apply-ing alternative treeing methods. Multifurcations indicate that acommon branching order was not significantly supported afterapplying alternative treeing approaches. Detailed branchingorders are shown if supported by at least 50% of the “treeings”performed in addition to the maximum-likelihood approach.Given that the focus is on the higher (taxonomic) ranks,rather restrictive variability filters were applied. Consequently,resolution power is lost for lower levels. Of special importance,relationships within genera lack the resolution that would beobtained with genus/family level analyses. Furthermore, thetype strain tree, which is available online at www.bergeys.org, isan extract of comprehensive trees comprising some thousandsequences. Thus, trees for the specific groups in subsequentchapters, which are based upon smaller datasets and includevariable sequence positions, may differ with respect to detailedtopology, especially at levels of closer relationship within andbetween genera. In the trees shown here, branch lengths – inthe first instance – indicate significance and only approximateestimated number of substitutions.Starting with the second edition of Bergey’s Manual of SystematicBacteriology, the arrangement of content follows a phylogeneticframework or road map based largely on analyses of nucleotidesequences of the ribosomal small subunit RNA rather than onphenotypic data (Garrity et al., 2005). Implicit in the use ofthe road map are the convictions that prokaryotes have a phy-logeny and that phylogeny matters. However, phylogenies, likeother experimentally derived hypotheses, are not static but maychange whenever new data and/or improved methods of analy-sis become available (Ludwig and Klenk, 2005). Thus, the largeincreases in data since the publication of the taxonomic out-lines in the preceding volumes have led to a re-evaluation of theroad map. Not surprisingly, the taxonomic hierarchy has beenmodified or newly interpreted for a number of taxonomic units.These changes are described in the following paragraphs.The taxonomic road map proposed in volume 1 and updatedand emended in volume 2 was derived from phylogenetic andprincipalcomponentanalysesofcomprehensivedatasetsofsmall-subunit rRNA gene sequences. A similar approach is continuedhere. Since the introduction of comparative rRNA sequencing(Ludwig and Klenk, 2005; Ludwig and Schleifer, 2005), therehas been a continuous debate concerning the justification andpower of a single marker molecule for elucidating and estab-lishing the phylogeny and taxonomy of organisms, respectively.Although generally well established in taxonomy, comparableanalyses of multiple genes cannot currently be applied becauseof the lack of comprehensive datasets for other marker mole-cules. Even in the age of genomics, the datasets for non-rRNAmarkers are poor in comparison to more than 400,000 rRNAprimary structures available in general and specialized data-bases (Cole et al., 2007; Pruesse et al., 2007). Nevertheless, thedata provided by full genome sequencing projects have iden-tified a small set of genes representing the conserved core ofprokaryotic genomes (Cicarelli et al., 2006; Ludwig and Schle-ifer, 2005). Furthermore, comparative analyses of core genesequences globally support the small-subunit rRNA-derived viewof prokaryotic evolution. Although the tree topologies recon-structed from alternative markers differ in detail, the majorgroups (and taxa) are verified or at least not disproved (Ludwigand Schleifer, 2005). Consequently, this volume is organized onthe basis of updated and curated databases of processed small-subunit rRNA primary structures (http://www.arb-silva.de; Lud-wig et al., 2004).Data analysisThe current release of the integrated small-subunit rRNA data-base of the SILVA project (Pruesse et al., 2007) provided thebasis for these phylogenetic analyses. The tools of the arb soft-ware package (Ludwig et al., 2004) were used for data evalu-ation, optimization, and phylogenetic inference. A subset ofabout 33,000 high-quality sequences from Bacteria was extractedfrom the current SILVA SSU Ref database. Among the criteriafor restrictive quality analyses and data selection were: cover-age of at least positions 18 to 1509 (Escherichia coli 16S rRNAnumbering), no ambiguities or missing sequence stretches, nochimeric primary structures, low deviation from overall andgroup-specific consensus and conservation profiles, and goodagreement of tree topologies and branch lengths with pro-cessed sequence data. Unfortunately, not all of the type strainsequences successfully passed this restrictive quality check. Thealignment of the sequences of this subset, as well as all type strainsequences initially excluded given incompleteness or lower quality,was manually evaluated and optimized. Phylogenetic treeing wasfirst based on the high quality dataset and performed apply-ing phylum-specific position filters (50% positional identity).The partial or lower quality type strain sequences were subse-quently added using a special arb-tool allowing the optimallypositioning of branches to the reference tree without admittingtopology changes (Ludwig and Klenk, 2005). The consensustrees used for evaluating or modifying the taxonomic outline
2 Road mapwere based on maximum-likelihood analyses (RAXML, imple-mented in the arb package; Stamatakis et al., 2005) and furtherevaluated by maximum-parsimony and distance matrix analyseswith the respective arb tools (Ludwig et al., 2004).Taxonomic interpretationIn order to ensure applicability and promote acceptance, theproposed taxonomic modifications were made following aconservative procedure. The overall organization follows thetype “taxon” principle as applied in the previous volumes. Taxadefined in the outline of the preceding volumes were only uni-fied, dissected, or transferred in the cases of strong phylogeneticsupport. This approach is justified by the well-known low signifi-cance of local tree topologies (also called “range of unsharp-ness” around the nodes; Ludwig and Klenk, 2005). Thus, manyof the cases of paraphyletic taxa found were maintained in thecurrent road map if the respective (sub)-clusters rooted closelytogether, even if they were separated by intervening clustersrepresenting other taxa. While reorganization of these taxamay be warranted, it was not performed in the absence of con-firmatory evidence. The names of type strains of species withvalidly published names that are phylogenetically misplaced arealso generally maintained. These strains are mentioned in thecontext of the respective phylogenetic groups. In case of para-phyly, all concerned species or higher taxa are assigned to therespective (sub)-groups. New higher taxonomic ranks are onlyproposed if species or genera – previously assigned to differenthigher taxonomic units – are significantly unified in a mono-phyletic branch.Phylum “Actinobacteria”The phylum “Actinobacteria” is well supported by analyses ofthe 16S and 23S rRNA genes, presence of conserved insertionsand deletions (or indels) in certain proteins, and characteris-tic gene rearrangements (as reviewed by Goodfellow and Fie-dler, 2010). The current road map builds upon the thoroughtaxonomic analyses of Zhi et al. (2009) and Stackebrandt et al.(1997). However, in the classification proposed herein, the tax-onomic ranks of subclass and suborder are not used, and theclades previously represented by these ranks are mostly elevatedto the ranks of class and order, respectively. This modificationmakes the taxonomy of the “Actinobacteria” more consistent withthat of other prokaryotes and thereby facilitates comparisonsbetween phyla and the development of a unified classifica-tion across all bacteria and archaea. Moreover, it reduces thenumber of subdivisions within the higher ranks from six (class,subclass, order, suborder, family, and genus) to four. The lowernumber of higher ranks is more realistic given the limited abil-ity to distinguish phylogenetic relationships among this largeand complex group. This change has a number of importantconsequences. The class “Actinobacteria” now excludes thesubclasses Acidimicrobidae, Coriobacteridae, Nitriliruptoridae, andRubrobacteridae, with the elevation of these subclasses to classes.In addition, with the elevation of suborders to orders, the orderActinomycetales is now restricted to members of the family Actin-omycetaceae and many suborders that are well established in theliterature, such as Micrococcineae and Pseudonocardineae, are notused. Nevertheless, the possible confusion that might resultfrom these changes is out-weighed by the advantages of a sim-pler classification which more closely resembles that found inother prokaryotic phyla.The proposed classification does not preclude the use of theterm “actinomycetales” in its conventional sense. This practiceis common in other prokaryotic groups where the implementa-tion of a natural classification based upon phylogeny made theearlier terminology inappropriate. Thus, “bacillus” refers to amember of a large number of genera that encompass aerobic,endospore-forming rods and not the genus Bacillus.With the class “Actinobacteria” defined in this manner, sixclasses are proposed (Figure 1). In addition to the class “Acti-nobacteria”, which is now restricted to the clades formerlyclassified within the subclass Actinobacteridae, the classes “Acidi-microbidiia”, “Coriobacteriia”, “Nitriliruptoria”, “Rubrobacteria”, and“Thermoleophilia” are proposed. This last class includes many ofthe genera formerly classified within the Rubrobacteria that werenot closely related to it (see below).Class “Actinobacteria”As shown in Figure 2, the class “Actinobacteria” comprises theorders previously classified as suborders within the order Actino-mycetales by Zhi et al. (2009), as well as Bifidobacteriales, whichwas previously classified as an order, and the new order Jian-gellales (Tang et al., 2011). Upon publication of the class “Acti-nobacteria”, a type order was not designated. Hence, this nameis not validly published under Rule 27 of the International Codeof Nomenclature of Bacteria (Euzéby and Tindall, 2001; Lapage"Actinobacteria""Acidimicrobiia""Coriobacteriia""Thermoleophilia""Rubrobacteria""Nitriliruptoria"FIGURE 1. Overview of the classes within the phyla Actinobacteria. Con-sensus dendrogram of the phylogenetic relationships of the 16S rRNAgenes based on various maximum-likelihood and maximum-parsimonyanalyses and corrected according to results obtained when applyingalternative treeing methods. Multifurcations indicate that a commonbranching order was not significantly supported after applying alter-native treeing approaches. Detailed branching orders are shown ifsupported by at least 50% of the “treeings” performed in addition tothe maximum-likelihood approach. For additional methods, please seethe text.
