Microbial inefficient substrate use through the perspective of resource allocation models

Microbial inefficient substrate use through the perspective of resource allocation models Alberte Regueira, Juan M Lema, Miguel Mauricio-Iglesias Accepted Manuscript How to cite: Current Opinion in Biotechnology, Volume 67, February 2021, Pages 130-140, https://doi.org/10.1016/j.copbio.2021.01.015 Copyright information: © 2021 Elsevier Ltd. This manuscript version is made available under the CC-BY-NC-ND 4.0 license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Microbial inefficient substrate use through the perspective of resource allocation models A. Regueira1*, J. M. Lema1, M. Mauricio-Iglesias1 1CRETUS Institute, Department of Chemical Engineering, Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain *Corresponding author (e-mail: alberte.regu[email protected]) Abstract Microorganisms extract energy from substrates following strategies that may seem suboptimal at first glance. Beyond the so-called yield-rate trade-off, resource allocation models, which focus on assigning different functional roles to the limited number of enzymes that a cell can support, offer a framework to interpret the inefficient substrate use by microorganisms. We review here relevant examples of substrate conversions where a significant part of the available energy is not utilized and how resource allocation models offer a mechanistic interpretation thereof, notably for open mixed cultures. Future developments are identified, in particular, the challenge of considering metabolic flexibility towards uncertain environmental changes instead of strict fixed optimality objectives, with the final goal of increasing the prediction capabilities of resource allocation models. Finally, we highlight the relevance of resource allocation to understand and enable a promising biorefinery platform revolving around lactate, which would increase the flexibility of wasteto-chemical biorefinery schemes. Introduction The most common interpretation of the competitive exclusion principle [1] is that microbial metabolism must tend to optimality in the use of the limiting substrates. This quest for optimality would have a different expression for pure cultures, co-cultures, open mixed-cultures, and would be subjected to spatial and temporal heterogeneity. Focusing on the prediction of intracellular fluxes for E. coli, Schuetz et al. [2] showed that the maximisation of biomass (or ATP) yield, the most common way to translate metabolic optimality, was indeed consistent with the experimental results in substrate (i.e. carbon source) limited conditions. Microorganisms are assumed to behave like efficient scavengers that extract as much energy as possible from the substrate. However, in batch cultures, temporarily provided with limitless substrate, the best predictions were given by maximising the ATP yield per flux unit, or equivalently, maximising the energy yield while minimising the enzyme use. Actually, how to express optimality becomes even more complex when several microorganisms are present and the observed experimental behaviour may seemingly depart further from the expected efficient metabolic paradigm. If an efficient metabolism is the one capable of extracting the most energy (i.e. ATP) from the substrate, a large number of experimental results [3–8] prove that efficiency is not a fixed condition for dominating in natural or engineered environments. In this review, we first briefly summarise previous explanations of this inefficient use of substrate and then resource allocation modelling is proposed as the most satisfying mechanistic framework to explain optimality under different environmental conditions. (Apparently) inefficient microbial behaviours From the experiments, we observe that microorganisms change how efficiently they use the substrate (i.e. carbon source) depending on its availability (Fig. 1). For instance, yeasts such as S. cerevisiae, which rely on fermentation for growth, yield ethanol under high substrate availability conditions, even with excess supply of oxygen, leading to a sixteenth fraction of the ATP produced under complete mineralisation [3,4], which