Quantification of GTPase Cycling Rates of GTPases and GTPase:Effector Mixtures Using GTPase Glo Assays

In different cellular activities such as signal transduction, cell division, and intracellular transportation, small guanosine triphosphatases (GTPases) take on a vital role. Their function involves hydrolysis of guanosine triphosphate (GTP) to guanosine diphosphate (GDP). In this article, we explain the application of a commercially available GTPase assay—the GTPase Glo assay by Promega—for investigation of GTPase‐effector interactions. We provide experimental protocols together with an analysis model and software to obtain GTPase cycling rates of GTPases and GTPase:effector mixtures. GTPase cycling rates refer to the rates by which a GTPase completes an entire GTPase cycle. These rates enable quantification of the strength of GTPase effectors in a concentration‐dependent fashion, as well as quantification of the combined effect of two effectors, independent of which GTPase cycle step they are affecting. © 2024 The Authors. Current Protocols published by Wiley Periodicals LLC.


INTRODUCTION
Small guanosine triphosphatases (GTPases) are a class of enzymes that play a fundamental role in various cellular processes, including signal transduction, cell division, and intracellular transport.GTPases regulate the hydrolysis of guanosine triphosphate (GTP) to guanosine diphosphate (GDP).Mechanistically, this activity involves three steps (Fig. 1): (1) binding of GTP to the GTPase, (2) hydrolysis of GTP to GDP and free phosphate, and (3) release of GDP from the GTPase.GTPase activity is often regulated by effector proteins: GTPase activating proteins (GAPs) that boost step 2 and GDP/GTP exchange factors (GEFs) that enhance step 3 (Bos et al., 2009;Cherfils & Zeghouf, 2013;Vetter & Wittinghofer, 2001).Several assays exist that examine single GTPase cycle Figure 1 Schematic illustration of assay steps and the GTPase cycling reaction in the GTPase Glo assay (Promega).A GTPase, alone or in combination with effector proteins, is incubated with GTP for a specified time (e.g., 1.5 hr) during which GTPase cycling occurs.This is followed by two processing steps, in which GTP (blue ovals) is first converted to ATP (blue squares) and then made luminescent (orange stars) using an ATP-specific detection reagent.Finally, luminescence, reflecting the remaining GTP, is measured.GTP hydrolysis cycling rates can be extracted by fitting the data with an exponential model.
steps, which helps us understand the mechanistic details of these cycle steps and their regulation.On the other hand, they do not allow study of the interplay of effectors acting on different GTPase cycle steps.Furthermore, approaches that reconstitute the complex cellular functions of GTPases in vitro (Bezeljak et al., 2020;Kohyama et al., 2022;Loose et al., 2008;Vendel et al., 2019) are sensitive to variations in protein batch activities and benefit from easy and accessible assays assessing protein purification batches activities.
We here describe how a commercially available GTPase assay (the Promega GTPase Glo assay, henceforth referred to as the GTPase assay) can be used to quantitatively study GTPase-effector interactions and effector interplay and to easily test activity of GTPase and effector purification batches.In the GTPase assay, proteins of interest are incubated with GTP for a specified amount of time for GTPase cycling to occur.The reaction is then stopped and the amount of remaining GTP is determined as luminescence signal.GTPase cycling rates can be obtained from the resulting data using a coarse-grained exponential fitting model (Fig. 1).This article describes the workflow for conducting GTPase assays using GTPase:effector protein mixtures, and provides analysis software to obtain GTPase cycling rates.GTPase cycling rates describe the rate by which the GTPase completes entire GTPase cycles.Our protocols enable a quantitative analysis of the single and combined effect of one or two effectors on the entire GTPase cycle in a concentrationdepended fashion, thus allowing the study of effectors (e.g., GEFs and GAPs) acting together on different steps of the GTPase cycle.We do not advise using this protocol if single GTPase steps are to be examined in-depth, but rather when the effects of a protein on the entire GTPase cycle are of interest.This method is illustrated using a basic GTPase assay with human Ras (GTPase) (example 1) and an elaborate GTPase assay using Saccharomyces cerevisae Cdc42 (GTPase), Cdc24 (GEF), and Rga2 (GAP) (example 2) (Tschirpke, Daalman et al., 2023).Table 1 summarizes how these examples are used to illustrate the protocols described.The protocols can be applied to other GTPases and effectors.In the Basic Protocol, we outline the workflow for conducting GTPase Glo assays.In Support Protocol 1, we describe how the GTPase assay data (luminescence) are converted to amounts of remaining GTP.In Support Protocol 2, we describe how

of 29
Current Protocols data obtained by Support Protocol 1 can be fitted with a GTPase cycling model (and the openly available Matlab code developed by us) to obtain GTPase cycling rates.

STRATEGIC PLANNING
If the GTPase assay is being performed for the first time, we advise first practicing the pipetting of the small volumes required into the wells using colored water.It is best to conduct smaller assays using a larger number of replicates per sample (e.g., eight sample rows per assay with five replicates each, as described in the Basic Protocol) before moving on to larger assay sets with more proteins.
Once you feel more comfortable with the assay, larger assay sets can be conducted, for example, to look at the interaction between a GTPase and one or several effectors (e.g., GTPase-GEF, GTPase-GEF-GAP).The following section applies predominantly (but not solely) to such larger assays sets involving multiple GTPases and/or effector proteins (example 2).We advise users to determine how many proteins will be assayed well before starting the assays.Specifically, we recommend the following steps.
1. Verify that the proteins and protein buffers do not interfere with the assay signal.
The GTPase assay measures luminescence, which correlates with the amount of remaining GTP.Proteins that interact with and can alter luminescence (e.g., fluorescent tags, see Supporting Information S3) are therefore not suitable for this assay.Buffers containing ADP, ATP, GDP, GTP, and guanosine phosphate analogs are also not suitable, as they interfere with GTPase assay step reactions.Consider the buffer components carefully and verify that the effector proteins used do not exhibit GTPase activity.

