Introduction

The genus Amycolatopsis1 belongs to the family Pseudonocardiaceae and contains nocardioform actinomycetes that lack mycolic acids and contain meso-diaminopimelic acid, arabinose and galactose in their cell wall peptidoglycan. The genus is known for its antibiotic-producing strains, producing among others the ansamycin-type antibiotic rifamycin (by Amycolatopsis mediterranei and Amycolatopsis rifamycinica)2, 3, 4 and the glycopeptide antibiotic vancomycin (by Amycolatopsis orientalis).4 Of the antibiotics produced by this genus, the ansamycin and glycopeptide classes are the most important to medicine. With the continued emergence of antibiotic resistance in bacterial pathogens, there is always a need for the discovery of new antibiotics to combat the resistance. As several members of the genus Amycolatopsis produce antibiotics, the genus provides an ideal source of strains to screen for such novel compounds. With the descriptions of Amycolatopsis cihanbeyliensis5 and Amycolatopsis jiangsuensis,6 the number of members in the genus with validly published names has risen to 62.7 Recent advances in the taxonomy of the genus involve using phylogenetic and genetic distance analyses based on the gyrB and recN genes for distinguishing between genomic species.8, 9 Here we describe the characterization of two novel members of the genus isolated from soil samples collected from two locations in South Africa.

Materials and methods

Isolation

Strain M29T was isolated from a soil sample collected from a suburban garden in Roodepoort, Gauteng Province, South Africa, and strain JS72T from soil collected on Table Mountain above the Upper Campus of the University of Cape Town, Cape Town, South Africa. The soil sample collected from Roodepoort was subjected to microwave pretreatment. One gram of soil was vortexed in 10-ml sterile distilled water for 1 min. One milliliter of the suspension was transferred to a sterile glass Petri dish and exposed to microwave radiation (950 W output microwave oven on full power) for 20 s. After the microwave treatment the suspension was allowed to cool, serially diluted in sterile distilled water, spread-plated onto Modified Czapek Solution (MC) agar10 and incubated at 30 °C for 21 days. The Table Mountain soil was not subjected to any pretreatment. A soil suspension was prepared as mentioned above before being serially diluted and spread-plated onto Difco Middlebrook 7H9 agar (7H9; Becton, Dickinson, Sparks, MD, USA) containing 10 mmol l−1 glucose (albumin-dextrose-catalase supplement omitted) and incubated at 30 °C for 14 days. After subculturing, both strains were maintained on yeast extract-malt extract agar (International Streptomyces Project (ISP) medium 2).11

Genomic and phylogenetic analyses

Genomic DNA was extracted as previously described.12 The rapid identification of the isolates to the genus level was achieved by 16S-rRNA gene amplification and restriction endonuclease digestion,13 using single digestions with MboI (Sau3AI isoschizomer), VspI (AsnI isoschizomer), SphI, KpnI, HindIII, SalI and Psp1406I. The gyrB gene was amplified using the 7G-gyrB-F and GgyrB-R1, and GgyrB-F1 and 7G-gyrB-R primer combinations and the standard PCR conditions described by Everest and Meyers.8 The recN gene was amplified as described in Everest et al.9 Strains were screened for antibiotic biosynthetic genes involved in the production of ansamycin (Type I polyketide), glycopeptide and Type II (aromatic) polyketide antibiotics as per Wood et al.14 The PCR products were purified using an MSB Spin PCRapace kit (Invitek, Berlin, Germany) and sequenced. Sequence analysis, generation of concatenated gene sequences, genetic distance calculations and phylogenetic analyses were performed as detailed in Everest et al.15 DNA hybridization analysis was performed as a service by the BCCM/LMG culture collection as described in Everest et al.,15 with hybridizations being performed at 51 °C. Data are displayed as an average DNA–DNA hybridization value with the difference between the means of the reciprocal values given in parentheses.

Antibacterial analysis

Antimicrobial testing was performed using the standard agar overlay method as detailed in Wood et al.14 Strains were stab inoculated onto agar plates of MC, 7H9, ISP 2, Hacène’s medium,16 a medium for the enhancement of antibiotic-production by Nocardia strains17 and a medium for the enhancement of antibiotic production by Streptosporangium strains17 and incubated at 30 °C for 7–10 days. Tests for growth inhibition were performed against Escherichia coli ATCC 25922, Enterococcus faecalis (vancomycin sensitive), Enterococcus faecium (clinical isolate; VanA), Enterococcus phoeniculicola JLB-1T, Mycobacterium aurum A+ and Staphylococcus aureus ATCC 25923.