is usually named the Crabtree effect. Some aerobic bacteria (e.g. E. coli) present a similar behaviour and excrete acetate when the substrate concentration is high, usually referred as acetate overflow [5]. Mammalian cells can also consume glucose inefficiently and convert it to lactate in presence of oxygen, which is named in this case as the Warburg effect. Particularly, this behaviour is usually shown by rapidly proliferating cells (e.g. cancer cells) or by highly active striated muscle cells [9]. These three examples have in common that glucose is only partially metabolised but differ in that only in acetate overflow oxygen consumption is still present to maintain the electron balance and regenerate NAD+ for glycolysis. Other difference is that while acetate overflow and the Crabtree effect occur in response to a change in environmental conditions (i.e. substrate availability), the Warburg effect is linked to a change in regulation: cancer cells rely on glycolysis to proliferate faster and muscle cells when contraction activity requirements are high [9]. Fig. 1. Microorganisms express different phenotypes depending on substrate availability. When it is low (e.g. a continuous reactor operated at low substrate flux), an efficient metabolism is promoted that squeezes substrate ATP production potential (as a hybrid car Open anaerobic microbiome E. coliS. cerevisiaeL. lactis Substrate availability Efficiency Performance Mammalian cells Acetate ( 7 ATP) Lactic acid (2 ATP) Ethanol (2 ATP) Acetate & butyrate (3.3 ATP) Acetate & ethanol (3 ATP) CO2 (30 ATP) + ─ Uptake Rate + ─ Lactic acid (2 ATP) Lactic acid (2 ATP) CO2 (30 ATP) CO2 (30 ATP) makes the most of a litre of fuel). On the contrary, at high substrate availabilities, microorganisms generally opt for strategies that do not optimise ATP production from the substrate but allow for a faster substrate uptake (as a sports car is designed to optimise its performance regardless of fuel consumption or efficiency). Circle areas are proportional to the ATP yield per glucose of each phenotype. In complete and partial (i.e. acetate overflow in E. coli) aerobic phenotypes, NADH and UQH2 were estimated to produce 2.5 and 1.5 ATP per molecule, respectively, and the cost of transporting NADH into the mitochondria to consume 1 ATP per molecule. In the case of anaerobic bacteria, lactic acid bacteria abandon their typical lactate production and shift towards a higher energy yielding acetate and ethanol conversion, which provides 50% more ATP per unit of substrate, at low dilution rates in a continuous reactor [6]. Anaerobic open microbiomes (i.e. mixed communities that are permeable to the entry of new strains from the surroundings), that could be considered to have a higher driving force to behave efficiently due to the fierce competition among their constituents, also may behave inefficiently (Fig. 1). It was recently shown that, in a discontinuous reactor, lactate is the main product of glucose anaerobic fermentation even though its ATP yield is the lowest of all possible products of the process [7]. These evidences have puzzled microbiologists for decades as it is striking that competitive exclusion principle selects clear inefficient behaviours that do not extract as much energy as possible from the substrate. Hypotheses for inefficient behaviours A recurrent hypothesis was that yeasts started to produce ethanol at a certain growth rate due to limitations in the cellular membrane for accommodating the electron transfer chain [10,11]. At a certain catalysed substrate flux, the maximum capacity is reached, and oxygen consumption could not be increased further. An anaerobic version of this limitation was proposed by González-Cabaleiro et al. [12], which limits the rate of electron transport in catabolic reactions. However, experiments show that oxygen consumption rates actually decrease at increasing growth rates, indicating that the respiration capacity is not fully utilised and, therefore invalidating this hypothesis [13,14]. Another competing explanation, the chemical warfare hypothesis, states that the motivation of producing ethanol or carboxylic acids is to displace other competing