Verify that incubation times lie in the exponential decline region of the remaining GTP.
In the GTPase assay, the proteins of interest are incubated with GTP for a specified amount of time during which GTPase cycling can occur.The reaction is then stopped and the amount of remaining GTP is determined as a luminescence signal.GTPase cycling rates can be obtained from these data through a coarse-grained exponential fitting model (Fig. 1 and Supporting Information S8).The model is based on our observation that the amount of remaining GTP declines exponentially with time (Supporting Information S4).
It is advisable to verify that this is true for the proteins being studied (GTPase enzymes and GTPase:effector mixtures) and incubation times to be used.To obtain such data, prepare one batch of serial dilutions of the proteins you will use.Use exactly these dilutions to conduct several assays with different incubation times.Check whether the amount of remaining GTP in these assays follows an exponential decline.Further, analyze each assay individually using our GTPase cycling model.Verify that different incubation times (different assays) yield the same rates (e.g., Supporting Information S4).
3. Dialyze all proteins in same buffer.In the assay, protein activity is determined by normalizing the assay readout (luminescence) to a well containing buffer only.Thus, differences in ion concentrations, detergents, or other additives (such as glycerol) can lead to differences in luminescence.This is easily avoided if all proteins used in an assay set are prepared in the same buffer.We advise against using a buffer containing glycerol, as the increased viscosity may affect protein activities (unless protein activities in a denser environment are of interest).Further, buffers should not contain DTT; GDP, GTP, or other guanosine analogs; ADP, ATP, and other nucleotide triphosphates; or EDTA or other chelating agents that can complex magnesium.If a GTPase-GEF interaction is to be investigated, the buffer usually requires magnesium salts (e.g., 10 mM MgCl 2 ).
Many proteins require glycerol for storage.In this case, keep the protein at a high concentration (at least six times of the concentration that ought to be used in the assay) in a buffer containing glycerol (e.g., 50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 10 mM MgCl 2 , 1 mM 2-mercaptoethanol, 10% glycerol) and then dilute it into the same buffer without glycerol.GTPase assays usually require small concentrations of proteins (in our experience, 0.5-5 μM).If the protein is stored at a high concentration, the amount of glycerol in the final reaction will be negligible.
If it is not possible to prepare all proteins in the same buffer, one must verify that the differences in buffer composition do not affect luminescence intensities.A GTPase assay containing only the different buffers can elucidate this.If all buffer mixtures exhibit the same luminescence values, then the differences in buffer composition do not affect luminescence.It should be noted that it is still possible that the buffer components affect protein behavior.

Use the same serial dilution for all assays.
The assay is sensitive to concentration changes, especially for highly active GTPases and effectors that strongly boost GTPase activity, and thus is also sensitive to pipetting errors when preparing protein dilutions.Prepare serial dilutions of every protein and use exactly the same serial dilution for the entire assay series (Fig. 2).Different serial dilutions of the same protein can exhibit slightly different rates.For example, when comparing the effect of effector protein X on GTPase A and GTPase B, the same serial dilution of effector X must be used for both GTPases.

5.
Include a reference sample if data from different assays will be compared.GT-Pase assays can be sensitive to GTPase activity changes resulting, for example, from small pipetting errors or changes in conditions (e.g., temperature, shaker speed).As this

of 29
Current Protocols can lead to slight changes in protein activity between assays, it is best to include all samples to be compared in the same assay.This is not always possible.If the strength of several effectors on one GTPase will be compared, we recommend always including one sample containing only the GTPase in each assay (Fig. 2).The assay-specific activity of the GTPase sample can then be used to account for any assay-specific variability (through fitting the correction factor c corr , as will be explained in Support Protocol 2 and in Eqn. 9 of Supporting Information S8). Assay variations are typically very small (reflected as c corr ≈ 1.0) (Tschirpke, Daalman et al., 2023;Tschirpke, van Opstal et al., 2023).We recommend excluding assays that show vastly different protein activities (reflected in a c corr much larger or much smaller than 1), as these indicate that the GTPase behavior/assay conditions are unusual.
6. Determine useful concentration ranges before conducting an assay series.In our experience, a good readout regime is 5%-90% remaining GTP.Higher values (>90% remaining GTP) have a larger error (due to normalization) and lower values (<5% remaining GTP) may show a saturation effect due to the small amount of remaining GTP.
In the recommended regime, a saturation of the fitted cycling rates may occur.If an assay with multiple proteins (e.g., a GTPase and two effector proteins) is conducted, it is useful to determine beforehand which concentration ranges of each protein lead to optimal readouts to reduce the time it takes to conduct the entire assay series.It is also advisable to stick to similar incubation times for all assays, as this streamlines the selection of protein concentrations (for assays containing protein mixtures) that yield suitable readout regimes.
7. Make an assay plan.Use the same serial dilutions when quantifying and comparing multiple protein interactions.Since many proteins are not stable for weeks at room temperature or 4°C and require glycerol for cryopreservation (which is not recommended), all required assays must be conducted within a rather short time period to obtain reliable and comparable results.We strongly recommend careful planning of which assays will be conducted using which proteins, and calculating the required volumes of protein needed.Our recommended procedure is explained below using as an example a GTPase assay series investigating the cycling of a GTPase and the effects of both a GEF and a GAP, alone and in combination (example 2; Fig. 2; Tschirpke, Daalman et al., 2023).

