Proposed minimal standards for describing new genera and species of the suborder

Correspondence P. Schumann psc@dsmz.de DSMZ-Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Inhoffenstraße 7B, 38124 Braunschweig, Germany Institut für Angewandte Mikrobiologie, Justus-Liebig-Universität, 35392 Giessen, Germany Institut für Bakteriologie, Mykologie und Hygiene, Veterinärmedizinische Universität, A-1210 Wien, Austria All-Russian Collection of Microorganisms (VKM), G. K. Skryabin Institute of Biochemistry and Physiology of Microorganisms, RAS, Pushchino, Moscow Region 142290, Russia


INTRODUCTION
The suborder Micrococcineae was established by Stackebrandt et al. (1997) on the basis of a characteristic set of 16S rRNA gene signature nucleotides. Subsequent reclassifications and descriptions of additional genera have necessitated the dynamic adaptation of the originally proposed sets of signature nucleotides since new members added to the suborder have given a fresh insight into its genotypic diversity (Stackebrandt & Schumann, 2006;Zhi et al., 2009). Consequently, an emended description of the suborder Micrococcineae was proposed on the basis of a revised set of 16S rRNA gene signatures (Zhi et al., 2009) Stackebrandt & Schumann, 2000b;Li et al., 2005Li et al., , 2008Zhi et al., 2009) and 86 genera, including two pairs of genera with names that are heterotypic synonyms (Microbacterium/ Aureobacterium and Rothia/Stomatococcus) and one pair of homotypic synonyms (Pseudoclavibacter/Zimmermannella) (Table 1 and http://www.the-icsp.org/taxa/ micrococcineaelist.htm).
Many of the genera presently considered to be members of the suborder Micrococcineae were formerly dealt with by the Subcommittees No. 8 on Arthrobacter and Related Organisms and No. 9 on Microbacterium and Related Organisms of the International Committee on Systematic Bacteriology (now the International Committee on Systematics of Prokaryotes, ICSP). Minimal standards for the description of taxa related to the genera Arthrobacter and Microbacterium were proposed by Schumann & Prauser (1995). These minimal standards are in need of updating in the light of current taxonomic classifications, the numerous descriptions of novel taxa and the new methodological approaches to bacterial systematics. In order to comprehensively describe a new genus or species of the suborder Micrococcineae, to allow a clear identification of the novel taxon and to maintain the taxonomic soundness of the suborder, it is recommended that the following minimal standards are adopted.

General considerations
The general prerequisites for the description of new taxa and their designation of type genera, type species and type strains have been outlined in the Bacteriological Code (1990 Revision) (Lapage et al., 1992), with several proposals for modifications (e.g. Labeda, 1997Labeda, , 2000Tindall, 1999Tindall, , 2008Tindall et al., 2006;Tindall & Garrity, 2008), in previous proposals for minimal standards (e.g. Bernardet et al., 2002) and in the Instructions to Authors of the International Journal of Systematic and Evolutionary Microbiology (http://ijs.sgmjournals.org). Type strains of novel species must be deposited in two publicly accessible culture collections in different countries, as recommended by the Judicial Commission on Prokaryote Nomenclature. For further details regarding the practice of valid publication of names, authors should consult Tindall et al. (2006), Tindall (2008) and Tindall & Garrity (2008).
It is essential that the source of the strains representing the new taxon is described comprehensively and unambiguously with respect to geographical location (country, region, mountain, lake etc., preferably with geographical coordinates) and origin of the sample (host organism, environmental sample such as type of soil, sediment, depth of water etc.). The conditions of isolation, maintenance and preservation must be reported.
The polyphasic approach, which integrates genomic and phylogenetic information with phenotypic and chemotaxonomic data (Vandamme et al., 1996), is currently considered the general principle for taxonomic conclusions. Minimal standards for the description of new taxa according to Recommendation 30b of the Bacteriological Code (1990 Revision) (Lapage et al., 1992) have to be established in accordance with this principle.
The differentiation of genera of the suborder Micrococcineae is primarily based on 16S rRNA gene sequence comparisons in combination with chemotaxonomic characteristics. Members of the suborder are often difficult to differentiate by morphology, but show a broad chemotaxonomic diversity (Table 2). In many cases, the chemotaxonomic characteristics are consistent with the phylogenetic position of organisms ( Fig. 1) and are regarded as important phenotypic traits in the differentiation of genera of this suborder. Therefore chemotaxonomic characteristics need to be carefully analysed and evaluated for descriptions of novel taxa (Stackebrandt & Schumann, 2006).
Many 'classical' phenotypic characteristics, including morphological, physiological and biochemical features used for the description of the phenotype may be affected by cultural or test conditions. A comparative experimental study of phenotypic characteristics in representatives of putative novel species and of type strains of recognized taxa in the authors' own laboratory rather than comparisons with data reported in the literature is recommended (see also paragraph on cellular fatty acids). This is of particular importance when the organisms to be compared appear highly related on the basis of 16S rRNA gene sequences and where the species description is based on only a single strain. The number and significance of differentiating phenotypic characteristics must be sufficiently high in order to delineate the species borders as reliably as possible and to ensure that a new isolate of the species will be phenotypically recognizable.
Several genera of the suborder Micrococcineae (Bogoriella, Demequina, Humibacter, Humihabitans, Ruania, Serinicoccus, Terracoccus etc.) are represented by the type strains of their type species only. The description of new taxa should preferably be based on many isolates from different sources if possible since single strains cannot reflect the full biological diversity of species and genera. However, the description of a novel taxon with extraordinary combinations of characteristics should not be inhibited by the fact that intensive isolation programs have failed to provide additional strains. The description of taxa represented by only a few isolates or even a single strain poses a special challenge as regards the selection of appropriate techniques for characterization.

