Blackwell Science, LtdOxford, UKEMIEnvironmental Microbiology1462-2920Blackwell Publishing Ltd, 2003510828841Original ArticlePCR primers for the class ActinobacteriaJ. E. M. Stach et al . Environmental Microbiology (2003) 5(10), 828–841 doi:10.1046/j.1462-2920.2003.00483.x New primers for the class Actinobacteria: application to marine and terrestrial environments James E. M. Stach,1* Luis A. Maldonado,2 cytosine and form a distinct phyletic line in the 16S rDNA Alan C. Ward,2 Michael Goodfellow2 and Alan T. Bull1 tree (Embley and Stackebrandt, 1994; Stackebrandt et al., 1Research School of Biosciences, University of Kent, 1997). Members of the taxon are of interest primarily Canterbury, Kent CT2 7NJ, UK. because of their importance in agriculture, ecology, indus- 2School of Biology, University of Newcastle, Newcastle try and medicine (McNeill and Brown, 1994; Strohl, 2003). upon Tyne NE1 7RU, UK. Actinobacteria are widely distributed in terrestrial (McVeigh et al., 1996; Heuer et al., 1997; Hayakawa et al., 2000), freshwater (Goodfellow et al., 1990; Wohl and Summary McArthur, 1998) and marine (Goodfellow and Haynes, In this study, we redesigned and evaluated primers for 1984; Takizawa et al., 1993; Colquhoun et al., 1998) hab- the class Actinobacteria. In silico testing showed that itats where they are involved in the turnover of organic the primers had a perfect match with 82% of genera matter (McCarthy, 1987; Schrempf, 2001) and xenobiotic in the class Actinobacteria, representing a 26–213% compounds (Kastner et al., 1994; Bunch, 1998; De Schr- improvement over previously reported primers. Only ijver and De Mot, 1999). Some actinobacteria are serious 4% of genera that displayed mismatches did so in the pathogens of animals, including humans, and plants terminal three bases of the 3¢¢¢¢ end, which is most (Locci, 1994; McNeill and Brown, 1994; Trujillo and critical for polymerase chain reaction success. The Goodfellow, 2003), whereas others form nitrogen-fixing primers, designated S-C-Act-0235-a-S-20 and S-C- associations with non-leguminous plants (Benson and Act-0878-a-A-19, amplified an ªªªª640 bp stretch of the Silvester, 1993). 16S rRNA gene from all actinobacteria tested (except Currently, actinobacteria, especially spore-forming act- Rubrobacter radiotolerans) up to an annealing tem- inomycetes, represent the most economically and bio- perature of 72∞∞∞∞C. An Actinobacteria Amplification technologically valuable prokaryotes, producing over half Resource (http://microbe2.ncl.ac.uk/MMB/AAR.htm) the bioactive compounds present in the Antibiotic Litera- was generated to provide a visual guide to aid the ture Database (Lazzarini et al., 2000), notably antibiotics amplification of actinobacterial 16S rDNA. Application (Lazzarini et al., 2000; Strohl, 2003), antitumour agents of the primers to DNA extracted from marine and ter- (Zheng et al., 2000; Dieter et al., 2003), enzymes restrial samples revealed the presence of actinobac- (Peczynska-Czoch and Mordarski, 1988; Oldfield et al., teria that have not been described previously. The use 1998) and enzyme inhibitors and immunomodifiers of 16S rDNA similarity and DNA–DNA pairing correla- (Umezawa, 1988). However, the rediscovery rate of bio- tions showed that almost every actinomycete clone active compounds from microorganisms currently in cul- represented either a new species or a novel genus. ture has been estimated to be 95% (Fenical et al., 1999). The results of this study reinforce the proposition that In order to isolate novel actinobacteria for biotechnology, current culture-based techniques drastically underes- we need first to understand their ecology, which encom- timate the diversity of Actinobacteria in the environ- passes diversity, species richness and distribution. Molec- ment and highlight the need to evaluate taxon- ular techniques overcome culture bias and can be used specific primers regularly in line with improvements to investigate actinobacterial ecology; this approach has in databases holding 16S rDNA sequences. been used to detect actinobacteria in environmental sam- ples where corresponding culture-based procedures have been unsuccessful (Relman et al., 1992; Heuer et al., Introduction 1997; Rheims et al., 1999), and has highlighted novel The class Actinobacteria encompasses bacteria that are actinobacterial lineages (McVeigh et al., 1996; Rheims diverse with respect to their biochemistry, morphology and et al., 1996; 1999; Rheims and Stackebrandt, 1999; Lude- relationship to oxygen, but have DNA rich in guanine plus mann and Conrad, 2000). In contrast, there are instances where actinobacteria have been isolated from environ- mental samples but have not been detected in clone librar- Received 4 April, 2003; accepted 22 May, 2003. *For correspon- ies generated from the same sample (Felske et al., 1997; dence. E-mail [email protected]; Tel. (+44) 1227 823336; Fax (+44) 1227 763912. Li et al., 1999). Actinobacteria-specific/biased primers © 2003 Society for Applied Microbiology and Blackwell Publishing Ltd PCR primers for the class Actinobacteria 829 have been designed to increase the likelihood of detecting specific primers based on an alignment of an equal