Experiments are performed in 96-well half-area plates, containing 50 µL final assay volume. Inhibitor dilutions are performed in transparent conical bottom plates prior to transfer to assay plates, while substrate and enzyme stock solutions are prepared in plastic microcentrifuge tubes. All compound stocks solutions and dilution series are prepared in assay buffer from concentrated DMSO solutions, so that final DMSO assay concentration does not exceed 1%. For continuous assays, substrate, NAD+ co-substrate (for sirtuin assays), inhibitor, and trypsin are added to the assay plate, followed by addition of enzyme and immediate readout at room temperature every 30 s until the assay is completed. All experiments are performed as internal duplicates, and control wells without inhibitor and without enzyme or inhibitor (background) are included.

Determination of continuous assay conditions:

1. For a set concentration of substrate higher than the *K*M (ideally 3 times higher, use 200 µM substrate concentration if *K*M unknown), perform time course experiments at varying enzyme and trypsin concentration until linear progression curves are obtained with appropriate signal-to-noise ratio (optimally, rates of 3–5% of the uninhibited reaction should be discernable from baseline). Fine-tune enzyme concentration if necessary.

*Example: incubate enzyme at 1000, 100, 10 and 1 nM concentration with trypsin (16, 8, 4 and 2 µg/mL), substrate (200 µM), and NAD**+ **(500 µM; only for sirtuins) for 60 min with fluorescence reads every 30 s.*

2. For the preferred concentration of enzyme, screen concentrations of trypsin to find the lowest possible concentration at which the conversion rate is kept as high as possible. This ensures that AMC release by trypsin is not the rate-limiting step, while minimizing the effect of trypsin on enzyme stability. Assay progression curves should be linear for at least 30 min (see *Troubleshooting *section).

*Example: incubate enzyme (10 nM), substrate (200 µM), and NAD**+**(500 µM; only for sirtuins) with trypsin at 8, 4, 2, 1, 0.5 and 0.25 µg/mL concentration for 90 min with fluorescence reads every 30 s.*

3. Optional: determine the Michaelis-Menten constant (*K*M) of the substrate. This will be useful for ensuring that substrate concentration is appropriate for maintaining steady-state conversion, and for determining inhibitor *K*i values from the obtained IC50.

*Example: incubate enzyme (10 nM), NAD**+**(500 µM; only for sirtuins), and trypsin (4 µg/mL) with substrate at 195, 130, 87, 58, 38, 26, 17, 11, 7.6, 5.1, 3.4 and 2.3 µM concentration (1.5-fold dilution) for 60‒90 min with fluorescence reads every 30 s, measure initial conversion rates and fit to the Michaelis-Menten equation to determine the substrate K**M**.*

Continuous inhibition assays:

1. Estimate the inhibitor potency (IC50,est) with a small number of end-point or continuous assays.

*Example (end-point): incubate enzyme and substrate (and NAD**+**) with inhibitor at 10, 1, 0.1 and 0.01 µM concentration for 30 min, add trypsin developer (200 µg/mL) and read after 10 min assay development. Estimate the inhibitor IC**50 **by comparing to the control without inhibitor.*

2. Perform dose-response continuous inhibition assays (2-fold dilutions) with the preferred concentrations of enzyme and trypsin (and NAD+) starting at 20 times the IC50,est.

*Example: for IC**50,est **= 1 nM, incubate enzyme (10 nM) and substrate (200 µM (and 500 µM NAD**+**)) with inhibitor at 20, 10, 5, 2.5, 1.25, 0.62, 0.31, 0.16, 0.078, and 0.039 nM concentration for 90 min with fluorescence reads every 30 s, or until the progression curve from control wells without inhibitor deviate from a straight line.*

3. Visually assess final substrate conversion at each concentration of inhibitor, and repeat experiments with alternative inhibitor concentrations until covering the range from no inhibition to full inhibition.

4. Perform final experiments twice in order to report data as average ± SD of two individual experiments.

Data processing and fitting (instructions for GraphPad Prism 7.0):

1. Paste data in Prism as two replicate values in side-by-side sub columns, and enter time points as “X” values. Place baseline data in “Group A”, and label all other groups with the concentration of inhibitor employed.