Genus
Diamino      DThe monounsaturated 12-methyl-tetradecenoic acid (ai-C 15 : 1 ) may also be present in organisms of other genera when grown at temperatures below 10 uC. dGlycine or glycine amide bound to a-carboxyl group of glutamic acid at position 2 of peptide subunit. §Peptidoglycan structure supposed on the basis of data from Sasaki et al. (1998) and Behrendt et al. (2002). ||D-Alanine amide bound to a-carboxyl group of glutamic acid at position 2 of peptide subunit.

P. Schumann and others 1832
International Journal of Systematic and Evolutionary Microbiology 59 ined in most laboratories are recommended in order to support the differentiation of special groups (Table 4). Taxonomists should feel positively encouraged to apply additional taxonomic methods to supplement those specified here and to continue to search for novel tools that are useful for the differentiation of taxa within this suborder.

CULTURAL AND BIOCHEMICAL CRITERIA Colony appearance
Most of the 'classical' taxonomic characteristics used and recommended for the description of genera and species of the suborder Micrococcineae are the same as those employed for many other bacteria and are available from the literature (e.g. Smibert & Krieg, 1994). Some characteristics and methods which reflect the specificity of the suborder under consideration are outlined below.
The colony characteristics, including colour, shape, diameter, edge, elevation, surface, consistency and transpar-ency, penetration into agar, etc., may serve for the (preliminary) differentiation of species when observed under standardized conditions (Bergey's Manual;Sneath et al., 1986). The composition of media and growth conditions must be specified, since the colony characteristics (e.g. colony colours of related species) tend to change and depend on the growth conditions. Growth on agar media supplemented with vitamins, Casitone (milk) and yeast extract, as well as incubation in full daylight, often induce or enhance pigment formation (Koyama et al., 1974;Vidaver & Davis, 1988;Holt & Krieg, 1994;Trutko et al., 2005). Attention must be paid to controlling pH values since deviations from a given pH may markedly alter the colony colour (Turner & Messenger, 1986). When organisms with alkaliphilic (alkalitolerant) or acidophilic (acidotolerant) properties and their relatives that prefer neutral pH values are compared, the colony colour should be tested at different pH values. For those representatives of the suborder which form aerial mycelia, the use of diagnostic ISP media (Shirling & Gottlieb, 1966) is often appropriate.   (are) phenotypically clearly different from the most closely related species on the basis of at least five traits which must have been studied concurrently. Clear evidence may be provided by examination of several strains for which non-clonal relatedness was demonstrated. Species identity and nonclonal relatedness may be demonstrated by genomic fingerprinting (high degree of similarity but differences still present in the banding patterns).

Morphology
The notion of the taxonomic value of morphological characters in the classification and differentiation of genera and species in the group under consideration has undergone changes in recent decades.
The importance of morphological characteristics seems to increase for genera since the range of available chemotaxonomic generic criteria appears insufficient for the discrimination of some phylogenetically neighbouring coherent groups for which generic status is claimed (for examples see IJSEM, recent publications). Generalized terms such as 'cocci', 'irregular rods', 'coryneform morphology' and 'nocardioform morphology' are currently widely used in the descriptions of novel genera.
However, for both genera and species, detailed descriptions of the morphology must be provided, including cell size and shape, both at earlier and later growth phases, the type of cell division, cell arrangements, the formation of substrate (primary) and aerial (secondary) mycelium and their fragmentation, and the production and characterization of arthrospores or, if appropriate, other resting cells. As some representatives of the suborder Micrococcineae form spores (arthrospores), e.g. species of the genera Arthrobacter (Tanaka et al., 2001) and Myceligenerans (Cui et al., 2004), it is recommended that it should be stated precisely which kind of spores or resting cells are observed and that the term 'non-endospore-forming' rather than 'non-spore-forming' is used in relevant cases. It is recommended that morphological changes during growth should be tested and noted and that the presence (or absence) of a specific life cycle should be determined. All of the above characteristics should be examined on at least two appropriate agar media of different compositions (e.g. rich in organics as well as chemically defined). The use of relatively old cultures or resting cells is preferable for inoculation in order to synchronize the development of subsequent populations. It is recommended that light or electron micrographs of the culture are presented at different growth phases, particularly for the type strains of the type species of newly described genera.

Motility
The simplest method to examine motility is the 'hangingdrop method'. Motility can be concluded from the presence of flagella as demonstrated by electron microscopy. Additional methods to test for motility and the formation of flagella were proposed by Holt & Krieg (1994). The inoculation and testing conditions may be important for some groups. Good results are usually obtained for cultures on semi-solid agar media (agar, 0.2-0.4 %) obtained after inoculation with relatively old cells (1-2 weeks) from agar media. Since not all cells in a population produce flagella, a sample taken from the periphery of colonies is preferable. It should be taken into account that in many representatives of the suborder Micrococcineae, flagella are formed at a later growth phase, i.e. when the cells are at the division stage and/or when the first generations of cells have already consumed the nutrient compounds around the colonies. Even for rapidly growing organisms, it is recommended that motility is examined for a growth period of up to 7-10 days. It should also be taken into account that high temperatures, the presence of glucose or low humidity may inhibit the formation of flagella (Holt & Krieg, 1994) and that cells may lose flagella during sample preparation for electron microscopy. It should be emphasized that the absence of flagella does not necessarily mean that the cells are nonmotile. Members of the suborder may possess other, not yet fully elucidated mechanisms of motion, e.g. via pili ('twitching motility'). The method used for testing motility and the type of movement observed must be clearly stated.

