Scholarly article on topic ' Diversity of epothilone producers among S orangium strains in producer-positive soil habitats '

Diversity of epothilone producers among S orangium strains in producer-positive soil habitats Academic research paper on "Industrial Biotechnology"

0
0
Share paper
Academic journal
Microbial Biotechnology
OECD Field of science
Keywords
{""}

Academic research paper on topic " Diversity of epothilone producers among S orangium strains in producer-positive soil habitats "

Diversity of epothilone producers among Sorangium strains in producer-positive soil habitats

Shu-guang Li,1§ Lin Zhao,1§ Kui Han,1§ Peng-fei Li,1 Zhi-feng Li,1 Wei Hu,1 Hong Liu,1 Zhi-hong Wu1 and Yue-zhong Li1*

1 State Key Laboratory of Microbial Technology, School of Life Science, Shandong University, Jinan 250100, China.

Summary

Large-scale surveys show that the anti-tumour compounds known as epothilones are produced by only a small proportion of Sorangium strains, thereby greatly hampering the research and development of these valuable compounds. In this study, to investigate the niche diversity of epothilone-producing Sorangium strains, we re-surveyed four soil samples where epothilone producers were previously found. Compared with the < 2.5% positive strains collected from different places, epothilone producers comprised 25.0-75.0% of the Sorangium isolates in these four positive soil samples. These sympatric epothilone producers differed not only in their 16S rRNA gene sequences and morphologies but also in their production of epothilones and biosynthesis genes. A further exploration of 14 soil samples collected from a larger area around a positive site showed a similar high positive ratio of epothilone producers among the Sorangium isolates. The present results suggest that, in an area containing epothilone producers, the long-term genetic variations and refinements resulting from selective pressure form a large reservoir of epothilone-producing Sorangium strains with diverse genetic compositions.

Introduction

Myxobacteria are a special type of bacteria that adopt complicated multicellular social lifestyles (Whitworth, 2008). These bacteria are also known fortheir biosynthesis of diverse and novel secondary metabolites (Reichenbach,

Received 10 July, 2013; revised 24 October, 2013; accepted 30 October, 2013. *For correspondence. E-mail lilab@sdu.edu.cn; Tel. (+86) 531 88564288; Fax (+86) 531 88564288. §These authors contributed equally to this work. Microbial Biotechnology (2014) 7(2), 130-141 doi:10.1111/1751-7915.12103

Funding Information No funding information provided.

2001). Among the different myxobacterial taxa, Sorangium cellulosum is an intriguing organism for drug-screening efforts because the secondary metabolic compounds discovered from S. cellulosum strains comprise up to 48.4% of the total metabolites obtained thus far from myxobacteria (Gerth etal., 2003). For example, epothilones (Fig. 1A), which act on cancer cells by mimicking the mechanism of Taxol, i.e. stabilizing microtubules (Bollag etal., 1995), are produced by S. cellulosum (Gerth etal., 1996). Biochemical, pharmacological and clinical studies have shown that epothilones are highly promising for cancer treatment (Reichenbach and Hofle, 2008). Some epothilones and their chemically modified derivatives are being used in clinical studies or trials (Larkin and Kaye, 2006), and one has been approved for clinical use by the U.S. Food and DrugAdministration. However, incontrastto increasing progress in their applications, the production of epothilones in Sorangium strains is not optimal. Sorangium cells grow slowly, and they possess multiple antibiotic resistance capabilities, have abundant extracellular polysaccharides and exhibit a tendency to aggregate (Shimkets etal., 2006), all of which impede the isolation and cultivation of Sorangium strains as well as their genetic manipulation. For example, since the conjugation method was first developed in S. cellulosum in 1992 (Jaoua etal., 1992), genetic methods have improved (Pradella etal., 2002; Julien and Fehd, 2003; Kopp etal., 2004; Xia etal., 2008), but metabolic engineering in Sorangium remains inefficient. Many heterologous hosts have been employed for epothilone biosynthesis, including Streptomyces coelicolor (Tang etal., 2000), S. venezuelae (Park etal., 2008), Escherichia coli(Mutka et al., 2006), Pseudomonas putida (Fu etal., 2008) and Myxococcus xanthus (Julien and Shah, 2002; Lau etal., 2002; Fu etal, 2008; Opwald etal., 2012), but epothilone production is very low in these heterologous hosts. The ratio of epothilone producers is also low among Sorangium strains (Gerth et al., 2001), and screening for producers of epothilones has to be based on large libraries of Sorangium strains. For example, taking advantage of their massive collection of Sorangium strains isolated from all over the world over the past decades, the former German Research Centre for Biotechnology (GBF) has identified dozens of epothilone producers (39 out of 1600) (Gerth etal., 2001; 2003). Therefore, searching for more epothilone producers is necessary not onlyto identify new strains with potentially suitable characteristics but also

„.OH

Epothilone A (R = H) Epothilone B (R = CH3)

Region I

Region II

Region

Region IV

ß-lactamase

Genomic region encoding Hypothetical proteins

Epothilone biosynthesis gene cluster

Transposase tRNA-Glu

Ieposa Nl eposp "N|EPOSB^>r

~EF BTE-

Module Q NRPS Module 1 Module 2 Module 3 Module 4 Module 5 Module 6 Module? Module 8 P450

—----- -■-1 №------1 h— -,J=, -^1=-II— —31-------Ii" I C=

epoC-at2 or AT

Hypothetical proteins

Putative post-modification enzymes

Fig. 1. The molecular structures of the epothilone analogues A and B (A) and their biosynthetic gene cluster and its flanking region (B) in S. cellulosum So0157-2 according Han and colleagues (2013). The region in (B) has a length of 128 kb and can be divided into four independent subregions: Region I (length = 12 kb, GC% = 70%) encodes 12 hypothetical proteins and 1 p-lactamase, and none of the genes had synteny with the reported So ce56 genome. Region II (length = 61 kb, GC% = 69%) is responsible for epothilone biosynthesis and post-modification. Region III (length = 51 kb, GC% = 72%) is a non-ribosomal peptide synthetase (NRPS) gene cluster with no clear function, and all the genes arose from Burkholderia-sourced sequences. In region IV (length = 4 kb), the genes encoding a putative transposase and a tRNA gene (glutamic acid, Glu) are identified. ACP, acyl carrier protein; AT, acyltransferase; ATModC2, the second acyltransferase domain of the epoC module; KR, p-ketoreductase; KS, p-ketoacyl synthase; NRPS, non-ribosomal peptide synthetases.

to obtain more strains for further genetic modification, as well as for studies of their biosynthetic mechanisms.

