Kinetic and Structural Characterization of the Self-Labeling Protein Tags HaloTag7, SNAP-tag, and CLIP-tag

The self-labeling protein tags (SLPs) HaloTag7, SNAP-tag, and CLIP-tag allow the covalent labeling of fusion proteins with synthetic molecules for applications in bioimaging and biotechnology. To guide the selection of an SLP–substrate pair and provide guidelines for the design of substrates, we report a systematic and comparative study of the labeling kinetics and substrate specificities of HaloTag7, SNAP-tag, and CLIP-tag. HaloTag7 reaches almost diffusion-limited labeling rate constants with certain rhodamine substrates, which are more than 2 orders of magnitude higher than those of SNAP-tag for the corresponding substrates. SNAP-tag labeling rate constants, however, are less affected by the structure of the label than those of HaloTag7, which vary over 6 orders of magnitude for commonly employed substrates. Determining the crystal structures of HaloTag7 and SNAP-tag labeled with fluorescent substrates allowed us to rationalize their substrate preferences. We also demonstrate how these insights can be exploited to design substrates with improved labeling kinetics.


Supplementary figures
S28-49 Figure S1: Chemical structure of SLP substrates S28-29 Figure S2: Labeling kinetics of HT7 with fluorescent CA substrates S30 Figure S3: Comparison of model 1 and model 2 fitted to HT7 labeling kinetics S31 Figure S4: Modeling of HT7 labeling kinetics using measured parameters to compare the kinetic models 1 and 2 S32 Figure S5: Labeling kinetics of HT7 and HOB with CA-TMR and CA-Alexa488 S33 Figure S6: Rate and equilibrium constants of HT7 labeling with various fluorescent CA substrates S34 Figure S7: Affinity of the dead mutant HT7 D106A to fluorescent CA substrates S35 Figure S8: Labeling kinetics of HT7 with non-fluorescent CA substrates S36 Figure S9: Additional structural information on HaloTag proteins S37 Figure S10: Biochemical study of the interaction of HT7 with chloroalkane-fluorophores S38 Figure S11: Labeling kinetics of SNAP with fluorescent BG and CP substrates S39 Figure S12: Labeling kinetics of SNAP measured by stopped flow anisotropy S40 Figure S13: Comparison of fluorophore substrate affinities between the dead mutants SNAP C145A and SNAPf C145A S41 Figure S14: Comparison of non-derivatized core substrate affinities to the dead mutant SNAP C145 S42 Figure S15: Sequence and additional structural information related to SNAP. S43 Figure S16: Labeling kinetics of SNAPf with fluorescent BG and CP substrates S44 Figure S17: Labeling kinetics of CLIP and CLIPf with fluorescent BC S45 Figure S18: Labeling kinetics of hAGT, SNAP and CLIP with the non-respective BG-, CP-and BC-TMR substrates S46 Figure S19: Labeling kinetics of SNAP with non-fluorescent BG and CP substrates S47 Figure S20: Labeling kinetics of SNAP and SNAPf with BG-5-TMR and BG-5-CPY S48 Figure S21: Correlation between SNAPf labeling kinetics and substrate affinity S48 Figure S22: Labeling kinetics of SsOGT-H 5 with BG-Alexa488 and BG-TMR S49

Supplementary tables
S50-52 Table S1: Data collection and refinement statistics the X-ray crystal structures S50 Table S2: Kinetic parameters of HT7 labeling with fluorescent CA substrates S51 Table S3: Comparison kapp of HT7 labeling kinetics analyzed using models 1 and 2 S51 Table S4: Comparison of HT7 and HOB labeling kinetics with fluorescent CA substrates S51 Table S5: Kinetic parameters of SNAP and SNAPf labeling with fluorescent substrates analyzed using model 1.2 S52 Table S6: Kinetic parameters of SNAP labeling with BG-/CP-TMR measured via stopped flow S52 Table S7: Comparison of SNAP/CLIP with SNAPf/CLIPf labeling kinetics with fluorescent substrates S52 Table S8: Comparison of SNAP labeling with 5-and 6-fluorophores. S52 Table S9: Kinetic parameters of SsOGT-H 5 labeling S52

biotin-1)
The compound was synthesized according to the procedure from Masharina et al. 2012 15 .