Gram reaction
Although members of the suborder Micrococcineae belong to the Gram-positive bacteria in terms of cell-wall structure, the staining reaction may be negative under special circumstances and differentiation between the terms 'Gram reaction' and 'Gram type' is necessary (Wiegel, 1981). Negative or indistinct Gram-stain has been reported for some members of the suborder, e.g. the family Microbacteriaceae, even in relatively young cultures. It is recommended that cultures should be subjected to the Gram staining procedure just after growth has become visible and that staining is tested again at a later exponential growth phase. As Gram-positive bacteria are in general not lysed by 3 % KOH, this simple test is recommended as an additional criterion for the classification of the Gram-reaction status of a novel strain (Ryu, 1938). The test method must be clearly stated.

Growth requirements
The data on growth requirements characterize an organism and are essential for the determination of its optimal cultivation. Suitable media that are as different as possible, the necessary additives (vitamins and other growth factors, extracts, salt, etc.) and the range and optima of pH and growth temperature must be precisely documented. Some useful recommendations for the description of alkaliphilic and alkalitolerant organisms are outlined by Sorokin (2005).

Relationship with oxygen and oxidative or fermentative metabolism
Most members of the suborder Micrococcineae are strict aerobes and are characterized by the oxidative type of metabolism. There are also organisms which grow both under aerobic and anaerobic conditions and display the fermentative type of metabolism in addition to the oxidative type (e.g. representatives of the genus Oerskovia). Besides the aforementioned groups, there are also the microaerophilic representatives, which require low concentrations of oxygen for optimal growth (e.g. Agromyces ramosus) and this characteristic should be noted whenever appropriate. The examination of the capability of an organism to grow in an anaerobic chamber is recommended to demonstrate its relationship with oxygen. Anaerobic chambers are also suited to test the influence of alternative electron acceptors. For testing fermentative properties (oxidative or fermentative utilization of glucose), the oxidation-fermentation test (O/F-test, Hugh-Leifson test) is generally accepted (Leifson, 1962;Smibert & Krieg, 1994). For testing microaerophilic growth, stab cultures in semi-liquid media (0.2-0.4 % agar) can be used. Microaerophilic growth occurs primarily as a distinct or diffuse band below the agar surface (Gledhill & Casida, 1969;Smibert & Krieg, 1994).

Catalase activity
Almost all tested organisms of the considered group possess catalase activity as the main factor inactivating reactive oxygen. To test for presence of catalase, the standard method employing 3 % H 2 O 2 is recommended. Although the method is very simple, it is necessary to follow exactly the general recommendations outlined by Smibert & Krieg (1994). In the case of a negative catalase test, it is advisable to characterize other factors that may be preventing oxygen damage as they might turn out to be differentiating characteristics for the organism under study.

Oxidase activity
The classical 'oxidase test' (Smibert & Krieg, 1994) with N,N,N9,N9-tetramethyl-p-phenylenediamine (most frequently used in this group) or with N,N-dimethyl-pphenylendiamine differentiates two types of aerobic respiratory chains (Unden, 1999). As some organisms grown under different conditions or at different growth stages may show different reactions with N,N,N9,N9tetramethyl-p-phenylenediamine, oxidase activity should be examined in young, actively growing cultures from agar media. The term 'negative oxidase reaction with N,N,N9,N9-tetramethyl-p-phenylenediamine' is recommended instead of 'oxidase-negative', since all aerobic bacteria have oxidases in the respiratory chains. The test method and growth conditions must be clearly stated.
Compared with the simple 'oxidase test', the determination of the terminal oxidase pattern may provide more information on respiratory chains. This characteristic has not been used regularly for taxonomic purposes. However, available data show that it may be helpful for the differentiation of some species and genera of the family Microbacteriaceae. For instance, species of the genus Plantibacter produce only the cytochrome oxidase aa 3 , while the members of the genus Rathayibacter synthesize cytochrome oxidase bb 3 or both bb 3 and aa 3 , depending on growth conditions. Some quinole oxidases, e.g. bo 3 or bb 3 , may be also indicative of species (Trutko et al., 2003). On the other hand, the bd-type oxidases seem to be of little significance for differentiation as they are synthesized at relatively low oxygen concentrations in all organisms tested so far. Growth conditions for testing this property should be standardized. The methods of determination of the terminal oxidase patterns are based on spectrophotometry (preliminary spectral characteristics) and chromatography (Trutko et al., 2003).