Considerations (as described earlier):
(1) Different serial dilutions of the same protein may exhibit slightly different rates due to small pipetting errors.Hence, the same dilution should be used for the entire assay set.(2) For larger assays, it is advisable to know in advance which concentration ranges give good signal to reduce the time the proteins need to be stored.(3) A reference sample should be included (a sample of the GTPase alone, at the concentration used in the particular assay).( 4) The incubation times used should be in the exponential region.(5) It is advisable to use similar incubation times for all assays, as this will make it easier to choose protein concentrations for assays containing protein mixtures that will result in suitable readouts.
Protein dilutions: To conduct the assay, protein dilutions are needed for the three proteins used in the entire assay set (GTPase, GEF, and GAP; Fig. 2, left column).These dilutions will be used to create the protein mixtures for one GTPase assay (Fig. 2, middle column).
The concentration of the dilutions depends on two factors: (1) The number of protein species in the assay (N).For a mixture containing a GTPase, GEF, and GAP, N is 3.As it is easiest to prepare the various mixtures by combining proteins in 1:1:1 volume ratios, each protein will be diluted by a factor of three in setting up the assay.
(2) To initiate the GTPase reaction, a 2× GTP solution is mixed in a 1:1 volume ratio with each protein mixture.Together, the protein dilutions (left column) should be prepared at a 2×N-fold higher concentration than the final concentration in the assay (right column).Thus, for an assay set with a GTPase, GEF, and GAP, all protein dilutions should be prepared at

of 29
Current Protocols 6× concentration.For assay sets investigating mixtures of four or five proteins, N will be 4 or 5 and the dilutions must be 8× or 10×, respectively.
Procedure: First, an assay is conducted with a serial dilution of only the GTPase (Fig. 2, top row).From this assay, pick one or two GTPase concentrations to conduct all subsequent assays.Because the subsequent assays involve a GEF and a GAP, both of which boost GTPase activity, it is wise to choose a GTPase concentration that leads to a high amount of remaining GTP (e.g., 80%-90%).Next, an assay is conducted with the chosen concentration of GTPase and a serial dilution of the GEF (Fig. 2, second row).A similar assay is conducted with GTPase and the GAP (Fig. 2, third row).Finally, an assay is conducted with all three proteins in multiple combinations (GTPase alone, GTPase:GEF, GTPase:GAP, and GTPase:GEF:GAP) all using the same GTPase concentration.For this final assay, choose GEF and GAP concentrations that resulted in a medium amount of remaining GTP (e.g., 60%-70%) to leave room for even lower values when both proteins are combined (Fig. 2, bottom row).Ideally, additional assays will be added to control for potential non-canonical effects (see below).

Control for non-canonical effects by adding an inert protein.
We have observed several potential artefacts during GTPase cycling in this assay.These effects are generally small and can be accounted for.For example, we observed that both bovine serum albumin (BSA) and casein cause a slight boost in GTPase activity of both Ras and Cdc42.BSA and casein are considered inert and have no known interaction with either Ras or Cdc42.We suspect that they seemingly boost GTPase activity by increasing the effective GTPase concentration.A few GTPase molecules might stick to the well in each reaction chamber, being rendered inactive.When another protein is added, the chamber wall will be covered with some molecules of both GTPase and the other protein, thereby increasing the effective GTPase concentration.To ensure that the effect of a protein on GTPase cycling is not only due to these non-canonical effects, we advise conducting additional assays with the GTPase and an inert protein (e.g., BSA or casein).The effect of the protein of interest will need to exceed that of an equimolar concentration of the inert protein to be of non-canonical origin.It should be noted that BSA is strongly negatively charged, which could lead to non-specific protein interactions.Therefore, casein might be more suitable as a control.For example 2 with a GTPase, GEF, and GAP (Fig. 2), control assays for non-canonical effects include: (1) GTPase:casein serial dilution, (2) GTPase:GEF:casein, and (3) GTPase:GAP:casein.

CONDUCTING GTPase Glo ASSAYS
In the GTPase Glo assay, proteins of interest (GTPase enzymes with or without effectors) are incubated with GTP for a specified amount of time for GTPase cycling to occur (Fig. 1).Once the reaction is stopped, the remaining GTP is translated into a luminescence signal and measured.The following steps are from the assay manual with only minor modifications.The volumes provided are for assays in 384-well microplates.Assays using larger volumes can be conducted using different plates (see the assay manual).In addition to following the steps here, we recommend carefully reading the assay manual.
The following steps are for a GTPase assay with a serial dilution for six concentrations of Ras GTPase (example 1; Fig. 3).An illustration of the required assay planning steps is shown in Supporting Information S2.The assay results in data that are found in ex-ample1.xlsx,tab: 'E1', 'E2'.

Materials
Proteins of interest: GTPase (e.g., human Ras, EMD Millipore, cat.no.553325) Optional effector proteins (GEP, GAP) Figure 3 Schematic of steps for a GTPase assay using a serial dilution of six Ras GTPase concentrations.The preparatory steps are further illustrated in Supporting Information S2.

of 29
Current Protocols Prepare protein 1. Dialyze GTPase in protein buffer following the instructions accompanying the chosen dialysis device.
If the GTPase is at a high concentration (e.g., 50-100 μM) and is stored in buffer that does not differ significantly from protein buffer, it can be sufficient to dilute the protein using protein buffer.
For more details on buffer considerations, see Strategic Planning.
In general, shorter dialysis times at lower temperatures (e.g., 4 °C) are favorable for most proteins.

Prepare for assay (mise-en-place)
Before beginning, decide on the assay scope and prepare all materials needed.For templates, see Supporting Information S1; for an example assay using Ras GTPase, see Supporting Information S2 and Figure 3.
2. Decide which samples and how many replicates are needed.
We recommend three to five replicates per sample.In the beginning, we recommend one GTPase per assay and five replicates of each GTPase concentration.We suggest a total of eight samples per assay, including six GTPase concentrations plus two protein buffer samples for normalization.
3. Calculate the volumes (V) of protein dilutions, 2× GTP solution, Glo solution, and detection reagent needed to conduct the assay (for example calculations, see Supporting Information S2).
We recommend preparing a small excess of each solution.In particular, we recommend preparing 5 μl extra of each protein dilution to ensure that there is sufficient solution for the last well.For five replicates per sample, prepare 5 μl × 5 + 5 μl = 30 μl per dilution.
When calculating V GTP , the number of samples includes buffer samples as well.For the example shown in Fig. 3, the number of samples = 6 dilutions + 2 buffer = 8.
4. Calculate the dilution series steps for the GTPase, keeping in mind that the solutions must be 2× the final concentration in the assay (for example calculations, see Supporting Information S2).
5. Set up the 384-well assay plate as in Figure 3C.Label wells that will be used in the assay and seal wells that will not be used with Parafilm to prevent contamination with dirt.
a.For one assay, designate five columns of a 96-well plate to give five replicates per sample.b.Arrange the two protein buffer samples in the top and bottom rows of the plate.
In rare cases, luminescence values drift slightly towards lower values throughout the assay.Placement of a buffer row as the first and last rows ensures that this drift can be detected and be accounted for in the analysis.
c. Arrange the six GTPase dilutions between the buffer rows.d.Leave an empty row between all buffer/sample rows.
This will avoid spill-over of luminescence signal between samples (see Supporting Information S5).
6. Cut four pieces of Parafilm that are big enough to cover the area used for the assay.a The ADP solution provided in the kit must first be diluted from 10 mM to 1 mM (e.g., mix 2 μl of 10 mM ADP with 18 μl ultrapure water).

