Enzyme Kinetics Calculator: Km, Vmax, kcat with Fits

You ran a lactase activity assay at eight substrate concentrations, recorded initial velocities, and now need Km and Vmax to characterize the enzyme. An enzyme kinetics calculator fits your rate-versus-substrate data to the Michaelis–Menten equation by nonlinear regression and returns Km, Vmax, and — if you supply enzyme concentration — kcat. Those three parameters tell you how tightly the enzyme grabs substrate, how fast it turns it over, and how efficient it is compared to other enzymes or mutants.

The mistake that ruins the most kinetics experiments: using progress-curve endpoints instead of initial velocities. If you measure product at 30 minutes and call that the rate, substrate depletion and product inhibition have already bent the curve. The Michaelis–Menten model assumes steady-state conditions where [S] is essentially constant. Take readings within the first 5–10% of substrate consumption, or measure a full time course and fit only the linear initial segment.

Km is the substrate concentration at which the reaction runs at half its maximum velocity. It is not a binding affinity in the strict thermodynamic sense — it conflates binding (k₁, k⁻₁) and catalysis (k₂) into one apparent constant. A low Km means the enzyme reaches half-max speed at low [S], which usually (but not always) means tight binding. Compare Km to the physiological [S]: if [S] >> Km, the enzyme runs near Vmax in vivo; if [S] << Km, it operates in the first-order regime where small changes in [S] proportionally change the rate.

Vmax is the maximum velocity when all enzyme molecules are saturated. It depends on enzyme concentration: double [E] and Vmax doubles. By itself, Vmax is not very useful for comparing enzymes because it bundles together how much enzyme you added and how fast each molecule works.

kcat (turnover number) separates the two: kcat = Vmax / [E]ₜ. It tells you how many substrate molecules one enzyme molecule converts per second. Typical kcat values range from 1 s⁻¹ for slow enzymes to 10⁶ s⁻¹ for catalase. If you do not know [E]ₜ, you cannot get kcat — and [E]ₜ means total active enzyme, not just total protein (inactive enzyme does not count).

The ratio kcat/Km combines catalytic speed and substrate recognition into a single number with units of M⁻¹ s⁻¹. It is the best single metric for comparing two enzymes acting on the same substrate, or one enzyme acting on different substrates. Higher kcat/Km means the enzyme is better at finding and converting that substrate under conditions where [S] is well below Km.

The theoretical upper limit is the diffusion rate — about 10⁸–10⁹ M⁻¹ s⁻¹ — because the enzyme cannot grab substrate faster than substrate diffuses to the active site. Enzymes that approach this limit (like superoxide dismutase or triosephosphate isomerase) are called “diffusion-limited” or “catalytically perfect.”

When comparing mutants, a kcat/Km ratio that drops 10-fold usually means the mutation disrupted either substrate binding or the catalytic mechanism (or both). Report which parameter changed: a mutation that halves kcat but does not change Km affects catalysis, not binding. A mutation that doubles Km but keeps kcat intact weakens binding without slowing the chemical step.

A curve that visually passes through your data points does not guarantee a good fit. Always check the residuals — the difference between each observed rate and the fitted value. If the Michaelis–Menten model is appropriate, residuals should scatter randomly around zero with no pattern. A U-shaped or arched residual pattern means the model is systematically missing the data, which happens with substrate inhibition, cooperative binding, or two-site kinetics.

R² is tempting but misleading for nonlinear fits. An R² of 0.99 does not mean the model is correct — it just means the curve explains most of the variance. A substrate-inhibition model will often give R² > 0.99 on data that follows simple Michaelis–Menten, because more parameters always improve the fit. Use the residual plot, not R², to decide whether the model is right.

If your highest [S] point sits far from the rest and drags the fit, check whether it is genuinely at saturation or whether substrate inhibition is kicking in. A single outlier at high [S] can pull Vmax up and inflate Km. Try fitting with and without that point — if Km changes by more than 20%, investigate that data point.