Utilization of different carbon sources
Records of the utilization of different carbon sources, including sugars, organic acids, amino acids, etc., as sole carbon source for growth are valuable phenotypic traits in the characterization and delineation of most species within the genera of the suborder Micrococcineae. Several slightly varying basal media can be recommended for different groups of organisms (see e.g. Shirling & Gottlieb, 1966;Smibert & Krieg, 1994; recent publications on particular genera). Although the use of commercial test systems for the examination of the physiological characteristics of novel organisms has been controversial (Whiley & Kilian, 2003;Stackebrandt et al., 2002b), their use is considered acceptable as these kits have been employed in the descriptions of many novel taxa of the suborder due to their ease of use and speed of handling. For comparative analyses it is, however, necessary to study all organisms concurrently with the same tests. Otherwise the data cannot be compared.
The utilization of organic acids is examined either by testing growth on carbon-free basal media supplemented with the compound to be tested or by classical Simmons' citrate tests (basal medium free of additional carbon sources; bromothymol blue indicator) (Smibert & Krieg, 1994). For some groups, the basal medium should be supplemented with required additives (vitamins, salts, traces of amino acids, etc.) in order to obtain unambiguous results.
The number and type of carbon sources tested may vary depending on the group of organisms.

Acid production from carbohydrates and other carbon sources
The methods for the examination of these properties have been described by Smibert & Krieg (1994), Gordon & Mihm (1957) and in publications on particular genera. The test for acid production is performed under aerobic conditions. The following sources can be recommended: D-glucose, L-arabinose, raffinose, L-rhamnose, D-ribose, salicin, sucrose, myo-inositol, D-mannitol, glycerol.
The test media should meet the growth requirements of the particular organism. It is recommended that the minimal concentration of the pH indicator is used since the usually recommended concentrations (developed for other groups of organisms) may inhibit the growth of representatives of the suborder Micrococcineae, in particular of those of the family Microbacteriaceae.
With respect to this and the groups of characteristics mentioned above, it should be noted that both the terms 'utilization of carbon sources' and 'acid production from ...', which are usually used to circumscribe the respective features of members of the suborder Micrococcineae, are covered by the term 'utilization' and there is some confusion over this issue (and hence with the interpretation of results). Though the tests 'growth on sole carbon source' and 'acid production from ...' (carbon-free basal medium) often give the same positive or negative result, they reflect different features of the organisms. Besides, the terms 'utilization of sugars' and 'acid production from ...' are employed to report the results obtained using basal media supplemented with additional carbon sources, e.g. tryptone (API Coryne system, bioMérieux). In addition, the pH indicators (phenol red or bromothymol blue) most frequently used to test acid production show clear positive reactions at somewhat different pH values (6.8 and 6.0, respectively).
Comparisons of data generated by different methods or of author's own data with those from the literature often lead to confusion because of conflicting results. The principles of the tests must be taken into account and the results properly interpreted and described.

Enzymic activities and decomposition of selected substrates
Enzymic activities are used in almost all groups of the suborder for the circumscription of species and also of genera (e.g. Cellulomonas) and are determined by using particular organic substrates (e.g. aesculin, cellulose, starch, Tweens). It must be clearly indicated which method or commercial test system was applied for testing enzymic activities. The number of activities to be examined for unambiguous delineation of species may vary depending on the group of organisms.
The following minimal set of activities is recommended: $ Nitrate reduction determined by the sulfanilic acid and alpha-naphthylamine method and confirmed with zinc dust (Barrow & Feltham, 1993).
$ Urease activity determined on Christensen agar (Christensen, 1946). The red-violet colour of inoculated agar is interpreted as a positive result.
$ Arginine dihydrolase activity tested in tubes containing Møller medium (Møller, 1955), both with and without a covering layer of sterile paraffin oil. Tests should be read daily for a period of 5 days. Development of a violet colour is recorded as a positive result.

H 2 S production
The ability to produce H 2 S is determined by using a growth medium with peptone (rich in organic sulfur) or liquid medium containing 0.05 % cysteine. The latter medium should be preferred as some members of the suborder Micrococcineae display a positive result only if cysteine is used as substrate. Sterile filter strips impregnated with lead acetate are used as the indicator (blacking is regarded as a positive reaction due to the formation of lead sulfide) (Smibert & Krieg, 1994).