fre- actinobacterial 16S rDNA in community DNA extracted quency of 16S rRNA genes from all genera; (ii) evaluate from environmental samples (McVeigh et al., 1996; Heuer and compare the newly designed primers in silico; (iii) test et al., 1997; Ludemann and Conrad, 2000). To this end, primer specificity on a comprehensive library of actinobac- McVeigh et al. (1996) found that 46 out of 53 clones gen- teria and non-actinobacteria type strains; (iv) evaluate the erated from amplified DNA of community DNA extracted in situ specificity of the primers in terrestrial and marine from a temperate forest soil were actinobacterial in origin. environments; and (v) produce an Actinobacteria Amplifi- When the same primers were used to amplify 16S rDNA cation Resource that will allow investigators to adjust the from a rhizosphere soil, only 2% of the clones were of primer set in order to amplify members of specific actino- actinobacterial origin (Macrae et al., 2001). Amplification bacterial genera. of 16S rDNA from environmental samples has shown that The actinobacteria primers designed in this study gave primers specific for actinobacteria have had a success a 26–213% increase in the coverage of actinobacteria rate of between 2% and 87% (McVeigh et al., 1996; over corresponding primers and allowed the detection of Ludemann and Conrad, 2000; Peters et al., 2000; Macrae many novel actinobacterial lineages that have been unde- et al., 2001). tected previously (McVeigh et al., 1996; Heuer et al., The utility of actinobacteria-specific primers is defined 1997; Ludemann and Conrad, 2000). It is apparent from by both their specificity (i.e. minimal hybridization to non- these results that actinobacterial diversity based on target DNA) and their coverage (i.e. how many members detected or cultivated species is underestimated by at of the class Actinobacteria are amplified by the primers). least an order of magnitude. These properties are directly influenced by the quality and quantity of sequences used to design the primers, e.g. if three actinobacterial sequences are used, primers will Results have high specificity but low coverage. Secondly, primers Testing of primers in silico designed from alignments dominated by specific genera or species will be biased towards those species and have The theoretical specificities of primers S-C-Act-0235-a-S- reduced coverage. The rapid growth in the number of 16S 20 and S-C-Act-0878-a-A-19 were tested by submission rDNA sequences available in the Ribosomal Database to the CHECK_PROBE algorithm of the RDP, using default Project (RDP; Maidak et al., 2001) serves as a warning parameters and allowing zero mismatches. Previously to evaluate regularly and, if necessary, redesign primers; described actinobacteria-specific primers were also sub- when the first sets of actinobacteria-specific primers mitted for comparison. The sequence, target position and were designed (McVeigh et al., 1996; Heuer et al., 1997), CHECK_PROBE results for each of the primers are given in the RDP contained ª250 actinobacterial 16S rDNA Table 1. Primer S-C-Act-0235-a-S-20 displayed a 26% sequences (McVeigh et al., 1996; Maidak et al., 2001). increase in the number of perfect actinobacteria matches Presently, the RDP contains ª2300 actinobacterial over primer ACT283F, and displayed a 213% increase sequences (Maidak et al., 2001), while the forthcoming over primer F243. Primer S-C-Act-0878-a-A-19 gave an release will comprise 7500 sequences (Cole et al., 2003). 18% increase over primer ACT1360R and 13% over Clearly, primers designed from a small set of sequences primer AB1165r. Primers S-C-Act-0235-a-S-20 and S-C- will no longer reflect the diversity of those currently Act-0878-a-A-19 amplify the V3 to V5 regions of the 16S present in the databases and, therefore, an evaluation of rRNA gene (Brosius et al., 1981). Representatives of 168 actinobacteria-specific primers is warranted. genera were used in the alignment, and only 18% of these The aims of this study were to: (i) design actinobacteria- showed one or more mismatches with primer S-C-Act- Table 1. Comparison of actinomycete-specific primers. Identical Identical Percentage actinomycete Primera Sequence (5¢Æ3¢) matchesb actinomtycetesb matches Reference AB1165r ACCTTCCTCCGAGTTRAC 2013 99.5 2003 Ludemann and Conrad (2000) ACT1360R CTGATCTGCGATTACTAGCGACTCC 1731 99.7 1725 McVeigh et al. (1996) ACT283F GGGTAGCCGGCCUGAGAGGG 2632 88.3 2324 McVeigh et al. (1996) F243 GGATGAGCCCGCGGCCTA 836 100 836 Heuer et al. (1997) S-C-Act-235-a-S-20 CGCGGCCTATCAGCTTGTTG 2646 99.7 2639 This study S-C-Act-878-a-A-19 CCGTACTCCCCAGGCGGGG 2321 87 2019 This study a. E. coli numbering (Brosius et al., 1981). b. Information obtained using the PROBE_MATCH function of the RDP including sequences posted as unaligned by the RDP. © 2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 828–841 830 J. E. M. Stach et al. 0235-a-S-20; only 4% displayed mismatches in the termi- Fourteen clones belonged to the recently described bac- nal three bases of the 3¢ end that is most crucial to terial phylum Gemmatimonadetes (Zhang et al., 2003). polymerase chain reaction (PCR) success (Sommer and The remaining four clones belonged to the phylum Tautz, 