Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Enzyme expression kinetics by Escherichia coli during transition from rich to minimal media depends on proteome reserves

Nature Microbiology volume 8, pages 347–359 (2023)Cite this article

Subjects

Abstract

Bacterial fitness depends on adaptability to changing environments. In rich growth medium, which is replete with amino acids, Escherichia coli primarily expresses protein synthesis machineries, which comprise ~40% of cellular proteins and are required for rapid growth. Upon transition to minimal medium, which lacks amino acids, biosynthetic enzymes are synthesized, eventually reaching ~15% of cellular proteins when growth fully resumes. We applied quantitative proteomics to analyse the timing of enzyme expression during such transitions, and established a simple positive relation between the onset time of enzyme synthesis and the fractional enzyme ‘reserve’ maintained by E. coli while growing in rich media. We devised and validated a coarse-grained kinetic model that quantitatively captures the enzyme recovery kinetics in different pathways, solely on the basis of proteomes immediately preceding the transition and well after its completion. Our model enables us to infer regulatory strategies underlying the ‘as-needed’ gene expression programme adopted by E. coli.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Steady-state protein abundances.
Fig. 2: AA downshift with various carbon supplements.
Fig. 3: Enzyme recovery kinetics for individual AAB groups.
Fig. 4: Model of dynamic proteome allocation for AA downshift.
Fig. 5: Data and model for AA downshift without serine.

Similar content being viewed by others

Data availability

The data underlying the figures are presented as Source Data. Other data are available from the corresponding authors upon request.

All mass spectrometric raw data files as well as the data analysis output files have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository with the dataset identifier PXD035278. The E. coli spectral library used for DIA/SWATH data analysis has been published previously19 and is available via SWATHAtlas: http://www.peptideatlas.org/PASS/PASS01421.

Code availability

The code used to compute protein intensities (xTop v2) is available at https://gitlab.com/mm87/xtop. The code used to model the AA downshift growth kinetics on the basis of the mathematical model presented in the main text is available at https://gitlab.com/ch47/model-of-aa-downshift.

References

  1. Yao, C. K., Muir, J. G. & Gibson, P. R. Review article: insights into colonic protein fermentation, its modulation and potential health implications. Aliment. Pharmacol. Ther. 43, 181–196 (2016).

    Article  CAS  Google Scholar 

  2. Oliphant, K. & Allen-Vercoe, E. Macronutrient metabolism by the human gut microbiome: major fermentation by-products and their impact on host health. Microbiome 7, 91 (2019).

    Article  Google Scholar 

  3. Kellerman, A. M., Dittmar, T., Kothawala, D. N. & Tranvik, L. J. Chemodiversity of dissolved organic matter in lakes driven by climate and hydrology. Nat. Commun. 5, 3804 (2014).

    Article  CAS  Google Scholar 

  4. Sniatala, B., Kurniawan, T. A., Sobotka, D., Makinia, J. & Othman, M. H. D. Macro-nutrients recovery from liquid waste as a sustainable resource for production of recovered mineral fertilizer: uncovering alternative options to sustain global food security cost-effectively. Sci. Total Environ. 856, 159283 (2023).

    Article  CAS  Google Scholar 

  5. Koch, A. L. The adaptive responses of Escherichia coli to a feast and famine existence. Adv. Microb. Physiol. 6, 147–217 (1971).

    Article  CAS  Google Scholar 

  6. Mori, M., Schink, S., Erickson, D. W., Gerland, U. & Hwa, T. Quantifying the benefit of a proteome reserve in fluctuating environments. Nat. Commun. 8, 1225 (2017).

    Article  Google Scholar 

  7. Schaechter, M., MaalØe, O. & Kjeldgaard, N. O. Dependency on medium and temperature of cell size and chemical composition during balanced growth of Salmonella typhimurium. Microbiology 19, 592–606 (1958).

    CAS  Google Scholar 

  8. Neidhardt, F. C. & Magasanik, B. Studies on the role of ribonucleic acid in the growth of bacteria. Biochim. Biophys. Acta 42, 99–116 (1960).

    Article  CAS  Google Scholar 

  9. Maaløe, O. & Goldberger, R. F. Biological regulation and development. Biol. Regul. Dev. 1, 487–542 (1979).

    Article  Google Scholar 

  10. Scott, M., Gunderson, C. W., Mateescu, E. M., Zhang, Z. & Hwa, T. Interdependence of cell growth and gene expression: origins and consequences. Science 330, 1099–1102 (2010).

    Article  CAS  Google Scholar 

  11. You, C. et al. Coordination of bacterial proteome with metabolism by cyclic AMP signalling. Nature 500, 301–306 (2013).

    Article  CAS  Google Scholar 

  12. Liebermeister, W. et al. Visual account of protein investment in cellular functions. Proc. Natl Acad. Sci. USA 111, 8488–8493 (2014).

    Article  CAS  Google Scholar 

  13. Li, Z., Nimtz, M. & Rinas, U. The metabolic potential of Escherichia coli BL21 in defined and rich medium. Microb. Cell Fact. 13, 45 (2014).

