BAC Sudoku Sequencing Strategy for in silico Screening of Large-Insert Soil Metagenomic Libraries

Soil microorganisms express diverse bioactive natural products; however, the majority of soil microbes are recalcitrant to cultivation. We are using a metagenomic approach to bypass cultivation and directly capture the DNA from diverse microbial genomes in natural environments such as soils. A metagenomic library from an agricultural soil (Cullars Rotation, Auburn, AL) was constructed in a broad host-range BAC vector that contained 19,200 clones with an average insert size of 110kb. Screening was accomplished using strategy termed BAC Sudoku sequencing, wherein a pooling strategy is used to multiplex the results, while still providing the ID of individual clones. BAC clones were sequenced in pools (row, column, and plate) using indexed primers and paired end reads on an Illumina HiSeq. Contigs were assembled for each pool and screened for secondary metabolite gene clusters using antiSMASH, resulting in identification of  >1000 novel PKS/NRPS pathway-containing clones. The cloned pathways are very divergent from known pathways, with the GC content varying from 41 to 76% and the amino acid identity of the KS domains ranging from 32 to 83% to the best matching BLAST hit. Expression of these PKS pathway-containing clones in E. coli strain BTRA engineered for polyketide expression has resulted in multiple clones with evidence for heterologous expression of a cloned PKS pathway. These results indicate a high degree of unique sequence space has been recovered from large-insert metagenomic clones and a subset of these clones are capable of being heterologously expressed to produce secondary metabolites, thereby expanding our available resources for natural product discovery.