CoChat
Sign Up
PreprintCited by 1Protease and Inhibitor MechanismsAlzheimer's disease research and treatmentsProtein Structure and Dynamics

Identification of Substrate-specific Allosteric Fingerprints in Matrix Metalloprotease-1 on Amyloid-beta Peptide Aggregates and Drug Screening With Single Molecule Insights

Open access
Kamboj, Sumaer, Harms, Chase, Wright, Derek, Nash, Anthony, Klein-Seetharaman, Judith, Sarkar, Susanta K.·2022·Research Square
DOI 10.21203/rs.3.rs-1258797/v1
Open this paper in CoChat — free accountOpen PDF

Abstract

Amyloid-beta peptide (Aβ) is the primary component of water-insoluble extracellular plaques, one of the critical hallmarks of Alzheimer's disease (AD). Matrix metalloproteases (MMPs) are broad-spectrum proteases with diverse functions, including interactions with Aβ. Here we report single molecule measurements of MMP1 dynamics on Aβ-induced aggregates by calculating Forster Resonance Energy Transfer (FRET) between two dyes attached to the catalytic and hemopexin domains. We show that the two domains of MMP1 prefer closed conformations on Aβ-induced aggregates, in contrast to the preference for open conformations on collagen fibril, fibrin, and alpha-synuclein aggregates. We approximated the MMP1 dynamics by a two-state Poisson process and determined the kinetic rates of interconversion between the two states from histograms and correlations of FRET values. We performed molecular docking of MMP1 with Aβ using ClusPro, simulated MMP1 dynamics using different docking poses, and matched the experimental and simulated interdomain dynamics to identify an appropriate pose. We used simulations to create a two-dimensional map of correlations between every pair of MMP1 residues, which shows allosteric communications between the two MMP1 domains. We calculated a Gray Level Co-occurrence Matrix from the two-dimensional map of correlations and quantified MMP1 fluctuations by Shannon entropy. We identified the allosteric residues in the hemopexin domain by identifying residues having strong correlations with the catalytic motif residues. We identified that the residues I364, G369, P409, G410, and D418 in MMP1 have Aβ-specific allosteric correlations with the MMP1 catalytic motif by comparing residues for free and Aβ-bound MMP1. We used these Aβ-specific allosteric residues to select small molecule ligands after the virtual screening of molecules against Aβ-bound MMP1. Molecular understanding of interactions between MMP1 and Aβ-induced aggregates and identification of substrate-specific allosteric residues may enable controlling MMP1 function selectively on Aβ.

AI snapshot

Written from the paper’s full text.

TL;DR

Single-molecule FRET and simulations show that MMP1 prefers closed interdomain conformations on Aβ-induced aggregates, unlike its open preference on collagen, fibrin, and alpha-synuclein aggregates. The study identifies Aβ-specific allosteric residues and uses them to guide virtual screening of FDA-approved compounds.

Key findings

  • On Aβ-induced aggregates, active MMP1 without ligand has two states S1=0.47 and S2=0.52, with autocorrelation decay rate 0.02 s⁻¹; by comparison, collagen gives S1=0.44/S2=0.55 and 0.13 s⁻¹, fibrin S1=0.42/S2=0.51 and 0.08 s⁻¹, and alpha-synuclein aggregates S1=0.46/S2=0.52 and 0.08 s⁻¹. Tetracycline shifts both active and inactive MMP1 toward closed conformations.
  • Docking and MD simulations suggest pose 3 is functionally relevant, with MMP1 binding between Aβ residues V12 and L17; active MMP1 prefers closed conformations more than inactive MMP1, and catalytic pocket opening positively correlates with interdomain separation for active MMP1 in poses 1 and 3.
  • Allosteric analysis identified residues I364, G369, P409, G410, and D418 in the hemopexin domain as having Aβ-specific correlations >0.8 with the catalytic motif HELGHSLGLSH, common across three repeats of Aβ-bound active MMP1.
  • Virtual screening of ~9000 FDA-approved ZINC15 compounds yielded ~600 drugs with at least 10% difference between top two binding modes, then 23 lead molecules binding within ±2 residues of the Aβ-specific allosteric residues; Trosec had the highest predicted affinity and bound near K362.