    Article  Google Scholar 

  14. Scott, M. & Hwa, T. Shaping bacterial gene expression by physiological and proteome allocation constraints. Nat. Rev. Microbiol. https://doi.org/10.1038/s41579-022-00818-6 (2022).

  15. Hui, S. et al. Quantitative proteomic analysis reveals a simple strategy of global resource allocation in bacteria. Mol. Syst. Biol. 11, 784 (2015).

    Article  Google Scholar 

  16. Peebo, K. et al. Proteome reallocation in Escherichia coli with increasing specific growth rate. Mol. Biosyst. 11, 1184–1193 (2015).

    Article  CAS  Google Scholar 

  17. Schmidt, A. et al. The quantitative and condition-dependent Escherichia coli proteome. Nat. Biotechnol. 34, 104–110 (2016).

    Article  CAS  Google Scholar 

  18. Caglar, M. U. et al. The E. coli molecular phenotype under different growth conditions. Sci. Rep. 7, 45303 (2017).

    Article  CAS  Google Scholar 

  19. Mori, M. et al. From coarse to fine: the absolute Escherichia coli proteome under diverse growth conditions. Mol. Syst. Biol. 17, e9536 (2021).

    Article  CAS  Google Scholar 

  20. Wick, L. M., Quadroni, M. & Egli, T. Short- and long-term changes in proteome composition and kinetic properties in a culture of Escherichia coli during transition from glucose-excess to glucose-limited growth conditions in continuous culture and vice versa. Environ. Microbiol. 3, 588–599 (2001).

    Article  CAS  Google Scholar 

  21. Erickson, D. W. et al. A global resource allocation strategy governs growth transition kinetics of Escherichia coli. Nature 551, 119–123 (2017).

    Article  Google Scholar 

  22. Wang, X., Xia, K., Yang, X. & Tang, C. Growth strategy of microbes on mixed carbon sources. Nat. Commun. 10, 1279 (2019).

    Article  CAS  Google Scholar 

  23. Basan, M. et al. A universal trade-off between growth and lag in fluctuating environments. Nature 584, 470–474 (2020).

    Article  CAS  Google Scholar 

  24. Molenaar, D., van Berlo, R., de Ridder, D. & Teusink, B. Shifts in growth strategies reflect tradeoffs in cellular economics. Mol. Syst. Biol. 5, 323 (2009).

    Article  Google Scholar 

  25. Weiße, A. Y., Oyarzún, D. A., Danos, V. & Swain, P. S. Mechanistic links between cellular trade-offs, gene expression, and growth. Proc. Natl Acad. Sci. USA 112, E1038–E1047 (2015).

    Article  Google Scholar 

  26. Dourado, H. & Lercher, M. J. An analytical theory of balanced cellular growth. Nat. Commun. 11, 1226 (2020).

    Article  CAS  Google Scholar 

  27. Bruggeman, F. J., Planqué, R., Molenaar, D. & Teusink, B. Searching for principles of microbial physiology. FEMS Microbiol. Rev. 44, 821–844 (2020).

    Article  CAS  Google Scholar 

  28. de Groot, D. H., Hulshof, J., Teusink, B., Bruggeman, F. J. & Planqué, R. Elementary growth modes provide a molecular description of cellular self-fabrication. PLoS Comput. Biol. 16, e1007559 (2020).

    Article  Google Scholar 

  29. Roy, A., Goberman, D. & Pugatch, R. A unifying autocatalytic network-based framework for bacterial growth laws. Proc. Natl Acad. Sci. USA 118, e2107829118 (2021).

    Article  CAS  Google Scholar 

  30. Gillet, L. C. et al. Targeted data extraction of the MS/MS spectra generated by data-independent acquisition: a new concept for consistent and accurate proteome analysis. Mol. Cell. Proteom. 11, O111.016717 (2012).

    Article  Google Scholar 

  31. Ludwig, C. et al. Data-independent acquisition-based SWATH-MS for quantitative proteomics: a tutorial. Mol. Syst. Biol. 14, e8126 (2018).

    Article  Google Scholar 

  32. Li, G.-W., Burkhardt, D., Gross, C. & Weissman, J. S. Quantifying absolute protein synthesis rates reveals principles underlying allocation of cellular resources. Cell 157, 624–635 (2014).

    Article  CAS  Google Scholar 

  33. Neidhardt, F. C., Bloch, P. L. & Smith, D. F. Culture medium for enterobacteria. J. Bacteriol. 119, 736–747 (1974).

    Article  CAS  Google Scholar 

  34. Keseler, I. M. et al. The EcoCyc database: reflecting new knowledge about Escherichia coli K-12. Nucleic Acids Res. 45, D543–D550 (2017).

    Article  CAS  Google Scholar 

  35. Karp, P. D. et al. The EcoCyc Database. EcoSal Plus https://doi.org/10.1128/ecosalplus.ESP-0006-2018 (2018).

  36. Dai, X. et al. Reduction of translating ribosomes enables Escherichia coli to maintain elongation rates during slow growth. Nat. Microbiol. 2, 16231 (2016).

