Search

Corynebacterium glutamicum serves as important production host for small molecular compounds that are derived from precursor molecules of the central carbon metabolism. It is therefore a well-studied model organism of industrial biotechnology. However, a deeper understanding of the regulatory principles underlying the synthesis of central metabolic enzymes under different environmental conditions as well as its impact on cell growth is still missing. We studied enzyme abundances in C. glutamicum in response to growth on: (i) one limiting carbon source by sampling chemostat and fed-batch cultivations and (ii) changing carbon sources provided in excess by sampling batch cultivations. The targeted quantification of 20 central metabolic enzymes by isotope dilution mass spectrometry revealed that cells maintain stable enzyme concentrations when grown on d-glucose as single carbon and energy source and, most importantly, independent of its availability. By contrast, switching from d-glucose to d-fructose, d-mannose, d-arabitol, acetate, l-lactate or l-glutamate results in highly specific enzyme regulation patterns that can partly be explained by the activity of known transcriptional regulators. Based on these experimental results we propose a simple framework for modeling cell population growth as a nested function of nutrient supply and intracellular enzyme abundances. In summary, our study extends the basis for the formulation of predictive mechanistic models of bacterial growth, applicable in industrial bioprocess development.

Wild-type Corynebacterium glutamicum has no endogenous metabolic activity for utilizing the lignocellulosic pentose d-xylose for cell growth. Therefore, two different engineering approaches have been pursued resulting in platform strains harbouring a functional version of either the Isomerase (ISO) or the Weimberg (WMB) pathway for d-xylose assimilation. In a previous study we found for C. glutamicum WMB by-product formation of xylitol during growth on d-xylose and speculated that the observed lower growth rates are due to the growth inhibiting effect of this compound. Based on a detailed phenotyping of the ISO, WMB and the wild-type strain of C. glutamicum, we here show that this organism has a natural capability to synthesize xylitol from d-xylose under aerobic cultivation conditions. We furthermore observed the intracellular accumulation of xylitol-5-phosphate as a result of the intracellular phosphorylation of xylitol, which was particularly pronounced in the C. glutamicum ISO strain. Interestingly, low amounts of supplemented xylitol strongly inhibit growth of this strain on d-xylose, d-glucose and d-arabitol. These findings demonstrate that xylitol is a suitable substrate of the endogenous xylulokinase (XK, encoded by xylB) and its overexpression in the ISO strain leads to a significant phosphorylation of xylitol in C. glutamicum. Therefore, in order to circumvent cytotoxicity by xylitol-5-phosphate, the WMB pathway represents an interesting alternative route for engineering C. glutamicum towards efficient d-xylose utilization.

Biomass-derived d -xylose represents an economically interesting substrate for the sustainable microbial production of value-added compounds. The industrially important platform organism Corynebacterium glutamicum has already been engineered to grow on this pentose as sole carbon and energy source. However, all currently described C. glutamicum strains utilize d -xylose via the commonly known isomerase pathway that leads to a significant carbon loss in the form of CO 2 , in particular, when aiming for the synthesis of α-ketoglutarate and its derivatives (e.g. l -glutamate). Driven by the motivation to engineer a more carbon-efficient C. glutamicum strain, we functionally integrated the Weimberg pathway from Caulobacter crescentus in C. glutamicum . This five-step pathway, encoded by the xylXABCD -operon, enabled a recombinant C. glutamicum strain to utilize d -xylose in d -xylose/ d -glucose mixtures. Interestingly, this strain exhibited a tri-phasic growth behavior and transiently accumulated d -xylonate during d -xylose utilization in the second growth phase. However, this intermediate of the implemented oxidative pathway was re-consumed in the third growth phase leading to more biomass formation. Furthermore, C. glutamicum pEKEx3- xylXABCD Cc was also able to grow on d -xylose as sole carbon and energy source with a maximum growth rate of μ max = 0.07 ± 0.01 h −1 . These results render C. glutamicum pEKEx3- xylXABCD Cc a promising starting point for the engineering of efficient production strains, exhibiting only minimal carbon loss on d -xylose containing substrates.

Besides the well-established (13)C-metabolic flux analysis ((13)C-MFA) which characterizes a cell's fluxome in a metabolic and isotopic stationary state a current area of research is isotopically non-stationary MFA. Non-stationary (13)C-MFA uses short-time isotopic transient data instead of long-time isotopic equilibrium data and thus is capable to resolve fluxes within much shorter labeling experiments. However, a comparison of both methods with data from one single experiment has not been made so far. In order to create a consistent database for directly comparing both methods a (13)C-labeling experiment in a fed-batch cultivation with a Corynebacterium glutamicum lysine producer was carried out. During the experiment the substrate glucose was switched from unlabeled to a specifically labeled glucose mixture which was immediately traced by fast sampling and metabolite quenching. The time course of labeling enrichments in intracellular metabolites until isotopic stationarity was monitored by LC-MS/MS. The resulting dataset was evaluated using the classical as well as the isotopic non-stationary MFA approach. The results show that not only the obtained relative data, i.e. intracellular flux distributions, but also the more informative quantitative fluxome data significantly depend on the combination of the measurements and the underlying modeling approach used for data integration. Taking further criteria on the experimental and computational part into consideration, the current limitations of both methods are demonstrated and possible pitfalls are concluded.