Sorangium is a genus of cellulolytic myxobacteria, and it is proposed to contain at least two species (Yan etal., 2003; Shimkets etal., 2006; Jiang etal., 2008). In classical isolation techniques, one or two clones of one myxobacterial species with obviously different morphological characteristics are normally isolated from a given soil sample (Dawid, 2000; Li etal., 2000; Shimkets etal., 2006). However, molecular surveys (Wu etal., 2005; Jiang etal., 2007) indicate that there are many myxobacterial 16S rRNA gene sequences with < 2% differences between them. For example, among the 85 randomly sequenced clones in one molecular ecological survey, five different sequences belonged to the

Sorangium branch (Jiang etal., 2007). In other words, except of one or two isolates, many Sorangium linkages in soils may be ignored during normal large-scale isolation procedures. It is known that the growth of myxobacteria is cell density-dependent (Shimkets etal., 2006), and different strains of the same myxobacterial species normally separate to form respective colonies (Vos and Velicer, 2009), which suggest that separate colonies might represent different Sorangium strains. In fact, intraspecies diversity of M. xanthus was once reported in centimetre-scale soil samples (Vos and Velicer, 2008), which was later showed to be diverse in biosynthesis of secondary metabolites (Krug etal., 2008). In our previous studies, we performed a large-scale screen of more than 800 Sorangium strains isolated from 425 soil specimens

Niche-diversity of Sorangium epothilone producers 132 Table 1. Information of the soil samples analysed in this study and the positive ratios of epothilone producers among the Sorangium isolates.

Soil Isolation Detected Epothilone Positive

samples* Soil information number number producer ratio

0003 A wheat field in Jiangsu Province, collected in February, 1996 13 13 4 30.8%

0007 A vegetable field in Jiangsu Province, collected in February, 1996 8 8 3 37.5%

0087 A paddy field, in Fujian Province, collected in August, 1996 24 24 18 75.0%

0157 An alkaline soil in Yunnan Province, collected in September, 2002 12 8 2 25.0%

0155 An alkaline soil near the 0157 site, collected in September, 2002 14 14 6 42.9%

0156 An alkaline soil near the 0157 site, collected in September, 2002 22 2 1 50.0%

0558 An alkaline lake mud near the 0157 site, collected in August, 2008 73 51 13 25.5%

0559 An alkaline lake mud near the 0157 site, collected in August, 2008 10 10 1 10.0%

Total 176 130 48 36.9%

*Other assayed 10 soils near the 0157 site: nine produced no Sorangium isolates (0146, 0147, 0148, 0158, 0555, 0556, 0557, 0560 and 0561); while one had six isolates but none was epothilone producers (0145).

collected in different places of China to identify the epothilone producers (Dong etal., 2004; Hu etal., 2004; Li eta/., 2007), and the positive ratio was no more than 3%. In this study, to acquire more epothilone producers and investigate the diversity of Sorangium isolates in single soil samples, we re-surveyed four soil samples from which epothilone-producing strains or strains potentially possessing epothilone biosynthesis genes were discovered. Furthermore, the survey was extended to 14 specimens collected within approximately 10 km2 near one positive site.

Results

Diversity of Sorangium strains in soil niches

In our previous screening studies of hundreds of Sorangium strains (Dong etal., 2004; Hu etal., 2004), three epothilone producers (designated So0003-3, So0007-3 and So0157-2) were obtained from three soil samples (soil ID numbers 0003, 0007 and 0157). An additional strain, So0087-5, from the 0087 soil sample, did not produce epothilones, but the requisite biosynthesis genes might be present, as an almost identical ketoacyl synthase (KS) domain (704 bp) for the biosynthesis of epothilones was detected in So0087-5 (Identities = 98%, Positives = 98%) (Li etal., 2007). These four soil samples (Table 1) were regarded as positive soil samples and were re-surveyed for niche diversity of epothilone producers in this study. We picked and purified 8 to more than 20 separated Sorangium colonies from the four soil samples, depending on the number of colonies that appeared on the isolation medium. These Sorangium isolates grew well on mineral medium with filter paper as the only carbon source (CNST medium). The isolates from the same soil samples had either similar or different morphological characteristics (some representative Sorangium morphologies from each sample are shown in Fig. 2A). It was noted that some isolates, such as So0003-22, So0007-6-3-1 and So0157-24, could barely form fruiting body structures on

CNST medium. Isolation of these non-fruiting Sorangium strains was based on the characteristics of Sorangium swarms. Phylogenetic analysis demonstrated that most of these separate Sorangium clones from one soil sample differed in their 16S rRNA gene sequences (Fig. 2B). The sympatric Sorangium strains often clustered into single phylogenetic groups but also included separate members. For example, nine of the 12 sequenced 0003 isolates were in group S1, and three were in group S7; whereas except for one in group S5, 9 of the 10 sequenced 0087 isolates were in group S1. However, these isolates showed significant phylogenetic differences. The largest phylogenetic distance between sympatric isolates was between So0003-31-1 and So0003-22 and was approximately 2.2%, which is similar to the phylogenetic distance observed for different soil samples (Yan etal., 2003). These results indicated that Sorangium strains were morphologically and phylogenetically diverse in single soil samples.

Production of epothilones in the Sorangium isolates

As determined by high-pressure liquid chromatography (HPLC)-mass spectrometry (MS), the positive ratio of epothilone producers ranged from 25.0% to 75.0% of the Sorangium isolates in the four soil samples (Table 1; the production of epothilone A in different strains is listed in Table 2; some HPLC and MS values are shown in Supplementary material S1). The positive ratio for the production of epothilones (27/53) in these four positive samples was significantly higher than the previous reports (39/ 1600) from unselected soil samples (Gerth etal., 2001; 2003) (P < 0.001, Fisher's exact test). We checked some soil samples, i.e. 0081, 0085, 0133, 0139, 0184 and a campus garden soil sample in Shandong University, from which epothilone-producing Sorangium strains were not found (Dong etal., 2004; Hu etal, 2004; Li etal., 2007). In the resurveys, 5 to more than 10 separated Sorangium colonies were isolated and purified from these six 'nega-