    Article  Google Scholar 

  37. Maaløe, O. In Goldberger, R.F. (ed) Biological Regulation and Development Vol. 1 487–542 (Springer, 1979).

  38. Bremer, H. & Dennis, P. P. Modulation of chemical composition and other parameters of the cell at different exponential growth rates. EcoSal Plus https://doi.org/10.1128/ecosal.5.2.3 (2008).

  39. Milo, R. What is the total number of protein molecules per cell volume? A call to rethink some published values. Bioessays 35, 1050–1055 (2013).

    Article  CAS  Google Scholar 

  40. Oldewurtel, E. R., Kitahara, Y. & van Teeffelen, S. Robust surface-to-mass coupling and turgor-dependent cell width determine bacterial dry-mass density. Proc. Natl Acad. Sci. USA 118, e2021416118 (2021).

    Article  CAS  Google Scholar 

  41. Balakrishnan, R. et al. Principles of gene regulation quantitatively connect DNA to RNA and proteins in bacteria. Science 378, eabk2066 (2022).

    Article  CAS  Google Scholar 

  42. Björkeroth, J. et al. Proteome reallocation from amino acid biosynthesis to ribosomes enables yeast to grow faster in rich media. Proc. Natl Acad. Sci. USA 117, 21804–21812 (2020).

    Article  Google Scholar 

  43. Okano, H., Hermsen, R., Kochanowski, K. & Hwa, T. Regulation underlying hierarchical and simultaneous utilization of carbon substrates by flux sensors in Escherichia coli. Nat. Microbiol. 5, 206–215 (2020).

    Article  CAS  Google Scholar 

  44. Pavlov, M. Y. & Ehrenberg, M. Optimal control of gene expression for fast proteome adaptation to environmental change. Proc. Natl Acad. Sci. USA 110, 20527–20532 (2013).

    Article  CAS  Google Scholar 

  45. Smith, C. A. Physiology of the bacterial cell. A molecular approach. Biochem. Educ. 20, 124–125 (1992).

    Article  Google Scholar 

  46. Elf, J., Berg, O. G. & Ehrenberg, M. Comparison of repressor and transcriptional attenuator systems for control of amino acid biosynthetic operons. J. Mol. Biol. 313, 941–954 (2001).

    Article  CAS  Google Scholar 

  47. Wu, C. et al. Cellular perception of growth rate and the mechanistic origin of bacterial growth law. Proc. Natl Acad. Sci. USA 119, e2201585119 (2022).

    Article  CAS  Google Scholar 

  48. Hwa, T. In Gross, D. et al. (eds) Proc. 27th Solvay Conference on Physics. The Physics of Living Matter: Space, Time and Information, 86–97 (World Scientific, 2020).

  49. Potrykus, K. & Cashel, M. (p)ppGpp: still magical? Annu. Rev. Microbiol. 62, 35–51 (2008).

    Article  CAS  Google Scholar 

  50. Lemke, J. J. et al. Direct regulation of Escherichia coli ribosomal protein promoters by the transcription factors ppGpp and DksA. Proc. Natl Acad. Sci. USA 108, 5712–5717 (2011).

    Article  CAS  Google Scholar 

  51. Scott, M., Klumpp, S., Mateescu, E. M. & Hwa, T. Emergence of robust growth laws from optimal regulation of ribosome synthesis. Mol. Syst. Biol. 10, 747 (2014).

    Article  Google Scholar 

  52. Hauryliuk, V., Atkinson, G. C., Murakami, K. S., Tenson, T. & Gerdes, K. Recent functional insights into the role of (p)ppGpp in bacterial physiology. Nat. Rev. Microbiol. 13, 298–309 (2015).

    Article  CAS  Google Scholar 

  53. Sanchez-Vazquez, P., Dewey, C. N., Kitten, N., Ross, W. & Gourse, R. L. Genome-wide effects on Escherichia coli transcription from ppGpp binding to its two sites on RNA polymerase. Proc. Natl Acad. Sci. USA 116, 8310–8319 (2019).

    Article  CAS  Google Scholar 

  54. Yang, Y. et al. Relation between chemotaxis and consumption of amino acids in bacteria. Mol. Microbiol. 96, 1272–1282 (2015).

    Article  CAS  Google Scholar 

  55. Zampieri, M., Hörl, M., Hotz, F., Müller, N. F. & Sauer, U. Regulatory mechanisms underlying coordination of amino acid and glucose catabolism in Escherichia coli. Nat. Commun. 10, 3354 (2019).

    Article  Google Scholar 

  56. Gordon, W. G., Semmett, W. F., Cable, R. S. & Morris, M. Amino acid composition of α-casein and β-casein2. J. Am. Chem. Soc. 71, 3293–3297 (1949).

    Article  CAS  Google Scholar 

  57. Zaslaver, A. et al. Just-in-time transcription program in metabolic pathways. Nat. Genet. 36, 486–491 (2004).

    Article  CAS  Google Scholar 

  58. Cayley, S., Lewis, B. A., Guttman, H. J. & Record, M. T. Characterization of the cytoplasm of Escherichia coli K-12 as a function of external osmolarity: implications for protein-DNA interactions in vivo. J. Mol. Biol. 222, 281–300 (1991).