Some enzymes show a rate that rises with [S], peaks, and then drops at high substrate concentrations. This is substrate inhibition — excess substrate binds a second site (or binds non-productively to the active site) and shuts down catalysis. If you fit a standard Michaelis–Menten curve to substrate-inhibition data, the fitter will compromise between the rising and falling parts and give you garbage values for both Km and Vmax.

The fix: look at your raw data first. If the rate clearly decreases at the highest [S] values, use a substrate-inhibition model that includes a Ki,sub term: v = Vmax × [S] / (Km + [S] + [S]²/Ki). If you do not have enough points in the inhibited region to fit Ki reliably, exclude the inhibited points and fit only the ascending portion. Note this in your report.

Outliers at low [S] are even more dangerous because Km is most sensitive to the early part of the curve. A single bad point at [S] = Km can shift the fitted Km by 50% or more. Always run at least 8 substrate concentrations spanning 0.2× to 5× the expected Km, and run duplicates or triplicates at each concentration.

My Km changes every time I repeat the experiment. How much variation is normal?
A coefficient of variation (CV) under 20% across independent experiments is typical for well-behaved enzymes. If Km swings by 2-fold or more, check whether your substrate stock concentration is accurate, whether the enzyme is partially inactivated between runs, and whether you are measuring true initial rates.

Lineweaver–Burk gave me a different Km than nonlinear regression. Which is right?
Nonlinear regression is more reliable. The Lineweaver–Burk double-reciprocal plot distorts the error structure by inverting the data — low-rate points (which have the most relative error) become the most influential. Use Lineweaver–Burk for visual pattern recognition (inhibition type), not for extracting Km and Vmax.

I only have 4 data points. Is that enough?
Technically the Michaelis–Menten equation has two free parameters (Km and Vmax), so 4 points can produce a fit. But the confidence intervals will be enormous and you cannot detect systematic deviations. Eight points minimum, ideally with 3–4 points below Km, 1–2 near Km, and 2–3 well above Km.

My enzyme has a kcat of 0.01 s⁻¹. Is it broken?
Not necessarily. Some enzymes are genuinely slow. Lysozyme has a kcat of about 0.5 s⁻¹. If your enzyme is a protease or a nuclease and kcat is below 0.01, double-check that the enzyme is active (positive control substrate) and that [E]ₜ is correct. An overestimate of [E]ₜ will underestimate kcat proportionally.

Four equations cover the standard kinetic analysis workflow:

Units note: v and Vmax are in concentration per time (e.g., µM/min). [S] and Km are in the same concentration unit (µM, mM). kcat is in s⁻¹ (convert Vmax to per-second units before dividing by [E]ₜ). kcat/Km ends up in M⁻¹ s⁻¹ — convert Km to molar if it is in µM or mM.

Scenario: You measured β-galactosidase (lactase) activity with ONPG as substrate at eight concentrations. Enzyme concentration is 10 nM. You need Km, Vmax, and kcat.

Data: [S] (µM): 25, 50, 100, 200, 400, 800, 1600, 3200. Rates (v, µM/min): 4.8, 8.9, 15.2, 23.1, 31.4, 38.7, 43.1, 45.6.

Step 1 — Visual inspection.
The rate rises steeply at low [S] and levels off above 1600 µM. No drop at high [S], so no substrate inhibition. Michaelis–Menten model looks appropriate.

Step 2 — Nonlinear fit.
The calculator returns: Km = 210 µM, Vmax = 48.3 µM/min. Residuals scatter randomly — no systematic pattern.

Step 3 — kcat.
Vmax in per-second units: 48.3 / 60 = 0.805 µM/s.
kcat = 0.805 µM/s / 0.010 µM = 80.5 s⁻¹.
Literature value for β-galactosidase with ONPG is roughly 50–100 s⁻¹, so this is in range.

Step 4 — Catalytic efficiency.
kcat/Km = 80.5 / (210 × 10⁻⁶) = 3.83 × 10⁵ M⁻¹ s⁻¹.
Well below the diffusion limit, as expected for a glycosidase. The enzyme is efficient but not diffusion-limited.