Conduct assay
7. Thaw protein samples and assay solutions on ice.
8. Make six 2× serial dilutions of GTPase with protein buffer according to your calculations (Fig. 3A).Vortex to mix and spin down gently to collect all volume at the bottom of the tubes.9. Prepare the 2× GTP solution (Fig. 3A).Vortex the GTP and DTT stock solutions, then combine as then combine with protein buffer according to Table 2 and vortex again.
2× GTP should always be prepared immediately before starting the assay and should never be stored or frozen (see Supporting Information S6).We also advise aliquoting the GTP and DTT stocks to reduce the number of freeze-thaw cycles.In our experience, up to three freeze-thaw cycles will not impact assay performance.
12. Seal wells with two sheets of Parafilm and incubate for the designated time on an orbital shaker.
We recommend sealing the wells by placing one sheet of Parafilm on top of the plate (without stretching it), pressing it firmly into the plate with a rounded implement (e.g., scissor handles), and then repeating the process with the second sheet of Parafilm.This ensures that the wells are tightly closed.Using only one sheet can result in insufficient sealing (i.e., small holes form in the Parafilm from too much pressure).
The incubation time and temperature depend on the protein of interest and concentrations used.We recommend choosing a temperature in accordance with the protein's environment in vivo.For incubation times, we recommend 60-90 min.
13. Several minutes before the incubation is finished, prepare the Glo solution (Fig. 3E).
a. Vortex Glo buffer and 10 mM ADP to mix. b.Tap the tube of Glo reagent to mix (do not vortex).c.Dilute 10 mM ADP to 1 mM using ultrapure water.Vortex to mix.d.Combine reagents according to Table 3 and vortex to mix.We advise aliquoting the 10 mM ADP, Glo buffer, and Glo reagent for future use to reduce the number of freeze-thaw cycles.The aliquots can be reused (see Supporting Information S6).The 1 mM ADP and the prepared Glo solution can only be used fresh and should not be stored or frozen.
The Glo reagent usually runs out first.To increase the number of assay runs per kit, it is important to prepare only the volume needed.
14.At the designated time, add 10 μl Glo solution to all wells (Fig. 3E), seal with two sheets of Parafilm (as above), and incubate 30 min at room temperature on an orbital shaker.
If an incubation at room temperature is not possible, this incubation step can, in our experience, also be conducted at temperatures of up to 30°C.
We advise aliquoting the detection reagent to reduce the number of freeze-thaw cycles and decrease the time required for thawing.
Calculate how much detection reagent is required per assay run.We have observed cases where two separately stored aliquots from the same batch lead to distinct luminescence values (see Supporting Information S7).To avoid such a shift within an assay, place a sufficient volume in one tube (e.g., take several aliquots) and vortex to mix before adding the reagent to the wells.
16. Measure luminescence of each well using a plate reader for 20 min.

ANALYZING GTPase ASSAYS TO CORRELATE LUMINESCENCE WITH REMAINING GTP
This protocol describes basic analysis steps to calculate the amount of remaining GTP from the GTPase assay readout (luminescence).The amount of remaining GTP is a simple measure used to compare the activity of different GTPases (of the same concentration, incubated with GTP for the same amount of time).The protocol illustrates steps for example 1: an assay including six Ras GTPase serial dilutions and two buffer reference samples (see Basic Protocol and Supporting Information S2), with three replicates each.The protocol uses and generates data in example1.xlsx,tab: 'E1', 'E2'.A spreadsheet editor or other analysis software is needed.The provided Python script (see Supporting Information S9) is needed if cycling rates will also be determined.

Calculate remaining GTP and its error
1. Calculate the average luminescence for all time points from 5 to 20 min for each well (Fig. 4).
2. Calculate the average luminescence for each sample (Lum s ) by averaging the wells (replicates) for that sample (Fig. 5, row a).
Each sample refers to a different dilution of GTPase.
3. Calculate the average luminescence for protein buffer (Lum b ) by averaging the wells/replicates from both buffer rows in a single calculation (Fig. 5, row a).Do not distinguish between the first and last buffer row.The luminescence value of buffer corresponds to 100% remaining GTP.The amount of remaining GTP for each protein sample is determined as a percent of the buffer value and thus normalizes the protein samples.In principle, the protein samples also need to be normalized to 0% GTP.Given that we observed basically no luminescence in the blank samples (see Supporting Information S5), this step is not necessary.
6. Calculate the error ( remaining GTP; Fig. 5, row d) by error propagation using the equation:

