Accessing biodiversity by means of shotgun cloning environmental DNA is an exciting new technology that holds promise for natural products drug discovery. To facilitate this process, we expanded on our previous technology by devising new strains and vectors that enhance the chances of detecting activities encoded by environmental DNA.
In the course of this work, we developed new tools that are useful in both this and other applications. Our gene replacement vector contains a counterselectable marker for Streptomyces
) that allows positive selection of rare genetic events that lead to loss of plasmid sequences. This is an improvement over a previous Streptomyces
gene replacement plasmid, pRHB514 (14
), because the new vector does not leave a drug resistance marker in the chromosome. This attribute is critical for many applications, including defining structure-function relationships and the production of vaccine candidates. In addition, the new vector can be used in successive rounds of gene replacement in the same strain without the need to use multiple drug resistance markers. Finally, the presence of the selection marker in the plasmid allows the excised molecule to be recovered, thus permitting the replaced allele to be isolated.
We have constructed S. lividans
and P. putida
host strains that are optimized to express environmental libraries. Our S. lividans
strains contain complete and unmarked deletions of one or both pigmented antibiotic gene clusters (act
), providing a cleaner background for heterologous expression, and no residual antibiotic resistance markers. Prior published Streptomyces
strains (S. coelicolor
CH999 [Δact::ermE redE60
]) and K4-114 and K4-115 (S. lividans
]) do not have deletions of both clusters. Our P. putida
MBD1 host strain uses the
C31 integration system to integrate BAC vectors into the chromosome. We anticipate that this system can be extended to other host strains in the future, which will augment the panel of expression hosts available.
We developed BAC vectors that can be used to construct libraries containing large DNA inserts in E. coli
and that can then be transferred by high-throughput conjugation to both S. lividans
and P. putida
MBD1. Importantly, we have shown that an environmental library generated in pMBD14 can be efficiently transferred to S. lividans
and P. putida
by conjugation, including clones containing inserts of up to 85 kb. To our knowledge, this is the first example of conjugative transfer of such high-molecular-weight plasmids from E. coli
. Sosio et al. (32
) have constructed E. coli-Streptomyces
shuttle BACs that also use
C31-mediated site-specific recombination to integrate in the Streptomyces
chromosome, and they showed that inserts up to 120 kb can be introduced and maintained in S. lividans
. Their vectors, however, do not contain the oriT
sequence and thus have to be transferred into Streptomyces
by protoplast transformation, which is not amenable to high throughput. The simplicity and high efficiency of the conjugative transfer method described here makes feasible the transfer and screening of entire large-insert DNA libraries in Streptomyces
. Environmental DNA clones can be transferred on a one-to-one basis using this process, enabling the E. coli
counterpart of any interesting Streptomyces
clone to be easily identified.
Environmental libraries offer a potentially rich source of novel and useful natural products. However, converting this intriguing idea into a realistic discovery program is a challenging endeavor. Prior data concerning the frequency of genes and gene clusters in environmental libraries of various sizes have been published by us and others (9
). For example, one E. coli
BAC library generated a hit rate for antibacterial activities of roughly 1 antibacterial clone per 60 Mb of soil-derived DNA (17
). Another library of 5,000 cosmid clones yielded 11 partial clusters with homologies to type I polyketide synthases (9
). Based on these frequencies, it is our view that in order to maximize the chances of discovery, environmental libraries need to be generated continuously and screened in a high-throughput fashion, using as many different expression hosts as practical. The strains, vectors, and technologies reported here provide an important step forward by offering practical solutions to increasing both the host range and the throughput for screening environmental libraries. The data presented here demonstrate that the three expression hosts (E. coli
, S. lividans
, and P. putida
) differ in their abilities to express gene clusters encoding chemically diverse small molecules and should, thus, facilitate the capture of increasingly numerous and diverse natural product activities, greatly increasing the chances of success for this innovative technology.