1989). Planctomycetes. Clones within the class Actinobacteria were phylogen- etically diverse, representing the suborders Corynebac- Testing of primers with pure cultures and terineae and Frankineae and the families Actinosyn- environmental templates nemataceae, Micrococcaceae, Micromonosporaceae, Optimal primer annealing conditions for the new primer Nocardioidaceae and Pseudonocardiaceae. The majority set were investigated by gradient PCR using DNA from (64%) of actinobacterial clones belonged to the subclass 10 type strains (see Table 2); the gradient was made over Acidimicrobidae. The average pairwise similarity of all act- an annealing temperature range of 68–72∞C. All 10 type inobacterial clones was 87%. Comparison of 16S rRNA strains yielded amplicons of the predicted size over the gene similarity using the region amplified by primers S-C- entire gradient. A gradient PCR approach was also Act-0235-a-S-20 and S-C-Act-0878-a-A-19 and the whole applied to the 14 non-actinomycete type strains over a 16S rRNA gene for a large number of actinobacteria 50–65∞C range; none of these organisms gave amplifica- showed that 16S rDNA similarities assessed from the tion products. Further testing of the actinobacteria type amplified region were conservative by ª0.7%. strain library was made with the two-step amplification protocol described above, thus enabling a positive identi- Discussion fication of actinobacteria in under 1.5 h. All 147 actinobac- teria tested yielded an amplicon of the predicted size The primary aim of this study was to improve the detection except the type strain Rubrobacter radiotolerans; the fail- and identification of actinobacteria, either those in culture ure of the primers to amplify 16S rDNA from this organism or those represented in 16S rDNA clone libraries derived was expected (see Fig. 1). DNA extracted from environ- from DNA extracted from environmental samples. mental sources contains co-extracted contaminants that Although it was not our intention to estimate the full extent affect both PCR specificity and efficiency (Stach et al., of actinobacterial diversity in any of the environmental 2001); hence, a ‘touchdown’ protocol was implemented to samples, an unexpectedly high diversity was observed in amplify DNA extracted from the marine sediments and relatively small clone libraries. We used only one 16S soils. All environmental DNA samples gave a single ampl- rRNA gene sequence from each actinobacteria genus in icon of the predicted size using this protocol. the initial alignments to prevent specific genera that were well represented in sequence databases from biasing the primer design. This strategy simplified the identification of Screening and phylogenetic analysis of clones regions in the 16S rRNA gene conserved in the class One hundred clones were generated from environmental Actinobacteria. Previous examples of actinobacteria- DNA and, using a novel rapid clone screening method, specific primers (McVeigh et al., 1996; Heuer et al., 1997; were dereplicated to 85 in 50 h. An example of the single- Ludemann and Conrad, 2000) were designed using dif- strand conformation polymorphism (SSCP) dereplication ferent strategies. In silico testing confirmed the utility of of 25 clones generated from marine sediment D is shown our approach as the S-C-Act-0235-a-S-20 and S-C-Act- in Fig. 2. Phylogenetic analysis was continued with the 0878-a-A-19 primers showed a distinct (26–213%) dereplicated clones. Clones were assigned codes based improvement in the number of exact actinobacteria on the sample site and position in the microtitre plate. matches when compared with previous primers. Code ASb10, for example, indicates that the clone was The Actinobacteria Amplification Resource (AAR; isolated from the Alston soil sample and that it is archived Fig. 1) provides a visual tool that allows researchers to in row B, well 10 of the microtitre plate. Dereplicated adjust the S-C-Act-0235-a-S-20 primer in order to pro- clones (20 Alston soil; 15 Canterbury soil; 13 sediment A; vide a better match for actinobacteria that initially do not and 33 sediment D) were sequenced directly to obtain at show a perfect match (such as Rubrobacter sp.). It is least 500 bp of information from the V3 to V5 region of unlikely that the adjustment of the primer will be neces- 16S rDNA, and the returned sequences were subject to sary for members of genera showing two or fewer mis- chimera analysis. Twelve out of 85 clones (14%) were matches, especially when these are not located within found to be chimeras using the methods described below. the terminal three bases of the 3¢ end (Sommer and The remaining 73 clones were subjected to phylogenetic Tautz, 1989). The AAR can also be used to design analysis (Fig. 3). degenerate primers although such primers can pro- Fifty-five of the 73 clones belonged to the class Actino- duce artifacts in clone libraries as a result of primer bacteria based on analysis of 615 sequence positions. exhaustion (McVeigh et al., 1996). The fact that primers © 2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 828–841 PCR primers for the class Actinobacteria 831 Table 2. Strains tested using primers S-C-Act-0235-a-S-20 and