species as they are likely to have a lower tolerance to their toxicity [15]. However, this hypothesis is not consistent with these chemicals being produced only during substrate abundance, i.e. when substrate availability is not the limiting growth factor. The metabolic division of labour hypothesis affirms that in environments with high substrate fluxes (i.e. where substrate is highly available), substrate conversion is done in several steps (and performed by different microbial populations) rather than being completely converted by a single microbial species, as in low substrate flux conditions [16– 18]. This hypothesis is based on the theory of optimal pathway length which states, qualitatively, that longer pathways generate a higher ATP yield and that the total enzymes concentration is limited, which result in short pathways having a higher enzyme concentration for each metabolic step [19]. Therefore, shorter pathways can attain higher substrate uptake rates, but at the expense of a lower ATP yield. In this sense, it is already suggested that in microbial systems there is a trade-off between attaining a high substrate uptake flux or using the substrate efficiently (which is also named the rate vs yield trade-off [15,20–22]). Other authors consider that the apparent trade-offs between rate and yield are not necessarily an inescapable physical constraint and that are evolved cellular properties. In replicated long-term chemostat experiments under substrate limitation, it was reported that around half of E. coli strains developed spontaneously cross-feeding phenotypes [23] and Meijer et al. [24] simulated evolutionary trajectories showing that, even for substrate- limited chemostats, the emergence of metabolic labour division was an “evolutionary contingency”. The resource allocation theory The key aspect came when thinking of cells as self-replication systems needing a certain machinery (i.e. enzymes) to function, as factories need machines to produce goods, and that models should consequently consider this factor [25–27]. From this conception the theory of resource allocation emerged, which is at the present time the most convincing theoretical framework to mechanistically explain the previously mentioned inefficient substrate use. This theory states that cells are constrained by having a limited available protein (i.e. enzymes) concentration [13]. The different cellular processes, e.g. catabolism, membrane transport, anabolism, compete for a finite protein pool that should be allocated carefully to maximise fitness, which can be defined as the success of replication of organisms competing for the same resources [15]. Models including concepts from the resource allocation theory include self-fabrication models [25,28], balance-growth models taking into account proteome allocation [29] or approaches minimizing the enzyme cost for the maximum biomass production rate in comprehensive formulations of cellular metabolism [26], but the most usual modelling approaches are Flux Balance Analysis (FBA) with additional constraints related with the limits of the protein pool [13,30–37], which is therefore the main focus of this review. These models include one or more constraints imposing an upper limit on the global protein concentration or sections of the proteome (e.g. protein concentration allocated to membrane processes), which is determined using experimentally determined values of enzymatic activities and the metabolic fluxes values determined in silico by the model (BOX 1). The determined cellular fluxes (i.e. the model solution) are then constrained by the concentration of the enzymes catalysing them, including the own enzyme synthesis (i.e. the self-replicating anabolism). In this case, the models are usually referred as FBA with molecular crowding (FBAwMC) [37], as they place an upper bound on the enzyme crowdedness within cells, or Constrained Allocation FBA (CAFBA) [30], since fluxes are additionally constrained by proteome allocation. <BOX 1 should be place approximately here> Resource allocation modelling identifies a microbial trade-off between efficiency and flux Resource allocation models were used successfully to explain microbial behaviours that are not correctly captured with other metabolic modelling approaches (Table 1). The results of the models have a common thread: there exists a trade-off between efficiency and flux. Assuming that the ATP requirements to form biomass are relatively constant at different environmental conditions, to maximise the specific growth rate cells have to either maximise the specific substrate uptake rate (qS in Eq. 1) or the ATP yield on the substrate (YATP/S in Eq. 1). 