    Article  CAS  Google Scholar 

  59. Steinier, J., Termonia, Y. & Deltour, J. Smoothing and differentiation of data by simplified least square procedure. Anal. Chem. 44, 1906–1909 (1972).

    Article  CAS  Google Scholar 

  60. Doellinger, J., Schneider, A., Hoeller, M. & Lasch, P. Sample Preparation by Easy Extraction and Digestion (SPEED) - a universal, rapid, and detergent-free protocol for proteomics based on acid extraction. Mol. Cell. Proteom. 19, 209–222 (2020).

    Article  CAS  Google Scholar 

  61. Gilis, D., Massar, S., Cerf, N. J. & Rooman, M. Optimality of the genetic code with respect to protein stability and amino-acid frequencies. Genome Biol. 2, research0049.1 (2001).

    Article  Google Scholar 

  62. Klipp, E., Heinrich, R. & Holzhütter, H.-G. Prediction of temporal gene expression. Eur. J. Biochem. 269, 5406–5413 (2002).

    Article  CAS  Google Scholar 

Download references

Acknowledgements

We thank X. Dai, V. Patsalo and R. Balakrishnan for helpful discussions and technical assistance during the course of this work, and O. Diego for comments and suggestions. This research was supported by NSF Grant MCB 1818384 and NIH Grant R01GM109069 to T.H. C.L. and M.A. were supported by EPIC-XS funded by the Horizon 2020 programme of the European Union (project number 823839). R.A. acknowledges the support of the European Research Council (Proteomics4D: AdvG grant 670821 and Proteomics v3.0: AdvG-233226).

Author information

Author notes

  1. These authors contributed equally: Chenhao Wu, Matteo Mori.

Authors and Affiliations

  1. Department of Physics, U.C. San Diego, La Jolla, CA, USA

    Chenhao Wu, Matteo Mori, Hiroyuki Okano & Terence Hwa

  2. Bavarian Center for Biomolecular Mass Spectrometry, Technical University of Munich, Freising, Germany

    Miriam Abele & Christina Ludwig

  3. Chair of Proteomics and Bioanalytics, Technical University of Munich, Freising, Germany

    Miriam Abele

  4. Department of Biology, Institute of Molecular Systems Biology, ETH, Zurich, Switzerland

    Amir Banaei-Esfahani & Ruedi Aebersold

  5. Division of Biological Sciences, U.C. San Diego, La Jolla, CA, USA

    Zhongge Zhang & Terence Hwa

  6. Faculty of Science, University of Zurich, Zurich, Switzerland

    Ruedi Aebersold

Authors
  1. Chenhao Wu
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Matteo Mori
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Miriam Abele
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Amir Banaei-Esfahani
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Zhongge Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Hiroyuki Okano
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Ruedi Aebersold
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Christina Ludwig
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Terence Hwa
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

C.W. and T.H. designed the experimental and theoretical studies; C.W. performed most of the measurements other than proteomics, with contributions from H.O. and Z.Z.; M.A. and A. B.-E. performed proteomic experiments with supervision from R.A. and C.L.; C.W. and M.M. performed the quantitative modelling and theoretical analysis; M.M. developed the improved xTop2.0 algorithm and performed quantitative analysis of proteomic data. All authors contributed to the writing of the manuscript.

Corresponding authors

Correspondence to Chenhao Wu, Christina Ludwig or Terence Hwa.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Microbiology thanks Tobias Bollenbach and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Steady state protein abundance of key functional groups.

a–c, Breakdown of the steady state protein abundances of the translational machinery (Fig. 1) into (a) ribosomal proteins, (b) affiliated translational apparatus and (c) tRNA synthetases. In all cases, the protein mass fractions are plotted against the growth rate under carbon limitations (wild type: black open circles; titratable ptsG mutant: black open diamonds) and four rich conditions (green symbols as shown in the legend table). The same symbols are used in Fig. 1 and Extended Data Fig. 2. d–k, Steady state protein abundance of key metabolic enzyme groups. These panels are the same as Fig. 1c-j except with the inclusion of conditions for 18AAs with 0.4% glycerol supplement (red circles) and 0.4% glycerol as the sole carbon source (yellow squares). See Supplementary Table 2 for the details of growth conditions and Supplementary Table 3 for gene classifications. In all panels, error bars are propagated from the estimated errors for individual proteins (Methods).

