The transcriptional landscape of a rewritten bacterial genome

In previous work, we synthesized and assembled the rewritten bacterial genome of Caulobacter ethensis (C. eth-2.0 thereafter). The genome of C. eth-2.0 comprises 676 protein-coding and 54 non-coding genes of freshwater alpha-proteobacterium Caulobacter crescentus (Caulobacter thereafter) compiled into a 786 kb genome design. These genes were extracted from the native sequence as biological blocks (Designed Change 1), defined chunks of DNA that contain one or several genes that belong together from a gene regulatory point of view. Here, we compared transcripts originating from the genome of Caulobacter and from the plasmid-borne, segmented genome of C. eth-2.0. As interspersed sequences were omitted, genes are present in the same order, but with different neighbouring elements (Designed Change 2). The genome has been rewritten - for one, we substituted bases to optimize the nucleotide sequence for chemical synthesis and DNA assembly. Second, we seeded 123.562 synonymous codon changes in protein coding sequences (CDS) (Designed Change 3), collectively recoding 56.1% of all codons. Taken together, the rewritten genome of C. eth-2.0 incorporates the transcriptional and translational control elements that are annotated. Any feature that has not been annotated will have been erased upon sequence rewriting. For the present study, we reasoned that the in-depth comparison of the transcriptional landscape of the rewritten and native genomes would reveal gene regulatory features and, due to the scale of the analysis, allow us to formulate principles to improve design schemes that allow to rewrite DNA molecules whilst preserving biological functionality.