2. Optional: transform data from fluorescence units to concentration of AMC. A standard AMC concentration curve must be obtained for the plate reader employed, which can be then included in Prism under “Transform” – “User-defined Y functions”.

3. Use “Remove Baseline and Column Math” to remove baseline fluorescence (Group A) from all data sets.

4. Plot data from duplicate wells as discrete data points, in order to identify outliers and experimental errors (“New Graph of Existing Data…”, with the option “Create a new graph for each data set”). Data with low signal-to-noise ratio (for example, due to full inhibition of the enzyme) may be discarded, especially in the case of slow-binding inhibitors.

5. Use control data without inhibitor to determine the time frame of analysis in which the progression curve follows a straight line. Discard all data outside of this time frame.

*Example: discard data from 0 to 3 min if steady-state conversion is not reached and all data after 60 min if the conversion rate starts to decrease.*

6. Plot all remaining data in a single figure, as mean values with error bars for each concentration of inhibitor and assess the shape of each progression curve. If all experiments follow straight lines (__fast-on/fast-off kinetics__), go to **point 7**, otherwise continue from **point 8** (__slow-binding kinetics__).

7. __Fast-on/fast-off kinetics__: fit each progression curve to a straight line (“Analyze” – “Nonlinear regression” – “Straight line”) and select “Create summary table and graph” to obtain a XY type of graph with the Slopes and the concentrations of inhibitor. Transform inhibitor concentrations into logarithms, and use **Eq. 1 **(“Analyze” – “Nonlinear regression” – “log(inhibitor) vs. response – Variable slope”) to calculate the IC50 of the inhibitor. IC50 values can be transformed into *K*i values using the Cheng-Prusoff equation (**Eq. 2**).

8. __Slow-binding kinetics__: fit control data without inhibitor to a straight line, and all other progression curves to **Eq. 3**, which takes into account initial and final conversion rates (*V*in* *and *V*ss, respectively) and provides the observed rate at which the equilibrium is reached (*k*obs). This equation can be added to Prism under “Analyze” – “Nonlinear regression”in the form: Y=(Vss*X)+((Vin-Vss)*(1-exp(-kobs*X))/kobs)+Po, with *V*ss, *V*in, and *k*obs constrained to be greater than 0. Create a XY summary table with *k*obs data and the concentration of inhibitor, and exclude conditions for which the standard error exceeds the mean *k*obs value.

9. Collect *k*obs data from two individual experiments, create a data sheet with both data sets, and fit *k*obs data (“Analyze” – “Nonlinear regression”) to either **Eq. 4 **(linear correlation, “Straight line”) or **Eq. 5**(hyperbolic correlation, added as Y = kminus2+k2*X/(X+Ki1*(1+S/Km)), with *K*M constrained to be the value calculated before). **Eq. 4 **corresponds to __mechanism A__ of slow-binding kinetics, in which there is a single slow binding step (E + I **⇌**EI) with kinetic constants *k*1 (direct) and *k*–1 (inverse). **Eq. 5 **corresponds to __mechanism B__ of slow-binding kinetics, where there is a fast binding step followed by a slow transition between enzyme-inhibitor complexes (E + I **⇌**EI **⇌**EI*). Mechanism B is characterized by the *K*i,1 equilibrium constant for the first step and kinetic constants *k*2 and *k*–2 for the second step.6,13Summary tables provide values for all constants when possible.

10. Calculate the potency of the inhibitor (*K*i) using **Eq. 6 **for inhibitors following mechanism A and **Eq. 7 **for mechanism B. This will not be possible for irreversible inhibitors (*k*–1 or *k*–2 constants close to 0).6

11. Calculate dissociative half-life values (t1/2diss) using **Eq. 8 **for inhibitors following mechanism A and **Eq. 9 **for mechanism B.2,18 Since *k*–1 constants are not elucidated for mechanism B-type of inhibitors, the limiting scenario in which *k*–1 is much larger than *k*–2 may be considered, which provides the lowest t1/2diss possible.