Format data for subsequent analysis (optional)
This process is only required if the data analysis will be continued to determine GTPase cycling rates in Support Protocol 2. It involves organizing the data in a spreadsheet with a specific format (Fig. 6) and the following column headers: Run = assay number.This must consist of an E followed by a number, which can be followed by letter(s) with or without an underscore.For example: E1, E2, E100, E1a, E100abc, E1_a, E100_abc, E1a_abc, E100f_abc Time = incubation time.Values must be in hours.GTP_remaining = amount of remaining GTP normalized to 1  Error = error values normalized to 1 Buffer_error: error of buffer normalized to 1 [GTPase]_conc = name of the GTPase followed by '_conc' (e.g., Ras_conc in Fig. 6).Values must be in μM.
[Effector]_conc (optional) = name of the effector followed by '_conc'.Values must be in μM.This column should only be present if it contains more than one unique value per assay.
Reformatting the data can be done using the Python script Ras_example.ipynb(see Supporting Information S9 ).This will be the input for the Matlab script used in Support Protocol 2.
10. Use the find/replace option of the spreadsheet editor to replace all points (.) with commas (,).
The Python script generates numbers in the format 1.00, but the Matlab script requires the format 1,00.
We advise grouping only assays of the same type (i.e., containing the same proteins and dilutions) in one spreadsheet tab, as they must follow the same structure.For example, for an assay set examining the GTPase Cdc42 and its GEF Cdc24, two types of assays were conducted: (1) assays using only serial dilutions of the GTPase and (2) assays using a constant GTPase concentration and serial dilutions of the GEF.All assay data from (1) should be grouped in one tab and all assay data from (2) should be grouped in a second tab.
Additional examples are given in Support Protocol 2.

FITTING GTPASE ASSAY DATA TO OBTAIN GTPASE CYCLING RATES
This protocol describes how the data (preprocessed in Support Protocol 1) can be fitted using a GTPase activity model (see Supporting Information S8) to obtain GTPase cycling rates k, which describe the rate with which a GTPase completes GTPase cycles.It can be used to fit data of GTPases and GTPase:effector mixtures (with up to two effectors, X1 and X2).
In short, data from the GTPase serial dilutions are first fitted with an exponential to obtain GTPase cycling rates k 1 and k 2 : Tschirpke, Daalman, and Laan

of 29
Current Protocols [GTP] t = [GTP] 0h exp(−Kt ) using [GTP] 0h = 1 and Equation 1 where K refers to the overall GTP hydrolysis rate and depends on the concentration of the GTPase.We call the concentration-independent rates (k) GTPase cycling rates, referring to the fact that they describe GTPase cycling and not specific GTPase cycle steps.The overall hydrolysis rate K is composed of an unaided hydrolysis contribution by an individual GTPase (K * 1 ) and a cooperative contribution by two GTPase molecules (K * 2 ).
Next, mixtures of one GTPase and one effector (X1) are fitted with: and mixtures of one GTPase and two effectors (X1, X2) are fitted with: Equation 3 using values of 1 or 2 for the exponents n and m (i.e., n = 1 and m = 1, n = 1 and m = 2, n = 2 and m = 1, or n = 2 and m = 2).This fits yield correction factors (c corr ) and cycling rates (k 3,X1 , k 3,X2 , k 3,X1,X2 ).c corr is dimensionless and accounts for variability between assays.It should be close to 1.0 (see Supporting Information S8).K 1 and K 2 represent corrected versions of K * 1 and K * 2 obtained from the GTPase serial dilutions (K 1 = c corr K * 1 , K 2 = c corr K * 2 ).K 3,X1 , K 3,X2 , and K 3,X1,X2 represent the overall hydrolysis rate contribution connected to effectors.They depend linearly on the corrected GTPase concentration, but also on the respective effector concentrations, either linearly or quadratically.
Note that, in order to analyze GTPase:effector mixtures, k 1 and k 2 must be known.It is thus required to fit the data of GTPase (serial) dilutions first.The model allows one to fit GTPase:effector mixtures with up to two effectors, with each effector showing either a linear or quadratic concentration dependence.If the effectors show neither of these concentration dependencies (e.g., due to saturation), we advise either including only the linear/quadratic regimes into the analysis or extending the fitting model to match the specific case.With the provided software, it thus is possible to fit the following protein (mixtures): GTPase serial dilution (example 1) GTPase + effector serial dilution (linear or quadratic conc.dependence, Eqn. 2) (example 2) GTPase + effector X1 serial dilutions + effector X2 serial dilutions (Eqn.3) (example 2, Fig. 2) The model is based on our observation that the amount of remaining GTP declines exponentially with time (see Supporting Information S4).It is advisable to verify that this is true for the proteins (GTPase and effectors) and incubation times to be used.
The model code analyzes data of each GTPase assay individually and then generates pooled rate values (for details on weighting of k for pooling and error propagation, see Supporting Information S11).The analysis code allows one to exclude GTPase assays conc_corr_bounds states which range of c_corr allows assays to be included in the pooling.
As c corr is expected to be close to 1.0, it is generally advisable to set the lower bound not too close to 0, as such low values are indicative that the inferred activity of the GTPase is unusually low in the assay.
k_low_err_filt.'GTPase-name'states whether assays containing the particular GTPase that have low standard errors (1 -10 or lower) for k values are excluded from pooling.This parameter must be declared individually for each GTPase.
In general, it is advisable to set this parameter to 'true' (i.e., assays with low standard errors will be excluded).If, however, a GTPase is dominated by k 2 (and has k 1 ≈ 0 and thus std.err.(k 1 ) ≈ 0), this parameter must be set to 'false' to allow for pooling of k.In many cases this cannot be known before an initial analysis.If chosen inappropriately, the parameter simply needs to be changed and the code re-run.
Below, we first describe the general protocol steps (case 1), which we then apply to example 1 (a simple assay containing one GTPase) (case 2) and example 2 (an elaborate assay set containing assay data for the GTPase alone and GTPase:GEF, GTPase:GAP, and GT-Pase:GEF:GAP mixtures) (case 3).Finally, we provide analysis steps for the combined analysis of example 1 and example 2 data (case 4).