3Road mapet al., 1992). Nevertheless, because of its wide use, the name isadopted for this volume. If so designated by the Judicial Com-mission, the type order is likely to be Actinomycetales, which isthe convention followed herein.Within the class, two large clades are observed in the rRNAgene trees prepared for this volume (Figure 2). The first cladeincludes the orders “Actinopolysporales”, “Corynebacteriales”,“Glycomycetales”, “Jiangellales”, “Micromonosporales”, “Propionibac-teriales”, and “Pseudonocardiales”. The second clade includesthe orders Actinomycetales, Bifidobacteriales, “Kineosporiales”, andMicrococcales. The orders “Catenulisporales”, “Streptomycetales”, and“Streptosporangiales” form deep lineages radiating from the baseof the class. Lastly, the families of the order “Frankiales” do notform a clade, but appear as independent lineages at the baseof the tree. However, these relationships were not observed inrRNA gene trees of Zhi et al. (2009). Thus, in the absence ofconfirmatory evidence, they were not used to make taxonomicdecisions in the road map.Order Actinomycetales and family ActinomycetaceaeWith the elevation of the suborders of Zhi et al. (2009) toorders, this order now comprises only the family Actinomyceta-ceae. Thus, the taxonomic designation is no longer congruentwith the common term “actinomycetales”. The family appearsas an independent clade somewhat related to the family Jonesi-aceae of the order Micrococcales (Figure 3). It comprises the dif-fuse type genus Actinomyces and the four well-defined generaActinobaculum, Arcanobacterium, Mobiluncus, and Varibaculum.The genus Actinomyces is represented by five clades within thefamily. The largest clade includes the type species Actinomycesbovis and Actinomyces bowdenii, catuli, dentalis, denticolens, gerenc-seriae, graevenitzii, howellii, israelii, johnsonii, massiliensis, naeslun-dii, oricola, oris, radicidentis, ruminicola, slackii, urogenitalis, andviscosus. The remaining clades are no more closely related tothe type species than to the members of other genera withinthe family, which might be grounds for their reclassificationin the future. The second largest clade includes the speciesActinomyces canis, cardiffensis, funkei, georgiae, hyovaginalis, meyeri,odontolyticus, radingae, suimastitidis, turicensis, and vaccimaxil-lae. It is loosely related to two additional clades, one of whichincludes Actinomyces coleocanis and Actinomyces europaeus and theother Actinomyces neuii and the monospecific genus Varibaculum(type species Varibaculum cambriense). With only 90% sequencesimilarity and some phenotypic differences to Varibaculum,Actinomyces neuii may warrant reclassification in a new genus(Hall et al., 2003). The last clade clusters close to the root ofthe family tree and includes Actinomyces marimammalium, hong-kongensis, hordeovulneris, and nasicola. The taxonomic signifi-cance of these groups is discussed in further detail by Schaaland Yassin (2012).The other genera in the family all form monophyletic clades.The genera Actinobaculum and Arcanobacterium are related toeach other. Actinobaculum comprises the type species Actinobac-ulum suis and Actinobaculum massiliense, schaalii, and urinale.Arcanobacterium comprises the type species Arcanobacterium hae-molyticum and Arcanobacterium abortisuis, bernardiae, bialowiezense,bonasi, hippocoleae, phocae, pluranimalium, and pyogenes. Thegenus Mobiluncus comprises the type species Mobiluncus curtisiiand Mobiluncus mulieris. This genus has precedence over Falcivi-brio, which appears to be a later heterotypic synonym (Hoyleset al., 2004).Order “Actinopolysporales” and familyActinopolysporaceaeIn the rRNA gene trees described herein, this order is relatedto “Corynebacteriales” and “Pseudonocardiales”. These three ordersare members of a larger clade that also includes “Glycomycetales”,FrankiaceaeAcidothermaceaeCryptosporangiaceaeNakamurellaceaeGeodermatophilaceaeSporichthyaceae"Frankiales""Catenulisporales""Streptomycetales""Kineosporiales"Actinomycetales − Micrococcales"Propionibacteriales""Glycomycetales − Micromonosporales""Actinopolysporales""Jiangellales""Pseudonocardiales""Streptosporangiales""Corynebacteriales"BifidobacterialesFIGURE 2. Orders of the class Actinobacteria. Analyses were performedas described for Figure 1.
4 Road map“Jiangellales”, “Micromonosporales”, and “Propionibacteriales”(Figure 2). In the analyses of Zhi et al. (2009), only the relation-ship of “Actinopolysporales” and “Glycomycetales” was observed,although without bootstrap support. The monogeneric familycomprises the type species Actinopolyspora halophila and Acti-nopolyspora mortivallis. The species Actinopolyspora iraqiensis wasmisclassified and is now classified as a heterotypic synonym ofSaccharomonospora halophila.Order Bifidobacteriales and family BifidobacteriaceaeThe order Bifidobacteriales comprises the family Bifidobacteriaceae,which encompasses seven closely related genera (Figure 4). Inthe rRNA gene trees used here, the entire group is separated bya long branch that arises within a clade formed by the Actinomyc-etales, Micrococcales, and “Kineosporiales” (Figure 2). However, theroot of the order Bifidobacteriales is unstable and this relation-ship was not observed by Zhi et al. (2009). In their gene trees,based upon a different method of analysis, the order Bifidobac-teriales appears as a deep branch of the class “Actinobacteria”. Ineither case, both methods agree that the genera assigned to thisgroup are well separated from other orders.The family Bifidobacteriaceae comprises two clades. One cladeincludes the genus Bifidobacterium and the monospecific genusGardnerella (type species Gardnerella vaginalis). The secondclade comprises five additional monospecific genera with typespecies Aeriscardovia aeriphila, Alloscardovia omnicolens, Metascar-dovia criceti, Parascardovia denticolens, and Scardovia inopinata.The genus Bifidobacterium encompasses nine subclades, eachabout as related to one another as they are to Gardnerella vagi-nalis. These subclades include the type species Bifidobacteriumbifidum; Bifidobacterium adolescentis, angulatum, catenulatum,dentium, merycicum, pseudocatenulatum, and ruminantium; Bifi-dobacterium animalis, choerinum, cuniculi, gallicum, magnum, andpseudolongum; Bifidobacterium asteroides, bombi, coryneforme, indi-cum, minimum, mongoliense, psychraerophilum, and thermophilum;Bifidobacterium boum and thermacidophilum; Bifidobacterium breve,longum, and subtile; Bifidobacterium gallinarum, pullorum, andsaeculare; Bifidobacterium scardovii; and Bifidobacterium tsurumi-ense. The significance of these subclades is not certain and theyonly correspond imperfectly to similarities in cell-wall composi-tion and in 16S–23S intergenic spacer regions (Biavati and Mat-tarelli, 2012; Leblond-Bourget et al., 1996).Order “Catenulisporales” and families Catenulisporaceaeand ActinospicaceaeThis order represents a deep clade within the class “Actinobac-teria” and comprises two monogeneric families (Figure 5). Thefamily Catenulisporaceae contains the type species Catenulisporaacidiphila and three related species, Catenulispora rubra, subtrop-ica, and yoronensis. The family Actinospicaceae includes the typespecies Actinospica robiniae and Actinospica acidiphila.Order “Corynebacteriales”In the rRNA gene trees described herein, this order is relatedto the orders “Actinopolysporales” and “Pseudonocardiales”. In theanalyses of Zhi et al. (2009), only the relationship between“Corynebacteriales” and “Pseudonocardiales” was observed, albeitwithout bootstrap support. All three orders have the cell-wallchemotype IV, which includes the presence of meso-diamin-opimelate, arabinose, and galactose. However, their quinonesScardoviaParascardoviaAlloscardoviaBifidobacterium − GardnerellaMetascardoviaAeriscardoviaFIGURE 4. Genera of the order Bifidobacteriales. Analyses were performed as described for Figure 1.JonesiaceaeActinobaculumArcanobacteriumActinomycesMobiluncusVaribaculumActinomycetaceaeActinomycetalesFIGURE 3. Genera of the order Actinomycetales and Jonesiaceae of the order Micrococcales. Analyses wereperformed as described for Figure 1.
5Road mapand major fatty acids are different, and species belonging to the“Pseudonocardiales” lack mycolic acids. Therefore, the biologicalimportance of this deep relationship requires further examina-tion. These orders are also members of the larger clade thatalso includes “Glycomycetales”, “Jiangellales”, “Micromonosporales”,and “Propionibacteriales”, whose biological significance is alsonot known (Figure 2).Six suprageneric taxa can be recognized in the current 16SrRNA gene tree, the families Corynebacteriaceae, Dietziaceae, Myco-bacteriaceae, Nocardiaceae, Segniliparaceae, and Tsukamurellaceae(Figure 6). While most of these families are monophyletic,many of the genera of the Nocardiaceae appear as deep lineagesin the rRNA gene tree, no more closely related to the typegenus Nocardia than members of other families (see below).For that reason, this family appears to be paraphyletic. Whilethe status of some of the families, notably the Mycobacteriaceaeand Tsukamurellaceae, is strongly supported by chemotaxonomicdata, others such as the family Nocardiaceae are markedly het-erogeneous in this respect. Indeed, a case can be made for therecognition of the family Gordoniaceae to encompass the gen-era Gordonia, Millisia, Williamsia, and Skermania (Goodfellowand Jones, 2012). The assignment of the recently described,mycolateless genera Amycolicicoccus (Wang et al., 2010), Hoyo-sella (Jurado et al., 2009), and Tomitella (Katayama et al., 2010)further complicates the delineation of families classified in theorder Corynebacteriales (Goodfellow and Jones, 2012). Conse-quently, the biological significance of several families withinthis order requires further study.Family CorynebacteriaceaeThis family comprises the large and complex genus Corynebacte-rium and the monospecific genus Turicella (type species Turicellaotitidis). On the basis of the rRNA gene tree, many of the speciesof the genus appear equally related to the type species Coryne-bacterium diphtheriae and Turicella otitidis and, on this basis, thegenus is paraphyletic. However, most Corynebacterium speciespossess mycolic acids and the menaquinones MK-8(H2), MK-9(H2), or a mixture of both. Turicella and some Corynebacteriumspecies, such as Corynebacterium amycolatum, lack mycolic acids.Turicella also possesses the menaquinones MK-10 and MK-11(Goodfellow and Jones, 2012). It appears that these specieshave lost the capacity to produce mycolic acids, so the signifi-cance of this chemotaxonomic marker is ambiguous. Given theuncertainties in the relationships within this group, the currenttaxonomy is retained at this time.In addition to the type species, the genus Corynebacteriumcomprises a large number of species which, while related toCorynebacterium diphtheriae, do not form specific associationswith other species of the genus. These include Corynebacteriumaccolens, aquilae, argentoratense, atypicum, auris, camporealensis,capitovis, caspium, confusum, doosanense, felinum, flavescens, freibur-gense, halotolerans, kutscheri, lipophiloflavum, lubricantis, macgin-leyi, maris, massiliense, mastitidis, mycetoides, renale, spheniscorum,testudinoris, timonense, tuberculostearicum, and vitaeruminis. Inaddition, the rRNA gene tree identifies nine small clades of spe-cies, including: Corynebacterium afermentans, appendicis, coyleae,KineosporiaQuadrisphaeraKineococcusStreptacidiphilusKitasatosporaStreptomycesBifidobacteriales − Actinomycetales − MicrococcalesStreptomycetaceae − "Streptomycetales"Kineosporiaceae − "Kineosporiales"Catenulispora − CatenulisporaceaeActinospica − Actinospicaceae"Catenulisporales"FIGURE 5. Genera and families of the orders “Catenulisporales”, “Kineosporiales”, and “Streptomycetales”. Analyseswere performed as described for Figure 1.