Tolerance of NaCl and salt requirement
Representatives of many species and genera of the suborder Micrococcineae are well differentiated with respect to their salinity requirements. Tolerance of high concentrations of NaCl has been reported for some members of the suborder, e.g. members of the genus Nesterenkonia, Yaniella halotolerans (up to 25 % NaCl; Li et al., 2004Li et al., , 2008 and members of the genus Brevibacterium (up to 18 % NaCl; Gavrish et al., 2004). The tolerance of lower salt concentrations clearly differentiates species of many genera commonly known to be non-halotolerant. The ability to grow at different concentrations of NaCl should be tested in organisms of all new taxa, particularly when strains have been isolated from marine environments, saline and alkaline soils and various desert biotopes, as well as ice and frozen soils or sediments. It is also advisable to evaluate accurately whether particular organisms are either halotolerant or halophilic. A liquid growth medium with a minimal content of essential salts is preferable for testing tolerance to NaCl and the salt requirement. The highly diverse peptidoglycan structures of members of the suborder Micrococcineae (Table 2) can be subdivided according to the following criteria: $ type of diagnostic diamino acid $ type of cross-linkage (A or B according to Schleifer & Kandler, 1972) $ detailed structure, including amino acid sequence of the interpeptide bridge (see Schleifer & Kandler, 1972 and http://www.dsmz.de/microorganisms/main.php? content_id535).
Peptidoglycan structures are an important tool for the differentiation of members of the suborder Micrococcineae at the genus level and also at the species level within several genera (e.g. Arthrobacter, Microbacterium, Cellulomonas, Oerskovia, Brachybacterium and Nesterenkonia). On the other hand, the elucidation of the detailed peptidoglycan structure according to Schleifer & Kandler (1972) and Schleifer (1985) is a rather demanding task that requires specific experience and cannot be performed by all laboratories. For differentiation at the genus level, it is necessary to examine as a minimal requirement, the type of the diagnostic diamino acid (ornithine, lysine, diaminobutyric acid, diaminopimelic acid isomers, hydroxylated diamino acids) and other amino acids included in the peptidoglycan. The molar ratio of amino acids in peptidoglycan hydrolysates should be determined by using automated amino acid analysers (Schleifer & Kandler, 1972), HPLC (e.g. Lin et al., 2004), GC (e.g. MacKenzie, 1987) or GC/MS (e.g. MacKenzie, 1984) since quantitative data allow the discrimination between many peptidoglycan structures. Additional investigations such as enantiomeric analysis of amino acids by GC (e.g. Frank et al., 1980) or by HPLC (e.g. Sasaki et al., 1998), the amino acid sequence of the interpeptide bridge and the determination of the Nterminal amino acid of the interpeptide bridge (Schleifer, 1985) also provide helpful information for the elucidation of the detailed peptidoglycan structure and the type of cross-linkage. Taxonomists are encouraged to study the detailed peptidoglycan structure for differentiation at the genus as well as at the species level or to seek cooperation for the respective analytical methods.

Acyl type of peptidoglycan
Although most of the genera of the suborder Micrococcineae have not yet been studied with respect to the type of acylation of the muramic acid residues in the peptidoglycan, the examination of this feature is recommended for differentiation at the genus and at the species level. Members of some genera (Microbacterium and Okibacterium) contain glycolyl (usually as well as acetyl) residues in their peptidoglycan, whereas the majority of examined genera contain only acetyl residues. The genus Brachybacterium is inconsistent as regards the type of acylation of the muramic acid residues. Glycolylation can be detected by the procedure of Uchida & Aida (1984) or by a miniaturized method (Uchida et al., 1999).

Cell-wall sugars
Only a few genera of the suborder Micrococcineae are consistent with respect to cell-wall sugar composition. In particular, the presence of rhamnose seems to be diagnostic for some genera. The occurrence of specific sugars in some species is desirable information in the genus description. The determination of sugars is recommended for differentiation at the species level. Since whole cells may contain incorporated sugars from the cultivation media and typically also glucose, ribose and other sugars originating from the cytoplasm or capsules, it is recommended that cell walls are isolated and purified according to Schleifer (1985) prior to the analysis. Hydrochloric acid is not suited for hydrolysis because chloro-deoxy sugars may be formed. Instead, 1 M sulfuric acid has to be used for cell-wall hydrolysis. This can be removed easily prior to sugar analysis by extraction with N,N-dioctylmethylamine in chloroform as described by Hancock (1994). The identification of the sugars can be performed by TLC, e.g. as described by Komagata & Suzuki (1987). Quantification of the sugars can be performed by GC or HPLC as described by Saddler et al. (1991) and Takeuchi & Yokota (1989), respectively.

Respiratory quinone patterns
The length and degree of saturation of the isoprenoid side chain is very variable ( Table 2). The structural diversity of menaquinones and demethylmenaquinones as exclusive isoprenoid quinones in the suborder Micrococcineae can be analysed quantitatively by HPLC and mass spectrometric methods (Collins, 1994).

Polar lipids
Analyses of polar lipids of bacteria are performed in most laboratories by two dimensional TLC as described by Lechevalier et al. (1977), Komagata & Suzuki (1987) and Tindall (1990) for example. Many publications report 'unknown' phospholipid and glycolipid components (see Table 2) which could not be identified at the time of analysis. It is suggested that authors who cannot identify taxonomically relevant polar lipid components include a figure of the chromatogram in the manuscript in order to demonstrate the position of the respective spots for comparative purposes.