S-C-Act-0878-a-A-19. Species Strain code Species Strain code Actinobacteria: Nonomuraea rubra DSM 43768T Acrocarpospora corrugata IFO 13972T Nonomuraea spiralis DSM 43555T Acrocarpospora macrocephala IFO 16266T Nonomuraea turkmeniaca DSM 43926T Acrocarpospora pleiomorpha IFO 16267T Planobispora longispora DSM 43041T Actinomadura latina DSM 43382T Planomonospora venezuelensis DSM 43178T Actinomadura longicatena DSM 44361T Prauserella rugosa DSM 43194T Actinomadura madurae DSM 43067T Pseudonocardia autotrophica DSM 43088 ‘Actinomadura malachitica’ DSM 43462 Pseudonocardia thermophila DSM 43027 Actinomadura nitritigenes DSM 44137T Pseudonocardia sp. UN 1088 Actinomadura pelletieri DSM 43383T Rhodococcus equia DSM 20307T Actinomadura spadix DSM 43459T Rhodococcus erythropolis DSM 43066T Amycolatopsis japonica KCTC 9817T Rhodococcus globerulus DSM 43954T Amycolatopsis fastidiosa KCTC 9155T Rhodococcus pyridinovorans DSM 44555T Amycolatopsis lactamduransa UN AT3 Rhodococcus rhodnii DSM 43336T Amycolatopsis sulphurea KCTC 9428 Rhodococcus rhodochrous DSM 43241T Frankia sp.a DSM 44251 Rhodococcus wratislaviensis UN 805 Frankia sp. DSM 44263 Rhodococcus zopfii DSM 44108T Gordonia aichiensis DSM 43978T Rubrobacter radiotolerans DSM 5868 Gordonia alkanivoransa DSM 44369T Saccharomonospora caesia DSM 43068 Gordonia bronchialis DSM 43247T Saccharothrix australiensis DSM 43800T Gordonia desulphuricans DSM 44462T Saccharothrix espanaensis DSM 44229T Gordonia hydrophobica DSM 44015T Saccharothrix sp. UN 1253 Gordonia sinesedis DSM 44455T Spirillospora albida UN JC202 Kitasatospora azatica DSM 41650T Streptomyces afghaniensis ISP 5228T Kitasatospora cystarginea DSM 41680T Streptomyces azureus ISP 5106T Kitasatospora griseola DSM 43859T Streptomyces bellus ISP 5185T Kitasatospora phosalacinea DSM 43860T Streptomyces bicolor ISP 5140 Kitasatospora setae DSM 43861T Streptomyces chartreusis ISP 5085T Kutzneria viridogrisea DSM 43850T Streptomyces cinnabarinus ISP 5467T Kutzneria sp. UN CR5 Streptomyces coeruleorubidus ISP 5145T Lentzea sp. UN AUS10 Streptomyces cyaneusa ISP 5108T Microbispora chromogenes DSM 43165 Streptomyces fumanus ISP 5154T Microbispora rosea DSM 43839T Streptomyces griseorubiginosus ISP 5469T Micrococcus luteus DSM 20030T Streptomyces hawaiiensis ISP 5042T Micromonospora aurantiaca DSM 43813T Streptomyces iakyrus ISP 5482T Micromonospora brunnea ATCC 27334T Streptomyces indigocolor ISP 5432 Micromonospora carbonacea ATCC 27114T Streptomyces longisporus ISP 5166T Micromonospora chalceaa DSM 43026T Streptomyces luteogriseus ISP 5483T Micromonospora inositola DSM 43819T Streptomyces malaysiensis DSM 41697T Micromonospora olivasterosporaa DSM 43868T Streptomyces mirabilis DSM 40553T Microtetraspora fusca DSM 43841T Streptomyces neyagawaensis ISP 5588T Microtetraspora glauca ATCC 23057T Streptomyces pallidus ISP 5531 Microtetraspora niveoalba DSM 43174T Streptomyces pseudovenezuelae ISP 5212T Mycobacterium fortuitum DSM 46621T Streptomyces purpurascens ISP 5310T Mycobacterium senegalensea DSM 43656T Streptomyces roseoviolaceus ISP 5277T Mycobacterium sp. UN 1254 ‘Streptomyces sudanensis’ UN A1 Nocardia abscessus IMMIB D-1592T ‘Streptomyces sudanensis’ UN A4 Nocardia africanaa DSM 44491T ‘Streptomyces sudanensis’ UN A9 Nocardia africana DSM 44500 Streptomyces violatus ISP 5205T Nocardia asteroides ATCC 19247T Streptomyces violatus ISP 5209 Nocardia asteroides UN 97 Streptomyces violochromogenes ISP 5207 Nocardia asteroides UN 1135 Streptomyces yatensis DSM 41771T Nocardia brasiliensis ATCC 19296T Streptomyces sp. UN E38 Nocardia brevicatena DSM 43024T Streptomyces sp. UN I26 Nocardia carnea DSM 43397T Streptomyces sp. UN I27 Nocardia cerradoensis UN 1301T Streptomyces sp. UN I28 Nocardia corynebacteroides DSM 20151T Streptosporangium album DSM 43023T Nocardia crassostreae ATCC 700418T Streptosporangium fragile IFO 14311T Nocardia cummidelens DSM 44490T Streptosporangium non-diastaticum DSM 43848T Nocardia farcinica UN 1066 Streptosporangium pseudovulgare DSM 43181T Nocardia farcinica ATCC 3318T Streptosporangium roseum DSM 43021T Nocardia flavorosea JCM 3332T Streptosporangium vulgare DSM 43802T Nocardia fluminea DSM 44489T Tsukamurella paurometabola DSM 20162T Nocardia nova ATCC 33726T Williamsia muralea DSM 44343T Nocardia otitidiscaviarum DSM 43242T Nocardia paucivorans DSM 44386T Non-Actinobacteria: Nocardia pseudobrasiliensis ATCC 51512T Acetobacter aceti NCIMB 6656 Nocardia pseudobrasiliensis UN 1234 Agrobacterium tumefaciens DSM 30150 Nocardia soli DSM 44488T Azorhizobium caulinodans ATTC 43989T © 2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 828–841 832 J. E. M. Stach et al. Table 2. cont. Species Strain code Species Strain code Nocardia transvalensis DSM 43405T Azotobacter vinelandi DSM 2289T Nocardia uniformis DSM 43136T Bacillus cereus DSM 31T Nocardia vaccinii DSM 43285T Bacillus polymyxa DSM 36T Nocardia sp. UN J121 Clostridium sporogenes NCIMB 10196 Nocardia sp. UN 1275 Lactobacillus plantarum ATTC 8014 Nonomuraea africana DSM 43748T Neptunomonas naphthovorans ATTC 700637T ‘Nonomuraea asiatica’ UN A299 Pediococcus pentosaceus DSM 20336T Nonomuraea fastidiosa DSM 43674T Pseudomonas putida ATTC 17453 Nonomuraea flexuosa DSM 43186T