The O-mannosylation and production of recombinant APA (45/47 KDa) protein from Mycobacterium tuberculosis in Streptomyces lividans is affected by culture conditions in shake flasks

Background: The Ala-Pro-rich O-glycoprotein known as the 45/47 kDa or APA antigen from Mycobacterium tuberculosis is an immunodominant adhesin restricted to mycobacterium genus and has been proposed as an alternative candidate to generate a new vaccine against tuberculosis or for diagnosis kits. In this work, the recombinant O-glycoprotein APA was produced by the non-pathogenic filamentous bacteria Streptomyces lividans, evaluating three different culture conditions. This strain is known for its ability to produce heterologous proteins in a shorter time compared to M. tuberculosis. Results: Three different shake flask geometries were used to provide different shear and oxygenation conditions; and the impact of those conditions on the morphology of S. lividans and the production of rAPA was characterized and evaluated. Small unbranched free filaments and mycelial clumps were found in baffled and coiled shake flasks, but one order of magnitude larger pellets were found in conventional shake flasks. The production of rAPA is around 3 times higher in small mycelia than in larger pellets, most probably due to difficulties in mass transfer inside pellets. Moreover, there are four putative sites of O-mannosylation in native APA, one of which is located at the carboxy-terminal region. The carbohydrate composition of this site was determined for rAPA by mass spectrometry analysis, and was found to contain different glycoforms depending on culture conditions. Up to two mannoses residues were found in cultures carried out in conventional shake flasks, and up to five mannoses residues were determined in coiled and baffled shake flasks. Conclusions: The shear and/or oxygenation parameters determine the bacterial morphology, the productivity, and the O-mannosylation of rAPA in S. lividans. As demonstrated here, culture conditions have to be carefully controlled in order to obtain recombinant O-glycosylated proteins with similar quality in bacteria, particularly, if the protein activity depends on the glycosylation pattern. Furthermore, it will be an interesting exercise to determine the effect of shear and oxygen in shake flasks, to obtain evidences that may be useful in scaling-up these processes to bioreactors. Another approach will be using lab-scale bioreactors under well-controlled conditions, and study the impact of those on rAPA productivity and quality.