In addition, extractions were performed with equal volumes of organic solvents (chloroform, ethyl acetate and methanol) on the cell mass and culture filtrate fractions of 100-ml ISP 2 liquid cultures grown for 7 days at 30 °C with shaking. These extracts were concentrated 50 times and tested for activity using spot test bioautography18 against Bacillus subtilis var ING, E. faecium (clinical isolate; VanA), E. coli ATCC 25922, M. aurum A+, Mycobacterium bovis strain BCG (Tokyo), Mycobacterium smegmatis LR222, Mycobacterium tuberculosis H37RvT (=ATCC 27294T) and Pseudomonas aeruginosa ATCC 27853.

Characterization

Morphological, physiological and chemotaxonomic characterization (diagnostic diamino acid in the peptidoglycan, cell wall sugars and polar lipids) were performed as detailed in Everest et al.15 The analysis of respiratory quinones was carried out by the Identification Service of the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ, Braunschweig, Germany) on freeze-dried cells of cultures grown in ISP 2 broth at 30 °C for 3 days with moderate shaking. Fatty acid analyses were performed as a service by the BCCM/LMG Culture Collection, as per the recommendations of the commercial identification system MIDI (Microbial Identification System, Newark, DE, USA; MIDI Sherlock version 3.10; database TSBA 50 (rev 5.0)), on cells grown at 28 °C for 3 days on Tryptic Soy Agar (BBL 11768).

Results and Discussion

Strains JS72T and M29T were determined to belong to one of the genera Amycolatopsis, Pseudonocardia or Saccharopolyspora, based on the analysis of the 16S-rRNA gene restriction fragment patterns.13 BLASTN analysis,19 based on 1410 bp of 16S-rRNA gene sequence for JS72T and 1411 bp for M29T, showed that both strains belong to the genus Amycolatopsis. The closest relatives of strain JS72T were determined to be Amycolatopsis thailandensis CMU-PLA07T (99.15% 16S-rRNA gene sequence similarity as per BLAST), Amycolatopsis coloradensis DSM 44225T (99.08%) and Amycolatopsis umgeniensis UM16T (99.07%). The closest relatives of strain M29T were determined to be Amycolatopsis decaplanina DSM 44594T (99.22%), Amycolatopsis regifaucium GY080T (99.01%) and Amycolatopsis keratiniphila subsp. nogabecina DSM 44586T (99.01%). Analysis of the 16S-rRNA gene sequences using EzTaxon-e20 revealed Amycolatopsis lurida DSM 43134T (99.26% 16S-rRNA gene sequence similarity) and Amycolatopsis keratiniphila subsp keratiniphila DSM 44409T (99.12%) as additional close relatives of strains JS72T and M29T, with 99.48% and 99.05% similarity, respectively.

The construction of a 16S-rRNA gene maximum-likelihood phylogenetic tree21 showed that strain M29T grouped with A. regifaucium GY080T and strain JS72T with A. keratiniphila subsp keratiniphila DSM 44409T (with low bootstrap support), within cluster A as defined by Everest and Meyers8 (Figure 1). Strains JS72T and M29T share 99.15% 16S-rRNA gene sequence similarity.

Figure 1
figure 1

16S-rRNA gene phylogenetic tree showing the positions of strains JS72T and M29T within cluster A of the genus Amycolatopsis. The tree was constructed using the maximum likelihood method based on 1353 bp of common sequence. The percentage bootstrap values of 1000 replications are shown at each node (only values above 70% are shown), with asterisks (*) indicating the clades that were also formed in the trees constructed using the neighbor joining28 and maximum parsimony29 algorithms. Accession numbers are indicated in parenthesis after the strain numbers. The scale bar indicates 1 nucleotide substitution per 100 nucleotides.

Full size image

A phylogenetic tree based on the concatenated gyrB-16S-rRNA gene sequences (Supplementary Figure 1) was also constructed and supported the association depicted in the 16S-rRNA gene phylogenetic tree for strain M29T, with some minor rearrangements of the branches, but improved bootstrap support. The position of strain JS72T, however, changed and it now formed an association with A. umgeniensis, but with very low bootstrap support (Supplementary Figure 1). Strain JS72T was found to be most closely related to Amycolatopsis alba in the gyrB-recN concatenated gene phylogenetic tree (Figure 2 and Supplementary Figure 2), but the bootstrap support for this association was weak (51%). The 16S-rRNA gene sequence similarity between strain JS72T and the type strain of A. alba is 98.72%. The gyrB-recN concatenated gene tree showed that strain M29T clustered similarly to the other trees, but formed a close association with A. lurida with low (66%) bootstrap support (Figure 2).