S16
Reaction was conducted according to general procedure A using BG-NH2 and 2-azidoacetic acid (40.7 µmol; 5.7 μL) and 11.1 mg (23.8 μmol) of the desired product were obtained as a colorless TFA-salt in 64% yield.

Nor2)
Reaction was conducted according to general procedure A with a reduced reaction time of 15

S27
To a solution of TMR-5-COOH (2.5 mg, 5.    Figure S4: Modeling of HT7 labeling kinetics using measured parameters to compare the kinetic models 1 and 2. A. Modeling of the fluorescence anisotropy response at different reactant concentrations using model 1 and 2 with parameters determined for HT7 labeling with CA-TMR. At concentrations below Kd (327 nM for CA-TMR) both models yield a rather similar response. At concentrations higher than Kd (1000 nM) the response for model 2 shows a strong biphasic character as observed in the measured data, which is not matching the monoexponential behavior of model 1. At very high concentrations (10000 nM) the response for model 2 is again close to a monoexponential curve but the kinetic is much faster than the model 1 curve. This happens since the rise in fluorescence anisotropy for model 2 in the first milliseconds is not due to covalent reaction but mostly binding (k1). The binding rate constant k1 is faster than kapp if k-1 is not zero (kapp = k1 * k2 / (k2 + k-1)). Hence directly estimating kapp from fluorescence anisotropy traces by fitting model 1 to the data is only valid for concentrations below Kd or if k-1 << k2. B. Modeling the formation of covalently labeled product at different reactant concentrations using model 1 and 2 with parameters determined for HT7 labeling with CA-CPY. At concentrations below Kd (46 nM for CA-CPY) both models yield a rather similar behavior. At higher concentrations model 1 predicts a much faster product formation than model 2 since it does not account for enzyme saturation. C. Plot of the apparent first order reaction rate constant for product formation against substrate concentration for model 1 and 2 with parameters for CA-CPY. In contrast to model 1, model 2 accounts for enzyme saturation leading to a maximum reaction rate of kmax = k2 = 9.9 s -1 . The models start do diverge significantly once the substrate concentration exceeds Kd (46 nM). As a consequence, model 2 should be used for predicting formation of labeled HT7 if labeling is performed at high concentrations.  A. Alkane-TMR constraints by the crystal packing. Two monomers of HT7-TMR are displayed in grey and light-pink. The alkane-TMR (orange sticks) conformation is constrained by the light-pink monomer that was generated as a symmetry mate. A zoom is shown on the right panel. B. Structural comparison between 6U32 HT7-TMR (previously published 16 ) and 6Y7A HT7-TMR (PDB ID 6Y7A, this study). A zoom into the binding site with hydrogen bonds between 6Y7A HT7-TMR and 6U32 HT7-TMR and their respective reacted substrates are represented as black and dark-purple lines, respectively. C. Zoom into isolated catalytic aspartate and alkane-TMR substrate from both 6Y7A HT7-TMR and 6U32 HT7-TMR crystal structures. D. Structural comparison between HT7-TMR and HOB-TMR. Hydrogen bonds between HT7 and HOB and their respective reacted substrates are represented as black and gold dashed lines, respectively. The blue spheres represent differences between HT7 and HOB caused by the mutations.      Data analyzed using model 2.    Data analyzed using model 2  Data analyzed using model 1 or 1.2 (*) which included an additional phase (see Table S5).   [data] Delay 0.022 offset 0.0262 directory p ath/to/data sheet data.csv column 6 | conc P = 1 | conc S = 1 | resp onse Z = 1 * R | resp onse P.S = 1 * R | lab el c=1 column 5 | conc P = 0.75 | conc S = 0.75 | resp onse Z = 1.333 * R | resp onse P.S = 1.333 * R | lab el c=0.75 column 4 | conc P = 0.5 | conc S = 0.5 | resp onse Z = 2 * R | resp onse P.S = 2 * R | lab el c=0.5 column 3 | conc P = 0.25 | conc S = 0.25 | resp onse Z = 4 * R | resp onse P.S = 4 * R | lab el c=0.25 column 2 | conc P = 0.1 | conc S = 0.1 | resp onse Z = 10 * R | resp onse P.S = 10 * R | lab el c=0.