Sphingomonas yanoikuyae ATTC 51230T Nonomuraea helvata DSM 43142T Staphylococcus aureus NCIMB 10819 Nonomuraea polychroma DSM 43925T Streptococcus cremoris ATTC 11603 Nonomuraea pusilla DSM 43357T Nonomuraea recticatena DSM 43937T Nonomuraea roseola DSM 43551 Nonomuraea roseoviolacea DSM 43144T a. Strains used for PCR optimization. ‘ ’, not validated. T, type strain. ATCC, American Type Culture Collection, Manassas, VA, USA; DSM, Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Mascheroder Weg, Braunschweig, Germany; IFO, Institute of Fermentation, Osaka, Japan; ISP, International Streptomyces Project (strains from NRRL collection, Northern Regional Reference Laboratory, Peoria, USA); IMMIB, Institute of Medical Microbiology and Immunology of the University of Bonn, Germany; KCTC, Korean Collection for Type Cultures, Yusong, Republic of Korea; and UN, Collection of the University of Newcastle, Newcastle upon Tyne, UK. S-C-Act-0235-a-S-20 and S-C-Act-0878-a-A-19 gave an is also reported that SSCP screening of 16S rDNA gives actinobacteria-specific amplicon at an annealing temper- results that show a high degree of congruence with whole- ature of 72∞C facilitates their use as a diagnostic tool, as organism dereplication techniques such as pyrolysis mass a two-step PCR protocol can be used to confirm rapidly spectrometry (Brandao et al., 2002). Screening time was the presence of actinobacteria in environmental samples greatly reduced in the present study as a specific primer or to confirm the identity of actinobacteria tentatively set was used that does not amplify Escherichia coli 16S assigned to specific taxa. rRNA genes. The present protocol is easily adapted for The SSCP clone screening protocol developed for use with other primer sets, although amplicons >600 bp this study greatly reduces the time required to generate should be the subject of restriction digestion before SSCP dereplicated 16S rRNA gene clone libraries. Traditional analysis. approaches to screening 16S rDNA clone libraries have It is evident that the strategy used to design the primers relied on restriction fragment length polymorphism (RFLP) was successful as 76% of clones fell within the actino- using primers specific for the cloning vector. However, as mycete line of descent. The non-actinobacterial clones the amplified gene may be inserted in either direction, two belonging to the Gemmatimonadetes and Plancto- possible RFLP profiles may be generated as restriction mycetes are unsurprising; the type strain Gemmatimonas sites are not symmetrically distributed across the 16S aurantiaca displays one nucleotide mismatch in each of rRNA gene (Marchesi and Weightman, 2000). This prob- the new primers (the 16S rRNA gene sequence of G. lem may be overcome using modified primers (Marchesi aurantiaca was unavailable at the time of primer design), and Weightman, 2000) or restriction enzymes that cleave and planctomycete clone OM190 shows two mismatches in the cloning site of the vector (Vergin et al., 2001). with primer S-C-Act-0235-a-S-20 and one mismatch with Clones grouped by RFLP have been reported to show 52– primer S-C-Act-0878-a-A-19, neither of which are in the 99% similarity (Dunbar et al., 1999), although levels of terminal three bases of the 3¢ end. 79–100% can be achieved using a second round of RFLP Environmental contaminants can affect both the effi- (Vergin et al., 2001). The SSCP procedure can detect a ciency and the specificity of PCR (Stach et al., 2001) and, single basepair difference in 300 (Lee et al., 1996); hence, therefore, targets that do not exactly match the primer two dissimilar clones are unlikely to be grouped together sequences may be amplified. To overcome this drawback, using this method. A further advantage of SSCP is that it it may be possible to find a restriction site within the is straightforward to identify PCR products that have been amplified region of the Gemmatimonadetes and Plancto- generated from more than one template (see Fig. 2, sam- mycetes that is absent in the Actinobacteria, allowing the ples SDb02 and SDb07); hence, it is possible to omit removal of non-target 16S rDNA. Alternatively, a nested streak purification of transformants before dereplication. It PCR protocol could be used with an external primer that © 2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 828–841 PCR primers for the class Actinobacteria 833 Fig. 1. Actinobacteria Amplification Resource. A neighbour-joining tree showing phylogenetic relationships of genera within the class Actino- bacteria. Genera in blue display a perfect match with primer S-C-Act-0235-a-S-20, those in green a perfect match with primer S-C-Act- 0235-a-S-20 and primer F243 (Heuer et al., 1997), and those in red genera that mismatch with primer S-C-Act-0235-a-S-20. Letters and numbers in parenthesis indicate the adjust- ments that need to be made to primer S-C-Act- 0235-a-S-20 to provide a perfect match (e.g. 8ÆC indicates that base 8 in the S-C-Act-0235- a-S-20 primer should be changed from a T to a C). Changes are colour coded to indicate the frequency with which the base change was observed within members of the genera: black, conserved mismatch; green, mismatch found in <25% of species