Polyamines
Polyamine patterns are so far available for only a limited number of genera of the suborder Micrococcineae including representatives of the families Micrococcaceae, Intrasporangiaceae, Microbacteriaceae, Sanguibacteraceae, Brevi-P. Schumann and others bacteraceae, Bogoriellaceae, Promicromonosporaceae, Cellulomonadaceae and Dermabacteraceae. Species of the family Micrococcaceae have been shown to exhibit polyamine patterns which consist of the predominant compound spermidine often along with significant amounts of spermine (Hamana, 1994(Hamana, , 1995Altenburger et al., 1997Altenburger et al., , 2002aAltenburger et al., , 2002bGvozdiak et al., 1998;Busse & Schumann, 1999;Buczolits et al., 2003;Busse et al., 2003). Higher variability in polyamine patterns was detected among genera of the family Microbacteriaceae (Altenburger et al., 1997). Members of the genera Curtobacterium, Clavibacter and Rathayibacter show polyamine patterns with the major compounds spermidine and spermine, species of the genera Agrococcus and Plantibacter are characterized by the major compound spermine and Pseudoclavibacter helvolus contains the major compounds cadaverine and 1,3-diaminopropane. The single species of the genus Leifsonia analysed, Leifsonia poae, contains the major compound putrescine. Species of the genus Agromyces are characterized by very low polyamine contents. Four species of the family Intrasporangiaceae, each representing a different genus (Intrasporangium, Terrabacter, Terracoccus and Janibacter), were shown to share putrescine as a major compound.
Additional major compounds such as spermidine (Intrasporangium), spermidine and spermine (Terracoccus) or cadaverine (Janibacter) suggest the potential of polyamine patterns to distinguish at least between groups of this complex family. Species of the genera Promicromonospora and Cellulosimicrobium (family Promicromonosporaceae) as well as the genera Cellulomonas and Oerskovia (family Cellulomonadaceae) show polyamine patterns with the predominant compound spermidine and moderate amounts of spermine, but the overall polyamine concentrations may differ by two orders of magnitude. Members of the genus Brevibacterium (family Brevibacteriaceae) exhibit polyamine patterns predominated by cadaverine and putrescine and species of the genus Sanguibacter (family Sanguibacteraceae) show only spermine as the predominant polyamine. Two species of the genus Brachybacterium, representing the family Dermabacteraceae, show a polyamine pattern consisting of the major amines spermidine and spermine and the polyamine content is quite high. Georgenia muralis is the only representative of the family Bogoriellaceae which has been examined for polyamines. The major polyamines of this species are spermidine and spermine in similar concentrations, but the overall polyamine concentrations are approximately 4-12 orders of magnitude lower than in other representatives of the suborder.
This variability among representatives of the suborder Micrococcineae indicates that polyamine patterns are conserved at distinctive taxonomic levels and are therefore recommended characteristics for classification and identification. According to available data, the following taxa or groups of genera within a family share common characteristics in polyamine patterns: Clavibacter/Rathayibacter/ Curtobacterium, Leifsonia/Agromyces, Promicromonospora/ Cellulosimicrobium, Cellulomonadaceae and Intrasporangium/ Terrabacter/Terracoccus/Janibacter. Certain polyamine patterns are found in different families but, in polyphasic classification, this approach can play an important role in distinguishing taxa which cannot be differentiated by other traits. Hence analysis of polyamine patterns can be recommended at least for those taxa listed here and data from other taxa are highly desirable to complement the classification scheme for the suborder Micrococcineae.
Polyamines can be easily extracted in less than two hours without any special equipment and analysed by gradient HPLC after derivatization with dansyl chloride and detection with a fluorescence detector (Altenburger et al., 1997;Scherer & Kneifel, 1983) or after post-column derivatization with o-phthalaldehyde (Hamana, 1995).

Cellular fatty acids
The general type of cellular fatty acids (i.e. straight-chain saturated and monounsaturated type, 10-methylbranched-chain type, iso/anteiso-branched-chain type, vcyclohexyl fatty acid type, complex type; Suzuki & Komagata, 1983) is of significance for classification at the genus level whereas the quantitative profile of cellular fatty acids can be used for the differentiation of species. The discriminating power of cellular fatty acid profiles differs between families of the suborder Micrococcineae. For instance, most organisms of the family Microbacteriaceae display fatty acid profiles consisting of only few components of the iso-and anteiso-branched type and the discrimination of genera and species on the basis of such patterns is hardly possible. Only representatives of particular genera have been reported to contain characteristic components as v-cyclohexyl fatty acids (Suzuki & Komagata, 1983;Qiu et al., 2007;Vaz-Moreira et al., 2008), 2-hydroxy-tetradecanoic and monounsaturated octadecenoic acid (Zhang et al., 2007b) or 1,1-dimethoxy-alkanes (Kämpfer et al., 2000;Schumann et al., 2003) (Table 2). Members of the family Intrasporangiaceae, on the other hand, are characterized by complex patterns consisting of many components which are well suited for their differentiation. As a conclusion from these observations, cellular fatty acid profiles are recommended as a minimal requirement for differentiation at the genus and at the species level. The comparison of fatty acid profiles necessitates the standardization of the cultivation conditions since the composition of media, the age of the culture, the temperature and the availability of oxygen have a strong influence on the fatty acid patterns. Therefore it may be necessary to analyse the fatty acids of strains for comparison in one and the same laboratory rather than to compare one's own data with those available from literature. This is particularly important when quantitative data on individual fatty acids rather than overall patterns are compared. The preferable method for analysis of cellular fatty acid profiles is capillary GC of fatty acid methyl esters. If it is the case that some fatty acid methyl Minimal standards for the suborder Micrococcineae esters cannot be unambiguously identified on the basis of their retention times due to co-elution (e.g. C 18 : 1 v7c and cyclohexyl C 17 : 0 ; 2-OH C 14 : 0 and C 15 : 0 -DMA; 2-OH iso-C 15 : 0 and C 16 : 1 v7c), the identification should be supported by GC/MS or hydrogenation experiments.