Methods

  • smFRET with TIRF microscopy: active E219 and inactive Q219 MMP1 mutants with S142C/S366C labeled with Alexa555/Alexa647, measured on a thin layer of Aβ-induced aggregates; FRET was calculated as I_A/(I_A+I_D), and histograms/autocorrelations were fit to a two-state Poisson process.
  • All-atom MD: ClusPro docking of MMP1 (PDB 4AUO) and Aβ (PDB 1IYT) produced three top poses; Gromacs 2021.2 with Amber99 and SPC/E water, 20 ns production runs, 2 fs steps, sampled every 5 ps, three repeats at 22 °C or 37 °C; interdomain S142-S366 and catalytic pocket N171-T230 distances, residue-pair correlation matrices, and GLCM Shannon entropy were analyzed.
  • Virtual screening: AutoDock Vina against free MMP1 and three Aβ-bound poses, using ~9000 FDA-approved molecules from ZINC15; compounds were selected by ≥10% affinity difference between top two modes and proximity to Aβ-specific allosteric residues.

Limitations

  • Stated: experiments were performed on Aβ aggregates while simulations used Aβ monomers; no crystal structure exists for Aβ-bound MMP1, so binding poses were predicted; computational binding energy may not be the best indicator of appropriate docking poses; more studies are needed to explain substrate-dependent open/closed preferences; not all identified allosteric residues have associated lead molecules.
  • Evident from the text: MD simulations were short (20 ns) with three repeats and only three top docking poses; virtual screening used exhaustiveness 10 for quicker screening, and no experimental validation of the lead molecules is reported.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.21203/rs.3.rs-1258797/v1

Cites

51 of 51 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 14 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    In the presence of tetracycline, both active and inactive MMP1 prefer closed conformations more on Aβ-induced aggregates.

    “In the presence of tetracycline, a known MMP inhibitor, both active and inactive MMP1 prefer the closed conformations more (Figure 1F).”
    Quote in the paperSubstantial claim
  • Findingown data

    Residues I364, G369, P409, G410, and D418 in MMP1 have Aβ-specific allosteric correlations with the MMP1 catalytic motif.

    “Comparing the residues between Aβ-bound active MMP1 with free active MMP1, we identi ed that the residues I364, G369, P409, G410, and D418 are common among the three repeats of Aβ-bound active MMP1 simulations and have exclusive correlations for Aβ-bound active MMP1.”
    • Also stated in: § Abstract · plain
      “We identi ed that the residues I364, G369, P409, G410, and D418 in MMP1 have Aβ-speci c allosteric correlations with the MMP1 catalytic motif by comparing residues for free and Aβ-bound MMP1.”
    Quote in the paperSubstantial claim
  • Findingown data

    The two states are S1=0.47 and S2=0.52 on Aβ-induced aggregates for active MMP1 without ligand.

    “The two states are S1=0.47 and S2=0.52 on Aβ-induced aggregates for active MMP1 without ligand (Table S1).”
    Quote in the paperSubstantial claim
  • Findingown data

    Further selection identified 23 lead molecules binding within ±2 residues around the Aβ-specific allosteric residues (I364, G369, P409, G410, and D418).

    “Then, we further selected drugs by identifying drugs binding within ±2 residues around the Aβ-speci c allosteric residues (I364, G369, P409, G410, and D418) to only 23 lead molecules.”
    Quote in the paperSubstantial claim
  • Findingown data

    Virtual screening of approximately 9000 FDA-approved drugs selected about 600 drugs with at least 10% difference between binding affinities at the top two binding modes.

