Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Aug 5;15(1):6645.
doi: 10.1038/s41467-024-50647-9.

Rescaling protein-protein interactions improves Martini 3 for flexible proteins in solution

F Emil Thomasen #  1 , Tórur Skaalum #  2 , Ashutosh Kumar  3   4 , Sriraksha Srinivasan  3 , Stefano Vanni  5   6 , Kresten Lindorff-Larsen  7
Affiliations

Rescaling protein-protein interactions improves Martini 3 for flexible proteins in solution

F Emil Thomasen et al. Nat Commun. .

Abstract

Multidomain proteins with flexible linkers and disordered regions play important roles in many cellular processes, but characterizing their conformational ensembles is difficult. We have previously shown that the coarse-grained model, Martini 3, produces too compact ensembles in solution, that may in part be remedied by strengthening protein-water interactions. Here, we show that decreasing the strength of protein-protein interactions leads to improved agreement with experimental data on a wide set of systems. We show that the 'symmetry' between rescaling protein-water and protein-protein interactions breaks down when studying interactions with or within membranes; rescaling protein-protein interactions better preserves the binding specificity of proteins with lipid membranes, whereas rescaling protein-water interactions preserves oligomerization of transmembrane helices. We conclude that decreasing the strength of protein-protein interactions improves the accuracy of Martini 3 for IDPs and multidomain proteins, both in solution and in the presence of a lipid membrane.