Figure 2
figure 2

Subtree of the phylogenetic tree based on the concatenated gyrB-recN gene sequences showing the relationships of strains JS72T and M29T to closely related type strains. The tree forms part of a full maximum-likelihood tree, constructed based on 2520 bp of sequence for the 40 members of the genus for which gyrB and recN gene sequences are available (Supplementary Figure 2). The percentage bootstrap values of 1000 replications are shown at each node (only values above 70% are shown), with asterisks (*) indicating the clades that were also formed in the trees constructed using the neighbor joining and maximum parsimony algorithms. The strain numbers of the strains from which the gyrB and recN gene sequences were obtained are indicated. The scale bar indicates 1 nucleotide substitution per 100 nucleotides.

Full size image

The number of nodes with bootstrap values >70% (that is, moderate support for the depicted tree topology) in the clade containing strains JS72T and M29T was two in the 16S-rRNA gene tree (Figure 1), six in the gyrB-16S-rRNA gene tree (Supplementary Figure 1) and six in the gyrB-recN gene tree (Figure 2). Furthermore, the bootstrap values for the associations in Supplementary Figure 1 and Figure 2 were much higher than those in Figure 1 (the majority of which were <40%; data not shown).

Genetic distance values, based on the gyrB gene (1290 bp), were calculated between strains JS72T, M29T and all the type strains of Amycolatopsis species for which gyrB sequences are available (Supplementary Table 1), to assess whether these strains were likely to represent new species.8 The values for strain JS72T ranged from 0.026 to 0.213 (0.036–0.439 based on the 315 bp variable region8) and for strain M29T they ranged from 0.013 to 0.210 (0.018–0.428 based on the 315-bp variable region). The gyrB gene-based genetic distance values between strain JS72T and the type strains of 16S-rRNA gene cluster A (Figure 1) were all above the 0.02 threshold proposed to distinguish genomic species8 (0.026–0.055; 0.036–0.116 based on the 315-bp variable region). The values between strain M29T and A. keratiniphila subsp. keratiniphila, A. keratiniphila subsp. nogabecina and A. lurida were 0.013, 0.017 and 0.019 (0.021, 0.021 and 0.018 based on the 315-bp variable region). The values between strain M29T and all the other members of 16S-rRNA gene cluster A were above the 0.02 threshold (0.022–0.051; 0.028–0.100 based on the 315-bp variable region). These data show that strain M29T is likely a distinct genomic species from all Amycolatopsis-type strains except A. lurida (based on the 315-bp variable region). Thus, further evidence is needed to prove that strain M29T is a distinct species from A. lurida. The gyrB genetic distance between JS72T and M29T was 0.046 (0.081 based on 315-bp variable region), above the 0.02 threshold, which indicates they are likely separate species.8

The recN gene-based genetic distance values (1231 bp) were calculated between strains JS72T, M29T and all type strains for which recN gene sequences are available (Supplementary Table 1). For strain JS72T these genetic distances ranged from 0.056 to 0.317 (with the genetic distances against members of its 16S-rRNA gene cluster ranging from 0.056 to 0.095) and for strain M29T the values ranged from 0.055 to 0.318 (with the genetic distances against members of 16S-rRNA gene cluster A ranging from 0.055 to 0.107). The recN genetic distance between strains JS72T and M29T was 0.095. All these values are above the proposed 0.04 threshold to distinguish novel species in the genus.9 These data further suggest that strains JS72T and M29T are distinct species from all type strains of the genus Amycolatopsis. All gyrB- and recN-gene-based genetic distance values are presented in Supplementary Figures 3–5.

DNA–DNA hybridization experiments were performed between strains JS72T and M29T and their closest relatives based on 16S-rRNA gene sequence similarities. The results show that strain JS72T shared 50(15)% DNA relatedness to A. thailandensis JCM 16380T and 48(5)% to A. lurida NRRL 2430T. Strain M29T shared 29(25)% to A. lurida NRRL 2430T, 26(25)% to A. keratiniphila subsp. nogabecina NRRL B-24256T and 22(8)% to A. decaplanina NRRL B-24209T. Strains JS72T and M29T shared 25(5)% DNA relatedness. This proves that both strains JS72T and M29T are distinct genomic species.22

Screening for antibiotic biosynthetic genes revealed the presence of the P450 monooxygenase B (oxyB) gene, required for the production of glycopeptide antibiotics, in both strains JS72T (GenBank accession number: KF771261) and M29T (KF771265). Strain JS72T was also shown to contain the 3-amino-5-hydroxy-benzoic acid synthase gene (KF771260), required for the production of ansamycin antibiotics.