    “To this end, we screened ~9000 FDA-approved drugs (see methods) from the ZINC15 database and selected compounds for which the difference between the binding a nities at the top two binding modes is at least 10% because we want to nd drugs that bind preferably to a speci c site on MMP1.With a 10% difference, we found ~600 drugs out of ~9000 starting drugs.”
    Quote in the paperSubstantial claim
  • May combine two claims. 15
  • Classified as probably not a substantial claim. 16

  • May combine two claims.

    Findingown data

    Trosec has the highest predicted affinity and binds to residue K362.

    “Figure 6E shows how Trosec, the drug with the highest predicted a nity, binds to residue K362.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Virtual screening results change upon binding Aβ and depend on the binding pose.

    “We found that virtual screening results change upon binding Aβ and depend on the binding pose.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Active MMP1 prefers closed conformations more than inactive MMP1 in simulations, in agreement with experiments.

    “Figures 3D-F show the distributions of S142-S366 distance, which show that active MMP1 prefers closed conformations (high FRET) more than inactive MMP1, in agreement with experiments (Figure 1E).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    A sum of two Gaussians fits the FRET histograms, suggesting an equilibrium of the open and closed conformations.

    “A sum of two Gaussians ts the histograms, suggesting an equilibrium of the open and closed conformations.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The presence of noise increases the width of conformational histograms but subtly affects autocorrelations.

    “The presence of noise increases the width of conformational histograms but subtly affects autocorrelations.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    Both exponential and power law fit autocorrelations of simulated FRET values without noise for active and inactive MMP1.

    “Both exponential and power law t autocorrelations (active: dashed red line; inactive: solid green line).”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: The two-state model adequately describes MMP1 dynamics on Aβ.

  • Classified as probably not a substantial claim.

    Findingown data

    Substrate- and pose-dependent virtual screening of molecules against MMP1.

    “Substrate -and pose-dependent virtual screening of molecules against MMP1.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: AutoDock Vina docking scores reflect true binding affinities.

  • Classified as probably not a substantial claim.

    Findingown data

    MMP1-Aβ binding pose-dependent dynamics of MMP1 at 22 °C.

    “MMP1-Aβ binding pose-dependent dynamics of MMP1 at 22 °C.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The three top-scoring ClusPro binding poses between MMP1 and Aβ represent the biologically relevant binding modes.

  • Classified as probably not a substantial claim.

    Findingown data

    Binding affinities of lead compounds and closest MMP1 residues for Aβ-bound MMP1 in poses 1, 2, and 3.

    “Binding a nities of lead compounds and closest MMP1 residues for Aβ-bound MMP1 in (A) pose 1, (B) pose 2, and (C) pose 3.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: AutoDock Vina docking scores reflect true binding affinities.

  • Classified as probably not a substantial claim. May combine two claims.

    Findingnone given

    Conformational changes of the MMP1 catalytic motif and identification of Aβ-specific allosteric residues at 37 °C for pose 3.

    “Conformational changes of the MMP1 catalytic motif and identi cation of Aβ-speci c allosteric residues at 37 °C for pose 3.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The three top-scoring ClusPro binding poses between MMP1 and Aβ represent the biologically relevant binding modes.

  • Classified as probably not a substantial claim.

    Findingown data

    Interdomain dynamics of MMP1 on Aβ-induced aggregates at 22 oC with 100 ms time resolution.

    “Interdomain dynamics of MMP1 on Aβ-induced aggregates at 22 ºC with 100 ms time resolution.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Findingown data

    Quantification of MMP1 allostery on Aβ at 22 °C for pose 3.

    “Quanti cation of MMP1 allostery on Aβ at 22 °C for pose 3.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The three top-scoring ClusPro binding poses between MMP1 and Aβ represent the biologically relevant binding modes.

  • May combine two claims.

    Findingown data

    The decay rate of autocorrelations does not change as the noise level varies, and an exponential fit recovers the underlying decay rate with and without noise.

    “The decay rate of autocorrelations does not change as the noise level varies, and an exponential t recovers the underlying decay rate with and without noise.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In the absence of noise, both power-law and exponential distributions fit the autocorrelations; with noise, only an exponential distribution fits.

