Xylose is a common monosaccharide in lignocellulosic residues that Yarrowia lipolytica cannot naturally metabolise for lipid production and therefore, heterologous xylose metabolic pathways must be engineered in this yeast to facilitate its consumption. We have compared the metabolic efficiency of two xylose metabolic pathways by developing three recombinant Y. lipolytica strains: one harbouring a xylose reductase pathway, one with a xylose isomerase pathway, and one combining both pathways, and the strains were tested for xylose consumption and lipid production at different scales. The recombinant strain with the reductase pathway that was directly isolated in selective xylose medium showed the highest lipid yield, producing up to 12.8 g/L of lipids, or 43% of the biomass dry weight, without requiring any other xylose consumption adaptive evolution process. This strain achieved a lipid yield of 0.13 g lipids/g xylose, one of the highest yields in yeast reported so far using xylose as the sole carbon and energy source. Although the strain harbouring the isomerase pathway performed better under oxygen‐limiting conditions and led to higher lipid intracellular accumulation, it showed a lower xylose uptake and biomass production, rendering a lower yield under non‐limiting oxygen conditions. Unexpectedly, the combination of both pathways in the same strain was less effective than the use of the reductase pathway alone.

The 22‐hydroxy‐23,24‐bisnorchol‐4‐ene‐3‐one (4‐HBC) is a C22 steroid synthon of pharmaceutical interest that can be produced as a lateral end‐product of the catabolism of natural sterols (e.g., cholesterol or phytosterols). This work studies the role of an aldehyde dehydrogenase coded by the MSMEG_6563 gene of Mycolicibacterium smegmatis , named msRed, in 4‐HBC production. This gene is located contiguously to the MSMEG_6561 encoding the aldolase msSal which catalyses the retroaldol elimination of acetyl‐CoA of the metabolite intermediate 22‐hydroxy‐3‐oxo‐cholest‐4‐ene‐24‐carboxyl‐CoA to deliver 3‐oxo‐4‐pregnene‐20‐carboxyl aldehyde (3‐OPA). We have demonstrated that msRed reduces 3‐OPA to 4‐HBC. Moreover, the role of msOpccR reductase encoded by MSMEG_1623 was also explored confirming that it also performs the reduction of 3‐OPA into 4‐HBC, but less efficiently than msRed. To obtain a M. smegmatis 4‐HBC producer strain we deleted MSMEG_5903 ( hsd4A ) gene in strain MS6039‐5941 ( ΔkshB1, ΔkstD1 ) that produces 4‐androstene‐3,17‐dione (AD) from natural sterols (cholesterol or phytosterols). The triple MS6039‐5941‐5903 mutant was able to produce 9 g/L of 4‐HBC from 14 g/L of phytosterols in 2 L bioreactor, showing a productivity of 0.140 g/L h ⁻¹ . To improve the metabolic flux of sterols towards the production of 4‐HBC we have cloned and overexpressed the msSal and msRed enzymes in the MS6039‐5941‐5903 mutant rendering a production titter of 12.7 g/L with a productivity of 0.185 g/L h ⁻¹ , and demonstrating that the new recombinant strain has a great potential for its industrial application.

By exploring the use of plasmids to confer Rhodococcus jostii RHA1 the possibility of utilizing xylose to produce lipids we have observed that the plasmid used was not always maintained in the transformants as expected. Instead, we observed an illegitimate integration of the antibiotic resistance gene from the plasmid into the recombinant cells. Genome sequencing of the transformants has provided evidence that this illegitimate integration is not size-, site-or sequence-specific. But even more surprising, genome sequencing revealed the presence of an unexpected circular multicopy replicative element (75–80 kb) that appears to be excised from the chromosome as a consequence of the stress generated by the antibiotic used in the selection process. The excised fragment does not contain any of the typical features of genomic islands. These results provide evidence that the genome of this oleaginous strain is more plastic than initially anticipated and our findings open the option of developing new ways to genetically modify this strain by using illegitimate recombinant approaches. But even more remarkably, the discovery of this atypical replicative element raises new questions about the existence of novel mechanisms of evolution in bacteria.

Rhodococcus jostii RHA1 is an oleaginous bacterium that has attracted considerable attention due to its capacity to use different carbon sources to accumulate significant levels of triacylglycerols that might be converted into biofuels. However, this strain cannot transform xylose into lipids reducing its potential when growing on saccharified lignocellulosic biomass. In this work, we demonstrate that wild type R. jostii RHA1 can be evolved by adaptive laboratory evolution (ALE) to metabolize xylose without engineering heterologous metabolic pathways in the host. We have generated a phenotypically adapted ALE-xyl strain able to use xylose as the sole carbon and energy source more efficiently that an engineered recombinant strain harbouring heterologous xylA and xylB genes encoding a xylose isomerase metabolic pathway. The R. jostii RHA1 ALE-xyl strain accumulates lipids very efficiently using xylose as substrate, but even more importantly it can consume glucose and xylose at the same time. Transcriptomic analyses of ALE-xyl strain growing with glucose or xylose revealed the existence of a silent pentose metabolizing operon that is overexpressed in the presence of xylose. The detection of a xylose reductase activity together with the presence of xylitol in the cytoplasm of ALE-xyl strain suggests that xylose is consumed by a reductase pathway. This study demonstrates that, in cases where a clear phenotypic selection method is available, ALE can be used to improve very efficiently industrial microbial strains without using genetic engineering tools. Strategies focused to exploit the silent phenotypic flexibility of microorganisms to metabolize different carbon sources are powerful tools for the production of microbial value-added products using saccharified lignocellulosic wastes.