Source data

Extended Data Fig. 2 Central carbon metabolism related to growth on amino acids.

a, Summary of AA degradation pathways based on EcoCyc34,35. Five AAs (ala, ser, thr, gly, trp) degrades into pyruvate and six AAs (asp, asn, glu, gln, pro, arg) degrade into TCA intermediates. Degradation pathways are not known for the remaining nine amino acids (leu, ile, val, met, cys, lys, his, phe, tyr). b, The abundances of enzymes of Central Carbon Metabolism (CCM) in steady state growth, plotted against the growth rate. The symbols are the same as those used in Extended Data Fig. 1. The gene names shown with a black background are enzymes that increased by more than 50% or by more than 2‰ of proteome in CAA or RDM without glucose supplement compared to those with glucose supplement (that is, increased in filled green symbols compared to open green symbols). Error bars indicate uncertainties from the protein quantification method (Methods). c, Enzymes and metabolites involved in CCM. Metabolites that can be produced from AA degradation are boxed in yellow. Reactions are represented by colored arrows with the genes encoding the corresponding enzymes written next to it in italic. The color of the arrows corresponds to the box color in panel (b), with gray indicating reversible reactions involved in both glycolysis and gluconeogenesis, red indicating gluconeogenesis only, blue indicating glycolysis only, orange indicating TCA cycle, pink indicating glyoxylate shunt and green indicating pyruvate fermentation to acetate. The enzymes with increased abundances in rich medium without glucose supplement are clearly clustered in gluconeogenesis (red box in panel (b), five out of six enzymes) and in TCA cycle (orange box in panel (b), seven out of eight enzymes). This suggests that in the absence of glucose supplement, the AA degradation products are funneled through CCM to supply the cell’s non-AA carbon needs, for example, energy biogenesis and the biosynthesis of lipids and cell wall components, via TCA cycle and via gluconeogenesis.

Source data

Extended Data Fig. 3 Various amino acid downshifts with supplement of carbon sources.

a, Optical density during a nutrient shift from 20AA to no AA (red triangles), supplemented with glucose, as well as a control sample (grey circles), with the same medium before and after the shift. As described in Methods, when the pre-shift culture reached \(OD_{600}\sim 0.3\), we performed the shift by filtering and washing the pre-shift cell culture and re-suspending the pellets using the post-shift medium (Fig. 1b). The post-shift culture started at \(OD_{600}\sim 0.1\). From this raw growth curve, we match the start of the post-shift \(OD_{600}\) to the end of the pre-shift \(OD_{600}\) to make the growth curves continuous. Then we normalized the growth curve so that the normalized \(OD_{600}\) at the start of a shift is always 1. This leads to the triangles shown in panel (b). b, Comparison of AA downshift with all 20 AAs (triangles) and downshift with 18AA (excluding cysteine and tyrosine due to poor solubility and stability); see Methods and Supplementary Table 4. Both shifts are supplemented by 0.2% glucose before and after the shift. Note the presence or absence of cysteine and tyrosine made little difference to the growth transition kinetics. c, AA downshift experiments of panel (b) is repeated with glucose replaced by 0.4% glycerol as the carbon supplement. The growth transition is nearly the same for both 20AA-to-none and 18AA-to-none shifts. d, The grey filled circles are the same as those in panel (b), indicating the 18AA downshift with glycerol supplement. Open circles represent the result of the same downshift in strain NQ399 (see Supplementary Table 1), which has fully expressed glycerol uptake system (GlpFK). [1mM IPTG and 1mM 3MBA was added to the medium before and after the shift for the shift by NQ399 so that glpFK expression was fully induced throughout the shift11.] The lag period did not change, indicating that the glycerol uptake system was not the bottleneck of this growth transition. e, Repeatability of 18AA shift supplemented by glycerol. Symbols of different colors represent results of independent experimental runs. The CV of lag time across these runs is within 10%.

Source data

Extended Data Fig. 4 Time evolution of macroscopic protein groups during the AA downshift.

a–h, The time dependence of the total abundance of proteins from a number of key functional groups during AA downshift with glycerol as the carbon source. i–k, The eight functional groups in panels a–h are those introduced in Fig. 1; we additionally included (i) AA transport, (j) carbon catabolism and (k) glycerol uptake, with the lattest added given the use of glycerol as the carbon supplement. Three sets of data are included: set 1 (blue up-triangles) and set 2 (grey up-triangles) are replicates for 18AA downshift; the third dataset indicated by the red down-triangles are for downshift starting from 18AA minus serine; see Supplementary Table 2. Raw protein abundance data for each replicate are shown in Supplementary Table 9. Membership of each functional group is shown in Supplementary Table 3. All protein abundances are reported in mass fraction (%-proteome), except for glycerol uptake enzymes whose absolute abundance could not be fixed because these enzymes were not expressed in the reference condition where ribosome-profiling was performed32. Data corresponding to pre-shift are shown at time 0. Error bars for all protein groups (except the glycerol uptake group) are propagated from the estimated errors for individual proteins (Methods). The horizontal gold bars indicate data for steady state growth on glycerol and no AA, which is the final post-shift stead state; the thickness of the bars indicate estimated errors. l, Instantaneous growth rate (see Methods) at various time after the shift. For clarity, only one set of data is shown for 18AA downshift, as the repeatability of the growth curve is already shown in Extended Data Fig. 3e.