Necessary resources
Spreadsheet editor Matlab software license, analysis scripts, and example data files encompassing this protocol Case 1: General procedure 1a.Set up the Matlab environment.To run the code, ensure that all required files are present: The assay data (assaydata.xlsx)must follow the formatting described in Support Protocol 1 and Figure 6.
The core of the fitting of the model occurs in the function Crocodile_model.m.This function determines which model scenario of effectors applies and then generates fit results.To plot these, plot_rate_data.m is required, which uses colormaps from useful_colormaps.m (containing a colormap from Smith et al., 2015).When needed, an unequal variance t test (Unequal_var_t_test.m ; Ruxton, 2006) is used to test for non-zero interactions between effectors.From the main function Pro-cess_assay.m,a predefined spreadsheet list is read using Read_assay_list.m to determine which data sets to use and how they are built up.The data are then read using extract_activity_data.m and passed to Crocodile_model.m.

COMMENTARY Background Information
In the GTPase Glo assay, the proteins of interest (GTPase enzymes with or without effectors) are incubated with GTP for a specified amount of time for GTPase cycling to occur, hydrolyzing GTP to GDP and free phosphate.The reaction is then stopped by addition of Glo solution, which contains a nucleoside-diphosphate kinase and ADP.The kinase converts the remaining GTP to ATP.Addition of the detection reagent, containing a luciferase/luciferin mixture, makes the ATP

of 29
Current Protocols luminescent, allowing measurement on a plate reader in luminescence mode (Fig. 1).The luminescence signal correlates with the amount of remaining GTP and thus inversely correlates with GTPase activity.The higher the luminescence values, the more GTP remained in solution and the less activity the GTPase had.
Low or no luminescence corresponds to very little or no remaining GTP and a high GTPase activity.
The GTPase Glo assay examines the entire GTPase cycle of the respective GTPase.An assay that examines all steps of the GTPase cycle can be advantageous.For one, it allows investigation of GTPase activities and GTPase-effector interactions.Compared to other in vitro assays that examine only one specific step of the GTPase cycle, the rates obtained by this assay may be more comparable to those observed in vivo, as GTPase cycling occurs constantly in vivo.Further, this assay enables comparison of the strength of effectors that act on different steps of the GTPase cycle, such as GEFs and GAPs.In addition, effector interplay can be studied with this method.The reasons for these advantages are also causes for the assay's main disadvantage.Because GTPase cycling (multiple completions of the entire GTPase cycle) is studied, the origin of observed effects remains elusive.Thus, this method is less suited when detailed mechanistic features of a GTPase cycle step are being investigated.For this purpose, one may need to use other GTPase assays such as the MESG/phosphorylase system (Zhang et al., 1997) or N-methylanthraniloyl-GTP/GDP system (Rapali et al., 2017), which examine only the GTP hydrolysis or GDP release step.

Critical Parameters
The GTPase assay is sensitive to small pipetting errors as well as changes in buffer and assay components.We strongly recommend a mise-en-place procedure, in which each assay is carefully planned (e.g., use one of the templates provided in Supporting Information S1) and the materials are prepared and in place before the assay is started.This greatly reduces the likelihood of errors.Accurate pipetting is essential.
We further advise: 1. Aliquoting and vortex all components before use (except the Glo reagent).2. Dialyzing all proteins into the same buffer (lacking glycerol, if possible).3. Including control samples such as a buffer row at the beginning and end of the plate, a positive control (Ras GTPase), and potentially inert proteins to account for non-canonical effects.4. Preparing 2× GTP solution, 1 mM ADP, and Glo solution fresh before use (never reuse or store).

Troubleshooting
For a brief coverage of troubleshooting, see Table 4. Please consult the manufacturer's user manual (Promega) for more detailed discussion of basic problems such as (1) no change in luminescence with increasing/decreasing concentrations of GTPase, GAP, or GEF; (2) low signal-to-background ratio; and (3) high or low luminescence signals.

Understanding Results
The GTPase assay yields luminescence values that are translated into amount of remaining GTP.These values can be fitted with GTPase cycling rates k.The use of example 1 luminescence and cycling rates k 1 , k 2 and their pooling will be discussed.Example 2 discusses assay correction factors c corr and cycling rates k 3,X .

Example 1
In this example, data were collected for five serial dilutions of Ras (spreadsheet: exam-ple1.xlsx in the Data folder; see Data Availability).The assay steps are described in the Basic Protocol and Supporting Information S2, and the analysis is outlined in Support Protocols 1 and 2.
Luminescence.The luminescence values of the buffer rows should always be quite similar (Fig. 7A), although a small shift can be observed.If the shift becomes large, the error bars of the entire assay will increase, because the buffer wells are used for normalization.Addition of Ras should lower luminescence values, as Ras activity reduces GTP concentration (Fig. 7B).With increasing Ras concentrations, the luminescence should drop more.Replicates/wells containing the same Ras concentration should have the same luminescence.The bigger the spread between replicates, the bigger the error will be.In some cases, we observe a small decrease of luminescence over time.As long as this decrease is not drastic and occurs in all wells to roughly the same extent, it will not impact the analysis.Assays in which the luminescence decreases drastically within 20 min should  Use a buffer that does not contain chelating agents be discarded from further analysis, as well as wells that behave differently than the majority.
According to the assay manual, luminescence should be measured after 5 min.We average the luminescence values from 5-20 min for every assay in order to smooth out possible small deviations in the luminescence curve.As long as the time window begins after 5 min, the exact time window is less important than using the same time window for every assay.
The data from this assay can be used to compare the activity of different GTPases, the activity of two different effectors on a single GTPase, or the activity of one effector on different GTPases, as long as all samples are part of the same assay (to avoid small variability between assays) and are used at the same concentration.
GTPase cycling rates k 1 , k 2 .To increase comparability of GTPase and effector activities, we developed a GTPase cycling model (Supporting Information S8).Support Protocol 2 describes how the model can be used to fit data for Ras.In short, we fit an exponential to the amount of remaining GTP (Supporting Information S8) using Eqn. 1, where [Ras] is used for [GTPase].Here, K refers to the overall GTP hydrolysis rate and depends on the concentration of Ras.GTPase cycling rates k describe the rate of the entire GTPase cycle of Ras.The model fits two cycling rates (k 1 and k 2 ) for every GTPase.K 1 accounts for the linear contribution of the GTPase (K 1 = k 1 [GTPase]), while K 2 accounts for the quadratic contribution (K 2 = k 2 [GTPase] 2 ), which can be due to dimerization or other cooperative effects increasing the activity of the enzyme (for examples, see Zhang et al., 1999).The bigger the K 2 /K 1 ratio, the bigger the contribution of the nonlinear term.This can suggest that the GTPase forms a dimer that is more active than the monomer, but additional studies examining GTPase di-and oligomerization are required before making such a deduction.The data indicate that some effect leads to a nonlinear concentration dependence   of K, but does not reveal its origin.Ras does not show a linear concentration dependence (Fig. 8A) and is dominated by k 2 (Fig. 8B; k 2 >>> k 1 ).In this case, this is likely due to dimerization, as Ras GTPases are known to dimerize.
Output file structure and pooling of k rates.Estimates of Ras cycling rates k 1 and k 2 are summarized in Data_summary.xlsx(Fig. 8B).The file is structured in the following way: The Run column shows the assay number and is followed by two Varied Protein