within the genus; blue, mis- match found in 25–66% of species; red, mis- match in 66–77% of species. Genera enclosed in quotation marks have not been validated; asterisks indicate genera present in the TAXON- OMY server of GenBank that no longer have any standing in prokaryotic nomenclature (type strain transferred to another genus). © 2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 828–841 834 J. E. M. Stach et al. Fig. 2. Single-strand conformation polymor- phism dereplication of 25 clones from Norwe- gian marine sediment D (Raunefjorden) clones generated using the S-C-Act-0235-a-S-20 and S-C-Act-0878-a-A-19 primers. Dereplication was achieved by product–moment UPGMA clus- ter analysis; similarities (r-values) are expressed as percentages. does not amplify non-target 16S rDNA. We have used the (Micrococcaceae) an Arthrobacter species; CSb01 AAR to design such a strategy; semi-nested PCR using (Pseudonocardiaceae) a novel Pseudonocardia species; primer S-C-Act-0235-a-S-20 and primer ACT1360R in the Sda10 (Corynebacterineae) a novel Williamsia species; first round followed by primer S-C-Act-0235-a-S-20 and and CSc07 and SAa11 novel Mycobacterium species. The primer S-C-Act-0878-a-A-19 in the second round resulted remaining clones (76%) can be considered as either novel in 99% of 167 clones being of actinobacterial origin, species or genera with no closely related cultured including members of the genera Rhodococcus and representatives. Streptomyces not observed in this study; the remaining McVeigh et al. (1996) designed actinomycete-specific 1% of clones were most closely related to uncultured primers and applied them to DNA extracted from a tem- representatives in the phylum Gemmatimonadetes perate forest soil to produce an actinomycete 16S rDNA (J. E. M. Stach, unpublished). clone library. Their study revealed three distinct actino- The actinobacterial clones were distributed throughout mycete groups comprising new actinomycete species that the class Actinobacteria, showing that the primers are represented several new genera. In the present study, we appropriate for the detection of actinobacterial diversity. It detected approximately 10 times the number of novel gen- seems likely that this diversity is genuine as comprehen- era in a similar-sized library. Furthermore, when primers sive methods were used for chimera detection. However, S-C-Act-0235-a-S-20 and S-C-Act-0878-a-A-19 were it should be noted that chimera detection using the applied to a relative low-diversity, actinomycete-rich sam- CHIMERA_CHECK algorithm might be limited by the lack of ple (smear cheese surface microbiota), species were cultured representatives in many of the actinobacteria detected that matched the diversity expected from culture- clusters. based studies (Brennan et al., 2002). In contrast, the Using conservative estimates, it can be predicted that application of primers F243 and R513 (Heuer et al., 1997) clone pairs Sab04 and Sde01, Asb04 and Asc01 (Acidimi- resulted in the detection of only Arthrobacter species (N. crobidae) and CSc10 and Csa01 (Micromonosporaceae) Bora and A. Ward, personal communication), thus provid- represent species of novel genera. Similarly, clone Asb07 ing clear evidence of the reduction in bias and improve- (Frankineae) represents a novel Blastococcus species; ment in species coverage of primers designed in this Asa04 (Frankineae) a novel Sporichthya species; Asa07 study. © 2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 828–841 PCR primers for the class Actinobacteria 835 Fig. 3. Phylogenetic relationship of the partial 16S rDNA gene sequences generated in this study (for tree construction, see Experimental procedures). Tree shows the class Actinobacteria and the phyla Gemmatimonadetes and Planctomycetes. The scale bar represents the number of changes per base position. Bootstrap values are shown at 60–80% (indicated by the filled circle) and at 80–100% (indicated by the open circle). © 2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 828–841 836 J. E. M. Stach et al. The prediction that 76% of the clones represent new value in continuing efforts to isolate members of novel genera is conservative as actual correlations between actinomycete genera. The primers reported here will facil- DNA similarity and 16S rDNA homology show that no two itate the isolation of members of such genera by allowing organisms with <99% 16S rDNA homology showed >70% bioprospecting of habitats before isolation efforts, followed DNA similarity (Stackebrandt and Goebel, 1994). The act- by their use to monitor the efficacy of new isolation inobacterial clones, on average, showed between 86% strategies. and 93% homology with either their closest cultured rep- resentative or an uncultivated clone; Amycolatopsis and Experimental procedures Rhodococcus strains, which represent different suborders in the class Actinobacteria, share ª88% 16S rRNA gene Bacteria and environmental samples homology. The bacteria used to determine primer specificity are listed It is worth noting that microvariation artifacts can be in Table 2. Non-actinobacteria were grown in nutrient broth introduced into clone libraries when amplifying members (Oxoid), Frankia strains in DPM medium (DSM medium 737; of closely related species. Such artifacts are caused by http://www.dsmz.de) and R. radiotolerans in tryptone soya PCR-introduced errors, the formation of chimeric mole- broth (DSM medium 535). All other actinobacteria were cules and heteroduplex formation (Speksnijder et al., grown in glucose yeast extract broth. 