Source data

Extended Data Fig. 5 Time evolution of individual AA biosynthesis enzyme abundances during the AA downshift.

a–n, The abundance time courses for individual AA biosynthesis enzymes during ‘18AA-to-none’ shift (set 1) supplemented by glycerol. The enzymes in the same pathway are plotted in the same plot. Error bars represent standard errors obtained using the xTop protein quantification method (Methods). In each panel, the lines in the legend box next to the gene name indicates the operon structure. The same line style means the genes belong to the same operon. E.g., for panel (m), cysC and cysN are in one operon, cysHIJ are in another operon, while cysD, cysE, cysK, cysM are by themselves. The genes in different groups don’t share operons except serC-aroA and aroF-tyrA. In each panel, the similar color of the recovery curves indicating similar kinetic behavior. E.g., in (i), SerA and SerC (both in red) have similar kinetics, while SerB (in blue) is different. Similar kinetic behavior in panels c-d,f,j are all attributable to the operon structure. Similar kinetic behaviors are also seen in other panels where the genes are not in the same operons. More specifically, in the met group (a), changes in the abundance of MetE dominated the entire group. In the arg group (b), most enzymes showed large increase between 15 min and 50 min after the shift, except ArgA, ArgD, ArgH. The total abundance of this group was mainly determined by the abundance of ArgG and ArgI. In the ilv group (e), several enzymes (for example ilvC, ilvN) exhibited moderate increase in the first 50 min, while the rest increased only after 50 min. Since IlvC is much more abundant than the other enzymes, the group kinetics was largely determined by this enzyme alone. In the glt group (g) the GdhA enzyme had a different trend compared to GltB and GltD. In the lys group (h), three enzymes (DapA, DapE, DapF) were approximately constant throughout the shift. The rest started increasing after 50 min, most of which showed only small increases during the shift, except for LysC, with a big jump between 50 min and 100 mins. But since LysC is not the most abundant enzyme in the lys group, it did not significantly affect the group kinetics itself. In the aro group (k), most of the enzymes stayed constant or dropped during the shift. Three enzymes clearly increased during the shift, among which AroG was the most abundant. In the phetyr group (l), decrease in TyrA could likely be attributed to the fact that tyrosine was not provided in the pre-shift medium, so that TyrA was fully expressed before the shift and was then shutoff (with abundance decreasing due to dilution) during the growth recovery phase. (n) The time courses of proteins belonging to the ‘other’ group are shown for completeness.

Source data

Extended Data Fig. 6 Recovery kinetics of AAB pathways.

a–n, Panels a–i are the same as Fig. 3c-k, but showing both replicates of 18AA downshift (blue and grey triangles). Also shown are a number of AAB groups whose total enzyme abundances changed by < 2-fold (panels j–n). Colored bands indicate the abundances in the pre-shift conditions, with the dashed lines indicating the measured value and the width indicating the standard error. In all cases, errors on protein groups are propagated from the estimated errors for individual proteins (see Methods).

Source data

Extended Data Fig. 7 Relation between onset time and pre-expression levels.

a, The total abundance of the glt group relative to the pre-shift total abundance is plotted at different times after the shift (same data as Fig. 3f. Between the measured data points, we used linear interpolation (blue dashed lines) to connect the data points. The green arrow shows the fold-change in enzyme abundance in post-shift steady-state (yellow bar) from the pre-shift level for the glt pathway. The red arrow indicates the onset time, defined as the time when the enzyme abundance exceeds the pre-shift value by 25%. b, Same as panel (a), but with a piecewise 3rd degree polynomial interpolation of the time course of the enzyme abundance. c), The onset times obtained from the linear and thirddegree polynomial interpolation are hardly different. d–g, The onset time obtained from various definition for each AAB pathway is plotted against the inverse of fold change in the total protein abundance of that pathway (triangles). Panel (e) is the same as that shown in Fig. 3l. Panels (d) and (f) show the same but with onset threshold defined as 10% and 50% from the initial value, respectively. Panel (g)is with thirddegree polynomial interpolation. Circles are model predictions of onset time. h–j, show plots of the onset time (defined as in panel a) against the pre-shift, the post-shift abundances and the increase from pre- to post- shift for individual AAB enzymes, respectively. The Pearson’s correlation coefficient between the onset time and the pre-shift abundance (or the log of abundance) is 0.56 (or 0.61). The Pearson’s correlation coefficient between the onset time and the post-shift abundance (or the log of abundance) is −0.45 (or −0.56). The Pearson’s correlation coefficient between the onset time and the increase abundance from pre- to post-shift (or the log of abundance) is −0.50 (or −0.75). Due to the definition of the onset time in (a), the enzymes with fold change less than 1.25 or undetected in the pre-shift condition do not have defined onset times, and are thus excluded in the plots. k–m, show plots of the onset time with the pre-shift, the post-shift abundance and the increase abundance from pre- to post- shift for all enzymes of each pathway, respectively.

