The Four E. coli Growth Phases
The four E. coli growth phases — lag, log (exponential), stationary, and death — appear in every batch culture of the organism and structure every recombinant protein process built on it. Each phase is a different metabolic programme, activated by a different set of regulators, and demanding a different bioprocess action from the engineer.
In this guide we walk through each E. coli growth phase with the biology that defines it, the typical OD600 and doubling-time numbers you should expect at the bench and in the bioreactor, and the decision each phase drives: when to inoculate the production vessel, when to induce, when to harvest, and when to stop.
OD600 & Induction Timing Helper
Paste 2 or more live OD600 readings to compute the current μ and predict when the culture will hit your induction target.
Lag Phase — Adaptation Before Division
The E. coli lag phase is the interval after inoculation during which cells are metabolically active but not yet dividing. A healthy, well-matched inoculum produces a lag of 10 minutes to 1 hour in LB; poorly matched media, cold shock, cryo-recovery, or heavy starvation stretch the lag to several hours. Work by Rolfe et al. (2012) showed that the lag phase is a distinct, dynamic programme in which Salmonella and E. coli transiently accumulate intracellular manganese and iron before committing to division — not simply an idle pause.
What actually happens in lag phase:
- Ribosome biogenesis ramps up from the starvation-capped level to the exponential-phase target (30–50 % of cell mass in rich medium).
- Transient metal accumulation (Mn, Fe, Zn) primes metalloenzymes involved in DNA synthesis, oxidative defence, and translation.
- Outer-membrane porins and transporters for the new carbon source are synthesised.
- RpoS levels fall sharply as the general stress response is switched off.
Lag phase — at a glance
: 10 min – 4 h · : essentially flat · : ~0
Bioprocess decision: wait — do not induce, do not feed, do not scale. Investigate any lag > 2 h that was not expected.
Long lag phases on a repeated culture are almost always a red flag. Common causes include: inoculum taken from stationary phase (RpoS locked on), freeze-thaw damage from cryo, mismatched osmolality or salt between pre-culture and production medium, antibiotic carry-over, and trace-metal limitation. If lag is extending campaign to campaign, see the cell growth monitoring guide for diagnostic approaches.
Log (Exponential) Phase — Maximum Growth Rate
The E. coli log phase (also called exponential phase) is the period of balanced, maximum-rate growth, where the specific growth rate (μ) is constant and equal to the medium-specific μmax. In LB at 37 °C, MG1655 K-12 reaches μmax ≈ 1.4 h−1, doubling every ~30 minutes; in M9 glucose, μmax falls to 0.6–0.8 h−1 (td ~60 min).
Log-phase E. coli is a translation-dominated cell. Roughly 90 % of transcription is stable RNA (rRNA + tRNA), ribosomes occupy up to 40 % of cell dry mass, and bulk protein synthesis rates peak. This is why most recombinant protein production is run through the log phase — the host machinery for making protein is at its most capable. See the specific growth rate formula guide for the math behind μ and doubling time.
Exponential E. coli obeys the growth equation:
X(t) = X0 · eμt td = ln(2) / μ
Worked example — predicting OD at induction time
You inoculate an LB flask at OD600 = 0.05 at 37 °C. After 1 h the OD is 0.12; after 2 h it is 0.25. You want to induce at OD 0.6 with IPTG.
Step 1. Fit μ from the two post-lag timepoints using linear regression of ln(OD):
μ = (ln 0.25 − ln 0.12) / (2 − 1) = 0.734 h−1 → td = 0.94 h (57 min)
Step 2. Project time from OD 0.25 to target OD 0.6:
Δt = ln(0.6 / 0.25) / 0.734 = 1.19 h (71 min)
So the induction point is t = 2 + 1.19 = 3.19 h from inoculation, or about 71 min from the last OD reading. In practice, the culture starts to decelerate above OD ~0.5–0.7 in flasks due to oxygen limitation, so sample at 40–50 min to catch the target.
Log phase — at a glance
: 3 – 8 h · : 0.05 → 2–5 in flasks, higher in bioreactors · : 0.6 – 1.8 h−1
Bioprocess decision: induce at mid-log (OD 0.4–0.8) for standard IPTG, start fed-batch feed at end-of-batch, control μ below μmax to avoid acetate overflow.
The cost of running E. coli too hard in log phase is acetate overflow. When glucose uptake rate exceeds the respiratory capacity of the TCA cycle, excess acetyl-CoA is diverted to acetate via the Pta–AckA pathway. Above ~0.2 g/L, acetate inhibits both growth and recombinant protein yield. This is the reason fed-batch processes cap μ well below μmax — typically μset = 0.1–0.3 h−1 for high-cell-density fermentation. See the E. coli acetate overflow guide for the full mechanism.