6 Road mapglaucum, imitans, mucifaciens, riegelii, sundsvallense, thomssenii,tuscaniense, and ureicelerivorans; Corynebacterium ammoniagenes,casei, and stationis; Corynebacterium amycolatum, freneyi, hansenii,sphenisci, sputi, ulceribovis, and xerosis; Corynebacterium aurimu-cosum, minutissimum, phocae, simulans, and singulare; Corynebac-terium auriscanis, bovis, falsenii, jeikeium, kroppenstedtii, resistens,suicordis, terpenotabidum, urealyticum, and variabile; Corynebacte-rium callunae, efficiens, and glutamicum; Corynebacterium ciconiae,propinquum, and pseudodiphtheriticum; Corynebacterium cystitidis,glucuronolyticum, and pilosum; and Corynebacterium durum, matru-chotii, pseudotuberculosis, and ulcerans.The affiliations of some additional species of the genus areuncertain. Corynebacterium ilicis was originally thought to be ahomotypic synonym of Arthrobacter ilicis. However, this appearsnot to be the case and the rRNA gene sequence of Corynebac-terium ilicis is not known (Judicial Commission of the Interna-tional Committee on Systematics of Bacteria, 2008). The typestrain of Corynebacterium striatum has been lost and this speciesmight be declared nomen dubium (Coyle et al., 1993). Lastly,Corynebacterium beticola appears to be more properly classified asa strain of Erwinia herbicola (Collins and Jones, 1982).Family DietziaceaeThis monogeneric family comprises the type species Dietziamaris and the closely related species Dietzia aerolata, cercidiphylli,cinnamea, kunjamensis, natronolimnaea, papillomatosis, psychralcal-iphila, schimae, and timorensis.Family MycobacteriaceaeThis monogeneric family is well-defined in the 16S rRNA genetree and also characterized by the presence of mycolic acidswith large numbers of carbon atoms and the predominanceof dihydrogenated menaquinones with nine isoprene units(Magee and Ward, 2012). It comprises the type species Myco-bacterium tuberculosis and over 140 other species with validly pub-lished names. In the 16S rRNA gene tree, a major clade is easilyrecognized that corresponds to most of the “slow-growing spe-cies” (Magee and Ward, 2012). This clade is of special impor-tance because it includes many dangerous pathogens, as well asthe type species Mycobacterium tuberculosis. Other species includeMycobacterium africanum, arosiense, asiaticum, avium, bohemicum,botniense, bouchedurhonense, bovis, branderi, caprae, celatum, chi-maera, colombiense, conspicuum, cookii, florentinum, gastri, genavense,gordonae, haemophilum, heckeshornense, heidelbergense, interjectum,intermedium, intracellulare, kansasii, kubicae, kyorinense, lacus, lenti-flavum, leprae, malmoense, mantenii, marinum, marseillense, microti,montefiorense, nebraskense, noviomagense, palustre, parascrofulaceum,paraseoulense, parmense, pinnipedii, pseudoshottsii, riyadhense, sas-katchewanense, scrofulaceum, seoulense, shimoidei, shottsii, simiae, sto-matepiae, szulgai, triplex, ulcerans, vulneris, and xenopi. A second,unrelated clade is composed of other “slow-growing species”:Mycobacterium arupense, hiberniae, kumamotonense, nonchromogeni-cum, and terrae. Most of the remaining species, including mostof the “rapid-growing species”, are poorly differentiated fromeach other in the 16S rRNA gene tree.Family NocardiaceaeIn the current rRNA gene tree, the family Nocardiaceae is para-phyletic and is represented by seven genera comprising fivedeep lineages, none of which are more closely related to eachother than to other members of the order “Corynebacteriales”.In the previous road map, the family contained only the gen-era Nocardia and Rhodococcus (Garrity et al., 2005). Gordonia andSkermania were classified within the family Gordoniaceae and Wil-liamsia was classified within the family “Williamsiaceae”. Zhi et al.(2009) subsequently proposed combining all these genera andMillisia into a single family Nocardiaceae based upon the rRNAgene signature nucleotides. With the addition of Smaragdicoccus(Adachi et al., 2007), this taxonomy was used here.Chemotaxonomic criteria provide some modest supportfor combining these genera in a single family. Smaragdicoccuspossesses cyclic menaquinones, MK-8(H4, w-methylenecycl)and MK-8(H4, dicycl), which are similar in structure to thoseof the genera Nocardia and Skermania, which are dominatedby MK-8(H4, w-cycl). This feature is distinctive enough to sug-gest a close relationship between these genera. Likewise, bothGordonia and Williamsia contain the menaquinone MK-9(H2),which is consistent with their close relationship in the rRNAgene tree. In contrast, Rhodococcus and Millisia both possess MK-8(H2) even though the rRNA gene trees do not indicate a spe-cial affinity between them. Therefore, while the menaquinonecomposition supports the hypothesis that some members of thisCorynebacteriaceaeRhodococcusGordoniaSkermaniaNocardiaSmaragdicoccusMillisiaNocardiaceaeWilliamsiaMycobacterium − MycobacteriaceaeTsukamurella − TsukamurellaceaeDietzia − DietziaceaeSegniliparus − SegniliparaceaeFIGURE 6. Genera and families of the order “Corynebacteriales”. Analy-ses were performed as described for Figure 1.
7Road mapgroup are related to each other, it does not provide strong sup-port that they represent a single family. Somewhat clearer evi-dence for the placement of these seven genera in a single familycomes from similarities in their mycolic acids. Nocardia speciescontain mycolic acids with 46–64 C-atoms, and Rhodococcus spe-cies contain mycolic acids with 30–54 C-atoms. The other fivegenera all have mycolic acids with numbers of C-atoms withinthese ranges. Lastly, all genera within this family possess meso-diaminopimelate, but this feature is also common to otherfamilies of the “Corynebacteriales” and to other orders, such asthe “Pseudonocardiales”. In conclusion, the chemotaxonomicmarkers are evidence of additional phylogenetic relatednessnot apparent in the rRNA gene tree and offer some support forthe current classification. Nevertheless, the composition of thefamily Nocardiaceae should be considered tentative until a timewhen conclusive evidence is available.In the rRNA gene tree, the genus Nocardia forms a deep,monophyletic lineage near the base of the “Corynebacteriales”.The Nocardia tree itself exhibits a complicated branching pat-tern with many very short branches. Three clusters can be rec-ognized. The first cluster contains the type species Nocardiaasteroides and a large number of species whose relatedness is notwell differentiated: Nocardia abscessus, acidivorans, alba, altami-rensis, anemiae, blacklockiae, brasiliensis, caishijiensis, concava, cras-sostreae, cyriacigeorgica, exalbida, gamkensis, harenae, inohanensis,iowensis, jejuensis, jiangxiensis, lijiangensis, mexicana, miyunensis,neocaledoniensis, niigatensis, ninae, nova, polyresistens, pseudobrasil-iensis, pseudovaccinii, seriolae, takedensis, tenerifensis, terpenica, thai-landica, transvalensis, uniformis, vinacea, wallacei, xishanensis, andyamanashiensis. This cluster also contains three subclades thatare well differentiated from the other species: Nocardia brevicat-ena and paucivorans; Nocardia carnea, flavorosea, jinanensis, pig-rifrangens, sienata, speluncae, and testacea; and Nocardia coubleae,cummidelens, fluminea, ignorata, salmonicida, and soli. The secondcluster includes Nocardia africana, amamiensis, aobensis, araoensis,arthritidis, beijingensis, cerradoensis, elegans, kruczakiae, pneumoniae,vaccinii, vermiculata, and veterana. The third cluster is at the baseof the Nocardia lineage and includes the species Nocardia asiat-ica, farcinica, higoensis, otitidiscaviarum, puris, and shimofusensis.The genus Gordonia is represented by a well-defined cladecomprising the type species Gordonia bronchialis and Gordoniaaichiensis, alkanivorans, amarae, amicalis, araii, cholesterolivorans,defluvii, desulfuricans, effusa, hankookensis, hirsuta, hydrophobica,kroppenstedtii, lacunae, malaquae, namibiensis, otitidis, paraffiniv-orans, polyisoprenivorans, rhizosphera, rubripertincta, shandongensis,sihwensis, sinesedis, soli, sputi, terrae, and westfalica. It is closelyrelated to the genus Williamsia, a relationship that is also sup-ported by similarities in menaquinone content and otherchemotaxonomic markers (Goodfellow and Jones, 2012).In the rRNA gene tree, the monospecific genus Millisia (typespecies Millisia brevis) is related to the genera Gordonia and Wil-liamsia. This relationship is also consistent with similarities inthe chain lengths of their mycolic acids.The genus Rhodococcus is represented by a number of deep,paraphyletic lineages with short branch lengths at the baseof the radiation including the family Nocardiaceae and order“Corynebacteriales”. One clade comprises the type species Rho-dococcus rhodochrous and Rhodococcus aetherivorans, coprophilus,gordoniae, phenolicus, pyridinivorans, ruber, and zopfii. A second,poorly resolved clade encompasses Rhodococcus baikonurensis,erythropolis, fascians, globerulus, imtechensis, jialingiae, jostii, kore-ensis, kyotonensis, maanshanensis, marinonascens, opacus, percolatus,qingshengii, tukisamuensis, wratislaviensis, and yunnanensis. Lastly,a number of species are not resolved into clades and may rep-resent new genera. These include Rhodococcus corynebacterioides,equi, kroppenstedtii, kunmingensis, rhodnii, and triatomae. The casefor the recognition of a new genus for Rhodococcus equi is espe-cially strong (Jones and Goodfellow, 2012).The monospecific genera Skermania (type species Skermaniapiniformis) and Smaragdicoccus (type species Smaragdicoccus nii-gatensis) represent two deep lineages at the base of the rRNAgene tree of the family Nocardiaceae (Figure 6).The genus Williamsia forms a well-defined clade that isclosely related Gordonia. It is represented by the type speciesWilliamsia muralis and Williamsia deligens, marianensis, maris, andserinedens.Family SegniliparaceaeThis monogeneric family is represented by the type species Seg-niliparus rotundus and Segniliparus rugosus. These species form adeep lineage at the base of the radiation including the Nocar-diaceae and “Corynebacteriales” (Figure 6).Family TsukamurellaceaeThis monogeneric family comprises the type species Tsuka-murella paurometabola and the closely related species Tsukamurellacarboxydivorans, inchonensis, pseudospumae, pulmonis, spongiae, spu-mae, strandjordii, sunchonensis, and tyrosinosolvens. The presenceof a completely unsaturated menaquinone MK-9 detected in allspecies of this group supports placement in a single genus.Order “Frankiales”This order is formed by elevation of the suborder of Frankineae(Stackebrandt et al., 1997; Zhi et al., 2009) and comprises thefamilies Frankiaceae, Acidothermaceae, Cryptosporangiaceae, Geoder-matophilaceae, Nakamurellaceae, and Sporichthyaceae (Figure 7).The major difference between this order and the correspond-ing group described in the previous road map (Garrity et al.,2005) is the reclassification of the genera Kineosporia and Kine-ococcus to the novel order “Kineosporiales” (Zhi et al., 2009). Inaddition, the family Cryptosporangiaceae was proposed to includethe genus Cryptosporangium, which had previously been classi-fied within the family Kineosporiaceae but was retained in the“Frankiales”. Lastly, the previous road map included the mono-generic family Microsphaeraceae. However, this genus name wasruled illegitimate because of precedence of a fungal genus, andthe bacterial species was reclassified to the new family Naka-murellaceae (Tao et al., 2004).Although the group “Frankiales” was observed in early stud-ies of the phylogeny of the 16S rRNA gene in Actinobacteria(Normand et al., 1996; Stackebrandt et al., 1997), more recentstudies with different selections of outgroups and other meth-ods of phylogenetic analyses do not provide strong support forthis classification. For instance, although 16S rRNA gene sig-natures have been proposed for the order, there is little boot-strap support for this group in the 16S rRNA gene tree of Zhiet al. (2009). In the gene trees prepared for this volume, onlythe families Frankiaceae and Acidothermaceae are closely related(Figures 2 and 7), a relationship also observed by Zhi et al.