2001). However, it is unlikely that the diversity reported here is greatly influenced by such artifacts as a high- fidelity DNA polymerase was used to generate the clone DNA extraction and purification library, the PCR cycle number was kept low (25 cycles), A list of the environmental soil and marine sediment samples and clone sequences were subjected to comprehensive is given in Table 3. Genomic DNA was extracted from the chimera analysis. Microvariation artifacts are reported to strains using a DNeasy kit according to the manufacturer’s give a 0.2–1.4% sequence difference between parent and protocol (Qiagen). Microbial DNA from the environmental aberrant sequences (Speksnijder et al., 2001); the closely samples was extracted using a method modified from that of Saano and Lindstrom (1995): 1 g (dry weight) of soil/ related sequences in the present investigation showed a sediment sample was added to 2.5 ml of extraction buffer much higher level of divergence, hence it is unlikely that (120 mM NaHPO, pH 8.0, 1% SDS), lysozyme and 2 4 the majority of clones are PCR artifacts. However, confir- achromopeptidase were added to final concentrations of mation of the existence of novel species should be con- 5 mg ml-1 and 0.5 mg ml-1 respectively. Samples were incu- firmed using either isolation or stringent probing methods. bated for 1 h at 37∞C with occasional shaking. Proteinase K The high degree of novel actinobacteria detected in the (100 mg ml-1) was added and incubation continued (1 h, environmental samples is significant. A survey of the Anti- 37∞C). The salt concentration of the preparation was raised by the addition of 450 ml of 5 M NaCl before the addition of biotic Literature Database indicates that, of the 23 000 375 ml of 10% CTAB (CTAB in 0.7 M NaCl), and samples bioactive microbial products held, 57.8% are produced by were mixed by gentle vortexing and incubated for 20 min at members of the class Actinobacteria (Lazzarini et al., 65∞C. An equal volume of chloroform was added to the prep- 2000). It is evident from the present study that more than aration, which was vortexed and transferred to a 15 ml 50 novel species/genera were detected in the small clone polypropylene tube containing Phase Lock Gel™ (Eppen- library (73 clones) derived from four different environ- dorf). The samples were centrifuged (15 min, 9000 g, 4∞C), ments. It is reasonable to conclude that such new lineages and the aqueous phase was mixed with an equal volume of isopropanol and incubated for 1 h at 20∞C. Nucleic acids may represent taxa that will produce novel bioactive com- were precipitated by centrifugation (20 min, 16 000 g, 4∞C), pounds, as they share a common history (McVeigh et al., washed with ice-cold 70% ethanol and dissolved in 200 ml of 1996; Ward and Goodfellow, 2003). Clearly, the diversity doubled-distilled water. DNA was subject to two rounds of of Actinobacteria greatly exceeds that predicted by those purification using a Wizard DNA clean-up column, according already in culture and highlights the great biotechnological to the manufacturer’s instructions (Promega). Table 3. Terrestrial soils and marine sediments used as sources of environmental DNA. Sample Description Location AS Alston soil with history of metal contamination NY740470a CS UKC campus soil from rhizosphere of Betula pendula, top 5–10 cm TR140597a SA Norwegian fjord marine sediment A, By Fjorden, 316 m deep, top 10 cm N60∞23.798 E5∞13.296 SD Norwegian fjord marine sediment D, Raunefjorden, 187 m deep, top 10 cm N60∞16.512 E5∞11.043 a. UK Ordnance Survey grid reference. © 2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 828–841 PCR primers for the class Actinobacteria 837 Primers centre) and incubation at 37∞C for 15 min, followed by DNase I inactivation at 95∞C for 3 min. For cultured actinobacteria, Primers were designed to be diagnostic for actinobacterial DNA (ª 50 ng) was added, after DNase I treatment, and PCR 16S rDNA. In order to prevent primer bias in favour of mem- was done using a two-step protocol: initial denaturation at bers of genera with multiple 16S rRNA gene sequence 95∞C for 4 min, followed by 35 cycles of 95∞C for 30 s and entries in the available databases, one sequence for each 70∞C for 1 min. Amplification of environmental DNA was done genus in the class Actinobacteria was downloaded from the using a ‘touchdown’ protocol (Roux, 1995), which consisted TAXONOMY server of GenBank (Wheeler et al., 2002). The of an initial denaturation at 95∞C for 4 min, followed by dena- genera and species used are shown in Table 4 together with turation at 95∞C for 45 s, annealing at 72∞C for 45 s and GenBank accession numbers. Sequences were aligned extension at 72∞C for 1 min; 10 cycles in which the annealing using the CLUSTAL W algorithm (Thompson et al., 1994) avail- temperature was decreased by 0.5∞C per cycle from the