Fig. 2. Morphological characteristics (A) and phylogenetic relationships (B) of selected Sorangium isolates. Strains in (A) were incubated on CNST medium for 2 weeks. Bar = 5 mm. These Sorangium representatives showed different morphological characteristics, such as swarm shapes, structure and colour of fruiting bodies and sporangioles. The phylogenetic tree in (B) was constructed using 16S rRNA gene complete sequences. Chondromyces crocatus Cm c5 (GenBank accession number GU207874) was used as the root. Sorangium strains So ce26 (AF387629), So ce56 (AJ316014), So ce1871 (FJ457641) and the epothilone-producing strains, So ce90 (AJ316015) and KYC3013 (EU242515), were used as references (shown in italics). The epothilone producers are underlined for easy tracking. The bar is equivalent to one nucleotide change per 100 bp, The bootstrap support is from 1000 replicates. S1-S6 represent Sorangium subgroups 1-6 respectively.

tive' soil samples, but none of them were epothilone producers (data not shown), which is consistent with our previous screen. The results suggested that epothilone producers were normally restricted in some places, instead of being ubiquitous in soil. However, considering there were also high ratios of non-epothilone producers in those 'positive' soil samples, screening on two or three isolates probably missed epothilone producers in a soil

sample. For example, although our screening for the production of epothilones was negative with two strains isolated from the 0087 soil sample (Dong etal., 2004; Hu etal., 2004), many Sorangium isolates from the soil yielded epothilones. These results suggested that the previously reported positive ratio was an underestimation of the epothilone producers among the Sorangium strains isolated from different samples.

■— So0087-7-1|EU864256 So ce90\AJ316015 So0003-1|EU545483

- So0003-1-2|EU545484 So0087-3-3|EU545498

- So0003-2-1|EU545485 So0087-5|AF387628 So0087-17|EU545502

" KYC3013\EU242515 — So0558-17b|HQ829376 r So0003-3|DQ256395 I—So0157-2|DQ256394 >1- So0007-3|AY252114 ft So0157-18|EU545507

So0003-19 EU545487 So0003-21 |EU545490

— So0003-31-1 |EU545492 So0007-12|EU545495 So0003-7|EU545486

L- So0003-19-2|EU545489 -So0087-2-2|HQ829366

— So0157-1|EU240496 So0087-10|EU545500

4- So0087-18|EU545503 So0087-20-1 |EU545504 So0087-14|EU545501

i-So0558-F|HQ829389

"I-So0558-G4|HQ829391

— So0559-8|HQ829394

— So ce56\AJ316014 ' So0558-10a|HQ829369

■ So0558-13b|HQ829371

■ So0558-G|HQ829390

, So0558-18|HQ829377 L So0558-33|HQ829383

e°"n156-7c|HQ829404 i0156-5|HQ829399 30558-b|HQ829388 o0156-20b|HQ829409 9I So0156-22|HQ829410

-So0558-16b|HQ829374

■ So ce1871\FJ457641 — So ce26\AF387629

-So0087-23|EU240532

74,-So0558-13|HQ829370

500558-15a|HQ829372 So0558-31|HQ829382 So0558-35b|HQ829386

500559-2|HQ829392

So0558-17a|HQ829375

So0558-24b|HQ829379

— So0558-21hei|HQ829378

J-So0156-18|HQ829408

1— So0156-7|HQ829403

-So0558-38a|HQ829387

-So0156-12|HQ829406

I-So0558-16|HQ829373

- So0156-6n|HQ829402 So0156-6c|HQ829401

■ So0156-13|HQ829407

— So0156-6a|HQ829400 I- So0156-9a|HQ829405

— So0558-4|HQ829367 So0559-7|HQ829393

— So0558-8|HQ829368 So0558-27|HQ829381

-So0155-3|HQ829396

i So0155-13a|HQ829397 1 So0155-15b|HQ829398

99. So0003-19-1|EU545488 68 H So0003-39|EU545493

i-So0003-22|EU545491

rt,,1-So0558-35a|HQ829384

So0007-6-3|EU545494

So0007-16|EU240531 I- So0157-29|EU545510 j— So0558-35a-1 |HQ829385 P— So0155-1 |HQ829395 1 r So0157-24|EU240530 „H— So0157-26|EU545509

S764T-So0157-25|EU545508

-Ch.crocatus Cm c5|GU207874

Fig. 2. Continued.

Table 2. Production of epothilone A by some Sorangium isolates on solid CNST medium after 2-week fermentation.

Epothilone A Epothilone A

Isolate (mg l-1: ) Isolate (mg l-1)

So0003-1-2 0.53 ± 0.02 So0087-26 17.09 ± 1.34

So0003-2-1 0.34 ± 0.02 So0157-2 2.47 ± 0.09

So0003-21 4.66 ± 0.11 So0157-18 1.58 ± 0.04

So0003-31-1 13.78 ± 2.27 So0155-3 0.01

So0007-6-3 4.50 ± 1.08 So0155-4a 0.01

So0007-12-2A 15.71 ± 2.30 So0155-9 0.01

So0007-12-2B 15.08 ± 1.87 So0155-13 0.01

So0087-2-1 4.68 ± 0.89 So0155-13a 0.01

So0087-2-2 1.94 ± 0.24 So0155-15 < 0.01

So0087-2-3 < 0.01 So0156-13 0.01

So0087-3-2 5.38 ± 0.16 So0558-8a 0.01

So0087-7-1 9.39 ± 0.38 So0558-10a 0.02 ± 0.01

So0087-7-2-1 16.25 ± 3.32 So0558-11b 0.01

So0087-7-2-2 11.97 ± 1.80 So0558-13 0.01

So0087-10 9.18 ± 2.15 So0558-16a < 0.01

So0087-13 3.16 ± 0.62 So0558-17 < 0.01

So0087-13-2 10.19 ± 2.20 So0558-23 0.13 ± 0.04

So0087-14 17.05 ± 1.33 So0558-29b 0.01

So0087-15 20.43 ± 2.51 So0558-31a 0.01

So0087-15-2 6.25 ± 1.09 So0558-G 2.42 ± 0.83

So0087-16 8.89 ± 1.07 So0558-G1 0.52 ± 0.18

So0087-17 12.88 ± 1.25 So0558-G2 0.75 ± 0.26

So0087-18 0.54 ± 0.02 So0558-G3 1.54 ± 0.53

So0087-20-2 6.87 ± 0.50 So0559-4 0.01

*For strains that had average productions < 0.01, the standard deviations were even smaller, and weren't shown in the table.