PubMed Disclaimer

Conflict of interest statement

K.L.-L. holds stock options in and is a consultant for Peptone Ltd. All other authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Expected effects of proposed force field modifications.
Schematic overview showing the expected effects of rescaling protein-water and protein-protein interactions in Martini 3. Overestimated compactness of soluble IDPs and multidomain proteins and specific membrane interactions for peripheral membrane proteins have previously been reported,.
Fig. 2
Fig. 2. Starting structures for simulations of multidomain proteins.
Starting structures of multidomain proteins used for Martini simulations. See the Methods section for a description of the source of the structures and how they were assembled.
Fig. 3
Fig. 3. Agreement between simulations and SAXS data for multidomain proteins.
Mean radii of gyration (Rg) calculated from simulations plotted against Rg determined from Guinier fits to SAXS data for (a) simulations with unmodified Martini 3, (b) simulations with protein-water interactions in Martini 3 rescaled by λPW = 1.10, and (c) simulations with protein-protein interactions in Martini 3 rescaled by λPP = 0.88. Error bars on the experimental Rg-values were determined using Atsas AUTORG and represent a fitting error from the Guinier fit plus a standard deviation over possible selections of the Guinier region. The diagonal is shown as a dashed line and a linear fit with intercept 0 weighted by experimental errors is shown as a solid line. Pearson correlation coefficients (rP) with standard error of the mean from bootstrapping with 9999 resamples are shown on the plots. d Reduced χ2 to experimental SAXS intensities given by SAXS intensities calculated from unmodified Martini 3 simulations (blue) and Martini 3 simulations with protein-water interactions rescaled by λPW = 1.10 (orange) or protein-protein interactions rescaled by λPP = 0.88 (green). Mean and standard error of the mean over all 15 proteins are shown on the plot. Note the logarithmic scale for χr2. e Representative conformation of TIA1 with an Rg corresponding to the average Rg in (blue) simulations with unmodified Martini 3, (orange) simulations with protein-water interactions in Martini 3 rescaled by λPW = 1.10, and (green) simulations with protein-protein interactions in Martini 3 rescaled by λPP = 0.88. Simulations of hnRNPA1, hisSUMO-hnRNPA1 and TIA1 with λPW = 1.10 were taken from Thomasen et al.. Source data are provided as a Source Data file.
Fig. 4
Fig. 4. Agreement between simulations and SAXS or PRE data for IDPs.
Mean radii of gyration (Rg) calculated from simulations plotted against Rg determined from Guinier fits to SAXS data for (a) simulations with unmodified Martini 3, (b) simulations with protein-water interactions in Martini 3 rescaled by λPW = 1.10, and (c) simulations with protein-protein interactions in Martini 3 rescaled by λPP = 0.88. Error bars on the experimental Rg-values were determined using Atsas AUTORG and represent a fitting error from the Guinier fit plus a standard deviation over possible selections of the Guinier region. The diagonal is shown as a dashed line and a linear fit with intercept 0 weighted by experimental errors is shown as a solid line. Pearson correlation coefficients (rP) with standard error of the mean from bootstrapping with 9999 resamples are shown on the plots. d Reduced χ2 to experimental SAXS intensities given by SAXS intensities calculated from unmodified Martini 3 simulations (blue) and Martini 3 simulations with protein-water interactions rescaled by λPW = 1.10 (orange) or protein-protein interactions rescaled by λPP = 0.88 (green). Mean and standard error of the mean over all 12 proteins are shown on the plot. Note the logarithmic scale for χr2. e Reduced χ2 to experimental PRE NMR data given by unmodified Martini 3 simulations (blue) and Martini 3 simulations with protein-water interactions rescaled by λPW = 1.10 (orange) or protein-protein interactions rescaled by λPP = 0.88 (green). Note the logarithmic scale for χr2. f Representative conformation of TauK25 with an Rg corresponding to the average Rg in (blue) simulations with unmodified Martini 3, (orange) simulations with protein-water interactions in Martini 3 rescaled by λPW = 1.10, and (green) simulations with protein-protein interactions in Martini 3 rescaled by λPP = 0.88. All simulations with λPW=1.10 were taken from Thomasen et al.. Source data are provided as a Source Data file.
Fig. 5
Fig. 5. Protein self-association in solution.
a Fraction bound calculated from MD simulations of two copies of the folded proteins ubiquitin and villin HP36, and two copies of the IDPs α-synuclein, p15PAF, hTau40, and FUS LCD, with unmodified Martini 3 (blue), Martini 3 with protein-water interactions rescaled by λPW = 1.10 (orange) (taken from Thomasen et al.), and Martini 3 with protein-protein interactions rescaled by λPP = 0.88 (green). Box plots show the results of 10 replica simulations. The bound fraction in agreement with Kd = 4.9 mM for ubiquitin self-association is shown as a dashed line. The bound fraction in agreement with Kd > 1.5 mM for villin HP36 self-association is shown as a shaded gray area. α-synuclein, p15PAF, and hTau40 should not self-associate under the given conditions based on PRE, or SEC-MALLS data, while FUS LCD should transiently self-associate based on PRE data. Boxplots show the first quartile, median, and third quartile; whiskers extend from the box to the farthest data point lying within 1.5 times the inter-quartile range from the box, and points outside the whiskers are shown individually. b Interchain PREs calculated from the simulations of two copies of FUS LCD from panel a and comparison with experimental PREs (black). PREs are shown for the three spin-label sites at residues 16, 86, and 142 marked with dashed black lines. The rotational correlation time, τc, was selected individually for each λ to minimize χr2. Error bars represent the standard error of the mean over 10 replica simulations. c χr2 between calculated and experimental PRE data for two copies of FUS LCD shown in panel (b). Source data are provided as a Source Data file.
Fig. 6
Fig. 6. Protein-membrane interactions.
MD simulations (four replicas, each 3 μs long) were performed for peripheral membrane proteins, multidomain proteins, and intrinsically disordered regions with appropriate membrane composition (see Methods for details). Simulations were performed with unmodified Martini 3 (blue), protein-water interactions in Martini 3 rescaled by λPW = 1.10 (orange), and protein-protein interactions in Martini 3 rescaled by λPP = 0.88 (green). For each system, the corresponding atomistic structure of the protein is shown on top. Boxplots show the first quartile, median, and third quartile; whiskers extend from the box to the farthest data point lying within 1.5 times the inter-quartile range from the box, and points outside the whiskers are shown individually. Source data are provided as a Source Data file.
Fig. 7
Fig. 7. Radii of gyration of mTurq-mNeon and hnRNPA1LCD variants.
a Mean radii of gyration (Rg) calculated from simulations with unmodified Martini 3 (blue), Martini 3 with protein-water interactions rescaled by λPW = 1.10 (orange), and Martini 3 with protein-protein interactions rescaled by λPP = 0.88 (green) are plotted against Rg determined by SAXS for mTurquoise2 and mNeonGreen connected by a linker region with the insertion of 0, 8, 16, 24, 32, or 48 GS repeats. Error bars on the experimental Rg-values were determined using Atsas AUTORG and represent a fitting error from the Guinier fit plus a standard deviation over possible selections of the Guinier region. b Rg calculated from simulations with unmodified Martini 3 (blue) and Martini 3 with protein-protein interactions rescaled by λPP = 0.92 (pink) or λPP = 0.88 (green) are plotted against Rg determined by SAXS for wild-type hnRNPA1LCD and six sequence variants with varied composition of charged and aromatic residues. Error bars on the experimental Rg-values represent the fitting error of the molecular form factor model,. We show a zoom-in for each of the force fields along with the given Pearson correlation coefficient with standard error of the mean from bootstrapping with 9999 resamples. Source data are provided as a Source Data file.
Fig. 8
Fig. 8. Transmembrane protein self-association.
a, b Snapshots of simulation systems (left) and corresponding potential of mean force profiles for dimerization (right) for transmembrane domains of EphA1 and ErbB1. The two copies of the protein are shown in green and orange. Lipid headgroups are shown in gray and lipid tails are shown in yellow. Potential of mean force profiles for the transmembrane domains of EphA1 and ErbB1 were calculated from simulations with unmodified Martini 3 (blue), Martini 3 with protein-water interactions rescaled by λPW = 1.10 (orange), and Martini 3 with protein-protein interactions rescaled by λPP = 0.88 (green). Profiles were aligned to zero at the plateau region at r = 3.4 nm, indicated by the dashed black line. The dashed red line corresponds to the experimental values of association free energy (ΔG) from FRET experiments for EphA1 and ErbB1. The errors represents the standard deviation of four profiles calculated from 2 μs blocks. Source data are provided as a Source Data file.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Thomasen, F. E. & Larsen, K. L. Conformational ensembles of intrinsically disordered proteins and flexible multidomain proteins. Biochem. Soc. Trans.50, 541–554 (2022). - PubMed
    1. Bottaro, S. & Lindorff-Larsen, K. Biophysical experiments and biomolecular simulations: a perfect match? Science361, 355 LP – 360 (2018). 10.1126/science.aat4010 - DOI - PubMed
    1. Ingólfsson, H. I. et al. The power of coarse graining in biomolecular simulations. Wiley Interdiscip. Rev. Comput. Mol. Sci.4, 225–248 (2014). 10.1002/wcms.1169 - DOI - PMC - PubMed
    1. Marrink, S. J., Risselada, H. J., Yefimov, S., Tieleman, D. P. & De Vries, A. H. The MARTINI force field: coarse grained model for biomolecular simulations. J. Phys.Chem. B111, 7812–7824 (2007). 10.1021/jp071097f - DOI - PubMed
    1. Monticelli, L. et al. The MARTINI coarse-grained force field: extension to proteins. J. Chem. Theory Comput.4, 819–834 (2008). 10.1021/ct700324x - DOI - PubMed

LinkOut - more resources