of 29
Current Protocols columns stating which proteins were varied in the assay.If the assay contains only one varied protein, the second column is empty (as for this example).Next, Parameter 1 is named along with its values, 95% upper and lower bounds, and standard error.The same is repeated if there are additional parameters (Parameter 2 is shown).The last column lists the adjusted R 2 of the fit and a Boolean ('Used for pooling') if the experiment is used for pooling.For each assay (i.e., each run: E1, E2), the rate value estimates are shown.Below, in the 'pooled' row, the pooled values of k estimates are listed (details on weighting of k for pooling and error propagation are provided in Supporting Information S11).
An experiment can be excluded from pooling if c corr or standard errors of k are out of range.Decision criteria are set before analyzing the data using two parameters (see Support Protocol 2): (1) conc_corr_bounds states the range of c corr that allows assays to be included in pooling, and (2) k_low_err_filt.Ras states if assays with low standard errors (1 -10 or lower) on values are excluded from pooling.In Support Protocol 2 (example 1), we used conc_corr_bounds = [0.5 1.5]; k_low_err_filt.Ras = false; for analyzing Ras data.Although conc_corr_bounds states the range of c corr that allows assays to be included in the pooling, it does not influence pooling of k 1 and k 2 , as c corr is only used for fitting k 3,X (Eqns.2,3).We set k_low_err_filt.Ras to false so experiments with low standard errors on k values are not excluded from pooling.This makes sense for Ras, as Ras is dominated by k 2 and has k 1 ≈ 0 (and thus k 1 errors are close to 0).

Example 2
For this example, data are provided (spreadsheet: example2-matlab.xlsx in the Data folder; see Data Availability), for the following set of GTPase assays: (A) Cdc42 (GTPase) dilutions (B) Cdc24 (GEF) dilutions using a constant Cdc42 concentration (C) Rga2 (GAP) dilutions using a constant Cdc42 concentration (D) Cdc24 and Rga2 dilutions using a constant Cdc42 concentration The analysis is described in Support Protocol 2. In short, a GTPase cycling model (Supporting Information S8) is fitted using Eqn. 4. Here, K refers to the overall GTP hydrolysis rate and depends on the concentrations of Cdc42 and effectors Cdc24 and Rga2.First, the data for Cdc42 serial dilutions (group A) are fitted with an exponential to obtain estimates of GTPase cycling rates k 1 and k 2 for each experiment.These estimates are then bundled to get a single pooled estimate for k 1 and k 2 .Next, Cdc42:Cdc24, Cdc42:Rga2, and Cdc42:Cdc24:Rga2 mixtures (groups B, C, and D) are fitted to obtain estimates for the correction factors c corr and cycling rates k 3 (measuring the impact of the effector on GT-Pase cycling) for every experiment.These latter estimates can then be pooled to obtain a single estimate for the k 3 values (a pooled estimate for c corr has no function here).
Correction factors c corr .The correction factor c corr accounts for variability between assays, i.e., for the observation that the rates for the GTPase can vary between assays.Possible reasons for this include (1) small concentration differences introduced by pipetting of small volumes, (2) temperature and shaker speed fluctuations during incubation, (3) small changes in effective GTPase concentration through sequestration of GT-Pases in complexes with effectors, and/or (4) intrinsic changes in protein activities due to other external conditions.c corr maps all factors that lead to variations between assays onto the GTPase concentration.The correction factor is estimated for each experiment individually, relative to a predefined GTPase serial dilution assay.This is done by matching the overall hydrolysis rate at zero effector concentration inferred from the fit to the estimated overall hydrolysis rate based on the GTPase serial dilution assay.Concretely, c corr follows from: where k1 and k2 are estimates of cycling rates from the GTPase serial dilution assay and Kmix is the estimated median of the overall hydrolysis rate fit at zero effector concentration (i.e., Kmix = Kmix ([X] = 0)).By using random draws of the fitted overall hydrolysis rate distribution, we can obtain a distribution for c corr as well.Consequently, c corr = 1.0 means that the GTPase in this assay has the same intrinsic activity relative to the serial dilution assay with the same GTPase (reference assay), which is optimal.Similarly, c corr = 0.5 means that the GTPase in this assay is 50% less active relative to the reference assay, and c corr = 1.5 means that the GTPase in this assay is 50% more active.Very low correction values indicate that the GTPase is behaving differently in the assay than before (possibly because the conditions are different) or that a fundamental error has occurred (e.g., large pipetting error).To ensure consistency within an assay set, we advise disregarding such assays from further analysis.
To implement this recommendation, we used conc_corr_bounds = [0.5 1.5] for analyzing the data set in Support Protocol 2, example 1.This means that assays yielding c corr values <0.5 and >1.5 will be excluded from pooling. Figure 9  GTPase cycling rates k 3 .GTPase cycling rates k describe the rate of the entire GTPase cycle of the GTPase or the effect of an effector on the overall rate.Cycling rates k 3,X can thus describe how strongly effector X affects the GTPase cycle, but do not reveal which step of the GTPase cycle the effector is acting upon, how strongly the effector is effecting this step, or if multiple steps are affected.To investigate these mechanisms, other GTPase assays need to be conducted (see Background Information).
Assays of group B are used to fit cycling rates k 3,Cdc24 , which describe the effect of Cdc24 on the overall GTPase cycle of Cdc42.We use a quadratic fit (K 3,Cdc24 = k 3,Cdc24 [Cdc42][Cdc24] 2 ) because the data show a nonlinear dependence of the overall rate K on Cdc24 concentration (Fig. 10).The fit is a phenomenological description of the data and does not reveal the origin of the nonlinearity.One possibility could be dimerization, with Cdc24 dimers exhibiting increased activity (Tschirpke, Daalman et al., 2023) These rates should be close to those obtained in assays using only Cdc42:Cdc24 or Cdc42:Rga2 (Fig. 12).A large difference can indicate a problem with the fit or the assay.The cycling rate k 3,Cdc24,Rga2 accounts for any potential synergy between the two effectors.When k 3,Cdc24,Rga2 = 0, there is no interaction between the proteins.When k 3,Cdc24,Rga2 < 0, the proteins antagonize/inhibit each other, and when k 3,Cdc24,Rga2 > 0 there is synergy.Again, any interaction between the proteins refers to their combined effect on the overall rate.(The   significance of the deviation from zero is also tested using Ruxton, 2006).The p value for the null hypothesis of zero interaction is stored in the Data_assays.matoutput file, in the Assays_processed variable, in the field of that assay, in the subfield 'k_tot', under 'Interaction_p'.While a positive k 3,Cdc24,Rga2 indicates that there is some synergy, it does not reveal the origin of the synergy.It could be due to physical proteinprotein interactions or to rate-limiting steps in the cycle that are relieved when both proteins are added (Tschirpke, Daalman et al., 2023).One must thus be careful in interpretating this rate.
Comparison of cycling rates k.Overall GTP hydrolysis rates K are concentration dependent and have the unit hr −1 .Cycling rates k are concentration-independent, but the different rates have different units (μM −1 hr −1 for k 1 , μM −2 hr −1 for k 2 , μM −3 hr −1 for k 3,Cdc24 , μM −2 hr −1 for k 3,Rga2 , and μM −4 hr −1 for k 3,Cdc24,Rga2 ), making them difficult to com-pare.The easiest way to represent all rates in one plot is to plot overall GTP hydrolysis rates K for 1 μM of each protein (Fig. 12).However, one has to consider the different concentration dependencies when interpreting/comparing K rates.A rate that scales quadratically with protein concentration will be twice as big as a rate that scales linearly when the protein concentration is doubled.Some rates may also be valid only for the regime in which they were fitted in (e.g., they may show saturation in higher concentration regimes).Thus, one must consider which protein concentrations are used to calculate K in order to make cycling rates k comparable.

