Jacob Jacobsen:
Genetic engineering of cyanobacteria

Date: 30-11-2012    Supervisor: Niels-Ulrik Frigaard

Microbial metabolism is the results of 3 billion years of chemical experiments. Microbes therefore offer a tremendous diversity in chemical transformations, which can – and should – be exploited by us. Therefore we need to develop an understanding of microbial metabolism and how to manipulate it, including genetic tools that allow metabolic engineering.

The cyanobacterial phylum represents a diverse group of aerobic photosynthetic bacteria that are widespread in nature. Cyanobacteria shaped our atmosphere by oxygen evolution through the splitting of water using energy from sunlight. The sole carbon source for growth in autotrophic cyanobacteria is CO2, which is reduced to carbohydrates during photosynthesis. Simple input requirements, fast growth rates and tolerance of adverse environmental conditions make cyanobacteria attractive candidates for large scale production of energy or value added metabolites without competing with traditional food crops for arable land.

The marine cyanobacterium Synechococcus sp. PCC 7002 is tolerant of a wide array of environmental conditions and is one of the fastest growing photoautotrophic microorganisms. The available genome sequence of this cyanobacterium and its natural ability to take up and stably integrate heterologous DNA make Synechococcus sp. PCC 7002 a good candidate for metabolic engineering.

For targeted gene inactivation, a suite of vectors were made by adaptation of a system previously used in plants and fungi. The vectors include a cassette that allows ligase independent directional cloning of two PCR products in a single step employing the uracilspecific excision reagent (USER) cloning principle. PCR amplified regions of a gene to be inactivated are inserted on either side of an antibiotic resistance cassette marker for positive selection of mutant strains.Four antibiotic resistance markers were found to be useful (spectinomycin, gentamicin, chloramphenicol and kanamycin). They were used for generation of both single-locus or double-loci mutants were of Synechococcus sp. PCC 7002.

The USER cloning principle was also used in the design of the expression system. The system comprised a cassette for stable expression of multiple genes as a single operon. The cassette included a strong constitutive promoter, a USER cloning cassette for cloning of a single PCR product, a transcription terminator and restriction sites for promoter exchange and vector assembly. The system was designed to be used with specific PCR primers carrying overhangs such that the gene cloned by PCR would include an upstream ribosome binding site and a downstream USER cloning site. Thus the native USER cloning site consumed by the cloning of a gene would be replaced by that included in the PCR product. This in turn allowed for sequential assembly of the operon prior to targeted insertion into the cyanobacterial chromosome using the above described vectors for targeted gene replacement.

Genes involved in glycogen metabolism in Synechococcus sp. PCC 7002 were identified by homology search. Two putative glycogen phosphorylases (A0481/glgP; A2139/agpA) and two putative glycogen synthases (A1532/glgA1; A2125/glgA2) were targeted for inactivation. Mutant cyanobacterial strains lacking either both glycogen synthase genes or both glycogen phosphorylase genes were produced and characterized for growth phenotype and glycogen content. While no difference in growth rate or glycogen content was detected between the phosphorylase double mutant and wild type strain, we found that both glycogen phophyrylases must be genetically inactivated to eliminate glycogen phosphorylase activity in the cells. While the growth rate of the glycogen synthase double mutant strain was identical to that of wild type Synechococcus during the exponential growth phase, hardly any glycogen accumulated in the mutant strain during this growth phase. During stationary growth phase the glycogen synthase double mutant grew slower than the wild type. Surprisingly glycogen accumulation was seen in the mutant strain during stationary growth phase although significantly less so than in the wild type strain.

The valuable sugar alcohol mannitol was chosen for heterologous production in Synechococcus sp. PCC 7002. For that purpose we identified an irreversible two-step biosynthetic pathway from fructose-6-phosphate to mannitol. The pathway, comprised by mannitol-1-phosphate dehydrogenase (mtlD) from Escherichia coli and codon-optimized mannitol-1-phosphatase (mlp) from the protozoan chicken parasite Eimeria tenella, was assembled in our novel USER-cloning based expression vector. The construct was inserted in the cyanobacterial chromosome and expressed resulting in accumulation of mannitol in the cells and in the culture medium. Insertion of the biosynthetic pathway in the glycogen synthase double mutant increased the yield of mannitol, presumably by redirecting fixed carbon to mannitol under conditions where glycogen normally accumulates in the wild type strain.

Comparison of the mannitol yield from different cyanobacterial strains carrying the mannitol cassette highlighted the relevance of the site of insertion into the cyanobacterial chromosome as well as the genotype of the parental strain. This was exemplified by the fact that production of mannitol from the parental strain, deficient in glycogen biosynthesis, was 10 times higher than from wild type parental background. The combined intra and extracellular mannitol yield from the glycogen deficient parental background accounted for approximately 30% of the dry weight of the cells.