Total protein profiles
Whole-cell protein profiles have been proved to be useful tools for determining the coherency of species (Zlamala et al., 2002a, b) and for the differentiation of species within the genus Brachybacterium (Buczolits et al., 2003).
The structure and composition of cell-wall teichoic acids and other cell-wall-associated anionic polysaccharides appear to be indicative of most of the actinomycete species where such polymers have been analysed so far (e.g. Fiedler & Schäffler, 1987;Schubert et al., 1996;Ortiz-Martinez et al., 2004;Potekhina et al., 2003aPotekhina et al., , b, 2004Potekhina et al., , 2007Naumova et al., 2001;Shashkov et al., 2004). The structural components of these polymers (glycerol, ribitol, mannitol, erythritol, galactose, glucose, rhamnose, aminosugars etc.) have been shown to be useful for the delineation of some species, e.g. in the genera Brachybacterium (Schubert et al., 1996) and Brevibacterium (Gavrish et al., 2004) and can be recommended as phenotypic traits for the characterization and differentiation of organisms at the species level.
Cells for the analysis of anionic phosphorus-containing polymers of cell walls can be grown in any suitable liquid media supplemented with phosphorus and should be harvested at the exponential growth phase.
The simple method for detecting the cell-wall teichoic acids is the measurement of thermostable phosphate levels, e.g. by the method of Ames (1966), combined with determination of the main structural components of polymers, mainly polyols (Bergmeyer, 1974;Endl et al., 1984;Streshinskaya et al., 2002). For the isolation of anionic polymers, a native cell wall is obtained from cells by fractionated centrifugation after disruption, e.g. by ultrasonication, and purified by 2 % sodium dodecyl sulfate at 100 u C for 5 min). The polymers are extracted with trichloroacetic acid (10 %, 4 u C, 48 h). The methods for the determination of the main structural components are based on the analysis of products of acid hydrolysis (2M HCl, 100 u C, 3 h) by chemical and chromatographic methods (Archibald, 1972;Naumova et al., 1980;Takeuchi & Yokota, 1989).
The main structural components of teichoic acids can be determined directly in the hydrolysates of purified cell walls without the extraction of polymers (Naumova et al., 1980). Polyols can also be identified by thin-layer or paper chromatography as described for sugar analysis (e.g. Komagata & Suzuki, 1987;Becker et al., 1964), with additional staining by 5 % AgNO 3 in aqueous ammonia solution (sugar spots are stained brown, while polyol spots become black). GC or HPLC analysis can also be applied (Saddler et al., 1991;Takeuchi & Yokota, 1989).

Cytochrome and haem composition
Although cytochrome and haem patterns have been studied so far only in some genera of the suborder, including Micrococcus, Kocuria, Dermacoccus, Agreia, Plantibacter, Rathayibacter and Okibacterium, they appear to be valuable for taxonomic purposes and are recommended for the differentiation of organisms at the genus and, particularly, at the species level. The methods for testing cytochrome and haem composition are based on spectrophotometry (preliminary spectral characteristics) and chromatography (Faller et al., 1980;Trutko et al., 2003).

Pigments
Many representatives of the suborder Micrococcineae produce pigments which are typically yellow (of different tints and intensities) to orange or red. Some species of the genera Clavibacter and Arthrobacter are known to produce blue pigments. The chemical composition of pigments is not regularly determined in taxonomic studies of representatives of the suborder Micrococcineae. Nevertheless, the published data on the chemical nature and composition of pigments provide useful taxonomic information for species and subspecies characterization and differentiation as shown, for instance, in articles on the genera Brevibacterium, Clavibacter, Microbacterium and Leucobacter (Jones et al., 1973;Gavrish et al., 2004;Trutko et al., 2005;Muir & Tan, 2007).

P. Schumann and others
The methods for the determination of pigments are based on spectrophotometry (preliminary spectral characteristics) and chromatography (Weeks, 1981;Dufossé et al., 2001;Trutko et al., 2005). While analysing pigments, it should be taken into account that growth conditions may affect their formation and that the production of pigments may be strain-dependent. Qualitative and quantitative changes in the carotenoid pigments of Curtobacterium flaccumfaciens (basonym Corynebacterium poinsettiae) were observed, for instance, when the level of the required growth factor thiamine was altered in the basal medium (Starr & Saperstein, 1953).

Fourier-transform infrared (FT-IR) spectroscopy
The method and its taxonomic application were described in detail by Oberreuter et al. (2002). The method is particularly valuable for the differentiation of related species from similar ecological niches, which display almost identical phenotypic traits in terms of traditional physiological and biochemical criteria and similar (identical) colony colours. It should be noted that spectral similarities of strains do not necessarily correspond to their phylogenetic relationships (Kümmerle et al., 1998;Oberreuter et al., 2002). The method has been tested on many representatives of the suborder Micrococcineae (e.g. Oberreuter et al., 2002;Behrendt et al., 2002).

Matrix Assisted Laser Desorption Ionization Timeof-Flight mass spectrometry (MALDI/TOF)
MALDI/TOF mass spectrometry is a rapid and efficient identification method for intact whole bacteria (Holland et al., 1996) which has been proven to aid medical diagnostics and risk assessment. This technique may cover a broad taxonomic range depending on the instrument conditions chosen, from the genus level up to the authentication of strains. Although reports on the application of MALDI/ TOF mass spectrometry for members of the suborder Micrococcineae (Vargha et al., 2006) are still rare, this fast method has turned out to be of great taxonomic importance when compared with gene sequence phylogenies  and its application as a tool for characterization and differentiation at the species level is highly encouraged.