Biosynthesis genes for epothilones

Epothilones are biosynthesized via a seven modules of type I modular polyketide synthase (PKS) mixed with a non-ribosomal peptide synthetase (Julien etal., 2000; Molnar etal., 2000) (Fig. 1B). To determine the differences of epothilone biosynthesis genes between multiple producers, we sequenced ATModC2 in some of the epothilone-producing Sorangium isolates. Acyltransferase (AT) domains are fundamental elements of type I PKSs (Cheng etal., 2003; Ginolhac etal., 2004; Jenke-Kodama etal., 2005). Phylogenetic analysis indicated that the epothilone AT domains were distinct from those of other myxobac-teria and non-myxobacteria (Fig. S1). The AT protein sequences of all reported epothilone biosynthesis cluster in So0157-2 (Han etal, 2013), So ce90 (Molnar etal., 2000), SMP44 (Tang etal, 2000) and KYC3013 (Hyun etal., 2008) were phylogenetically analysed (Fig. 3), which showed that the epothilone ATs were sorted into two groups, malonyl-coenzyme A (CoA) specific and methylmalonyl-CoA specific. The distances of AT domains in the same modules were < 3.2%, with an average value of 1.3 ± 0.8%. It is noticed that the second AT domain of the epoC module (ATwodC2), which is supposed to indiscriminately accept malonyl-CoA for epothilone A or methylmalonyl-CoA for epothilone B (Gerth etal., 2000; Molnar etal., 2000; Yadav etal., 2003; Petkovic etal., 2008), was also located in the malonyl-CoA-specific group.

Similar to that of the 16S rRNA gene sequences, the amino acid sequences of ATm<xjc2 of sympatric Sorangium isolates were often located in the same branches but also included separate members (Fig. 3). For example, among the seven sequenced ATm<xjc2 domain from sample 0087, six were in the same branch (C1), whereas one was rather distant from the others (So0087-18 in C3). The largest distance between different ATModC2 was approximately 5.8%, between So0087-17 and So0157-18. These results suggested that the biosynthetic genes for epothilones might be different in those separate epothilone Sorangium producers obtained from single soil samples.

Screening for epothilone producers around a positive site

In the course of evolution, bacteria distributed globally; therefore, epothilone producers could be found on different continents (Bollag etal., 1995; Gerth etal., 1996; Dong etal., 2004; Hyun etal., 2008). To investigate the spreading and distribution capacity of epothilone-producing Sorangium strains, 14 soil specimens collected from different locations near the 0157 site were further assayed. The sampling locales were over approximately 10 km2, and the samples were all alkaline, similar to that of the 0157 soil sample. Sorangium strains were found in five samples, including 0145,0155,0156, 0558 and 0559, and the appearance ratio of Sorangium organisms in this region was significantly higher (35.7%, 5/14) than that described in the previous report (5.2%, 73/1398) (Dawid, 2000) (P< 0.001, Fisher's exact test). Six to seventy-three separate Sorangium clones were isolated from these five soil samples, depending on their appearance on isolation plates. Consistent with the alkaline characteristics of the soils, most of the Sorangium isolates preferred to grow at pH 9.0 (alkali-tolerant). Epothilone producers were found in four of the soil samples, and the percent of epothilone producers among the detected Sorangium isolates was 27.3% (21 out of 77) (Table 1; the production abilities of the epothilone producers are listed in Table 2), which had no significant difference with that in the initial positive 0157 sample (P = 1000 > 0.05, Fisher's exact test). Similarly, phylogenetic analysis of the 16S rRNA gene sequences of some new isolates showed that they differed, normally located in the same branches in the phylogenetic tree, but they also included some separate strains in different branches (Fig. 2B). The largest phylogenetic distance between isolates from the same soil sample was approximately 3.1%, between So0558-10a and So0558-13. These results suggest that epothilone producers frequently appear in a large area around the established positive site (0157 sample) where they are able to live.

So0087-2-1 epoC-at2|EU864255

- So0087-14 epoC-at2|EU554402 So0087-2-3 epoC-at2|EU554400 So0087-2-2 epoC-at2|EU554399

- So0087-17 epoC-at2|EU554403 So0087-7-1 epoC-at2|EU554401

- KYC3013 epoC-at1 |GU063811 SMP44 epoC-at1 |AF217189

L- So0157-2 epoC-at1 |CP003969

- So ce90 epoC-at1 |AF210843 So007-12-2A epoC-at2|EU554397

»L So007-12-2B epoC-at2|EU554398 ^r So0157-2 epoC-at2|CP003969 I So0157-18 epoC-at2|EU554404

- So ce90 epoC-at2|AF210843 SMP44 epoC-at2|AF217189

■ KYC3013 epoC-at2|GU063811 So0087-18 epoC-at2|EU554405

r So0157-2 epoA-at|CP003969 SMP44 epoA-at|AF217189 KYC3013 epoA-at|GU063811 L So ce90 epoA-at|AF210843 r So ce90 epoE-at|AF210843 So0157-2 epoE-at|CP003969 , KYC3013 epoE-at|GU063811 SMP44 epoE-at|AF217189 SMP44 epoC-at3|AF217189 - KYC3013 epoC-at3|GU063811 So0157-2 epoC-at3|CP003969 681 So ce90 epoC-at3|AF210843 |- So ce90 epoB-at|AF210843 - SMP44 epoB-at|AF217189 L So0157-2 epoB-at|CP003969 KYC3013 epoB-at|GU063811 So ce90 epoC-at4|AF210843 KYC3013 epoC-at4|GU063811 So0157-2 epoC-at4|CP003969 SMP44 epoC-at4|AF217189 831 So ce90 epoD-at1 |AF210843 So ce90 epoD-at2|AF210843 SMP44 epoD-at2|AF217189 ■ So0157-2 epoD-at1 |CP003969 - So0157-2 epoD-at2|CP003969 SMP44 epoD-at1 |AF217189 KYC3013 epoD-at1 |GU063811 KYC3013 epoD-at2|GU063811

malonyl-CoA-specific AT

methylmalonyl-CoA-specific AT

Fig. 3. Phylogenetic analysis of the protein sequences of AT domains in epothilone biosynthesis clusters. ATModC2 are underlined for easy tracking. The bar is equivalent to 10 amino acids change per 100 bp. The bootstrap support is from 1000 replicates, and only values greater than 50 are shown. C1-C3 represent three subgroups of ATModC2.