8 Road map(2009) and supported by similarities in recA genes (Marechalet al., 2000). The remaining families appear as deep lineages ofthe class Actinobacteria but are not specifically related to the fam-ily Frankiaceae. Chemotaxonomic evidence does not fully resolvethis issue (Normand and Benson, 2012). The cell-wall diaminoacid is meso-diaminopimelate in all families tested except for thefamily Sporichthyaceae, which contains ll-diaminopimelate. Inthe families Frankiaceae, Cryptosporangiaceae, and Sporichthyaceae,the abundant menaquinones are MK-9(H4), MK-9(H6), andMK-9(H8). However, Geodermatophilaceae contains MK-9(H4)and MK-8(H4); and Nakamurellaceae contains mostly MK-8(H4).The menaquinones of the Acidothermaceae have not been deter-mined. Lastly, the fatty acid composition is also diverse. Mem-bers of all families contain mostly saturated fatty acids in therange of 15–18 carbons, some or all of which are branched.Thus, while the chemotaxonomic evidence identifies some uni-fying features, it does not provide strong evidence for the order“Frankiales”. Hence, its composition should be considered ten-tative until conclusive evidence is available.Family FrankiaceaeThis monogeneric family comprises the type genus Frankia(type and only species Frankia alni). Common nitrogen-fixingsymbionts of dicotyledonous plants, at least 12 genospecieshave been isolated but their names have not been validly pub-lished. Hence, the diversity of the genus is certainly larger thanrepresented by the nomenclature.Family AcidothermaceaeThis monogeneric family comprises the type genus Acidother-mus (type and only species Acidothermus cellulolyticus). Its growthtemperature range of 37–70°C, inability to fix N2, and rapidgrowth distinguish this moderately thermophilic taxon fromthe Frankiaceae.Family CryptosporangiaceaeThis family comprises the genus Cryptosporangium. It containsthe type species Cryptosporangium arvum and three closelyrelated species Cryptosporangium aurantiacum, japonicum, andminutisporangium. Based upon similarities in its 16S rRNA genesequence, the monospecific genus Fodinicola (type speciesFodinicola feengrottensis) is tentatively classified in this family asa genus incertae sedis.Family GeodermatophilaceaeThe type for this family is the monospecific genus Geodermato-philus (type species Geodermatophilus obscurus). Other gen-era and species include Blastococcus aggregatus (type species),jejuensis, and saxobsidens, as well as Modestobacter multiseptatus(type species) and versicolor.Family NakamurellaceaeThis family comprises three monospecific genera: Nakamurella(type species Nakamurella multipartita), Humicoccus (type speciesHumicoccus flavidus), and Saxeibacter (type species Saxeibacterlacteus). This latter is a recently described genus which is notdescribed further in this volume (Lee et al., 2008).Family SporichthyaceaeThis monogeneric family encompasses two closely relatedspecies: Sporichthya polymorpha (type species) and Sporichthyabrevicatena.Order “Glycomycetales” and family GlycomycetaceaeThis order is formed by elevation of the suborder Glycomy-cineae (Stackebrandt et al., 1997; Zhi et al., 2009) and com-prises the single family Glycomycetaceae. In the 16S rRNA genetrees prepared for this volume, the “Glycomycetales” appears asa long branch specifically related to the genus Actinocatenisporawithin the order “Micromonosporales” (Figure 8). Some chemot-axonomic and phenotypic properties tend to support this asso-ciation (Labeda, 2012; Matsumoto et al., 2007; Seo and Lee,2009; Thawai et al., 2006). The family Glycomycetaceae and Acti-nocatenispora each have type II cell walls which contain meso-diaminopimelic acid and glycine as well as N-glycolylmuramicacid. They also produce chains of spores on aerial hyphae.However, some differences also exist. Glycomycetaceae and Acti-nocatenispora possess phospholipid types I and II, respectively.Moreover, in the Glycomycetaceae, MK-10, MK-11, and MK-12menaquinones are abundant, whereas in the Actinocatenispora,MK-9 (H4, H6) menaquinones predominate. For these reasons,the association between these two taxa is considered tentativeat this time.The family Glycomycetaceae contains three genera. The genusGlycomyces encompasses the type species Glycomyces harbinensisand nine other species: Glycomyces algeriensis, arizonensis, endo-phyticus, lechevalierae, mayteni, rutgersensis, sambucus, scopariae,and tenuis. Closely related is the recently described monospe-cific genus Haloglycomyces (type species Haloglycomyces albus),which is not described in this volume (Guan et al., 2009). Lastly,the genus Stackebrandtia contains the type species Stackebrandtianassauensis and Stackebrandtia albiflava.FodinicolaCryptosporangiumCryptosporangiaceaeNakamurellaSaxeibacterHumicoccusNakamurellaceaeGeodermatophilusModestobacterBlastococcusGeodermatophilaceaeSporichthya − SporichthyaceaeAcidothermus − Acidothermaceae"Streptosporangiales"Frankia − FrankiaceaeFIGURE 7. Genera and families of the order “Frankiales”. Analyses wereperformed as described for Figure 1.
9Road mapCouchioplanesPseudosporangiumKrasilnikoviaSalinisporaSpirilliplanesAsanoaCatelliglobosisporaHamadaeaVerrucosisporaCatenuloplanesActinoplanesLongisporaPolymorphosporaPilimeliaActinocatenisporaGlycomycesHaloglycomycesStackebrandtiaPlantactinosporaRugosimonosporaLuedemannellaVirgisporangiumDactylosporangiumCatellatosporaPlanosporangiumMicromonosporaJiangelaHaloactinopolysporaGlycomycetaceaeMicromonosporaceaeJiangellaceae − "Jiangellales""Micromonosporales""Glycomycetales"FIGURE 8. Genera of the orders “Glycomycetales”, “Jiangellales”, and “Micromonosporales”. Analyses wereperformed as described for Figure 1.
10 Road mapOrder “Jiangellales” and family JiangellaceaeIn the rRNA gene trees described herein, this order is a mem-ber of a larger clade which also includes the orders “Corynebacte-riales”, “Glycomycetales”, “Micromonosporales”, “Propionibacteriales”,and “Pseudonocardiales” (Figure 2). However, this clade was notobserved in the analyses of Zhi et al. (2009) and is also not sup-ported by chemotaxonomic markers. While these latter resultsare not surprising given the depth of this relationship, the bio-logical importance of this clade should be considered tentativeat this time.The order “Jiangellales” was formed by elevation of the subor-der Jiangellineae. This classification was proposed following thedescription of a novel genus Haloactinopolyspora, which is closelyrelated to the genus Jiangella (Tang et al., 2011). Previously, Jian-gella was classified within the family Nocardioidaceae based uponpolyphasic evidence (Song et al., 2005). Currently, the ordercomprises the family Jiangellaceae and two genera (Figure 8).The genus Jiangella comprises the type species Jiangella gansuen-sis and Jiangella alba and alkaliphila. The monospecific genusHaloactinopolyspora (type species Haloactinopolyspora alba) is alsoclassified in the family.Order “Kineosporiales” and family KineosporiaceaeThis order is formed by elevation of the suborder Kineospori-ineae, which was proposed to include some of the genera previ-ously classified in the suborder Frankineae (Garrity et al., 2005;Zhi et al., 2009). In the rRNA gene trees used here, this orderis within a clade that also includes the orders Actinomycetales,Bifidobacteriales, and Micrococcales (Figure 2). This clade is notobserved in the analyses of Zhi et al. (2009) and is consideredtentative at this time. The order comprises the family Kineospori-aceae, which is composed of three genera (Figure 5). The genusKineosporia includes the type species Kineosporia aurantiaca andthe closely related species Kineosporia babensis, mesophila, miku-niensis, rhamnosa, rhizophila, and succinea. The genus Kineococ-cus encompasses the type species Kineococcus aurantiacus andfour closely related species Kineococcus gynurae, radiotolerans,rhizosphaerae, and xinjiangensis. In addition, Kineococcus marinus,which in rRNA trees appears equally related to the type spe-cies of all three genera in the family, is also classified in theKineococcus. Lastly, the monospecific genus Quadrisphaera (typespecies Quadrisphaera granulorum) is also a member of the orderKineosporiales.Order MicrococcalesThis order is formed by elevation of the suborder Micrococcineae(Zhi et al., 2009). In the rRNA gene tree used here, the orderlies within a clade that also includes the orders Actinomycetalesand Bifidobacteriales (Figure 2). The order “Kineosporiales” alsoappears as an ancestral lineage to this clade. The order Actino-mycetales forms a subclade along with the Micrococcales familiesBrevibacteriaceae, Dermabacteraceae, Jonesiaceae, and Micrococcaceae(Figure 9). The family Yaniellaceae has been transferred to theMicrococcaceae and is not used here (Yassin et al., 2011). A sec-ond subclade encompasses the families Beutenbergiaceae, Bogoriel-laceae, Cellulomonadaceae, Promicromonosporaceae, Rarobacteraceae,Ruaniaceae, and Sanguibacteraceae. The third subclade comprisesthe families Dermacoccaceae, Dermatophilaceae, and Intraspo-rangiaceae, which appears to be paraphyletic (Figure 12).The family Microbacteriaceae and the order Bifidobacteriales repre-sent the two remaining subclades. Zhi et al. (2009) also foundevidence for the first three subclades, albeit without bootstrapsupport. However, the associations with Actinomycetales and Bifi-dobacteriales were not reproduced. Thus, in the absence of addi-tional evidence, the biological significance of this clade and itsconstituent subclades remain uncertain.Family MicrococcaceaeRegardless of the complex phylogeny between the families inthe order, the family Micrococcaceae is well-defined in rRNA genetrees (Figure 10). It includes all the genera that were classifiedwithin this family in the previous road map except for Stoma-tococcus, which has since been reclassified to the genus Rothia(Collins et al., 2000; Garrity et al., 2005). In addition, four gen-era have been added, including Acaricomes, Yaniella, Zhihengli-uella, and the recently described Sinomonas (Zhou et al., 2009).The genus Micrococcus forms a subclade within the familythat is closely related to the genus Citricoccus. This relationshipis supported by similarities in the cell-wall composition, majorfatty acids, and polar lipids, although significant differencesare present in the menaquinone composition (Busse, 2012).The genus Micrococcus comprises the type species MicrococcusJonesiaceaeDermabacteraceaeMicrococcaceaeBrevibacteriaceaeBeutenbergiaceaeBogoriellaceaePromicromonosporaceaeMicrobacteriaceaeRuaniaceaeRarobacteraceaeCellulomonadaceae − SanguibacteraceaeDermacoccaceae − Dermatophilaceae − IntrasporangiaceaeActinomycetaceaeActinomycetalesBifidobacteriaceaeBifidobacterialesFIGURE 9. Overview of the families of the order Micrococcales. Analyseswere performed as described for Figure 1.
11Road mapluteus and five closely related species Micrococcus antarcticus,endophyticus, flavus, lylae, and yunnanensis.In rRNA gene trees, the monospecific genus Acaricomes (typespecies Acaricomes phytoseiuli) is closely related to Arthobactersanguinis within the large radiation of Arthobacter species (seebelow). However, chemotaxonomic data which might supportthis relationship are not available, and it remains unproven(Busse et al., 2012).The rRNA gene tree of the genus Arthrobacter is complex withmany short branches and multifurcations that are difficult toresolve. Moreover, the genera Acaricomes, Renibacterium, andZhihengliuella appear within the radiation that includes the typespecies Arthrobacter globiformis, making the genus Arthrobacterparaphyletic (Figure 10). Reclassification of many Arthrobacterspecies may be necessary in the future to reduce the diversity ofthe genus. In addition, some species are more closely affiliatedKocuriaRothiaSinomonasYaniellaNesterenkoniaCitricoccusMicrococcusArthrobacterAcaricomesRenibacteriumZhiengliuellaHelcobacillusDermabacterDevrieseaBrachybacteriumActinomycetaceaeActinomycetalesBrevibacterium − BrevibacteriaceaeJonesia − JonesiaceaeDermabacteraceaeMicrococcaceaeFIGURE 10. Genera of the families Brevibacteriaceae, Dermabacteraceae, Jonesiaceae, and Micrococcaceae of theorder Micrococcales. Analyses were performed as described for Figure 1.