Time Considerations
GTPase assays usually take a few hours, some of which is consumed by incubation steps.For an assay involving eight samples (including six serial dilutions of GTPase) and using an incubation time of 1.5 hr, we estimate 30 min for preparation of materials and solutions, 2 hr for conducting the assay, and another 30 min for the luminescence readout.This amounts to a total of 3 hr, including 2 hr for incubation/measurement steps.Assays encompassing more samples may involve longer times (as more protein dilutions need to be prepared).If proteins need to be dialyzed into a suitable buffer first, an additional dialysis step is required.The analysis time depends on the level of automatization and can range from a few minutes to 30-40 min.If the first step is done manually in a spreadsheet editor, ∼30 min are needed for this step.The subsequent analysis in Matlab requires only a few minutes of runtime.

Figure 2
Figure 2 Schematic of protein dilutions required to conduct a GTPase assay set investigating GTPase, GEF, and GAP activity and interactions (example 2).

4.Figure 4
Figure 4 Data analysis part 1 for the assay in the Basic Protocol: time-averaging of the luminescence values highlighted in green.Data from example1.xlsx,tab: 'E1'.

Figure 5
Figure 5 Data analysis part 2 for the assay in the Basic Protocol: calculating the amount of remaining GTP from luminescence values.Data from example1.xlsx,tab: 'E1'.

Figure 6
Figure 6 Reformatting of data from Figure 5 using the Python script provided in Supporting Information S9.Data taken from example1.xlsx,tab: 'E1'.

Figure 7
Figure 7 Example 1. Measured luminescence values of buffer wells (A) and Ras dilutions (B) over the time course of the measurement.The green box indicates values that are used for averaging.

Figure 8
Figure 8 Example 1. (A) The overall GTP hydrolysis rate K does not scale linearly with Ras concentration.Figure generated automatically when running the Matlab code using plot_fits = true(1).(B) GTPase cycling rates k 1 and k 2 for Ras.

Figure 9
Figure 9 Example 2. Histogram of c corr values using a bin size of 0.1.Generated for all c corr values of example 2 using Plot_c_corr_histogram.m (Supporting Information S10) and example2histogram.xlsx in the example2 matlab output folder.
shows a histogram of all correction factors.Apart from one assay that yielded c corr ≈ 0.1, all c corr values were within range, with most values being close to 1. Data_summary.xlsxstates correction factors as Parameter 1 for assays in which Cdc24 and/or Rga2 effectors are varied (Varied Proteins 1 and 2).Correction factor errors are discussed in Supporting Information S11.

Figure 10
Figure 10 Example 2. The overall GTP hydrolysis rate K scales nonlinearly with Cdc24 concentration.Figure generated automatically when running the Matlab code using plot_fits = true(1).

Table 1
Summary of How Example Assays 1 and 2 are Used throughout the Protocols a a Data files available at data.4tu.nl(see Data Availability).

Table 2
Pipetting Scheme for 2× GTP Solution

Table 3
Pipetting Scheme for Glo Solutions Current Protocols from pooling if c corr values or standard errors of are out of range.Decision criteria are set in the code using two parameters:

Table 4
Troubleshooting Guide for GTPase Glo Assays