OUP user menu

Application of the variable region in 16S rDNA to create an index for rapid species identification in the genus Streptomyces

Masakazu Kataoka, Kumiko Ueda, Takuji Kudo, Tatsuji Seki, Toshiomi Yoshida
DOI: http://dx.doi.org/10.1111/j.1574-6968.1997.tb12578.x 249-255 First published online: 1 June 1997


Partial nucleotide sequences (120 bp) of the 16S rRNA gene (rDNA) containing a variable α region were compared in 89 strains of the genus Streptomyces belonging to eight major clusters of category I in Bergey's Manual of Systematic Bacteriology. Fifty-seven kinds of partial 16S rDNA sequences were observed among the 89 strains. Forty-three of the strains were grouped into 11 ‘identity groups’, based on the fact that the strains in each group shared an identical sequence in the 120-bp region. The results of a phylogenetic analysis based on the 16S rDNA 120-bp sequences revealed that 60 of the 89 strains could be categorized into seven clusters, each consisting of four or more strains. Based on these observations it was concluded that short nucleotide sequences bearing the variable α region are useful for Streptomyces species identification.

  • Phylogenetic analysis
  • Streptomyces
  • 16S rRNA gene

1 Introduction

In the conventional taxonomy of Streptomyces, its systematics is, in practice, based on characteristics of mycelial morphology, pigment production, and certain physiological properties. In 1964, the International Streptomyces Project (ISP), a collaborative study group on Streptomyces taxonomy, proposed an elaborate description of criteria for some of the authentic and extant type strains of Streptomyces and their related taxa, which is dependent on a limited number of taxonomic characteristics selected from those accumulated in conventional taxonomic tests [1, 2]. In 1983, a large-scale numerical phenetic survey of Streptomyces and related taxa was carried out by Williams et al. [3]; 394 strains of Streptomyces-type cultures were examined with respect to 139 taxonomic characteristics and the results were analyzed statistically. Based on the findings, Streptomyces-type strains were grouped into 23 ‘major clusters’ each containing four or more strains, 20 ‘minor clusters’ each containing two to three strains, and 25 ‘single member clusters’[3]. The results of the above work are compiled in a chapter of Bergey's Manual of Systematic Bacteriology, Vol. 4 [4]. The proposed classification, however, disagrees with that obtained by the conventional taxonomic methods in several respects.

Recent progress in molecular biology and population genetics has dramatically increased the amount of evolutionary information available for species classification. For example, a database of ribosomal RNA genes, particularly of small subunits of ribosomal RNAs (16S rRNAs), has been compiled and successfully applied to determining phylogenetic relationships in bacteria [5]. In the case of Streptomyces, these studies have revealed that, in some cases, phenetic classification better reflects the phylogenetic relationships between the strains examined [6, 7].

Here, 89 ISP standard strains belonging to eight major clusters of category I defined by numerical identification in Bergey's Manual of Systematic Bacteriology [4] were subjected to a sequence comparison of a part of their 16S rRNA gene containing a highly variable α region and to phylogenetic analysis. From the phylogenetic relationships of these strains, we propose the usefulness of a data base generated on the basis of partial rDNA sequences for the rapid identification of species of the genus Streptomyces.

2 Materials and methods

2.1 Strains and culture conditions

The 89 strains of the genus Streptomyces used in this study (see Fig. 3) were obtained from the Japan Culture Collection of Microorganisms (JCM; Riken, Wako, Saitama, Japan). All these strains belong to eight major clusters (species groups) of category I, defined by numerical classification, in Bergey's Manual of Systematic Bacteriology [4]: S. albidoflavus, S. anulatus, S. halstedii, S. exfoliatus, S. violaceus, S. fulvissimus, S. rochei and S. chromofuscus. They were cultivated by the methods recommended in the JCM Strain Catalogue [8].

Figure 3

Phylogenetic tree constructed from the 120-bp α region sequences of 89 Streptomyces strains belonging to eight major clusters of category I in Bergey's Manual of Systematic Bacteriology [4]. Strain names are followed by the description in JCM catalogue but are cited as subjective synonyms in Bergey's Manual except for the type strains of the major cluster. The number preceding the species name indicates a major cluster of category I used in Bergey's manual: 1, S. albidoflavus; 2, S. anulatus; 3, S. halstedii; 4, S. exfoliatus; 5, S. violaceus; 6, S. fulvissimus; 7, S. rochei; 8, S. chromofuscus. JCM strain numbers (J), ISP strain numbers (I) and DDBJ DNA sequence accession numbers (D) for the 120-bp α region sequences are given along with the strain names. The IG and PC group symbols are shown in bold letters. The bar labeled 0.05 indicates 5 base changes per 100 nucleotides.