Discussion

As a result of both genetic variation and environmental selection, bacteria have evolved into the most diversified taxonomic group on Earth (Giovannoni and Stingl, 2005). This biodiversity, derived during evolution, may also involve particular metabolic pathways of a single species, such as the biosynthesis of epothilones in Sorangium strains. The biosynthetic enzymes for epothilones have been determined to be encoded by a chromosome-associated gene cluster in Sorangium cells (Julien etal., 2000; Molnar etal., 2000). In fact, except for the pMF1 plasmid from M. fulvus 124B02 (Zhao etal., 2008), there has been no report of extrachromosomal genetic materials in myxobacteria up to now. The biosynthesis gene cluster for epothilones has been reported in several Sorangium producers (Bollag etal., 1995; Gerth etal., 1996; Gong etal., 2007; Hyun etal, 2008), but the organization of the gene cluster on chromosomes is still not known. Analysing the genome sequence of one of our epothilone producers, S. cellulosum So0157-2 (GenBank accession number CP003969) (Han etal., 2013), suggested that a 128 kb region (Fig. 1B) including the epothilone gene cluster possibly arose from horizontal gene transfer. First, the whole 128 kb region has no synteny with the S. cellulosum So ce56 genome (Schneiker etal., 2007) (Fig. S2), and all the genes in region III are predicted to be from Burkhoder/a-sourced sequences. Second, tRNA and putative transposase genes are present near the end of region III (Chien etal., 2004). Third, an antibiotic gene (P-lactamase), often the element of integrons contained in some transposons (Levy and Marshall, 2004), is present in region I. We compared the sequences of the complete epothilone gene clusters in strains So0157-2, So ce90 and SMP44, which were isolated from different continents. The results revealed the high similarity among these three gene clusters with an average 98.5% identity, even in the non-coding intergenic spacer sequences (Fig. S3), suggesting that the biosynthesis genes for epothilones are highly conserved and origination-monophyletic, even though they are usually totally absent in non-producers (Schneiker etal., 2007).

The lack of diversified producers and efficient genetic performance systems is a serious limiting factor for the genetic engineering of epothilone producers and the industrialization of epothilone production. Furthermore, because of social characteristics of Sorang/um strains, difficulties in their isolation and cultivation make this approach hard to achieve. In the current study, resur-veying epothilone-positive soil samples using classical colony-based isolation techniques, we determined that Sorang/um strains in small habitats were a mixture of different strains, not only based on 16S rRNA gene

sequences and morphologies but also the production abilities of epothilones and their biosynthesis genes. Such an intraspecies diversity of sympatric Sorang/um strains is potentially useful for various applications. For example, the strains synthesizing valuable metabolites are regarded as potential industrial producers; however, the yields of metabolites in these fresh isolates are usually too low to be industrialized. Normally, after the first round of isolation, the diverse producing strains that remain in the soil are ignored and are not further re-mined. Instead, many efforts have been made to improve the initial yield by metabolic engineering and fermentation optimization in laboratory. However, the ill-defined backgrounds of these organisms and the lack of efficient genetic manipulation systems make metabolic engineering difficult to achieve or time-consuming, especially in undomesticated micro-organisms, such as myxobacteria. Here, we demonstrated that compared with the low positive ratios (< 2.5%) in different soil samples, epothilone producers constituted the major proportion (25-75%) of the Sorangium isolates in positive soils. An analysis of epothilone production and the 16S rRNA gene sequences revealed that epothilone producers could be distributed in different branches of a Sorangium phylogenetic tree, even though the positive ratio of epothilone producers in group S1 is larger than other groups (Fig. 2B). Because of the marked differences in the ability to produce epothilones and the fine differences in the biosynthesis genes (demonstrated using ATModC2), we suggest that the sympatric epothilone-producing Sorangium strains are highly diverse. Furthermore, epothilone-producing Sorangium cells were able to spread, acclimate and inhabit neighbouring locales where, if they remained viable, they formed a large reservoir of diversified epothilone producers. In billions of years of natural evolution, micro-organisms may generate many changes within a metabolic pathway, and such changes are further refined by selection. The promiscuous genetic composition of epothilone-producing Sorangium strains in positive areas provides many opportunities for the selection of promising candidates with the desired characteristics. For example, in the producers isolated from the 0087 soil sample, the yield of epothilone A changed from < 0.01 mg l-1 of So0087-2-3 to 20.43 mg l-1 of So0087-15, increasing more than 2000 times. Although the production of epothilones was affected by many factors, such as the growth rate of the cells on solid fermentation medium, the marked variation showed marked differences in the production ability of epothilones by different strains. It is thus an alternative and clearly an efficient and easy way to select strains with desirable characteristics, especially for those microorganisms that are not easily genetically manipulated in the laboratory.

Experimental procedures

Isolation of Sorangium clones

Four soil samples from which epothilone-producing strains or strains possessing the epothilone biosynthesis genes had been discovered (Dong etal., 2004; Hu etal., 2004; Li etal., 2007) were collected from different regions and environments in China (Table 1). All of the soil samples were collected from a depth of 10-15 cm under the soil surface. After collection, the soils were air-dried immediately and stored at room temperature. CNST medium (Yan etal., 2003) was used to isolate the cellulolytic myxobacterium Sorangium strains. Before autoclaving, the pH of the medium was adjusted to 7.0-7.2. Because the 0157 soil sample was alkaline, the enrichment medium was prepared at two pH values, 7.0 and 9.0, after autoclaving. Small pieces (approximately 1 by 1 cm) of sterilized filter paper were placed on the medium surface as the carbon source. Soil samples were spread over the paper and incubated at 30°C. To allow for the appearance of discrete Sorangium clones on the filter paper, the soil was ground to avoid any clots and then spread in a thin layer. Ten plates were used for the isolation of Sorangium clones from each soil sample. Growth on the enriching plates was observed under a dissecting microscope. Clones with Sorangium phenotypes, including fruiting body structures and swarms, were carefully isolated with an inoculating needle and transferred to WAT plates smeared with autoclaved E. colior VY/2 plates (Shimkets etal., 2006) for further purification using standard techniques. The purified isolates were cultured routinely on CNST medium at 30°C and morphologically and phylogenetically characterized according to previously described methods (Yan etal., 2003).