Overlay experiments showed that strain JS72T exhibited antibacterial activity against M. aurum A+, E. faecalis (vancomycin sensitive), E. phoeniculicola JLB-1T and S. aureus ATCC 25923, but showed no activity against E. coli ATCC 25922 and E. faecium VanA (vancomycin resistant). Strain M29T showed activity against M. aurum A+, but no activity against E. coli ATCC 25922 and E. faecium VanA. Combined solvent extracts of the broth culture of strain M29T showed activity against B. subtilis var ING, M. bovis BCG (Tokyo), M. aurum A+, M. smegmatis LR222, M. tuberculosis H37RvT and P. aeruginosa ATCC 27853, but no activity against E. coli ATCC 25922 and E. faecium VanA.

Both strains JS72T and M29T grouped in 16S-rRNA gene cluster A of Everest and Meyers,8 which contains most of the glycopeptide-producing Amycolatopsis type strains.23 The presence of the genes for glycopeptide production in these strains and their inability to inhibit the vancomycin-resistant E. faecium VanA strain suggest that these strains produce a glycopeptide-type antibiotic. Further support for this is lent by the fact that strain JS72T shows activity against a vancomycin-sensitive E. faecalis strain.

The results of the phenotypic characterization of strains JS72T and M29T are presented in Table 1 and the species descriptions. These data showed that the strains are phenotypically distinct from their closest-type strain relatives. The chemotaxonomic characteristics of the strains are presented in the species description and are consistent with membership of the genus Amycolatopsis.1, 24, 25, 26

Table 1 Phenotypic differences between strains M29T, JS72T and closely related Amycolatopsis-type strains
Full size table

Description of Amycolatopsis speibonae sp. nov.

Amycolatopsis speibonae (spei.bo’nae. L. n. spes -ei hope; L. adj. bonus good; N.L. fem. adj. speibonae of good hope, to indicate Cape Town, the Cape of Good Hope, South Africa, the source of the soil from which the type strain was isolated).

Gram positive. Colonies appear convoluted with raised centers on most media. Vegetative mycelium appears cream-brown in color, fragmenting into short rod-shaped elements in both liquid and agar cultures. Aerial mycelium appears white on ISP 4. No diffusible pigments are produced. Melanin is not produced on peptone-yeast extract-iron agar (ISP 6) or tyrosine agar (ISP 7). Grows at 30 °C, but not at 37 °C. Grows at pH 4.3, 7 and 10 as well as in the presence of up to 4% (w/v) NaCl, but not in the presence of 7% (w/v) NaCl. Catalase positive. Oxidase negative. Nitrate is reduced to nitrite. Does not produce H2S. Starch is not hydrolyzed. Casein, gelatin, hypoxanthine, Tween 80, L-tyrosine and urea are degraded. Allantoin is weakly degraded. Adenine, guanine, xanthine and xylan are not degraded. Utilizes adonitol, L(+)-arabinose, D(+)-cellobiose, D(–)-fructose, D(+)-glucose, meso-inositol, D(–)-mannitol, D(+)-mannose and D(+)-xylose as sole carbon sources. Weakly utilizes raffinose, D(+)-melibiose, α-lactose, salicin, sodium acetate and sodium citrate. Unable to utilize inulin, L(+)-rhamnose and sucrose as sole carbon sources. Utilizes DL-α-amino-n-butyric acid, L-asparagine, L-histidine, potassium nitrate and L-threonine as sole nitrogen sources, with weak growth on L-arginine, L-cysteine, L-4-hydroxyproline, L-phenylalanine, L-serine and L-valine. Unable to utilize L-methionine as a sole nitrogen source. Contains a type IV cell wall (meso-diaminopimelic acid, arabinose and galactose in the cell wall).27 Mycolic acids are absent. The polar lipid profile includes diphosphatidylglycerol, phosphatidylethanolamine, phosphatidylmethylethanolamine, phosphatidylinositol, phosphatidylinositol-mannoside, two unidentified phospholipids, two unidentified aminolipids, two unidentified glycolipids, an unidentified phosphoglycolipid and an unidentified aminophospholipid (Supplementary Figure 6). The predominant menaquinone is MK-9(H4) (85%), with minor amounts of MK-8(H4) (11%). The major fatty acids of the type strain are i-C15:0 (17.36%), C17:1ω8c (16.93%), C17:0 (13.45%), ai-C15:0 (11.14%), C16:1ω7c and/or iso-C15:0 2-OH (9.44%) and i-C16:0 (9.7%), with minor amounts of C15:1ω6c (7.42%), i-C14:0 (6.5%), ai-C17:0 (3.28%), C17:1ω6c (1.7%), C13:0 (1.54%) and C16:0 (1.53%).