    “In the absence of noise, both powerlaw and exponential distributions t the autocorrelations (Figure 2F).However, with noise, the y-intercept of the autocorrelations decreases, and only an exponential distribution ts the autocorrelations.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Simulations of MMP1 dynamics on a collagen monomer agreed with experiments on collagen fibril when the collagen backbone was restrained, suggesting strains in collagen monomers inside collagen fibrils.

    “We took a similar approach for collagen and found that simulations of MMP1 dynamics on a collagen monomer agreed with experiments on collagen bril when the collagen backbone was restrained (34), suggesting the presence of strains in collagen monomers inside collagen brils.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [34]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The catalytic domain sequence across the MMP family is very similar, but the hemopexin domain sequence varies, suggesting a role of communications in substrate diversity and specificity.

    “The catalytic domain sequence across the MMP family is very similar, but the hemopexin domain sequence varies (31,32), suggesting a role of communications in substrate diversity and speci city (12,33).”
    Quote in the paperNot a substantial claim0 of 4 sources checked
    • cites [31]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites [32]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites [12]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites [33]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Study design and conductown data

    The two-state Poisson process recovers underlying parameters reasonably well despite noise; histograms and autocorrelations recover the underlying parameters.

    “Underlying noise can obscure the individual states (Figures 2A-B), but histograms (Figures 2C and 2E) and autocorrelations (Figures 2D and 2F) still recovers the underlying parameters reasonably well (Table S2).”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Inactive MMP1 acts as a control to distinguish the effects of enzymatic activity from the effects of the microenvironment and photophysical properties of the dyes.

    “Inactive MMP1 acts as a control to distinguish the effects of enzymatic activity from the effects of the microenvironment and photophysical properties of the dyes.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    We repeated simulations three times for each condition.

    “We repeated simulations three times for each condition.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    PDB ID 4AUO starts with FVLTEG and has 366 residues.

    “For reference, PDB ID 4AUO starts with FVLTEG (47) and has 366 residues.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [47]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisereasoning

    E219 is the same residue as E200, which differs due to the first amino acid used for counting.

    “E219 (40) is the same residue as E200 (47), which differs due to the rst amino acid used for counting.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [40]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites [47]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given

    Tetracycline holds the two MMP1 domains via hydrogen bonds, facilitating the preference for closed conformations.

    “Previously, we argued that tetracycline holds the two MMP1 domains via hydrogen bonds, facilitating the preference for closed conformations.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The error bars in Figures 3D-I represent the standard deviations of three repeats of simulations for each condition.

    “The error bars in Figures 3D-I represent the standard deviations of three repeats of simulations for each condition.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Substrate- and pose-dependent screening results suggest that the substrate, the binding pose, and the ligand-binding site should be considered when targeting MMP1.

    “Substrateand pose-dependent screening results suggest that we need to consider the substrate, the binding pose, and the ligand-binding site targeting MMP1.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisereasoning

    Counting from MHSFPP, the FVLTEG numbers are 100, 101, 102, 103, 104, and 106.

    “We counted from MHSFPP, and as such, FVLTEG numbers are 100, 101, 102, 103, 104, and 106.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Full-length MMP1 has 469 residues.
    • Rests on an assumption: The first six amino acids of full-length MMP1 are MHSFPP, and the last six are FNCRKN.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    MT-MMP1 cleaves Aβ between H14 and Q15.

    “MT-MMP1 cleaves Aβ between H14 and Q15 (25).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [25]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

26 of 26

  • The paper relies on this without support.

    Simulations on Aβ monomers can inform the behavior of MMP1 on Aβ aggregates.

    “However, there are some caveats because we performed experiments on Aβ aggregates and performed simulations on Aβ monomers.”
    “There are caveats because we performed simulations on Aβ monomer and experiments on aggregates.”

  • The paper relies on this without support.