Cultivation for the production of epothilones

Because most of the newly isolated S. cellulosum clones grew poorly in liquid, the production of epothilones was measured on solid CNST medium in plates (Gong etal., 2007). To prepare the inocula for cultivation, the cells were first inoculated onto filter paper placed on CNST plates and incubated at 30°C for 4-5 days (Hou etal., 2006). The cells and the destroyed filter paper were then scraped and collected separately from each plate using an inoculation shovel. The cells were suspended and gently homogenized with glass beads (3 mm in diameter) in sterilized water, cen-trifuged (5000 r.p.m., 5 min, 4°C) and resuspended in sterile water at approximately 1 x 107 cells ml-1 (Gong etal., 2007). An aliquot of 2 ml of the cell suspensions was spread over an entire filter paper (70 mm in diameter) on CNST plate. After 4-5 days of incubation at 30°C for cellular growth, Amberlite XAD-16 resin beads (Rohm and Haas, Philadelphia, PA, USA) were spread over the colonies to absorb the epothilone products to avoid their feedback inhibition on the production (Gerth etal., 1996). The cultures were incubated for an additional 9-10 days. Epothilone production was measured in triplicate and was calculated as the yield of epothilone A divided by the volume of the total medium (mg l-1).

Detection and identification of epothilones

The resin beads that were spread over the clones were harvested, washed with distilled water, air-dried and

extracted with 5 ml of methanol. The extracts were then dried in vacuo at 40°C and stored at -20°C. For HPLC-MS analysis, samples were redissolved in 100 |l of methanol. A 10 |l aliquot was injected into a Surveyor HPLC (Thermo Finnigan, Pittsburgh, PA, USA) interfaced with a Finnigan MSQ classic quadrupole mass spectrometer (ESI-positive) (Thermo Finnigan). The analysis was performed on a Shim-pack MRC-ODS analytical reverse phase column (4.6 mm x 250 mm, 4.60 |m; Shimadzu, Tokyo, Japan) at a column temperature of 28°C with a mobile phase of 60% methanol (HPLC grade, Merck, Darmstadt, Germany) and 40% buffer (0.2% A.P. acetate acid/18 Mß Millipore water) at a flow rate of 1.0 ml min-1. The production of epothilone A, eluted at 20 min with baseline resolution, was detected at 249 nm, and the titre was quantified based on a standard curve generated using purified epothilone A (Gong etal., 2007). The MS analysis was performed under the following conditions: ESI-positive, probe temperature of 450°C, cone voltage of 75 V, full scan mass range from 300 to 900 amu and SIM scan at 494 [M + H]+ for epothilone A.

DNA extraction and polymerase chain reaction (PCR) amplification

DNA extraction and PCR amplification were performed as described previously (Li etal., 2007). The 16S rRNA gene sequences were amplified with primer pair 27F (5'-AGTTTGATCCTGGCTCAG-3') and 1492R (5'-TACCTTG TTACGACTT-3'). After constructed in the sequencing plasmid pMD 19-T Vector (Takara Biotechnology, Dalian, China), the PCR products were sequenced in both directions (Yan etal., 2003). To amplify the second AT domain in the epoC module (ATModC2) of the epothilone biosynthesis genes, a nested PCR technique was employed to avoid false priming. The first primer set was 5'-ACGTCGATTTCGTGGAATGC-3' and 5'-AGTGGACGCATGACGCTGAC-3', locating in the upstream ß-KS and downstream ß-ketoreductase domains (Fig. 1B). The product is 2.7 kb, containing theATModC2domain. The primer set for the second round was 5'-CTGCGCGA GCACCTGGACATGC-3' and 5'-GCTGCCGCTGCCACGGA TAGGT-3', targeting a 1.5 kbATModC2-containing product within the product of the first amplification round.

Phylogenetic analysis

The 16S rRNA gene sequences from different Sorangium strains and the second AT domain in the epoC module (ATModC2) from different epothilone producer strains were analysed using the Neighbor-Joining programme in MEGA version 5.05 (Biodesign Institute, AZ, USA) (Tamura etal., 2011).

Screening for epothilone producers around a positive site

To investigate the spreading and distribution abilities of epothilone-producing Sorangium strains, 14 soil specimens (Table 1) were collected from different places around the 0157 site. The sampling locales were within approximately 10 km2, and the samples were all alkaline, similar to that

of the 0157 soil. Sorangium strains were isolated, and epothilone production was detected as described above.

Accession numbers in GenBank

The nucleotide sequence data are available at GenBank under the following accession numbers: AF387628 (So0087-5), AY252114 (So0007-3), DQ256394 (So0157-2), DQ256395 (So0003-3), EU240496 (So0157-1), EU240530 (So0157-24), EU240531 (So0007-16), EU240532 (So0087-23), EU240533 (So0157-52), EU545483 to EU545510, EU864256 (So0087-7-1), HQ829366 to HQ829379, HQ829381 to HQ829410 for 16S rRNA genes, EU554397 to EU554406 and EU864255 for the second AT domain in the epoC module (ATModC2) for epothilone biosynthesis in Sorangium strains. The GenBank accession number of the S. cellulosum So0157-2 genome is CP003969.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (NSFC) for Distinguished Young Scholars (No. 30825001), the NSFC Key Program (No. 31130004) and the National High-tech R&D Program of China (863 Program) (No. 2012AA02A701).

Conflict of interest

None declared.

References

Bollag, D.M., McQueney, P.A., Zhu, J., Hensens, O., Koupal, L., Liesch, J., etal. (1995) Epothilones, a new class of microtubule-stabilizing agents with a taxol-like mechanism of action. Cancer Res 55: 2325-2333. Cheng, Y.Q., Tang, G.L., and Shen, B. (2003) Type I polyketide synthase requiring a discrete acyltransferase for polyketide biosynthesis. Proc Natl Acad Sci USA 100: 3149-3154.

Chien, M., Morozova, I., Shi, S., Sheng, H., Chen, J., Gomez, S.M., etal. (2004) The genomic sequence of the accidental pathogen Legionella pneumophila. Science 305: 19661968.