Source data

Extended Data Fig. 8 Construction of the regulatory functions for ribosomes and total AAB enzymes.

a), shows the abundance \(\phi _{Rb}^ \ast\) of ribosomal proteins at different growth rate (\(\lambda ^ \ast\)) for cultures growing exponentially in different nutrient sources. [Same data as that shown in Extended Data Fig. 1a; errors are propagated from the errors for individual ribosomal proteins (Methods).] The data is well-captured by a linear fit (dashed line), that is, \(\phi _{Rb}^ \ast = \phi _{Rb,0} + \lambda ^ \ast /\gamma\), with \(\phi _{Rb,0} = 4.45{{{\mathrm{\% }}}}\) and inverse slope \(\gamma = 8.35\;h^{ - 1}\). In steady state, we have \(\chi _{Rb}^ \ast \equiv \phi _{Rb}^ \ast\), obtained by setting \(\frac{d}{{dt}}\phi _{Rb} = 0\) in Eq. (2). of the main text. Using the definition of σ in Eq. (4). of the main text, \(\sigma ^ \ast = \frac{{\lambda ^ \ast }}{{\phi _{Rb}^ \ast }} = \frac{{\lambda ^ \ast }}{{\phi _{Rb,0} + \lambda ^ \ast /\gamma }}\) (F8.1) we can invert Eq. (F8.1) to obtain \(\lambda ^ \ast (\sigma ^ \ast )\), and obtain the regulatory function as \(\hat \chi _{Rb}\left( \sigma \right) = \chi _{Rb}^ \ast \left( \sigma \right) = \phi _{Rb}^ \ast \left( {\lambda ^ \ast \left( \sigma \right)} \right) = \frac{{\phi _{Rb,0}}}{{1 - \sigma /\gamma }}.\) (F8.2) b, The function obtained is sketched as the gray line in panel (b) and in Fig. 4b. c, The form of the regulatory function for the total AAB enzymes, \(\hat \chi _{A,tot}(\sigma )\), is derived similarly to that of \(\hat \chi _{Rb}\left( \sigma \right)\). We start with the steady state abundance \(\phi _{A,tot}^ \ast\) of the AAB enzymes in different growth conditions: the value in the pre-shift steady-state is indicated in figure by the open circle, and that of the post-shift steady-state is indicated by the filled square. We connect these two points by a straight line (green dash line) to yield the relation \(\phi _{A,tot}^ \ast = \phi _{A,tot}^{max} - \alpha _A\lambda ^ \ast\) with \(\phi _{A,tot}^{max} = 18.1\%\) and \(\alpha _A = 0.11\;h\). Using Eq. (F8.1) to relate \(\lambda ^ \ast\) to \(\sigma ^ \ast\), we obtain \(\hat \chi _{A,tot}\left( \sigma \right) = \phi _{A,tot}^ \ast \left( {\lambda ^ \ast \left( \sigma \right)} \right) = \phi _{A,tot}^{max} - \alpha _A \sigma \hat \chi _{Rb}\left( \sigma \right)\!,\) (F8.3) where \(\hat \chi _{Rb}\left( \sigma \right)\) is given by Eq. (F8.2). The resulting function is plotted in panel (d). d, This form shows that when the translational activity \(\sigma\) is low (reflecting limitation by the shortage of AAs), AAB enzymes are expected to be up-regulated. This is consistent with our qualitative expectation that due to the global regulation of ppGpp, cells allocate more protein synthesis to AAB enzymes after AA downshift. Quantitatively, the expression for \(\hat \chi _{A,tot}\left( \sigma \right)\) arise from the hypothesized linear relation between \(\phi _{A,tot}^ \ast\) and \(\lambda ^ \ast\) shown in (a). The hypothesized linear relation shown in (c) is difficult to measure directly. Our hypothesis of the negative linear relation in (c) is based on previous characterization of protein allocation under various modes of metabolic limitations:11,15 For exponentially growing cells with growth rate limited by the influx of carbon substrate in minimal medium, it was shown that the abundance of carbon catabolic enzymes increased linearly with decreasing growth rate, a distinct phenotype referred to as the “C-line”. Similarly, under internal bottleneck in AA biosynthesis, the abundances of various AAB enzymes increased linearly with decreasing growth rate, referred to as the “A-line”. The linear response depicted in panel (c) is in fact a form of the A-line characterized in (ref. 11), a response to limitation in AA fluxes consumed by the ribosomes. Thus, the linear form described by Eq. (F8.3) is well motivated, and its two parameters can be fixed by the abundances of AAB enzymes in the two steady state conditions, before and after the shift.

Extended Data Fig. 9 Detailed analysis of the serine problem.