12 Road mapwith type species of other genera in the family Micrococcaceaeand should be reclassified on those grounds.A combination of rRNA gene sequence similarity andchemotaxonomic features have been used to further classifyArthrobacter species into four “rRNA clusters”, five “subclades”and two “groups” (Busse et al., 2012). The rRNA clusters com-prise species with similar chemotaxonomic features and highrRNA gene sequence similarity that do not form a discreteclade in the phylogenetic trees. rRNA cluster 1 includes thetype species Arthrobacter globiformis and Arthrobacter humicola,oryzae, and pascens. rRNA cluster 2 includes Arthrobacter aure-scens, histidinolovorans, ilicis, nicotinovorans, nitroguajacolicus, andureafaciens. rRNA cluster 3 includes Arthrobacter chlorophenolicus,defluvii, niigatensis, oxydans, phenanthrenivorans, polychromogenes,scleromae, and sulfonivorans. rRNA cluster 4 includes Arthrobacterardleyensis, arilaitensis, bergerei, creatinolyticus, mysorens, nicotianae,protophormiae, rhombi, soli, and uratoxydans.The rRNA subclades comprise species with similar chemot-axonomic features that form discrete clades in the rRNA genephylogenetic trees. Subclade I comprises Arthrobacter antarcti-cus, gangotriensis, kerguelensis, psychrophenolicus, and sulfureus.Subclade II comprises Arthrobacter agilis, flavus, parietis, subter-raneus, tecti, and tumbae. Subclade III comprises Arthrobacter cit-reus, gandavensis, koreensis, and luteolus. Subclade IV comprisesArthrobacter alpinus, psychrochitiniphilus, psychrolactophilus, andstackebrandtii. Subclade V comprises Arthrobacter albidus andechigonensis as well as Sinomonas atrocyanea and flava. This sub-clade is unrelated to the other species of Arthrobacter (Figure10). Thus, reclassification of Arthrobacter albidus and echigonensiswithin the genus Sinomonas appears warranted.The groups encompass species with similar chemotaxo-nomic properties that do not necessarily possess high rRNAgene sequence similarity. Group 1 comprises Arthrobacter castelli,monumenti, and pigmenti. Group 2 comprises Arthrobacter albusand cumminsii. In rRNA gene trees, this latter group is affiliatedwith the clade containing the genera Nesterenkonia, Sinomonas,and Yaniella. Possibly, they represent a novel genus within thisclade. Lastly, not all described species could be classified withinthis scheme, including Arthrobacter alcaliphilus, crystallopoietes,methylotrophus, nasiphocae, ramosus, roseus, russicus, sanguinis,and woluwensis. In addition, the sequence of the rRNA gene ofArthrobacter viscosus suggests that it is closely related to Rhizobiumand misclassified within Arthrobacter (Heyrman et al., 2005).The genus Citricoccus is closely related to the genus Micrococ-cus in rRNA gene trees. It comprises the type species Citricoccusmuralis and the closely related species Citricoccus alkalitolerans,parietis, and zhacaiensis.The genus Kocuria represents one of the deepest lineages inthe family Micrococcaceae. The genus comprises four clades. Thefirst contains the type species Kocuria rosea as well as Kocuriaaegyptia, flava, himachalensis, polaris, and turfanensis. The secondclade is closely related and comprises Kocuria halotolerans, kore-ensis, and kristinae, as well as the genus Rothia. The third cladeincludes only Kocuria palustris. The last clade appears as thedeepest lineage in the family and comprises Kocuria atrinae,carniphila, gwangalliensis, marina, rhizophila, and varians. How-ever, the biological significance of these clades is not currentlysupported by chemotaxonomic or other evidence, so theirimportance is not yet certain.The genus Nesterenkonia is in a clade containing the generaSinomonas and Yaniella and the Arthrobacter species Arthrobacteralbus and cumminsii. The genus contains two subclades. The firstcomprises the type species Nesterenkonia halobia and Nesterenko-nia aethiopica, alba, flava, halophila, lacusekhoensis, and xinjiangensis.The second, closely related subclade comprises Nesterenkoniahalotolerans, jeotgali, lutea, and sandarakina. Although the chemot-axonomic and physiological properties of the two subclades arevery similar, only members of this second subclade possess pep-tidoglycan containing l-Lys–Gly–d-Asp (Stackebrandt, 2012b).In the first subclade, the peptidoglycan contains l-Lys–Gly–d-Glu or l-Lys–d-Glu.The monospecific genus Renibacterium (type species Reni-bacterium salmoninarum) is related to Arthrobacter russicus andArthrobacter Subclade IV, which includes Arthrobacter psychrolacto-philus, stackebrandtii, and psychrochitiniphilus. Differences in themenaquinone and peptidoglycan composition of Renibacteriumand the Arthrobacter species do not provide support for thisaffiliation, although it is possible that Renibacterium was derivedfrom an Arthrobacter ancestor by changes in these and othercharacters.The genus Rothia includes a well-defined clade composedof the type species Rothia dentocariosa and Rothia aeria, amarae,mucilaginosa, nasimurium, and terrae. These taxa are also relatedto some species of Kocuria.The genus Sinomonas comprises the type species Sinomonasflava and Sinomonas atrocyanea. In rRNA gene trees, it is affili-ated with the clade containing the genera Nesterenkonia, Yan-iella, and the Arthrobacter group 2 species Arthrobacter albus andcumminsii.The genus Yaniella contains the type species Yaniella halo-tolerans and Yaniella flava. Although originally classified in itsown family (Li et al., 2008), the rRNA gene trees calculatedhere suggest it is closely related to Nesterenkonia, Sinomonas,and Arthrobacter group 2 species Arthrobacter albus and cummin-sii. This conclusion is consistent with similarities in cell wall,menaquinone and phospholipid compositions (Yassin et al.,2011). However, the DNA G+C content is quite different, 53–58mol% in Yaniella and 64–72 mol% in Nesterenkonia.The genus Zhihengliuella comprises the type species Zhiheng-liuella halotolerans and Zhihengliuella alba and is closely relatedto the subclades of Arthrobacter. Presumably, this relationshipreflects the heterogeneity of the genus Arthrobacter rather thanthe need for reclassification of the genus Zhihengliuella.Family BeutenbergiaceaeThis family comprises the monospecific genera Beutenbergia(type species Beutenbergia cavernae), Miniimonas (type speciesMiniimonas arenae), Salana (type species Salana multivorans),and Serinibacter (type species Serinibacter salmoneus) (Figure 11).When initially proposed, the family also included the genusGeorgenia (Zhi et al., 2009). However, subsequent analyses ledto the reclassification of this genus into the family Bogoriellaceae(Hamada et al., 2009).Family BogoriellaceaeInitially a monogeneric family, this taxon has been emended toinclude the genus Georgenia (Hamada et al., 2009; Stackebrandtand Schumann, 2000). Currently, it comprises the monospecific
13Road mapgenus Bogoriella (type species Bogoriella caseilytica) and Georgeniamuralis (type species), ruanii, and thermotolerans.Family BrevibacteriaceaeThis monogeneric family is well separated from other membersof the order Micrococcales in rRNA gene trees (Figure 9). Thegenus Brevibacterium contains four clades. The first clade com-prises the type species Brevibacterium linens and Brevibacteriumantiquum, aurantiacum, avium, casei, celere, epidermidis, iodinum,marinum, oceani, permense, picturae, sandarakinum, and sanguinis.The second clade comprises Brevibacterium luteolum and otitidis.The third clade comprises Brevibacterium massiliense, mcbrellneri,paucivorans, and ravenspurgense. The fourth clade comprisesBrevibacterium album, pityocampae, and samyangense. Lastly, on thebasis of unpublished rRNA gene sequences, the species Brevibac-terium halotolerans and frigoritolerans appear to be misclassifiedand represent strains of Bacillus.Family CellulomonadaceaeThe family Cellulomonadaceae was proposed by Stackebrandtand Prauser (1991) to include the genera Cellulomonas, Jonesia, Oer-skovia,andPromicromonospora.Subsequentproposalselevatedthegenera Jonesia and Promicromonospora to the family level (Raineyet al., 1995; Stackebrandt et al., 1997). Thus, in the previousroad map (Garrity et al., 2005), the family Cellulomonadaceaecomprised the genera Cellulomonas, Oerskovia, and Tropheryma.Since that time, three subsequently described genera have beenadded (see below). In the current rRNA gene analyses, this fam-ily is paraphyletic, and the family Sanguibacteraceae appears asa specific relative of Oerskovia (Figure 11). This relationship isconsistent with similarities in cell-wall composition and in themenaquinone and fatty acid profiles of these groups (Stack-ebrandt and Schumann, 2012). In addition, Tropheryma is notclosely related to the other genera classified in the family andParaoerskoviaOerskoviaCellulomonasActinotaleaDemequinaSalanaSerinibacterBeutenbergiaBeutenbergiaceaeGeorgeniaBogoriellaPromicromonosporaMyceligeneransXylanimicrobiumXylanibacteriumXylanimonasIsoptericolaCellulosimicrobiumRuaniaHaloactinobacteriumMiniimonasBogoriellaceaeRuaniaceaeCellulomonadaceaeSanguibacter − SanguibacteraceaePromicromonosporaceaeRarobacter − RarobacteraceaeFIGURE 11. Genera of the Beutenbergiaceae, Bogoriellaceae, Cellulomonadaceae, Promicromonosporaceae, Rarobacteraceae,Ruaniaceae, and Sanguibacteraceae of the order Micrococcales. Analyses were performed as described for Figure 1.