Antibacterial activity is exhibited against E. phoeniculicola JLB-1T, a vancomycin-sensitive Enterococcus sp., M. aurum A+ and S. aureus ATCC 25923.

The type strain, JS72T (=DSM 46660T=NRRL B-24958T), was isolated from soil from Table Mountain above the Upper Campus of the University of Cape Town, Cape Town, South Africa.

Description of Amycolatopsis roodepoortensis sp. nov.

Amycolatopsis roodepoortensis (roo.de.poort.en’sis. N.L. fem. adj. roodepoortensis, pertaining to Roodepoort, a suburb in Gauteng province of South Africa, the source of the soil from which the type strain was isolated).

Gram positive. Colonies appear convoluted with raised centers on most media. Vegetative mycelium appears cream-brown in color, fragmenting into short rod-shaped elements in both liquid and agar cultures. Aerial mycelium appears cream on ISP 4. No diffusible pigments are produced. Melanin is not produced on peptone-yeast extract-iron agar (ISP 6) or tyrosine agar (ISP 7). Grows at 30 °C and at 37 °C. Grows at pH 4.3, 7 and 10 as well as in the presence of up to 7% (w/v) NaCl. Catalase positive. Oxidase negative. Nitrate is not reduced to nitrite. Produces H2S. Starch is not hydrolyzed. Casein, gelatin, hypoxanthine, Tween 80, L-tyrosine, urea and xanthine are degraded. Xylan is weakly degraded. Adenine, allantoin and guanine are not degraded. Utilizes adonitol, L(+)-arabinose, D(+)-cellobiose, D(–)-fructose, D(+)-glucose, meso-inositol, lactose, D(–)-mannitol, D(+)-mannose, salicin and D(+)-xylose as sole carbon sources. Weakly utilizes sodium acetate, sodium citrate and sucrose. Unable to utilize inulin, D(+)-melibiose, raffinose and L(+)-rhamnose as a sole carbon source. Utilizes DL-α-amino-n-butyric acid, L-arginine, L-asparagine, L-histidine, L-serine, L-valine and L-phenylalanine as sole nitrogen sources, with weak growth on L-cysteine, L-4-hydroxyproline, L-methionine, potassium nitrate and L-threonine. Contains a type IV cell wall (meso-diaminopimelic acid, arabinose and galactose in the cell wall).27 Mycolic acids are absent. The polar lipid profile includes diphosphatidylglycerol, phosphatidylethanolamine, phosphatidylmethylethanolamine, phosphatidylinositol, two unidentified phospholipids, an unidentified aminolipid, six unidentified glycolipids, an unidentified aminophospholipid, three unidentified aminoglycolipids and an unidentified phosphoglycolipid (Supplementary Figure 7). The predominant menaquinone is MK-9(H4) (84%), with minor amounts of MK-8(H4) (12%) and MK-9(H2) (3%). The major fatty acids of the type strain are C17:1ω8c (22.99%), i-C16:0 (22.55%), C17:0 (9.5%) and C17:1ω6c (9%), with minor amounts of C16:1ω7c and/or iso-C15:0 2-OH (7%), i-C15:0 (6.75%), C15:1ω6c (6.72%), i-C14:0 (6.44%), ai-C15:0 (1.99%), C16:0 (1.85%), ai-C17:0 (1.81%), C15:0 2-OH (1.2%), i-C17:0 (0.76%), C18:1ω9c (0.53%), C13:0 (0.44%) and C14:0 (0.38%). Antibacterial activity is exhibited against B. subtilis var ING, M. bovis BCG (Tokyo), M. aurum A+, M. smegmatis LR222, M. tuberculosis H37RvT and P. aeruginosa ATCC 27853.

The type strain, M29T (=DSM 46661T=NRRL B-24959T), was isolated from soil from a suburban garden in Roodepoort, Gauteng Province, South Africa.