𝜇 = 𝑞𝑠· 𝑌 𝐴𝑇𝑃/𝑆 · 𝑌 𝑋/𝐴𝑇𝑃 (1) where µ is the specific growth rate (h-1), qS is the specific substrate uptake rate (molS·CmolX1·h-1), YATP/S is the ATP yield on the substrate (molATP·molS-1) and YX/ATP is the biomass yield on ATP (CmolX·molATP-1). Strains relevant in the biotechnological field present a broad range values of maximum substrate uptake rate and of ATP yield on the substrate, which illustrates the high phenotypic plasticity of metabolism. Values span from 0.2 molS·CmolX-1·h-1, shown by bacterial open microbiomes yielding butyrate [7], to values up to 0.8 molS·CmolX-1·h-1, displayed by E. faecalis [38], a lactic acid bacteria, which shows the specialisation degree of this bacterial group in consuming substrate at high rates. Eukaryotic cells present microbial competition lies in the anabolism and not in a premium and low-cost catabolism. Given that most common LAB are natural to environments where peptides are available (e.g. milk or grass) [50], it could be reasonably hypothesised that losing the ability to build these compounds was positively selected by competitive selection, as already suggested in some studies: D’Souza et. al [51] showed in propagation experiments that E. coli rapidly developed an auxotrophic genotype for amino acids when supplementing amino acids in the cultivation media. Limits to resource allocation models Resource allocation models helped us explain satisfactorily some challenging microbial behaviours and predict parts of their phenotypes (i.e. their main catabolic products). However, some predictions regarding the actual proteome distribution do not match experimental observations. For example, experiments with Lactococcus lactis showing a switch from a catabolism yielding acetate-ethanol (premium catabolism) to solely production of lactate (low-cost catabolism) at increasing dilution rates in a continuous reactor, do not show strict proteome regulation [52]. Metabolic regulation is apparently done, in this case, using post-translational modifications as it keeps enzymes of both catabolic branches highly expressed at all conditions, which is in detriment of cellular performance and growth rate according to the resource allocation theory (BOX 1) and to experimental evidences. Goelzer et al. [28] compared the predicted proteome of a resource allocation model for B. subtilis metabolism with absolute protein quantification and detected the expression of some gratuitous enzymes related to the biosynthesis of some amino acids that were already supplemented in the cultivation media, and therefore not produced de novo. To test whether this overexpression was detrimental to cell performance, additional experiments were performed with mutant B.subtilis strains with these enzymes deleted and cells showed up to 18% higher growth rates. We argue that the proteome regulation proposed in resource allocation models should be interpreted as the fittest phenotype possible for a given environment resulting from optimal selection through ecological competition. Studies analysing the behaviour of an isolated pure cultures, as the mentioned experiments, cannot be representative of the outcome of natural selective pressures as not regulating the proteome is not a penalising trait that could lead to outcompetition. Moreover, another possible reasoning is that expressing gratuitous proteins is the result of microorganisms having evolved mechanisms to ensure robustness and protection in the case of sudden and unforeseen environmental variations or against fluctuations in protein production [28]. Therefore strict optimality principles should account for uncertainty when describing metabolic strategies [53,54]. Actually, a versatile