14 Road mapappears as a deep lineage related to the family Microbacteriaceae(Figure 13). However, because of the low sequence similaritiesto Microbacteriaceae and differences in DNA G+C content (noth-ing else is known of its chemotaxonomic properties), the genusTropheryma may warrant reclassification into a novel family.The genus Cellulomonas appears as multiple short branchesat the base of the family tree and hence is not well resolved intoclades. It comprises the type species Cellulomonas flavigena andCellulomonas aerilata, biazotea, bogoriensis, cellasea, chitinilytica, com-posti, denverensis, fimi, gelida, hominis, humilata, iranensis, persica,terrae, uda, and xylanilytica.The monospecific genus Actinotalea (type species Actinotaleafermentans) was formed by reclassification of the Cellulomonasspecies based upon its unusual menaquinone composition[MK-10(H4)] and low rRNA gene sequence similarities.In rRNA gene trees, the genus Demequina is clearly separatedfrom other members of the family (Figure 11). It comprises thetype species Demequina aestuarii and Demequina lutea.The genus Oerskovia forms a subclade within the family Cel-lulomonadaceae that includes the genus Paraoerskovia and thefamily Sanguibacteraceae. It comprises the type species Oerskoviaturbata and three closely related species Oerskovia enterophila, jen-ensis, and paurometabola.The monospecific genus Paraoerskovia (type species Paraoer-skovia marina) is related to the genus Oerskovia and the familySanguibacteraceae.Although still currently classified within the family Cellulo-monadaceae, the monospecific genus Tropheryma (type speciesTropheryma whipplei) appears to warrant reclassification.Family DermabacteraceaeThis family was proposed by Stackebrandt et al. (1997) to includethe genera Dermabacter and Brachybacterium (Stackebrandt et al.,1997). This classification is unchanged apart for the addition oftwo monospecific genera, Devriesea (type species Devriesea agama-rum) (Martel et al., 2008) and Helcobacillus (type species Helcoba-cillus massiliensis) (Renvoise et al., 2009), which were describedafter the deadline for inclusion in this volume. The members ofthis family that have been tested possess an unusual peptidogly-can type (A4g), which contains meso-diaminopimelate and aninterpeptide bridge of d-dicarboxylic amino acids. The familyalso contains the monospecific genus Dermabacter (type speciesDermabacter hominis) and Brachybacterium faecium (type species),alimentarium, conglomeratum, fresconis, muris, nesterenkovii, para-conglomeratum, phenoliresistens, rhamnosum, sacelli, tyrofermentans,and zhongshanense. These four genera form a well-defined cladethat clusters with the families Actinomycetaceae and Jonesiaceae inthe rRNA gene trees (Figure 10).Family DermacoccaceaeThis family was proposed by Stackebrandt and Schumann (2000)to include the genera Dermacoccus, Demetria, and Kytococcus basedupon analyses of rRNA genes. This classification was unchangedin the previous road map (Garrity et al., 2005) and subsequentanalyses (Zhi et al., 2009). However, in the rRNA gene trees pre-pared for this volume, Kytococcus appears to be unrelated to theother members of the family and clusters instead with somegenera from the family Intrasporangiaceae, which is also paraphyl-etic (Figure 12). However, comparison of chemotaxonomicmarkers does not support the reclassification of Kytococcus,which possesses completely unsaturated menaquinones and aninterpeptide bridge composed of l-Lys–d-Glu2(Stackebrandt,2012a). While these features are not found elsewhere within thefamily Dermacoccaceae, they are also absent from the familyIntrasporangiaceae (Kämpfer and Groth, 2012). Thus, the classifi-cation of Kytococcus has not been changed here.The genus Dermacoccus comprises the type species Derma-coccus nishinomiyaensis and Dermacoccus abyssi, barathri, and pro-fundi. It is closely related to the monospecific genus Demetria(type species Demetria terragena). The species of Kytococcus forma separate and well-defined clade comprising the type speciesKytococcus sedentarius and Kytococcus aerolatus and schroeteri.Family DermatophilaceaeThe current family was proposed by Stackebrandt et al. (1997)and retained in the previous road map (Garrity et al., 2005).However, in the rRNA gene trees prepared for this volume,the relationship between Dermatophilus and Kineosphaera, thetwo genera in the family, is not well supported (Figure 12).Similarly, the physiological and chemotaxonomic propertiesof these genera are quite different, suggesting that this familymay warrant re-examination (Stackebrandt, 2012c). The genusDermatophilus comprises the type species Dermatophilus congolen-sis and Dermatophilus chelonae. The genus Kineosphaera comprisesonly the type species Kineosphaera limosa.Family IntrasporangiaceaeIn the original proposal, this family comprised the generaIntrasporangium, Sanguibacter, and Terrabacter (Stackebrandtet al., 1997). Subsequently, the genera Janibacter and Terracoc-cus were added, and Sanguibacter was moved to a new family(Stackebrandt and Schumann, 2000). The genera Knoellia,Ornithinicoccus, Ornithinimicrobium, “Candidatus Nostocoida”,and Tetrasphaera were then added in the previous version ofthe road map (Garrity et al., 2005). “Candidatus Nostocoida”was subsequently united with the genus Tetrasphaera (McKenzíeet al., 2006). In the current road map, the family includes theremaining genera plus the genera Arsenicicoccus, Fodinibacter,Humibacillus, Humihabitans, Kribbia, Lapillicoccus, Marihabitans,Oryzihumus, Phycicoccus, and Serinicoccus.However, in the rRNA gene trees prepared for this volume,the family Intrasporangiaceae appears to be paraphyletic andcomprises a major clade of genera closely related to the typegenus Intrasporangium; a second clade consisting of the generaJanibacter, Knoellia, Marihabitans, and Kytococcus of the familyDermacoccaceae; a third clade composed of the genera Orni-thinicoccus, Ornithinimicrobium, and Serinicoccus; and the gen-era Arsenicicoccus and Kribbia, each of which represent deeplineages unaffiliated with other genera (Figure 12). However,this alternative phylogeny is not strongly supported by othertaxonomic evidence. As discussed above, chemotaxonomic datadoes not provide support for the reclassification of the genusKytococcus with the Intrasporangiaceae genera in this clade. More-over, the chemotaxonomic properties of most genera in thefamily appear relatively uniform. In contrast, while most of thefamily possesses diaminopimelate as the cell-wall diamino acid,the genera Ornithinicoccus, Ornithinimicrobium, and Serinicoccuspossess ornithine (Kämpfer and Groth, 2012), providing somesupport for the biological significance of the third clade. Never-theless, in the absence of additional evidence, the genera of thefamily Intrasporangiaceae have not been reclassified at this time.
15Road mapThe first clade observed in the rRNA gene trees comprisesIntrasporangium calvum; Fodinibacter luteus; Humibacillus xantho-pallidus; Humihabitans oryzae; Lapillicoccus jejuensis; Oryzihumusleptocrescens; Phycicoccus jejuensis (type species), aerophilus, big-eumensis, and dokdonensis; Terrabacter tumescens (type species),aerolatus, lapilli, terrae, and terrigena; Terracoccus luteus; and Tet-rasphaera japonica (type species), australiensis, duodecadis, elon-gata, jenkinsii, remsis, vanveenii, and veronensis. The second cladeencompasses Janibacter limosus (type species), anophelis, coral-licola, hoylei, melonis, and terrae; Knoellia sinensis (type species),aerolata, and subterranea; and Marihabitans asiaticum. The thirdclade consists of Ornithinicoccus hortensis; Ornithinimicrobiumhumiphilum (type species), kibberense, and pekingense; and Serini-coccus marinus. Lastly, Arsenicicoccus bolidensis (type species) andArsenicicoccus piscis and Kribbia dieselivorans do not appear to beaffiliated with other genera.Family JonesiaceaeThis family was proposed by Stackebrandt et al. (1997) to rec-ognize the distinctiveness of the genus Jonesia (Rainey et al.,1995). This genus comprises the type species Jonesia denitrificansand Jonesia quinghaiensis. In the rRNA gene trees prepared forthis volume, this taxon appears as a deep lineage specificallyrelated to the order Actinomycetales (Figure 9).Family MicrobacteriaceaeThis large family represents a distinct lineage within the orderMicrococcales (Figure 9). In rRNA gene trees, the family containssome well-defined clades, as well as a number of genera whichare not specifically related to any other genus (Figure 13). Theclades include: Microbacterium, Mycetocola, Okibacterium, andPlantibacter; Clavibacter, Cryobacterium, and Klugiella; AgromycesTerrabacterIntrasporangiumPhycicoccusJanibacterKytococcusKnoelliaKribbia dieselivoransDermacoccusOrnithinimicrobiumDermatophilusTerracoccusHumihabitansHumibacillusLapillicoccusFodinibacterOryzihumusMarihabitansOrnithinicoccusArsenicicoccusKineosphaeraDermatophilaceaeDermacoccaceaeDermacoccaceaeTetrasphaeraDemetriaSerinicoccusIntrasporangiaceaeIntrasporangiaceaeIntrasporangiaceaeFIGURE 12. Genera of the families Dermacoccaceae, Dermatophilaceae, and Intrasporangiaceae of the order Micrococ-cales. Analyses were performed as described for Figure 1.
16 Road mapand Humibacter; Agreia and Subtercola; Gulosibacter and Pseudo-clavibacter; Microcella and Yonghaparkia; Microterricola and Phyci-cola; and Rhodoglobus and Salinibacterium. Although assigned tothe family Cellulomonadaceae, the genus Tropheryma appears as adeep lineage related to this family.The large genus of Microbacterium comprises many closelyrelated species. The type species Microbacterium lacticum formsa well-defined clade with Microbacterium aoyamense, aurum,awajiense, deminutum, flavum, fluvii, hatanonis, invictum, koreense,lacus, pumilum, pygmaeum, schleiferi, terregens, and terricola. Inaddition, specific clusters of the following species are formed:Microbacterium gubbeenense, indicum, and luticocti; Microbacteriumaurantiacum, chocolatum, and kitamiense; Microbacterium arborescensand imperiale; and Microbacterium aquimaris and halophilum. How-ever, the remaining species do not appear specifically relatedto any other species in rRNA gene trees. These taxa includeMicrobacterium aerolatum, agarici, arabinogalactanolyticum, barkeri,binotii, dextranolyticum, esteraromaticum, flavescens, foliorum, ginsen-gisoli, halotolerans, hominis, humi, hydrocarbonoxydans, insulae, ker-atanolyticum, ketosireducens, kribbense, laevaniformans, liquefaciens,luteolum, marinilacus, maritypicum, natoriense, oleivorans, oxydans,paludicola, paraoxydans, phyllosphaerae, profundi, pseudoresistens,resistens, saperdae, sediminicola, soli, terrae, testaceum, thalassium,trichothecenolyticum, ulmi, and xylanilyticum.The genus Agreia comprises the type species Agreia bicolorataand Agreia pratensis. It appears to be related to the genus Sub-tercola.The genus Agrococcus includes the type species Agrococcus jen-ensis and three closely related species of Agrococcus baldri, citreus,and lahaulensis. The genus also contains three more distantlyrelated species, Agrococcus casei, jejuensis, and versicolor.The genus Agromyces encompasses the type species Agromy-ces ramosus and the closely related species Agromyces albus, allii,aurantiacus, bracchium, cerinus, fucosus, hippuratus, humatus, itali-cus, lapidis, luteolus, mediolanus, neolithicus, rhizospherae, salenti-nus, subbeticus, terreus, and ulmi. It also appears to be related tothe genus Humibacter.The monospecific genus of Clavibacter (type species Clavibactermichiganensis) appears to be related to the genera Cryobacteriumand Klugiella.The genus Cryobacterium is formed by the type species Cryobac-terium psychrophilum and the closely related psychrophilic speciesCryobacterium psychrotolerans and roopkundense. The mesophilicspecies Cryobacterium mesophilum forms a deeper lineage relatedto Klugiella xanthotipulae.The genus Curtobacterium represents a deep lineage in thefamily Microbacteriaceae and comprises the type species Curto-bacterium citreum and the closely related species Curtobacteriumammoniigenes, flaccumfaciens, herbarum, luteum, plantarum, andpusillum. The species Curtobacterium ginsengisoli represents adeeper lineage in this genus.The genus Frigoribacterium comprises the type species Frigorib-acterium faeni and Frigoribacterium mesophilum. Although the rRNAgene sequences are similar, these species do not form a distinctclade independent of closely related genera. The conclusionthat Frigoribacterium mesophilum may represent a new genus issupported by differences in the cell-wall amino acids. Frigorib-acterium mesophilum peptidoglycan contains 2,4-diaminobutyricacid, alanine, glycine, glutamate, and lysine (Dastager et al.,2008), whereas that of Frigoribacterium faeni contains alanine,glycine, homoserine, and d-lysine (Kämpfer et al., 2000). Like-wise, Frigoribacterium mesophilum lacks the fatty acid C15:0anteisodimethylacetal, which is characteristic of Frigoribacterium faeni.The monospecific genus Frondihabitans (type species Frondi-habitans australicus) is related to Frigoribacterium faeni.The monospecific genus Gulosibacter (type species Gulosibactermolinativorax) represents a deep lineage in the family along withthe genus Pseudoclavibacter.MycetocolaPlantibacterOkibacteriumKlugiellaClavibacterMicrocellaYonghaparkiaFrigoribacteriumFrondihabitansLabedellaRhodoglobusSalinibacteriumAgreiaMicroterricolaPhycicolaLeifsoniaAgromycesHumibacterSchumannellaCurtobacteriumRathayibacterGulosibacterPseudoclavibacterTropherymaSubtercolaMicrobacteriumAgrococcusLeucobacterCryobacteriumFIGURE 13. Genera of the family Microbacteriaceae and Tropheryma ofthe family Cellulomonadaceae from the order Micrococcales. Analyses wereperformed as described for Figure 1.