able in the MEGALIGN program (DNASTAR). Sequences were preceding cycle; and then 15 cycles of 95∞C for 45 s, 68∞C chosen on the basis of length and sequence quality, and were for 45 s and 72∞C for 1 min, with the last cycle followed by a checked using the SEQUENCE_MATCH algorithm available 5 min extension at 72∞C. PCR products were separated on through the RDP (Maidak et al., 2001). Conserved regions 2% agarose gels stained with ethidium bromide (Sambrook within the alignment were tested for their actinobacteria-spe- et al., 1989). PCR products were purified using MinElute cific primer potential in silico by submission to the PCR purification columns according to the manufacturer’s CHECK_PROBE algorithm of the RDP. Two regions showing a instructions (Qiagen). Purified DNA was blunt-end ligated to high degree of actinobacteria specificity were selected as the plasmid vector pETBlue-1 and used to transform NovaB- sites to which primers were raised. The primers were named lue competent cells using a Perfectly Blunt® cloning kit according to the conventions proposed by the Oligonucle- according to the manufacturer’s protocol (Novagen). otide Database Project (Alm et al., 1996) and are S-C-Act- 0235-a-S-20 and S-C-Act-0878-a-A-19, where C indicates Single-strand conformation polymorphism screening of the class Actinobacteria (Stackebrandt et al., 1997). clone libraries Actinobacteria Amplification Resource Dereplicated clone libraries were generated using a novel rapid screening approach adapted from Vergin et al. (2001). The 16S rDNA sequences of the strains listed in Table 4 were Positive transformants were identified by blue/white selection uploaded to the CLUSTAL X interface and aligned to generate and inoculated into 0.2 ml PCR strip-tubes (Eppendorf) con- a guide dendrogram from which a final alignment was made taining 10 ml of 1/10th LB broth (Sambrook et al., 1989). Five (Thompson et al., 1997). All calculations were made using microlitres of the cell suspension was used to inoculate 5 ml the programs available in the phylogeny inference package of LB broth containing 50 mg ml-1 carbenicillin, and the cul- PHYLIP (Felsenstein, 1993). A distance matrix was con- tures were incubated for 16 h at 37∞C. Forty-five microlitres structed from the alignment using the DNADIST program. The of a PCR mixture (see above) was added to the tubes con- phylogenetic tree was produced using the neighbour-joining taining the remaining 5 ml of cell suspension, and PCR was method from the NEIGHBOR program with the Jukes–Cantor carried out as above. Amplicons from positive clones were correction parameter. Bootstrap analysis was conducted dereplicated by SSCP analysis (Stach and Burns, 2002). using the SEQBOOT and CONSENSE programs with 100 resa- Clones containing unique inserts were identified using mplings. The CHECK_PROBE results for primers S-C-Act-0235- GELCOMPAR software (version 4.0; Applied Maths) with the a-S-20 and F243 (Heuer et al., 1997) were used to interro- rolling-disk background subtraction method. Similarity matri- gate the actinomycete tree to determine which genera dis- ces were calculated using the pairwise Pearson’s product– played perfect matches (the F243 primer was chosen as it is moment correlation coefficient (r-value) (Häne et al., 1993). most commonly cited in the literature). Those actinobacteria Cluster analyses of similarity matrices were made using the not showing perfect matches were investigated further by unweighted pair group method with arithmetic averages aligning all available 16S rDNA sequences for the relevant (UPGMA). Clones were archived in 96-well plates in dimethyl genus from the TAXONOMY server of GenBank (Wheeler et al., sulphoxide/LB and stored at -80∞C (Vergin et al., 2001). 2002) and determining the position and frequency of the Plasmids containing unique inserts were extracted using a mismatches. The Actinobacteria Amplification Resource MiniPrep kit according to the manufacturer’s instructions (AAR) is given in Fig. 1 and is accessible via the internet (Qiagen). Sequencing was conducted commercially (Qiagen) (http://microbe2.ncl.ac.uk/MMB/AAR.htm) using the S-C-Act-0235-a-S-20 primer. PCR amplification and cloning Phylogenetic analysis PCR was performed in a 96-well gradient DNA thermal cycler Clone sequences were analysed using the CHIMERA_CHECK (Techne). The annealing temperature of primers S-F-Act- algorithm of the RDP. In addition, secondary structure predic- 0254-a-S-20 and S-F-Act-0894-a-A-19 was investigated over tion (http://www.genebee.msu.su/services/rna2_reduced.html) a 60–72∞C gradient. PCR amplification of cultured actinobac- and partial treeing methods were conducted on suspected chi- teria and environmental DNA was made using the Failsafe™ meras. Non-chimeric sequences were submitted to the BLAST PCR system (Epicentre) using buffer B in a final volume of function of GenBank (Altschul et al., 1990) and the 50 ml. Ten picomol of each primer was added, contaminating SEQUENCE_MATCH program of the RDP to identify closely DNA was removed by the addition of 1 U of DNase I (Epi- related reference sequences. The phylogenetic position of the © 2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 828–841