metabolism, understood as having the potential to address changes in the environment was demonstrated for nine wild-type bacteria [55]. It was seen by measuring 13C fluxes that microorganisms actually grow at suboptimal rates making a compromise between tuning their fluxes for growth maximisation and minimising the needed flux changes to adapt to new conditions. Resource allocation models allowed us to understand mechanistically cellular behaviours that use the substrate in a seemingly inefficient way, to better comprehend how cells pursue optimality and were a significant step forward from previous modelling approaches. However, it is clear that there are still gaps in our way to unravel how cell optimality objectives are driven by evolution and shape microbial communities phenotypes. Application in environmental biotechnology The former paradigm of waste treatment based on substrate mineralisation (e.g. activated sludge process) is shifting towards waste-to-chemical biorefinery paradigms, as the carboxylate platform, in which waste is, in first place, anaerobically fermented to volatile fatty acids in open microbiome reactors [56,57]. The typical acids in this platform are the products of efficient conversions in low substrate flux environments (i.e. acetate or butyrate). Waste conversion can also be driven through inefficient substrate transformations to yield lactate, creating thus an alternative and promising waste-tochemical biorefinery scheme, the lactate platform, as this compound has diverse and established applications as feed preservative, in the production of cosmetics or as precursor of bioplastics [58]. With the mechanistic insight provided by resource allocation models, we have a deeper understanding of the factors that provoke the shift in microbial communities of the substrate use efficiency. In this sense, we can now engineer microbial communities by designing reactors with the appropriate environmental conditions leading to the production of chemicals produced at different degrees of substrate use efficiency. Conclusions In the past years, different explanations were proposed to reconcile experimentally observed inefficient microbial conversions with the assumed pursue of metabolic optimality. We show here that resource allocation modelling provides on most occasions the most satisfying theoretical framework for mechanistically explain what drives microorganisms to modify their substrate use efficiency at different environmental conditions in the sake of optimality. Resource allocation models identified that at low substrate flux conditions membrane transport limits growth and that an efficient use of the substrate is promoted. In environments with high substrate fluxes, the limited enzyme capacity of the cytoplasm constraints growth and inefficient partial substrates conversions with lower enzyme requirements provide a competitive advantage. The mechanistic insight provided by resource allocation models increases the flexibility of waste-to-chemicals biorefineries as we can design reactors with the appropriate environmental conditions for steering waste conversion to chemicals resulting of conversion at different degrees of efficiency. Acknowledgements The authors would like to acknowledge the support of the Spanish Ministry of Education (FPU14/05457) and project CONSERVAL (INTERREG V-A Spain-Portugal, POCTEP), co-financed by the ERDF (Ref: 2352). The authors belong to the Galician Competitive Research Group (ED431C2017/029) and to the CRETUS Strategic Partnership (ED431E 2018/01), both programmes are co-funded by Xunta de Galicia and ERDF (EU). 1. Hardin G: The competitive exclusion principle. Science 1960, 131:1292–1297. 2. Schuetz R, Kuepfer L, Sauer U: Systematic evaluation of objective functions for predicting intracellular fluxes in Escherichia coli. Mol Syst Biol 2007, 3:119. 3. Pfeiffer T, Morley A: An evolutionary perspective on the Crabtree effect. Front Mol Biosci 2014, 1:1–6. 4. Recherches I De: The Crabtree Effect : A Regulatory System in Yeast. 1966, 5. Pinhal S, Ropers D, Geiselmann J, Jongb H de: Acetate Metabolism and the Inhibition of Bacterial Growth by Acetate. 2019, 201:1–19. 