2.2 PCR amplification and sequencing of a part of 16S rDNA

To achieve the reliable sequencing of the partial 16S rDNA obtained by PCR amplification from various Streptomyces strains, direct sequencing of PCR products were carried out. Chromosomal DNA for the PCR template was prepared from cells in a single colony formed on a plate using an Insta Gene kit (Bio-Rad) according to the supplier's protocol. DNA fragments covering the variable region of 16S rDNA was amplified using phosphorylated sense and non-phosphorylated anti-sense primers [10]. The sense primer was phosphorylated using T4 polynucleotide kinase (Takara Shuzo) and ATP. PCR was performed in 50 μl reaction mixture (10 mM Tris-HCl, pH 8.3; 50 mM KCl; 1.5 mM MgCl2; 0.001%, w/v, glycerol) for 30 or 40 cycles of denaturation (for 30 s at 97°C), annealing (for 1 min at 50°C) and extension (for 1 min at 72°C) with Amplitaq DNA polymerase (Perkin Elmer). Whole samples were fractionated by agarose gel electrophoresis, and PCR products of 500 bp were recovered using a Gene Clean II kit (Bio 101). The phosphorylated sense strand was digested for 1 h at 37°C with λ-exonuclease (BRL) in the recommended buffer (67 mM glycine-KOH, pH 9.4; 2.5 mM MgCl2[9]). After phenol/chloroform extraction, the remaining anti-sense strand was used as a sequencing template.

DNA sequences were determined using a 7-deaza dGTP Sequenase Version 2.0 kit (USB) according to the supplier's protocol, except that 1 μg of single-stranded DNA binding protein (SSB, Stratagene) was added to 15 μl of the reaction mixture. After termination of the reaction, SSB was digested with 20 μg proteinase K (Sigma) for 30 min at 37°C in 6 μl of the termination mixture.

2.3 Oligonucleotides

Nucleotide sequences of the synthesized oligonucleotides used were as follows. Sense primer for PCR: 5′-TCACGGAGAGTTTGATCCTG-3′; anti-sense primer for PCR: 5′-GCGGCTGCTGGCACGTAGTT-3′; sequencing primer: 5′-AGTAACACGTGGGCAATCTG-3′. The primer sequences were selected from the conserved region and corresponded to nucleotide positions 1–20, 481–500, and 105–124, respectively, according to the Streptomyces ambofaciens rDNA sequence [11].

2.4 Analysis of rDNA sequences and construction of phylogenetic trees

Nucleotide sequences of the amplified products were edited by the GENETYX program (Software Development) on a PC9801 personal computer (NEC). Evolutionary analysis of nucleotide sequences was performed using the ODEN program (National Institute of Genetics, Japan). For construction of the evolutionary genetic distance matrices, frequencies of base substitutions per nucleotide were estimated by the method of Kimura [12]. The phylogenetic tree was constructed by the neighbor-joining method [13].

3 Results and discussion

3.1 Selection of a variable region suitable for phylogenetic analysis in the 16S rRNA gene of Streptomyces

To identify the most informative region of 16S rDNAs for the phylogenetic study of Streptomyces species, the nucleotide sequences of 16S rDNAs of Streptomyces lividans TK21 [14], S. coelicolor A3(2) [15], S. griseus[16], and S. ambofaciens[11] were aligned. This enabled the selection of a 120-bp region from nucleotide positions 158–277 of the 16S rRNA gene of S. ambofaciens, as it includes the most variable α region [17].

To test the verity of a phylogenetic tree based on partial sequences of the 16S rRNA gene, we compared two phylogenetic trees constructed by the neighbor-joining method using published total and partial 16S rRNA sequences of seven Streptomyces strains [11, 1416, 18]. We found the general topologies of the trees from the partial and total sequences to be similar (Fig. 1), but the genetic distance to be more emphasized in the partial-sequence tree. This observation indicated that the variation in the partial 16S rDNA sequences covering the variable region is sufficient for deducing phylogenetic relationships of Streptomyces strains at the intra-species level, although these segments are probably too variable for comparison at the generic level.

Figure 1

Phylogenetic relationships among published nucleotide sequences of the 16S rRNA gene in seven Streptomyces strains. The phylogenetic trees were constructed by the neighbor-joining method using whole (left) and partial (right) sequences of the 16S rRNA gene. The bar labeled 0.05 indicates 5 base changes per 100 nucleotides.