17Road mapThe monospecific genus Humibacter (type species Humibacteralbus) is related to the genus Agromyces.The monospecific genus Klugiella (type species Klugiella xan-thotipulae) is related to the genera Cryobacterium and Clavibacterand was described after the deadline for inclusion in this vol-ume (Cook et al., 2008).The monospecific genus Labedella (type species Labedellagwakjiensis) represents a distinct lineage within the familyMicrobacteriaceae.The genus Leifsonia includes the type species Leifsonia aquat-ica and some closely related species: Leifsonia lichenia, naganoen-sis, poae, shinshuensis, and xyli. In addition, other species areincluded in this genus, each of which appear to representa distinct lineage within the family and may thereby warrantreclassification (Evtushenko, 2012b). These include Leifsoniaantarctica, bigeumensis, ginsengi, kafniensis, kribbensis, and pindar-iensis. Lastly, two species, Leifsonia aurea and rubra appear to beclosely related to the genus Rhodoglobus.The genus Leucobacter comprises the type species Leucobacterkomagatae and the closely related species Leucobacter albus, allu-vii, aridicollis, chironomi, chromiireducens, iarius, luti, and tardus.This genus represents a distinct lineage within the familyMicrobacteriaceae.The genus Microcella comprises the type species Microcellaputealis and Microcella alkaliphila and forms a clade with Yong-haparkia.The monospecific genus Microterricola (type species Microter-ricola viridarii) is closely related to Phycicola gilvus.The genus Mycetocola comprises the type species Myceto-cola saprophilus and the closely related Mycetocola lacteus andtolaasinivorans. The remaining species, Mycetocola reblochoni,forms a deeper lineage of the genus. This genus also formsa larger clade with the genera Okibacterium and Plantibacter(Figure 12).The monospecific genus Okibacterium (type species Okibacte-rium fritillariae) forms a larger clade with the genera Mycetocolaand Plantibacter.The monospecific genus Phycicola (type species Phycicola gil-vus) is closely related to Microterricola viridarii.The genus Plantibacter, which contains the type species Plan-tibacter flavus and the closely related Plantibacter auratus, forms alarger clade with the genera Mycetocola and Okibacterium.The genus Pseudoclavibacter comprises the type species Pseudo-clavibacter helvolus and Pseudoclavibacter soli. This lineage repre-sents one of the deepest in the family Microbacteriaceae alongwith the genus Gulosibacter. Moreover, species of the genusZimmermannella, Zimmermannella alba, bifida, and faecalis, arealso in this clade. However, the genus Zimmermannella is a laterhomotypic synonym of Pseudoclavibacter, and is not used in thisvolume. Therefore, these remaining species should be reclas-sified.The genus Rathayibacter comprises the type species Rathay-ibacter rathayi and the closely related species Rathayibacter caricis,festucae, iranicus, toxicus, and tritici. The genus forms a distinctand deep clade in the family Microbacteriaceae.The monospecific genus Rhodoglobus (type species Rhodoglo-bus vestalii) is closely related to Leifsonia aurea and rubra.The genus Salinibacterium includes the type species Salinibac-terium amurskyense and Salinibacterium xinjiangense, both of whichare closely related to Rhodoglobus vestalii.The monospecific genus Schumannella (type species Schuman-nella luteola) represents a deep lineage in the family Microbacte-riaceae but was described after the deadline for inclusion in thisvolume (An et al., 2008).The genus Subtercola, which comprises the type species Sub-tercola boreus and Subtercola frigoramans, is closely related to thegenus Agreia.The monospecific genus Yonghaparkia (type species Yong-haparkia alkaliphila) is closely related to Microcella.Family PromicromonosporaceaeIn rRNA gene trees, this family appears as a well-defined claderelated to the family Rarobacteraceae (Figure 11). While this rela-tionship is not observed in other analyses of the rRNA gene(Stackebrandt and Schumann, 2000; Zhi et al., 2009), similari-ties in the cell walls, fatty acids, and menaquinones providesome support. For instance, the Promicromonosporaceae possesscell-wall type A4a composed of l-Lys, d-Glu or d-Asp, l-Ala orl-Ser or l-Thr (Schumann and Stackebrandt, 2012), whereasthe Rarobacteraceae have the A4b chemotype composed of l-Orn,d-Glu, l-Ala, and d-Ala (and d-Ser) (Kämpfer, 2012b). Membersof each family are rich in iso- and anteiso-branching fatty acidswith C15:0anteiso predominating and have MK-9(H4) as a major,if not predominant, menaquinone. Thus, this relationship war-rants further investigation.Currently, the family Promicromonosporaceae comprises sevengenera. The genus Promicromonospora contains the type speciesPromicromonospora citrea and the closely related Promicromono-spora aerolata, kroppenstedtii, sukumoe, umidemergens, and vindobon-ensis. In rRNA gene trees, Promicromonospora flava represents adeep branch of this lineage. However, phenotypic similaritiesbetween it and the other species do not provide strong supportfor reclassification as a novel genus (Jiang et al., 2009).The genus Cellulosimicrobium comprises the type speciesCellulosimicrobium cellulans and two closely related speciesCellulosimicrobium funkei and terreum.In rRNA gene trees, the genus Isoptericola is represented by acluster of very similar sequences which do not form a distinctclade within the tree. Nevertheless, in the absence of additionalevidence, phenotypic properties support classification within asingle genus. The genus comprises the type species Isoptericolavariabilis and Isoptericola dokdonensis, halotolerans, hypogeus, andjiangsuensis.The genus Myceligenerans, which contains the type speciesMyceligenerans xiligouense and Myceligenerans crystallogenes, isclosely related to the genus Promicromonospora.The monospecific genus Xylanibacterium (type species Xylani-bacterium ulmi) forms a clade with the monospecific generaXylanimicrobium (type species Xylanimicrobium pachnodae) andXylanimonas (type species Xylanimonas cellulosilytica) (Figure 11).Family RarobacteraceaeThis family comprises the type species Rarobacter faecitabidus andRarobacter incanus. It appears to be closely related to the familyPromicromonosporaceae (Figure 11).Family RuaniaceaeThis family comprises two monospecific genera, Ruania (typespecies Ruania albidiflava) and Haloactinobacterium (type species
18 Road mapHaloactinobacterium album) which form a distinct clade withinthe order Micrococcales (Figure 11).Family SanguibacteraceaeAlthough it forms a distinct clade in some analyses (Stack-ebrandt and Schumann, 2000; Zhi et al., 2009) this family isclosely related to the family Cellulomonadaceae in rRNA genetrees prepared here, especially the genera Oerskovia and Parao-erskovia (Figure 11). This relationship is consistent with simi-larities in the cell-wall composition and menaquinone and fattyacid profiles (Stackebrandt and Schumann, 2012). If supportedby additional evidence, reclassification of these genera mightbe warranted. The family comprises the type genus and speciesSanguibacter keddieii and the closely related species Sanguibacterantarcticus, inulinus, marinus, soli, and suarezii.Order “Micromonosporales” and familyMicromonosporaceaeThis order is formed by elevation of the suborder Micromono-sporineae (Stackebrandt et al., 1997) and encompasses a singlefamily Micromonosporaceae. In the 16S rRNA gene trees preparedfor this volume, the order “Glycomycetales” appears as a longbranch specifically related to the genus Actinocatenispora, whichitself is a deep branch within the order “Micromonosporales” (Fig-ure 8). While some chemotaxonomic and other phenotypicproperties tend to support this association, the relationship isconsidered tentative at this time (see above). In addition, whilemost genera appear as deep branches radiating from the baseof the family tree, four clades are observed. These include thegenera Micromonospora, Catellatospora, and Planosporangium;Catelliglobosispora and Hamadaea; Couchioplanes, Krasilnikovia,and Pseudosporangium; and Dactylosporangium and Virgisporan-gium. However, in the absence of additional evidence, the bio-logical significance of these clades is uncertain.In the rRNA gene trees prepared for this volume, the genusMicromonospora is represented by many short branches thatappear throughout the family tree. The genus does not forma distinct clade and other genera are interspersed within theMicromonospora sequences. The type species Micromonosporachalcea appears in the clade containing Actinoplanes, Catella-tospora, Longispora, and Planosporangium species (Figure 8).Other Micromonospora species in this clade include Micromono-spora aurantiaca, carbonacea, chokoriensis, coxensis, halophytica, kra-biensis, lupini, matsumotoense, mirobrigensis, purpureochromogenes,rifamycinica, saelicesensis, and siamensis. A second large group ofMicromonospora species form a multifurcation at the base of thefamily tree. These include Micromonospora auratinigra, chaiyaphu-mensis, chersina, citrea, coerulea, coriariae, eburnea, echinaurantiaca,echinofusca, echinospora, endolithica, fulviviridis, inositola, inyonen-sis, narathiwatensis, nigra, pallida, peucetia, rosaria, sagamiensis,and viridifaciens. Three additional species are also at the baseof the family tree but outside this cluster: Micromonospora olivas-terospora, pattaloongensis, and pisi. In spite of the ambiguities ofthe rRNA gene trees, gyrB, rpoB, or atpD gene trees suggestsMicromonospora olivasterospora is related to Micromonospora carbo-nacea. Likewise, in gyrB gene trees, Micromonospora pattaloongen-sis, Micromonospora pisi and Polymorphospora rubra form a grouprelated to other Micromonospora species, a classification whichis supported by similarities in their chemotaxonomy. Lastly,for the species Micromonospora gallica, the rRNA gene sequenceis not available and the type strain appears to have been lost.In contrast, most of the other genera in this family form well-defined clades.The genus Actinocatenispora comprises the type species Acti-nocatenispora thailandica and Actinocatenispora rupis and sera. Thegenus represents a deep lineage in the family phylogenetic treethat is specifically related to the order “Glycomycetales” (Figure8). Although these taxa share some similarities in cell-wall com-position, this association is considered tentative at this time (seeabove).The genus Actinoplanes comprises the type species Actinoplanesphilippinensis and the closely related species Actinoplanes auranti-color, campanulatus, capillaceus, consettensis, couchii, cyaneus, decca-nensis, derwentensis, digitatis, humidus, italicus, liguriensis, lobatus,missouriensis, palleronii, rectilineatus, regularis, sichuanensis, utahen-sis, and xinjiangensis. This clade represents a deep lineage inthe family. In the rRNA gene trees prepared for this volume, asecond clade is also present that includes Actinoplanes brasilien-sis, durhamensis, ferrugineus, teichomyceticus, tereljensis, and toeven-sis and the monospecific genus Planosporangium (type speciesPlanosporangium flavigriseum). This clade is specifically relatedto Micromonospora chalcea, the type species of the genus. Actino-planes globisporus forms a third clade, which is a deep lineage inthe family. Similar rRNA gene trees were generated by Ara et al.(2010), but not by Wiese et al. (2008) and Vobis et al. (2012).Although high levels of DNA hybridization between members ofthe second clade and Actinoplanes globisporus tend to support thesignificance of this clade (Ara et al., 2010), it is not strongly sup-ported by chemotaxonomic or phenotypic markers and awaitsconfirmation by other methods (Vobis et al., 2012). Lastly, therRNA gene sequence is unavailable for Actinoplanes friuliensis.The genus Asanoa comprises the type species Asanoa fer-ruginea and Asanoa iriomotensis and ishikariensis. These closelyrelated species represent a basal lineage in the family.The genus Catellatospora comprises the type species Catel-latospora citrea and Catellatospora bangladeshensis, chokoriensis,coxensis, and methionotrophica. These closely related species arespecifically related to Micromonospora chalcea, the type speciesof the genus, in rRNA gene trees and possess similar chemot-axonomic features. However, their morphologies are very dif-ferent.The monospecific genus Catelliglobosispora (type species Catel-liglobosispora koreensis) is related to the monospecific genusHamadaea (type species Hamadaea tsunoensis) (Figure 8). Thesetaxa were described after the deadline for inclusion in this vol-ume and are not described further.The genus Catenuloplanes forms a deep lineage in this fam-ily. It comprises the type species Catenuloplanes japonicas andCatenuloplanes atrovinosus, castaneus, crispus, indicus, nepalensis,and niger.The monospecific genus Couchioplanes (type species Couchio-planes caeruleus) forms a clade with the monospecific generaKrasilnikovia (type species Krasilnikovia cinnamomea) and Pseu-dosporangium (type species Pseudosporangium ferrugineum) (Fig-ure 8).The genus Dactylosporangium comprises the type species Dac-tylosporangium aurantiacum and the closely related species Dac-tylosporangium fulvum, matsuzakiense, roseum, thailandense, andvinaceum. This genus forms a clade with Virgisporangium ochra-ceum (type species) and Virgisporangium aurantiacum.