Stationary Phase — The RpoS Switch
E. coli stationary phase begins when a nutrient is exhausted (typically glucose, phosphate, or nitrogen) or an inhibitor (acetate, low pH) accumulates enough to stop balanced growth. In LB it is reached at OD600 3–10 in flasks (higher in aerated bioreactors). Net cell number stops rising; production of some stable proteins continues for many hours.
The molecular trigger is the alternative sigma factor RpoS (σS). During log phase RpoS is kept low by rapid ClpXP-mediated proteolysis; starvation signals (ppGpp accumulation, membrane energization shifts) stabilise RpoS and allow it to compete with σ70 for RNA polymerase. RpoS then drives expression of ~500 genes collectively known as the general stress response: catalases (katE, katG), superoxide dismutases, trehalose biosynthesis (otsAB), DNA-binding protection protein (Dps), osmotic-shock resistance, and biofilm genes. Battesti et al. (2011) remain the definitive review on RpoS regulation.
Stationary-phase E. coli is not a dormant cell:
- Bulk protein synthesis drops ~80 % compared with exponential phase, but targeted translation of RpoS-regulated transcripts continues robustly.
- Cells shrink (volume drops by ~30–50 %) and adopt a rounded morphology; nucleoid is condensed by Dps.
- Outer membrane thickens; multidrug efflux pumps upregulate.
- Gefen et al. (2014) showed that individual stationary-phase E. coli cells can maintain constant protein-production activity for tens of hours, long past bulk growth cessation.
Stationary phase — at a glance
: 2 h – 2 days · : 3–12 (flask), 30–200 (bioreactor) · : → 0
Bioprocess decision: harvest soluble recombinant proteins before deep stationary (proteolysis rises). For secondary-metabolite or stress-activated promoter systems, stationary phase is the production window.
Death Phase — Viability Collapse
The E. coli death phase is the decline in viable (colony-forming) cells after extended stationary phase. It can begin as early as 12–24 h in nutrient-limited LB cultures and takes days in minimal medium. The decline is roughly exponential on log-CFU axes, mirroring the exponential rise of log phase but negative. Importantly, the total cell count measured by OD600 often remains elevated — dead cells still scatter light — which is why OD-only monitoring misses the onset of death phase.
Mechanistically, death results from a combination of protease release, ATP depletion, intracellular oxidative damage accumulated during stationary phase, and in some strains programmed cell death mediated by toxin–antitoxin systems (mazEF, relBE). Pletnev et al. (2015) review the "survival guide" for E. coli in stationary phase — cells that survive longest are those that switched on RpoS early and shut down anabolic pathways efficiently.
Death phase — at a glance
: 12 h – 3 days post-inoculation · : plateau or slowly falling (light-scattering dead cells) · : declining exponentially
Bioprocess decision: do not hold product in-reactor here. Protease release, endotoxin spike, and inclusion-body quality loss all accelerate. Use the Harvest Window Predictor to catch the onset.
Harvest Window Predictor
Paste your VCD (or OD600-derived density), viability, glucose, and acetate data to find the optimal harvest day before death phase starts.
Doubling Times by Medium & Strain
Doubling time is the single best number to pin down before designing anything around an E. coli process. Medium identity, strain background, and temperature all move it substantially.
| Strain / medium | Temp (°C) | μmax (h−1) | Doubling time |
|---|---|---|---|
| MG1655 K-12 / LB | 37 | 1.4–2.0 | 21–30 min |
| MG1655 K-12 / M9 + glucose | 37 | 0.6–0.8 | 55–70 min |
| MG1655 K-12 / M9 + acetate | 37 | 0.2–0.3 | 140–210 min |
| BL21(DE3) / LB | 37 | 1.4–1.8 | 23–30 min |
| BL21(DE3) / TB or 2×YT | 37 | 1.6–2.1 | 20–26 min |
| MG1655 K-12 / LB | 30 | 0.9–1.2 | 35–45 min |
| MG1655 K-12 / LB | 25 | 0.5–0.7 | 60–85 min |
| B-strain fed-batch, μ-controlled | 37 | set 0.1–0.3 | 140–420 min (by design) |
See the doubling time reference table for cross-organism comparisons and the Growth Curve Fitter to extract μ and td directly from your OD data.