6. Thomas TD, Ellwood DC, Longyear VMC: Change from homoto heterolactic fermentation by Streptococcus lactis resulting from glucose limitation in anaerobic chemostat cultures. J Bacteriol 1979, 138:109–117. 7. Rombouts JL, Kranendonk EMM, Regueira A, Weissbrodt DG, Kleerebezem R, van Loosdrecht MCM: Selecting for lactic acid producing and utilising bacteria in anaerobic enrichment cultures. Biotechnol Bioeng 2020, 117:1281–1293. 8. Vazquez A, Liu J, Zhou Y, Oltvai ZN: Catabolic efficiency of aerobic glycolysis: The Warburg effect revisited. BMC Genomics 2010, 4. 9. Vazquez A, Oltvai ZN: Molecular crowding defines a common origin for the warburg effect in proliferating cells and the lactate threshold in muscle physiology. PLoS One 2011, 6. 10. Andersen KB, Von Meyenburg K: Are growth rates of Escherichia coli in batch cultures limited by respiration? J Bacteriol 1980, 144:114–123. 11. Zhuang K, Vemuri GN, Mahadevan R: Economics of membrane occupancy and respiro-fermentation. Mol Syst Biol 2011, 7:1–9. 12. González-Cabaleiro R, Ofiţeru ID, Lema JM, Rodríguez J: Microbial catabolic activities are naturally selected by metabolic energy harvest rate. ISME J 2015, 9:2630–2641. 13. Nilsson A, Nielsen J: Metabolic Trade-offs in Yeast are Caused by F1F0-ATP synthase. Sci Rep 2016, 6:1–11. 14. Van Hoek P, Van Dijken JP, Pronk JT: Effect of specific growth rate on fermentative capacity of baker’s yeast. Appl Environ Microbiol 1998, 64:4226–4233. 15. Bachmann H, Molenaar D, Branco dos Santos F, Teusink B: Experimental evolution and the adjustment of metabolic strategies in lactic acid bacteria. FEMS Microbiol Rev 2017, 41:S201–S219. 16. Kreft JU, Griffin BM, González-Cabaleiro R: Evolutionary causes and consequences of metabolic division of labour: why anaerobes do and aerobes don’t. Curr Opin Biotechnol 2020, 62:80–87. 17. Kappler O, Janssen PH, Kreft J, Schink B: Effects of alternative methyl group acceptors on the growth energetics of the O-demethylating anaerobe Holophaga foetida. 1997, 143:1105–1114. 18. West SA, Cooper GA: Division of labour in microorganisms: An evolutionary perspective. Nat Rev Microbiol 2016, 14:716–723. 19. Costa E, Pérez J, Kreft JU: Why is metabolic labour divided in nitrification? Trends Microbiol 2006, 14:213–219. 20. Pfeiffer T, Schuster S, Bonhoeffer S: Cooperation and competition in the evolution of ATP-producing pathways. Science 2001, 293:1436. 21. Cheng C, O’brien EJ, McCloskey D, Utrilla J, Olson C, Lacroix RA, Sandberg TE, Feist AM, Palsson BO, King ZA: Laboratory evolution reveals a twodimensional rate-yield tradeoff in microbial metabolism. PLoS Comput Biol 2019, 15:1–17. 22. MacLean RC: The tragedy of the commons in microbial populations: Insights from theoretical, comparative and experimental studies. Heredity (Edinb) 2008, 100:233–239. 23. Treves DS, Manning S, Adams J: Repeated evolution of an acetate-crossfeeding polymorphism in long-term populations of Escherichia coli. Mol Biol Evol 1998, 15:789–797. 24. Meijer J, van Dijk B, Hogeweg P: Contingent evolution of alternative metabolic network topologies determines whether cross-feeding evolves. Commun Biol 2020, 3:401. 25. Molenaar D, Van Berlo R, De Ridder D, Teusink B: Shifts in growth strategies reflect tradeoffs in cellular economics. Mol Syst Biol 2009, 5:1–10. 26. Wang T, Sha R, Dreyfus R, Leunissen ME, Maass C, Pine DJ, Chaikin PM, Seeman NC: Self-replication of information-bearing nanoscale patterns. Nature 2011, 478:225–228. 27. Neidhardt FC: Bacterial Growth: Constant Obsession with dN/dt. J Bacteriol 1999, 181:7405–7408. 28. Goelzer A, Muntel J, Chubukov V, Jules M, Prestel E, Nölker R, Mariadassou M, Aymerich S, Hecker M, Noirot P, et al.: Quantitative prediction of genome-wide resource allocation in bacteria. Metab Eng 2015, 32:232–243. 29. Basan M, Hui S, Okano H, Zhang Z, Shen Y, Williamson JR, Hwa T: Overflow metabolism in Escherichia coli results from efficient proteome allocation. Nature 2015, 528:99–104. 30. Mori M, Hwa T, Martin OC, De Martino A, Marinari E: Constrained Allocation Flux Balance Analysis. PLoS Comput Biol 2016, 12:1–24. 31. van Hoek MJA, Merks RMH: Redox balance is key to explaining full vs. partial switching to low-yield metabolism. BMC Syst Biol 2012, 6. 