Dawid, W. (2000) Biology and global distribution of myxobacteria in soils. FEMS Microbiol Rev 24: 403-427. Dong, H., Li, Y.Z., and Hu, W. (2004) Analysis of purified tubulin in high concentration of glutamate for application in high throughput screening for microtubule-stabilizing agents. Assay Drug Dev Technol 2: 621-628. Fu, J., Wenzel, S.C., Perlova, O., Wang, J., Gross, F., Tang, Z., et al. (2008) Efficient transfer of two large secondary metabolite pathway gene clusters into heterologous hosts by transposition. Nucleic Acids Res 36: e113. Gerth, K., Bedorf, N., Hofle, G., Irschik, H., and Reichenbach, H. (1996) Epothilons A and B: antifungal and cytotoxic compounds from Sorangium cellulosum (Myxobacteria). Production, physico-chemical and biological properties. JAntibiot (Tokyo) 49: 560-563.

Gerth, K., Steinmetz, H., Höfle, G., and Reichenbach, H.

(2000) Studies on the biosynthesis of epothilones: the biosynthetic origin of the carbon skeleton. J Antibiot (Tokyo) 53: 1373-1377.

Gerth, K., Steinmetz, H., Höfle, G., and Reichenbach, H.

(2001) Studies on the biosynthesis of epothilones: the PKS and epothilone C/D monooxygenase. J Antibiot (Tokyo) 54: 144-148.

Gerth, K., Pradella, S., Perlova, O., Beyer, S., and Müller, R.

(2003) Myxobacteria: proficient producers of novel natural products with various biological activities-past and future biotechnological aspects with the focus on the genus Sorangium. J Biotechnol 106: 233-253.

Ginolhac, A., Jarrin, C., Gillet, B., Robe, P., Pujic, P., Tuphile, K., etal. (2004) Phylogenetic analysis of polyketide synthase I domains from soil metagenomic libraries allows selection of promising clones. Appl Environ Microbiol 70: 5522-5527.

Giovannoni, S.J., and Stingl, U. (2005) Molecular diversity and ecology of microbial plankton. Nature 437: 343-348.

Gong, G.L., Sun, X., Liu, X.L., Hu, W., Cao, W.R., Liu, H., etal. (2007) Mutation and a high-throughput screening method for improving the production of epothilones of Sorangium. J Ind Microbiol Biotechnol 34: 615-623.

Han, K., Li, Z.F., Peng, R., Zhu, L.P., Zhou, T., Wang, L.G., etal. (2013) Extraordinary expansion of a Sorangium cellulosum genome from an alkaline milieu. Sci Rep 3: 2101.

Hou, P.B., Li, Y.Z., Wu, B.H., Yan, Z.C., Yan, B.X., and Gao, P.J. (2006) Cellulolytic complex exists in cellulolytic myxobacterium Sorangium. Enzyme Microb Technol 38: 273-278.

Hu, W., Dong, H., Li, Y.Z., Hu, X.T., Han, G.J., and Qu, Y.B.

(2004) A high-throughput model for screening anti-tumor agents capable of promoting polymerization of tubulin in vitro. Acta Pharmacol Sin 25: 775-782.

Hyun, H., Chung, J., Kim, J., Lee, J.S., Kwon, B.M., Son, K.H., and Cho, K. (2008) Isolation of Sorangium cellulosum carrying epothilone gene clusters. J Microbiol Biotechnol 18: 1416-1422.

Jaoua, S., Neff, S., and Schupp, T. (1992) Transfer of mobilizable plasmids to Sorangium cellulosum and evidence for their integration into the chromosome. Plasmid 28: 157-165.

Jenke-Kodama, H., Sandmann, A., Müller, R., and Dittmann, E. (2005) Evolutionary implications of bacterial polyketide synthases. Mol Biol Evol 22: 2027-2039.

Jiang, D.M., Wu, Z.H., Zhao, J.Y., and Li, Y.Z. (2007) Fruiting and non-fruiting myxobacteria: a phylogenetic perspective of cultured and uncultured members of this group. Mol Phylogenet Evol 44: 545-552.

Jiang, D.M., Zhao, L., Zhang, C.Y., Li, J., Xia, Z.J., Wang, J., etal. (2008) Taxonomic analysis of Sorangium strains based on HSP60 and 16S rRNA gene sequences and morphology. Int J Syst Evol Microbiol 58: 2654-2659.

Julien, B., and Fehd, R. (2003) Development of a mariner-based transposon for use in Sorangium cellulosum. Appl Environ Microbiol 69: 6299-6301.

Julien, B., and Shah, S. (2002) Heterologous expression of epothilone biosynthetic genes in Myxococcus xanthus. Antimicrob Agents Chemother 46: 2772-2778.

Julien, B., Shah, S., Ziermann, R., Goldman, R., Katz, L., and Khosla, C. (2000) Isolation and characterization of the epothilone biosynthetic gene cluster from Sorangium cellulosum. Gene 249: 153-160.

Kopp, M., Irschik, H., Gross, F., Perlova, O., Sandmann, A., Gerth, K., and Müller, R. (2004) Critical variations of conjugational DNA transfer into secondary metabolite multiproducing Sorangium cellulosum strains So ce12 and So ce56: development of a mariner-based transposon mutagenesis system. J Biotechnol 107: 29-40.

Krug, D., Zurek, G., Revermann, O., Vos, M., Velicer, G.J., and Müller, R. (2008) Discovering the hidden secondary metabolome of Myxococcus xanthus a study of intraspecific diversity. Appl Environ Microbiol 74: 30583068.

Larkin, J.M., and Kaye, S.B. (2006) Epothilones in the treatment of cancer. Expert Opin Investig Drugs 15: 691-702.

Lau, J., Frykman, S., Regentin, R., Ou, S., Tsuruta, H., and Licari, P. (2002) Optimizing the heterologous production of epothilone D in Myxococcus xanthus. Biotechnol Bioeng 78: 280-288.

Levy, S.B., and Marshall, B. (2004) Antibacterial resistance worldwide: causes, challenges and responses. Nat Med 10: S122-S129.

Li, Y., Li, J., Zhou, L., Zhang, Y., Hu, W., and Chen, Q. (2000) Isolation and identification of myxobacteria. Acta Microbiol Sin 40: 652-656. in Chinese.

Li, Z.F., Zhao, J.Y., Xia, Z.J., Shi, J., Liu, H., Wu, Z.H., etal. (2007) Evolutionary diversity of ketoacyl synthases in cellulolytic myxobacterium Sorangium. Syst Appl Microbiol 30: 189-196.

Molnar, I., Schupp, T., Ono, M., Zirkle, R., Milnamow, M., Nowak-Thompson, B., etal. (2000) The biosynthetic gene cluster for the microtubule-stabilizing agents epothilones A and B from Sorangium cellulosum So ce90. Chem Biol 7: 97-109.