Harvest and Induction Decisions by Phase
Which E. coli growth phase you sit in decides what you should do next. Here is the practical decision table most process engineers use:
| Phase | Induction / feed decision | Rationale |
|---|---|---|
| Lag | Do nothing | Cells not yet dividing; inducer wastes and can arrest recovery |
| Early log (OD 0.1–0.3) | Too early for standard IPTG; OK for auto-induction systems | Low biomass; per-cell product is diluted |
| Mid log (OD 0.4–0.8) | Standard IPTG induction window | Translation machinery peak; acetate still low |
| Late log (OD 0.8–1.5) | Acceptable for soluble proteins; consider temperature drop to 18–25 °C to improve folding | Lower per-cell yield but more total biomass |
| Early stationary | Harvest soluble proteins; induce RpoS-dependent or nutrient-starvation systems | Proteolysis rising; some promoters activate |
| Late stationary / death | Harvest regardless of product state | Protease release, endotoxin spike, quality decline |
For a deeper walkthrough of the induction decision see IPTG induction optimization, and for the full strategy-selection framework see E. coli expression systems.
How to Monitor Each Phase
The hard part of using E. coli growth phases in practice is knowing which phase you are in now, not after the fact. Each phase has a characteristic signal pattern across the standard bioprocess sensors:
For the full monitoring toolkit across offline, at-line, and online methods, see cell growth monitoring in suspension culture.
Frequently Asked Questions
What are the four growth phases of E. coli?
E. coli grown in batch culture passes through four phases: lag (cells adapt to the medium, no net division, 10 min to 4 h), exponential or log (cells divide at μmax, doubling every 20–30 min in rich medium), stationary (growth stops because a nutrient is exhausted or an inhibitor accumulates), and death (viable count declines). The transitions are controlled by distinct regulatory programmes, most notably the sigma factor RpoS at the onset of stationary.
How long does E. coli take to double at 37 °C?
E. coli MG1655 K-12 doubles approximately every 30 min in LB at 37 °C, ~60 min in M9 minimal medium with glucose, and 2–3 h on acetate as sole carbon source. BL21 in LB is similar or slightly faster than K-12.
At what OD600 should I induce E. coli with IPTG?
Standard practice is to induce at mid-log phase, OD600 0.4–0.8. At OD 0.4–0.6 cells are actively dividing and the translation machinery is maximally active. Inducing too early (OD < 0.2) gives low biomass; too late (OD > 1.5) reduces per-cell expression due to nutrient depletion and acetate accumulation. Auto-induction media and arabinose systems have different optima.
What is the lag phase and why does it matter?
The lag phase is the interval after inoculation when cells adapt to the new medium without dividing. It is metabolically very active — cells synthesise new enzymes, repair starvation damage, and accumulate transient intracellular metals. Lag duration depends on inoculum size and health, medium match between pre-culture and production medium, and any shocks. An unexpectedly long lag on a repeated culture is almost always a red flag.
What happens during E. coli stationary phase?
Net growth stops, and cells adopt a stress-resistance programme controlled by RpoS (σS). RpoS activates genes for oxidative defence, osmoprotection, DNA repair, and trehalose accumulation. Cells shrink, bulk protein synthesis drops ~80 %, but targeted translation of stress-response genes continues strongly. Some recombinant proteins are produced more efficiently in stationary phase because host biomass accumulation no longer competes for resources.
What causes acetate overflow in E. coli?
Acetate overflow occurs when glucose uptake exceeds the respiratory capacity of the TCA cycle, diverting excess acetyl-CoA to acetate via the Pta–AckA pathway. Above ~0.2 g/L, acetate inhibits both growth and product formation, which is why fed-batch strategies cap μ below μmax by glucose-limited feeding.
References
- Rolfe MD, Rice CJ, Lucchini S, Pin C, Thompson A, Cameron ADS, Alston M, Stringer MF, Betts RP, Baranyi J, Peck MW, Hinton JCD. Lag phase is a distinct growth phase that prepares bacteria for exponential growth and involves transient metal accumulation. Journal of Bacteriology (2012) 194:686–701. DOI: 10.1128/jb.06112-11.
- Battesti A, Majdalani N, Gottesman S. The RpoS-mediated general stress response in Escherichia coli. Annual Review of Microbiology (2011) 65:189–213. DOI: 10.1146/annurev-micro-090110-102946.
- Gefen O, Fridman O, Ronin I, Balaban NQ. Direct observation of single stationary-phase bacteria reveals a surprisingly long period of constant protein production activity. PNAS (2014) 111:556–561. DOI: 10.1073/pnas.1314114111.
- Pletnev P, Osterman I, Sergiev P, Bogdanov A, Dontsova O. Survival guide: Escherichia coli in the stationary phase. Acta Naturae (2015) 7:22–33. PMC: PMC4717247.
- 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. DOI: 10.1038/nature15765.