    The cleavage sites of MT-MMP1 are relevant to the MMP1-Aβ binding pose.

    “Since the computational binding energy may not be the best indicator of appropriate docking poses (39), we leveraged known cleavage sites for a similar MMP.”

  • The paper relies on this without support.

    The three best scoring docking poses include the functionally relevant MMP1-Aβ binding pose.

    “To mimic experiments, we selected the three best scoring poses (Figures 3A-C) and performed all-atom MD simulations at 22 °C for active and inactive MMP1.”

  • The paper relies on this without support.

    Full-length MMP1 has 469 residues.

    “Full-length MMP1 has 469 residues.”

  • The paper relies on this without support.

    The first six amino acids of full-length MMP1 are MHSFPP, and the last six are FNCRKN.

    “The rst six amino acids are MHSFPP, and the last six amino acids are FNCRKN.”

  • The paper relies on this without support.

    The catalytic triad HEL residues are H218, E219, and L220.

    “The catalytic triad HEL residues are H218, E219, and L220.”

  • The paper relies on this without support.

    The three top-scoring ClusPro binding poses between MMP1 and Aβ represent the biologically relevant binding modes.

    “We selected the three best scoring binding poses between MMP1 and Aβ.”

  • The authors state this as an assumption.

    The two-state model adequately describes MMP1 dynamics on Aβ.

    “Stochastic simulation of MMP1 dynamics on Aβ as a two-state system.”

  • The paper relies on this without support.

    AutoDock Vina docking scores reflect true binding affinities.

    “We adjusted the box size to enclose the whole MMP1 to identify each compound's binding locations and a nities using an exhaustiveness of 10 for quicker screening.”
9 of 9

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Two decay rates reported for the same MMP1 condition

    Minor

    The paper reports both 0.02 s−1 and 0.08 s−1 for the correlation decay rate of active MMP1 without ligand on Aβ aggregates. The conflicting values affect its comparison of dynamics across substrates.

1 of 1

Cites

Flagged first, then by citations

Research works printed but not linked: 3

  1. Overall, C. M. 2002. Molecular determinants of metalloproteinase substrate speci city. Molecular biotechnology 22(1):51-86.
  2. Czerski, J., W. Colomb, F. Cannataro, and S. Sarkar. 2018. Spectroscopic identi cation of individual uorophores using photoluminescence excitation spectra. Journal of microscopy 270(3):261-271.
  3. Colomb, W., J. Czerski, J. Sau, and S. Sarkar. 2017. Estimation of microscope drift using uorescent nanodiamonds as ducial markers. Journal of microscopy 266(3):298-306.

Cited by

Sorted by citations

Related

Authors
Kamboj, S., Harms, C., Wright, D., Nash, A., Klein-Seetharaman, J., & Sarkar, S. K.
Affiliations
  • Colorado School of Mines (US)
  • University of Oxford (GB)
  • Colorado School of Mines
Corresponding author
Sarkar, Susanta K.
ORCID
Venue
Research Square
Publisher
Research Square Platform LLC
Published
2022-01-27
DOI
10.21203/rs.3.rs-1258797/v1
OpenAlex
W4210800620
Semantic Scholar
6c719e8b558f3b357466f600560a836a9b2dba05
Cited by
1
References
33
Citations by year
2025: 1
Type
posted content
Language
en
Open access
Licence
Locations
Research Square (acceptedVersion, cc-by)
Funders
  • National Institute for Health and Care Research
  • Oxford Health NHS Foundation Trust
  • National Institutes of Health
  • National Institute of General Medical Sciences (R15GM137295)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › Protease and Inhibitor Mechanisms
  • Health Sciences › Medicine › Physiology › Alzheimer's disease research and treatments
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
Keywords
MMP-1, amyloid-beta, Alzheimer's disease, allosteric communication, protein aggregation, hemopexin domain, catalytic domain, fluorescence resonance energy transfer, virtual screening, small-molecule ligands, Shannon entropy, protein-ligand docking

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-30.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-30.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review