3.2 DNA sequences of 89 independent Streptomyces strains

A 500-bp DNA sequence, partially encompassing the 16S rRNA gene, was amplified from a DNA sample prepared from a single colony of each of the 89 strains, and the 120-bp sequence bearing the variable α region was determined (see Fig. 3 for the DDBJ accession numbers of these sequences). The nucleotide sequences of the hyper-variable region from nucleotide positions 172–202 in 17 out of the 89 strains, as well as some variations outside of the region, are shown in Fig. 2. It has been suggested that the hyper-variable region forms a stem and loop structure in vivo in S. ambofaciens 16S rRNA [11]. The conservation of the secondary structure is particularly notable, but the sequences forming the stem are the most variable: the most of complementing nucleotides in the stem were changed together. The GCAT sequence in the loop was found to be highly conserved, indicating that it may have an important role in the ribosome function. Only three strains, S. alboniger (Fig. 2), S. niveus (not shown), and S. spheroides (not shown), differed from the others in having a TCCT sequence at the corresponding GCAT site.

Figure 2

Sequence variation in the variable region of 16S rRNA among 17 Streptomyces strains. Numerals indicate the nucleotide positions corresponding to the 16S rRNA gene of S. ambofaciens. The region of nucleotide positions 172–202 is the most variable region, and is believed to form a stem and loop structure (see text). Asterisks represent perfectly conserved bases in the 17 species so far examined.

Among the 89 strains, 57 kinds of 120-bp nucleotide sequences could be identified. In particular, 11 identical sequences were shared by 47 of the strains: each sequence was shared by two to eight strains. The groups containing these strains were designated as ‘identity groups’ (IGs). It is apparent from Fig. 3 that the strains in each IG belong to the same respective numerical major cluster, the only exception being IG-G which includes strains from the major clusters of S. exfoliatus and S. violaceus. From the fact that the α region is the most variable, it is thought that strains labeled as various different species names but sharing an identical sequence may actually belong to a single species even though minor differences in their phenotypic characteristics resulted in their being classified as different.

3.3 Phylogenetic relationship through the partial sequence

To clarify the phylogenetic relationships of the 89 Streptomyces strains, the 120-bp sequences bearing the variable α region were subjected to computer-assisted phylogenetic analysis for the construction of a phylogenetic tree (Fig. 3). This tree has seven major phylogenetic clusters (classified as PC-A to PC-G), each comprised of four or more strains; 60 of the 89 strains were classified into these seven clusters. Williams et al. [3] have shown that the numerical major clusters, S. albidoflavus, S. anulatus, and S. halstedii, are closely related, and can be considered to form a large ‘super cluster’. This is partially in line with our results, the exception being the strains of the S. albidoflavus major cluster; according to the phylogenetic tree generated (Fig. 3), these are not closely related to the strains belonging to the major clusters of S. anulatus, and S. halstedii. In our tree, all the strains of the S. albidoflavus major cluster are grouped into phylogenetic cluster PC-D with less than three base substitutions. The homogeneity of these strains has also been suggested by the results from whole 16S rRNA sequences [19], chromosomal DNA hybridization [20], pyrolysis mass spectrometry [21] and fatty acid analysis [22]. PC-B is the largest cluster in our tree, comprising 19 of the 25 strains of the S. anulatus major cluster and five of the seven strains of the S. halstedii major cluster but it is not closely related to PC-D. Strains of other numerical cluster were found to be distributed among other phylogenetic clusters or were independent. The identity groups IG-J and IG-K are closely related to PC-G. The two strains belonging to IG-K, S. flavisclerotiacus and S. minutisclerotiacus, were formerly classified as Chainia species, but were later transferred to Streptomyces[23].

In conclusion, we believe that the phylogenetic tree presented here will serve as a useful tool for rapid identification of the phylogenetic localization of newly isolated Streptomyces strains, and that it is likely to prove more effective than the conventional methods. Additionally, it should provide a means for easy and appropriate identification, which is especially important in the fermentation industry in order to avoid taxonomic confusion of strains dealt with in patent literature.


We are grateful to Tobias Kieser (John Innes Centre, UK) for helpful discussions. This work was partially supported by Grant-in Aid 07044149 to T.Y. under the International Scientific Research Project of the Japanese Ministry of Science, Education, Sports and Culture.


  • 1 M. Kataoka and K. Ueda contributed equally to this work.

  • 2 Mitsubishi-Kasei Institute of Life Sciences, 11 Minami-Ooya, Machida-shi, Tokyo 194, Japan.

  • 3 Institute for Fermentation, Osaka, Juso-honmachi, Yodogawa-ku, Osaka 532, Japan.


View Abstract