32. Mori M, Marinari E, De Martino A: A yield-cost tradeoff governs Escherichia coli’s decision between fermentation and respiration in carbon-limited growth. npj Syst Biol Appl 2019, 5. 33. Zeng H, Yang A: Modelling overflow metabolism in Escherichia coli with flux balance analysis incorporating differential proteomic efficiencies of energy pathways. BMC Syst Biol 2019, 13:1–18. 34. Regueira A, Rombouts JL, Wahl SA, Mauricio-Iglesias M, Lema JM, Kleerebezem R: Resource allocation explains lactic acid production in mixed-culture anaerobic fermentations. Biotechnol Bioengeneering 2021, 118:745–758. 35. da Silva LG, Tomás-Martínez S, van Loosdrecht MCM, Wahl SA: The environment selects: Modeling energy allocation in microbial communities under dynamic environments. bioRxiv 2019, doi:10.1101/689174. 36. Rugen M, Bockmayr A, Steuer R: Elucidating temporal resource allocation and diurnal dynamics in phototrophic metabolism using conditional FBA. Sci Rep 2015, 5:1–16. 37. Beg QK, Vazquez A, Ernst J, De Menezes MA, Bar-Joseph Z, Barabási AL, Oltvai ZN: Intracellular crowding defines the mode and sequence of substrate uptake by Escherichia coli and constrains its metabolic activity. Proc Natl Acad Sci U S A 2007, 104:12663–12668. 38. Sauer M, Russmayer H, Grabherr R, Peterbauer CK, Marx H: The Efficient Clade: Lactic Acid Bacteria for Industrial Chemical Production. Trends Biotechnol 2017, 35:756–769. 39. Rombouts JL, Mos G, Weissbrodt DG, Kleerebezem R, van Loosdrecht MCM: Diversity and metabolism of xylose and glucose fermenting microbial communities in sequencing batch or continuous culturing. FEMS Microbiol Ecol 2019, 95:fiy233. 40. Rombouts JL, Mos G, Weissbrodt DG, Kleerebezem R, Van Loosdrecht MCM: The impact of mixtures of xylose and glucose on the microbial diversity and fermentative metabolism of sequencing-batch or continuous enrichment cultures. FEMS Microbiol Ecol 2019, 95:fiz112. 41. Kreft JU: Biofilms promote altruism. Microbiology 2004, 150:2751–2760. 42. Komori T, Shibai A, Saito H, Akeno Y, Germond A, Horinouchi T, Furusawa C, Tsuru S: Enhancement of K-strategy evolution in histidine utilization using a container with compartments. Genes to Cells 2018, 23:893–903. 43. Rafrafi Y, Trably E, Hamelin J, Latrille E, Meynial-Salles I, Benomar S, GiudiciOrticoni MT, Steyer JP: Sub-dominant bacteria as keystone species in microbial communities producing bio-hydrogen. Int J Hydrogen Energy 2013, 38:4975–4985. 44. Pfeiffer T, Bonhoeffer S: Evolution of cross-feeding in microbial populations. Am Nat 2004, 163:E126–E135. 45. Schumacher R: Metabolic trade-offs arising from increased free energy conservation in Saccharomyces cerevisiae (Doctoral Thesis). 2018, doi:10.4233/uuid:177e9f4c-f847-436d-9fd4-9ed97ba709d9. 46. Möller P, Liu X, Schuster S, Boley D: Linear programming model can explain respiration of fermentation products. PLoS One 2018, 13:1–18. 47. Basan M: Resource allocation and metabolism: the search for governing principles. Curr Opin Microbiol 2018, 45:77–83. 48. Harris LK, Theriot JA: Surface Area to Volume Ratio: A Natural Variable for Bacterial Morphogenesis. Trends Microbiol 2018, 26:815–832. 49. Dennis PP, Bremer H: Modulation of Chemical Composition and Other Parameters of the Cell at Different Exponential Growth Rates. EcoSal Plus 2008, 3. 50. Carr FJ, Chill D, Maida N: The lactic acid bacteria: A literature survey. Crit Rev Microbiol 2002, 28:281–370. 51. D’Souza G, Kost C: Experimental Evolution of Metabolic Dependency in Bacteria. PLOS Genet 2016, 12:e1006364. 52. Goel A, Eckhardt TH, Puri P, de Jong A, Branco dos Santos F, Giera M, Fusetti F, de Vos WM, Kok J, Poolman B, et al.: Protein costs do not explain evolution of metabolic strategies and regulation of ribosomal content: Does protein investment explain an anaerobic bacterial Crabtree effect? Mol Microbiol 2015, 97:77–92. 53. Tourigny DS: Dynamic metabolic resource allocation based on the maximum entropy principle. J Math Biol 2020, 80:2395–2430. 54. Ferenci T: Trade-off Mechanisms Shaping the Diversity of Bacteria. Trends Microbiol 2016, 24:209–223. 55. Schuetz R, Zamboni N, Zampieri M, Heinemann M, Sauer U: Multidimensional