GENOTYPIC CRITERIA
Analysis of the gene sequences 16S rRNA gene sequence analysis is the current method of choice for the determination of phylogenetic relationships. The availability of the sequences of the type strains of most species with validly published names in comprehensive public databases is the main advantage of sequence analyses based on this gene. The necessity of using several different algorithms for the construction of phylogenetic dendro-grams and for the examination of the reliability of branching points has already been outlined (Bernardet et al., 2002). Whereas sequence analysis of the 23S rRNA gene provides similar phylogenetic information to 16S rRNA gene sequencing (Ward et al., 2000), less information should be expected from the analysis of the smaller 5S rRNA genes. Nevertheless, 5S rRNA sequencing was applied successfully for the proposal of the family Microbacteriaceae (Park et al., 1993). The degree of taxonomic resolution may be different when the phylogenetic analysis is based on other molecular targets, e.g. the DNA gyrase B subunit gene (Hatano et al., 2003;Richert et al., 2005), the tRNA intergenic spacer regions (Hinrikson et al., 2000), hrcA (Ahmad et al., 1999), hsp60 (Kwok et al., 1999), sodA (Poyart et al., 2001), rpoB (Morse et al., 2002) or tuf (Heikens et al., 2005). Though there are currently only a few complete genome sequences of members of the suborder Micrococcineae available (http:// www.genomesonline.org/), it can be expected that genome sequences will have increasing impact on the understanding of the phylogeny of the suborder Micrococcineae and in the definition of its species in future.

DNA base composition
The range of G+C values within a genus is an important taxonomic criterion. The indication of the G+C value of the type strain of the type species of a novel genus is mandatory; it is optional for type strains of novel species in established genera (Stackebrandt et al., 2002b). For precise determination of G+C values, the direct quantification of nucleosides (e.g. Mesbah et al., 1989;Tamaoka & Komagata, 1984) by HPLC methods is recommended. Since totally sequenced bacterial genomes with known base compositions have become available recently (http:// www.ncbi.nlm.nih.gov), the respective strains represent excellent reference organisms for calibration purposes.

DNA-DNA hybridization
Although there are ideas for novel promising methods for the definition of bacterial species (Stackebrandt et al., 2002b;Gevers et al., 2005), DNA-DNA hybridization is still the established technique for species delineation. Past experience has shown that DNA-DNA hybridization is recommended for the evaluation of species status when the value for 16S rRNA gene sequence similarity is above 97 % (Stackebrandt & Goebel, 1994). In order to evaluate the stringency of the DNA-DNA hybridization, it is necessary that the experimental conditions (buffer system, ionic strength and reassociation temperature) are reported. Different methods for DNA-DNA hybridization have been described and their results compared by Grimont (1988). The filter methods using radioactive DNA showed high reproducibility but lost their importance after the worldwide introduction of strict safety regulations. The reliability of results of the spectrophotometric method is comparable to those obtained by filter techniques, as revealed by the studies of Huß et al. (1983). Miniaturized methods using Minimal standards for the suborder Micrococcineae microtitre plates (e.g. Ezaki et al., 1989;Ziemke et al., 1998) offer the advantage of a higher sample throughput but can only be recommended when the reproducibility of their results is as high as those of filter or spectrophotometric techniques (Goris et al., 1998).

Genomic fingerprinting
Rapid DNA-typing methods are considered appropriate tools for the determination of inter-and intraspecies relatedness by the Ad Hoc Committee for the Re-Evaluation of the Species Definition in Bacteriology (Stackebrandt et al., 2002b). These methods investigate whole genomes (AFLP, RAPD, Rep-PCR, PFGE), rrn operons, the 16S rRNA gene (ARDRA), or the intergenic 16S-23S rRNA spacer regions (ISR). Ribotyping and methods based on whole genomes are especially recommended for the examination of whether strains belong to the same species. However, it has been shown recently for Microbacterium paraoxydans that ribotyping does not identify all strains of this species (Buczolits et al., 2008), demonstrating that, as with other fingerprinting techniques, highly similar banding patterns are useful for species identification whereas different patterns do not necessarily indicate another species. However, further standardization of these techniques is required in order to improve the reproducibility between laboratories before these methods can be proposed as alternative minimal criteria that are equivalent to DNA-DNA hybridization. A survey of DNA fingerprinting techniques and their application in bacterial systematics is given by Pukall (2006).

Sequencing of housekeeping genes and multilocus sequence typing (MLST)
The Ad Hoc Committee for the Re-Evaluation of the Species Definition in Bacteriology (Stackebrandt et al., 2002b) considers these techniques 'methods of great promise' for novel approaches to the species concept. Information for the differentiation of species should be obtained from the sequences of at least five genes coding for proteins of metabolic function. These genes should originate from diverse chromosomal loci and occur over a wide range of taxa. Sequencing of housekeeping genes and MLST have the advantage that discrimination at the species level could be performed by comparing the sequences available from databases instead of direct comparison of isolates to type strains by DNA-DNA hybridization or DNA-typing patterns. MLST (Maiden et al., 1998) is presently mainly used in epidemiology and for the investigation of microbial populations but will have an increasing importance for bacterial systematics at the interand intraspecific level (Stackebrandt, 2002).

NOTE ADDED IN PROOF
The genus name Frondicola  was recently deemed to be illegitimate as it is a later homonym of a fungal genus name Frondicola Hyde, 1992 (Fungi, Ascomycota, Sordariomycetes, Xylariomycetidae, Xylariales, Hyponectriaceae) [Principle 2 and Rule 51b(4) of the Bacteriological Code (1990 Revision)]. Therefore, a new genus name, Frondihabitans gen. nov., has been proposed for this taxon (Greene et al., 2009). A new name was proposed for the type species, Frondihabitans australicus sp. nov., to replace the illegitimate combination Frondicola australicus .