Mutka, S.C., Carney, J.R., Liu, Y., and Kennedy, J. (2006) Heterologous production of epothilone C and D in Escherichia coli. Biochemistry 45: 1321-1330.

Oßwald, C., Zipf, G., Schmidt, G., Maier, J., Bernauer, H.S., Mu 'ller, R., and Wenzel, S.C. (2012) Modular construction of a functional artificial epothilone polyketide pathway. ACS Synth Biol. doi:10.1021/sb300080t.

Park, S.R., Park, J.W., Jung, W.S., Han, A.R., Ban, Y.H., Kim, E.J., etal. (2008) Heterologous production of epothilones B and D in Streptomyces venezuelae. Appl Microbiol Biotechnol 81: 109-117.

Petkovic, H., Sandmann, A., Challis, I.R., Hecht, H.J., Silakowski, B., Low, L., etal. (2008) Substrate specificity of the acyl transferase domains of EpoC from the epothilone polyketide synthase. Org Biomol Chem 6: 500-506.

Pradella, S., Hans, A., Spröer, C., Reichenbach, H., Gerth, K., and Beyer, S. (2002) Characterisation, genome size and genetic manipulation of the myxobacterium Sorangium cellulosum So ce56. Arch Microbiol 178: 484-492.

Reichenbach, H. (2001) Myxobacteria, producers of novel bioactive substances. J Ind Microbiol Biotechnol 27: 149156.

Reichenbach, H., and Höfle, G. (2008) Discovery and development of the epothilones: a novel class of antineoplastic drugs. Drugs R D 9: 1-10.

Schneiker, S., Perlova, O., Kaiser, O., Gerth, K., Alici, A., Altmeyer, M.O., etal. (2007) Complete genome sequence of the myxobacterium Sorangium cellulosum. Nat Biotechnol 25: 1281-1289. Shimkets, L.J., Dworkin, M., and Reichenbach, H. (2006) The myxobacteria. In The Prokaryotes. Dworkin, M., Falkow, S., Rosenberg, E., Schleifer, K.H., and Stackebrandt, E. (eds). New York: Springer, pp. 31-115. Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., and Kumar, S. (2011) MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 28: 2731-2739. Tang, L., Shah, S., Chung, L., Carney, J., Katz, L., Khosla, C., and Julien, B. (2000) Cloning and heterologous expression of the epothilone gene cluster. Science 287: 640-642. Vos, M., and Velicer, G.J. (2008) Isolation by distance in the spore-forming soil bacterium Myxococcus xanthus. Curr Biol 18: 386-391. Vos, M., and Velicer, G.J. (2009) Social conflict in centimeter-and global-scale populations of the bacterium Myxococcus xanthus. CurrBiol 19: 1763-1767. Whitworth, D.E. (2008) Myxobacteria: Multicellular^ and Differentiation. Washington, DC: ASM Press. Wu, Z.H., Jiang, D.M., Li, P., and Li, Y.Z. (2005) Exploring the diversity of myxobacteria in a soil niche by myxobacteria-specific primers and probes. Environ Microbiol 7: 16021610.

Xia, Z.J., Wang, J., Hu, W., Liu, H., Gao, X.Z., Wu, Z.H., etal. (2008) Improving conjugation efficacy of Sorangium cellulosum by the addition of dual selection antibiotics. J Ind Microbiol Biotechnol 35: 1157-1163. Yadav, G., Gokhale, R.S., and Mohanty, D. (2003) Computational approach for prediction of domain organization and substrate specificity of modular polyketide synthases. J Mol Biol 328: 335-363. Yan, Z.C., Wang, B., Li, Y.Z., Gong, X., Zhang, H.Q., and Gao, P.J. (2003) Morphologies and phylogenetic classification of cellulolytic myxobacteria. Syst Appl Microbiol 26: 104-109. Zhao, J.Y., Zhong, L., Shen, M.J., Xia, Z.J., Cheng, Q.X., Sun, X., etal. (2008) Discovery of the autonomously replicating plasmid pMF1 from Myxococcus fulvus and development of a gene cloning system in Myxococcus xanthus. Appl Environ Microbiol 74: 1980-1987.

Supporting information

Additional Supporting Information may be found in the online version of this article at the publisher's web-site:

Supplementary material S1 The HPLC and MS figures of purified epothilone A standard (A) and epothilone A produced by some strains (B, C, D).

Fig. S1. Phylogenetic analysis of the protein sequences of AT domains retrieved from sequenced myxobacterial genomes. Some non-myxobacterial AT sequences were taken as reference. AT domains from epothilone biosynthesis clusters were compressed into two red triangles for easy tracking. The trees were constructed in a cycle using the MEGA 5.05 programme. Bootstrap support was based on

1000 replicates. One unit along the bar is equivalent to 10 nucleotides change per 100 bp. The outermost cycle was added after construction. Beside S. cellulosum So0157-2 and So ce56, other myxobacterial strains used in AT domains extracting are Anaeromyxobacter dehalogenans 2CP-1, Anaeromyxobacter dehalogenans 2CP-C, Anaeromyxobacter sp. Fw109, Anaeromyxobactersp. K, Haliangium ochraceum dsm14365, Myxococcus fulvus HW-1, Myxococcus xanthus DK1622, Stigmatella aurantiaca dw4/3. The two numbers linked by '. . .' showed the positions of the AT sequences in the genome.

Fig. S2. The upstream and downstream region of the '128 kb length epothilone gene cluster and its flanking region' in So 0157-2, when compared with So ce56. The biosynthesis gene cluster for epothilones and its flanking

regions were extracted from the completely sequenced genome of the So0157-2 strain (Han etal., 2013). The open reading frames (ORFs) were predicted using Glimmer 3.02 (Delcher etal., 2007) and GeneMark.hmm (Borodovsky and Lomsadze, 2011). All ORFs larger than 50 amino acid residues were blasted against the NR database of the NCBI with a cut-off of 1 E-5. The GC content was calculated by GEECEE of the EMBOSS programme (Rice etal., 2000). Fig. S3. Similarity of the epothilone biosynthesis gene clusters (about 56 kb length) from three different Sorangium strains So ce 90 (AF210843), SMP44 (AF217189) and So0157-2 (CP003969). Each domain of the three whole gene clusters was homologous with an average of 98.5% identity, indicated by colourful histograms. The similarities of the intergenic spaces were also shown by grey histograms.