Substrate dependent heterotrophic CO2-fixation as indicator for metabolic phenotypes

Abstract

Virtually all heterotrophic organisms incorporate carbon dioxide (CO2) into their biomass via anaplerosis. Despite the fact that the activity of anaplerotic enzymes, such as pyruvate carboxylase, depends on the utilised organic substrate(s), this relation has hardly been explored so far.

To study whether CO2-incorporation can reveal which substrates out of a pool of dissolved organic carbon (DOC) are utilised by environmental microorganisms, the model organism Bacillus subtilis W23 (B. subtilis W23) was grown in a minimal medium with different types of organic substrates: glucose, lactate, or malate, respectively, each together with 1 g/L NaH13CO3. Incorporation of H13CO3- was traced by elemental analysis-isotope ratio mass spectrometry (EA-IRMS) of bulk biomass and gas chromatography-mass spectrometry (GC-MS) of protein-derived amino acids after derivatisation. Until the late logarithmic phase, 13C-incorporation into the tricarboxylic acid (TCA) cycle increased over time and occurred via [4-13C]oxaloacetate formed by carboxylation of pyruvate. Levels of 13C-incorporation were highest for growth on glucose and lowest on malate. 13C-Incorporation into gluconeogenesis products was mainly detected in the lactate and malate experiment, whereas glucose down-regulated this path. Ratios of 13C-excess calculated from the 13C-excess values of the M+1 isotopomers of specific sets of amino acids served as a diagnostic tool to identify (i) substrates that initiate active anaplerosis and (ii) substrates that require active gluconeogenesis at high statistical significance. During growth of B. subtilis W23 on glucose or lactate the ratios of 13C-excess in anaplerosis-relevant amino acids vs. “baseline” amino acids (i.e. Asp/Val, Asp/Ala, Glu/Val and Glu/Ala) yielded values above 20 displaying an active anaplerosis. In contrast, values below 10 were obtained for the same sets of amino acids, when B. subtilis W23 grew on malate. To identify active gluconeogenesis, the 13C-excess in a second set of amino acids was considered: gluconeogenesis-derived vs. “baseline” amino acids, i.e. Tyr/Val, Tyr/Ala, Phe/Val and Phe/Ala. When growing on lactate, values clearly above 1 evidenced the presence of active gluconeogenesis, whereas growth on glucose resulted in values below 1. A proof-of-principle study with a natural groundwater community confirmed that incorporation of H13CO3- by natural communities could be traced and led to specific labelling patterns in the amino acids. Ratios of 13C-excess showed on the one hand no need for active anaplerosis (Asp/Val,… < 10) and on the other hand an active gluconeogenesis (Tyr/Val,… > 1). Remarkably, these ratios and labelling patterns exhibited a striking similarity to those ratios and patterns obtained from growth experiments with B. subtilis W23 and malate (Asp/Val,… < 10 and Tyr/Val,… > 1.5) as carbon source. This similarity suggests that groundwater microbes mainly fed on humic substances (i.e. a mixture of many molecules, with an aromatic centre and phenolic and carboxylic substituents) that are decomposed into short organic acids, such as succinate, entering the central carbon metabolism at the stage of the TCA cycle. This exemplifies that our approach may elucidate the type of the main organic carbon substrate metabolised by the majority of the heterotrophic bacterial community in an environmental sample. We explored whether this simple approach – using heterotrophic fixation of 13CO2/H13CO3- under in vivo conditions – could also answer questions concerning metabolic deficiencies and bacterial physiology. To investigate this capability, the metabolism of leucine was addressed, because this amino acid is an unfavourable substrate for B. subtilis W23. Again, 13C-incorporation of H13CO3- was traced by EA-IRMS of bulk biomass and GC-MS of protein- and cell wall-derived amino acids. Remarkably, no 13C-incorporaton into gluconeogenetic products was detected when leucine was offered as growth substrate. The amino acids’ 13C-labelling patterns were very similar to the patterns obtained from our experiments with B. subtilis W23 growing on glucose. The ratios (Asp/Val,… > 20 and Tyr/Val,… < 1) calculated from the 13C-excess values of the M+1 isotopomers of our chosen indicator amino acids proved this observation. This implies that, rather than leucine, the bacteria must have used organic matter leftovers from the inoculum, which mainly consisted of carbohydrates. Leucine metabolism presumably stopped at the level of 3-methylbutanoyl-CoA, if metabolised at all. Further, we tested whether our approach could be used to study the effect of carbon catabolite repression: we were able to confirm the strict repression of other carbon sources by malate in a co-substrate experiment conducted with malate and leucine.

Hence, the combined results from controlled experiments with model organisms/ model substrates, a proof-of-principle study with a natural groundwater community and a physiological case study on metabolic bacterial deficiency underline the potential of the labelling approach to (i) characterise carbon sources of heterotrophic microorganisms in their natural environments, (ii) elucidate bottlenecks in metabolism of heterotrophic organisms and (iii) study co-substrate metabolism with regard to carbon catabolite repression.