a–f, Our model did not quantitatively capture the growth kinetics of 18AA downshift as seen in Fig. 5a (compare blue line and blue circles) despite its success in predicting the recovery kinetics of most AAB pathways. The discrepancy between the observed and predicted growth curves likely has to do with the behavior of the serine pathway, whose recovery kinetics was also poorly predicted (Fig. 3j). a, The common recipe of AA composition for the “rich medium” we used (RDM) contained >tenfold higher serine than the natural AA composition in proteins (blue circles, with arrow pointing to serine); see (ref. 33) for recipe for RDM and (ref. 56) for the composition of the protein casein (brown diamond) from which casamino acids is derived, with the latter similar to the natural AA frequency across proteins61, plotted as the x-coordinate. b, Growth curve for ‘18AA-minus-ser’ downshift (red diamond) shows a significantly shorter lag than 18AA downshift (black circles). In comparison, if the shift starts from withholding another aa instead of ser, for example, thr, lys, his, glu, chosen because the onset time of these enzyme groups are the closest to that of the ser group (Extended Data Fig. 7), the corresponding growth curves (grey symbols) exhibited similar lag as the 18AA shift. This comparison shows that there is something special for AA downshift involving serine. In the main text, we showed that our model was able to capture accurately the shift if serine is excluded, that is, the 18AA-minus-ser downshift. c–f, Below, we provide a more detailed analysis of factors that lead to different predictions for the model with and without serine. We note that differences in model predictions with and without serine are completely dictated by the two different pre-shift states. It can be attributed to two aspects: First, fractional pre-shift reserves of the AAB groups, \(q_j\left( 0 \right)\), are somewhat different as shown in panel (c). Second, different form of the allocation function \(\chi _{A,tot}\) as shown in panel (e), constructed based on the relationship between total AAB enzyme abundance and growth rates panel (d). To analyze the contribution of each aspect, we mixed and matched the model input as shown in the table in panel (f) and calculated the lag time in each case. In each case, the height of the bar representing the average of n=1000 model runs obtained by randomly generating the input data (growth rates and protein abundances, assumed normally distributed) using the uncertainties reported in Supplementary Table 6; error bars represent the standard deviation of the simulated data. The lag difference predicted for the shift with and without serine (combinations A and D) is well beyond the uncertainty range. Although \(\chi _{A,tot}\) is constructed from only two data points, its form is reliable in predicting the lag time of different shift. Furthermore, since the lag duration of the combinations B and C are in the middle of the lag time for A and D, we conclude that the difference in model predictions for the two shifts received similar contributions from \(\chi _{A,tot}\) and from fractional pre-shift reserves \(q_j(0)\).

Source data

Extended Data Fig. 10 Applicability of the “just-in-time” program to enzyme recovery kinetics.

a, The ‘just-in-time’ program of enzyme recovery kinetics proposed by Zaslaver et al.57 states that during AA downshift, the response time of an enzyme in a linear pathway is shorter if it is located towards the beginning of the pathway (as indicated by the blue symbols), while the maximum abundance is higher towards the beginning of the pathway (as indicated by the red symbols). b, Examples of response time calculated based on the proteomic measurements of SerA, SerB, SerC (Extended Data Fig. 6i). Between the measured data points, we used linear interpolation (gray dash lines) to connect the data points. The x-coordinates of the solid dots represent the corresponding response times, defined as the time when the enzyme abundance reaches 50% of the maximum during the shift according to Zaslaver et al.57. For enzyme abundance exceeding 50% at the beginning of the shift (as in the case of SerB), we took the response time to be zero. c–f, The observed response times and maximum enzyme abundances for four linear segments of the AAB pathways are plotted in the order of enzyme locations for met in panel (c), ser in panel (d), arg in panel (e), lys in panel (f). Neither the response time nor the maximum abundance exhibited the predicted trends shown in (a) for any of the four pathways. While some of the differences may arise from the fact that the study of Zaslaver et al. limited a single amino acid while our study limited 18 amino acids, we note that the derivation of the ‘just-in-time’ program done in Zaslaver et al. assumed that the synthesis of an enzyme in a pathway depended only on the pool of its substrates. This assumption ignored the crucial condition that all amino acids, including the one being limited, are needed in order to synthesize the enzymes. Analysis based on this assumption may well be applicable to the generic onset of a metabolic pathway where it was first proposed62, but hard to justify for the onset of an AAB pathways where the end-product of the pathway is a limiting factor of AAB enzyme synthesis.

Source data

Supplementary information

Supplementary Information

Supplementary Notes 1–4, Tables 1–8, Table 9 legend and References.

Reporting Summary

Supplementary Table 9

Proteomic dataset for this study. Absolute protein mass fractions (ϕi, mass of the ith protein over the total proteome mass) of E. coli measured under the conditions listed in Supplementary Table 2, as well as their relative errors.

Source data

Source Data Fig. 1

Statistical source data.

Source Data Fig. 2

Statistical source data.

Source Data Fig. 3

Statistical source data.

Source Data Fig. 4

Statistical source data.

Source Data Fig. 5

Statistical source data.

Source Data Extended Data Fig. 1

Statistical source data.

Source Data Extended Data Fig. 2

Statistical source data.

Source Data Extended Data Fig. 3

Statistical source data.

Source Data Extended Data Fig. 4

Statistical source data.

Source Data Extended Data Fig. 5

Statistical source data.

Source Data Extended Data Fig. 6

Statistical source data.

Source Data Extended Data Fig. 7

Statistical source data.

Source Data Extended Data Fig. 9

Statistical source data.

Source Data Extended Data Fig. 10

Statistical source data.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wu, C., Mori, M., Abele, M. et al. Enzyme expression kinetics by Escherichia coli during transition from rich to minimal media depends on proteome reserves. Nat Microbiol 8, 347–359 (2023). https://doi.org/10.1038/s41564-022-01310-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41564-022-01310-w

This article is cited by

Search

Advanced search

Quick links

Nature Briefing Microbiology

Sign up for the Nature Briefing: Microbiology newsletter — what matters in microbiology research, free to your inbox weekly.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing: Microbiology