19Road mapThe following genera represent additional deep lineages inthe family tree: the monospecific genus Longispora (type speciesLongispora albida); Luedemannella, comprising the type speciesLuedemanella helvata and Luedemanella flava; Pilimelia, contain-ing the type species Pilimelia terevasa and Pilimelia anulata andcolumellifera; the monospecific genus Plantactinospora (type spe-cies Plantactinospora mayteni); the monospecific genus Polymor-phospora (type species Polymorphospora rubra); Rugosimonospora,encompassing the type species Rugosimonospora acidiphila andRugosimonospora africana; Salinispora, comprising the type spe-cies Salinispora arenicola and Salinispora tropica; the monospecificgenus Spirilliplanes (type species Spirilliplanes yamanashiensis);and Verrucosispora, comprising the type species Verrucosispora gif-hornensis and Verrucosispora lutea and sediminis.Order “Propionibacteriales”This order is formed by elevation of the suborder Propionibacter-ineae (Zhi et al., 2009). It forms a well-defined clade comprisingthe families Propionibacteriaceae and Nocardioidaceae (Figure 14).However, two genera currently classified within the familyNocardioidaceae appear as deep lineages outside of that familybut still within the order (see below). The current taxonomyclosely resembles that of the previous road map (Garrity et al.,2005) except that the genera Friedmanniella and Micropruinahave been transferred from the family Nocardioidaceae to thefamily Propionibacteriaceae. This reclassification is well supportedby rRNA gene trees (Figure 14; Zhi et al., 2009) and chemot-axonomic markers (Schumann and Pukall, 2012).Family PropionibacteriaceaeThis family comprises 13 genera grouped within five clades(Figure 14). The first clade includes the genus Propionibacterium,which forms two of the five subclades recognizable in this clade.The first subclade comprises the type species Propionibacteriumfreudenreichii and Propionibacterium acidifaciens, australiense, andcyclohexanicum. The second subclade comprises Propionibacte-rium acidipropionici, acnes, avidum, granulosum, jensenii, microaero-philum, propionicum, and thoenii. While the ‘classical’ and ‘dairy’Propionibacterium species are found in both subclades, the‘cutaneous’ species Propionibacterium acnes, avidum, and granu-losum are all in the second subclade (Patrick and McDowell,2012). Moreover, meso-diaminopimelic acid is the dominantcell-wall diamino acid only in members of the first subclade,while ll-diaminopimelic acid is the dominant diamino acid inall tested species of the second subclade, observations whichprovide further support for the significance of these subclades.The third and fourth subclades comprise the monospecificgenera Brooklawnia cerclae and Propionimicrobium lymphophilum,ActinopolymorphaKribbellaMicrolunatusFriedmanniellaPropionibacteriaceaeLuteococcusTessarococcusPropionibacteriumAeromicrobiumMarmoricolaNocardioidaceaePropioniferaxGranulicoccusMicropruinaBrooklawnia cerclaePropionimicrobiumAestuariimicrobiumPropionicicellaPropionicimonasNocardioides"Propionibacteriales"FIGURE 14. Genera and families of the order “Propionibacteriales”. Analyses were performed as described for Figure 1.
20 Road maprespectively. The genus Tessaracoccus forms the fifth subcladeand encompasses the type species Tessaracoccus bendigoensis andclosely related species Tessaracoccus flavescens and lubricantis.The genus Luteococcus forms a deep branch of first clade andincludes the type species Luteococcus japonicus and the closelyrelated species Luteococcus peritonei and sanguinis.The second clade comprises the genera Friedmanniella andMicrolunatus. The genus Friedmanniella is composed of the typespecies Friedmanniella antarctica and the closely related speciesFriedmanniella capsulata, lacustris, lucida, luteola, okinawensis,sagamiharensis, and spumicola. The genus Microlunatus com-prises the type species Microlunatus phosphovorus and the closelyrelated species Microlunatus aurantiacus, ginsengisoli, panaciter-rae, and soli.The monospecific genera Granulicoccus (type species Granuli-coccus phenolivorans) and Propioniferax (type species Propioniferaxinnocua) form the third clade.The monospecific genera Micropruina (type species Micro-pruina glycogenica), Propionicicella (type species Propionicicellasuperfundia), and Propionicimonas (type species Propionicimonaspaludicola) form the fourth clade.The monospecific genus Aestuariimicrobium (type species Aes-tuariimicrobium kwangyangense) represents an independent lin-eage within this family and the fifth clade.Family NocardioidaceaeIn the previous road map (Garrity et al., 2005) and in theanalyses of Zhi et al. (2009), this family included the generaNocardioides, Actinopolymorpha, Aeromicrobium, Kribbella, and Mar-moricola. However, in the rRNA gene trees prepared for thisvolume, the family is paraphyletic with the genera Actinopolymor-pha and Kribbella, appearing outside the taxon as deep lineageswithin the order “Propionibacteriales” (Figure 14). This relation-ship is quite robust and independent of whether phylum- ororder-specific filters (or nucleotide positions) are used duringtree construction and might suggest that these genera warrantreclassification. However, chemotaxonomic evidence does notprovide strong support for either the current classification ora new one. On one hand, all five genera possess similar cellwalls characterized by the tetrapeptide l-Ala–d-Glu–ll-diamin-opimelate–d-Ala, a single glycine residue in the interpeptidebridge and similar fatty acids (Evtushenko, 2012a). On theother hand, teichoic acids have been found in all species of thegenera Nocardioides and Aeromicrobium that have been analyzed.In contrast, some Kribbella strains possess unusual phosphate-less acidic glycopolymers called teichulosonic acids and lackteichoic acids. Moreover, unlike the other genera, Actinopo-lymorpha forms branched fragmenting mycelia. Thus, in theabsence of strong support for the reclassification of the generaActinopolymorpha and Kribbella, these taxa are retained in thisfamily at this time.The genus Nocardioides comprises a large clade of speciesclosely related to the type species, as well as a smaller numberof more distantly related species. The species closely related tothe type species Nocardioides albus include Nocardioides, aestuarii,agariphilus, aquaticus, aquiterrae, aromaticivorans, bigeumensis,caeni, exalbidus, fonticola, furvisabuli, ganghwensis, ginsengisoli,hankookensis, humi, hwasunensis, insulae, islandensis, kongjuensis,kribbensis, lentus, luteus, marinus, nitrophenolicus, oleivorans, panaci-humi, panacisoli, plantarum, pyridinolyticus, sediminis, simplex,terrae, terrigena, and tritolerans. The more distantly related speciesform seven clades: Nocardioides alkalitolerans; Nocardioides basal-tis, dokdonensis, marinisabuli, and salarius; Nocardioides daphniae;Nocardioides dilutus and halotolerans; Nocardioides dubius; Nocar-dioides jensenii; and Nocardioides koreensis. These species appearto be about equally related to Nocardioides albus and the genusMarmoricola (Figure 12).The genus Actinopolymorpha is composed of the type speciesActinopolymorpha singaporensis and the closely related speciesActinopolymorpha alba, cephalotaxi, and rutila.The genus Aeromicrobium comprises the type species Aeromi-crobium erythreum and the closely related species Aeromicrobiumalkaliterrae, fastidiosum, flavum, ginsengisoli, marinum, panaciter-rae, ponti, and tamlense.The genus Kribbella comprises the type species Kribbella flav-ida and the closely related species Kribbella alba, aluminosa, anti-biotica, catacumbae, ginsengisoli, hippodromi, jejuensis, karoonensis,koreensis, lupini, sancticallisti, sandramycini, solani, swartbergensis,and yunnanensis.The genus Marmoricola encompasses the type species Mar-moricola aurantiacus and the closely related species Marmoricolaaequoreus and bigeumensis.Order “Pseudonocardiales” and familyPseudonocardiaceaeThis order is formed by elevation of the suborder Pseudonocar-dineae (Zhi et al., 2009). In the previous road map (Garrity et al.,2005) and the analyses of Zhi et al. (2009), the suborder com-prised two families, Pseudonocardiaceae and Actinosynnemataceae.However, the taxonomic status of these families was recently re-examined (Labeda et al., 2011). Based on the rRNA gene trees,chemotaxonomic evidence, and morphology, it was concludedthat the retention of the family Actinosynnemataceae was not war-ranted and that the genera it contained should be united intothe family Pseudonocardiaceae. In the rRNA gene trees preparedfor this volume, this enlarged family contains two clades whichcorrespond to some of the genera assigned to the each of theoriginal families, as well as numerous lineages that arise fromthe base of the family tree (Figure 15). The first clade includesonly genera previously assigned to the Pseudonocardiaceae andcomprises the genera Pseudonocardia, Actinomycetospora, Amyco-latopsis, Kibdelosporangium, Prauserella, Saccharomonospora, Sac-charopolyspora, Sciscionella, and Thermocrispum. The second cladeincludes only genera previously assigned to the family Actinosyn-nemataceae and contains the genera Actinosynnema, Lechevalieria,Lentzea, Saccharothrix, and Umezawaea. Although not readily dis-tinguished by most chemotaxonomic criteria, the whole-cellsugars of these clades are different. The first clade possessesarabinose and galactose, and the second clade possesses galac-tose, mannose, and sometimes either rhamnose or ribose.Previously, Thermobispora was also assigned to the Pseudono-cardiaceae (Garrity et al., 2005). Re-examination of this relation-ship suggested that this genus was more closely related to theorder “Streptosporangiales” (Figure 16; Zhi et al., 2009). However,because of the absence of corroborating evidence for this affili-ation, this genus is classified as an order incertae sedis in the cur-rent road map.Four subclades are apparent in the clade comprised exclu-sively of genera from the original family Pseudonocardiaceae(Figure 15). The first subclade includes the genus Pseudonocardia