Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2018 Jun 22;5(8):1800471.
doi: 10.1002/advs.201800471. eCollection 2018 Aug.

Elucidating Self-Assembling Peptide Aggregation via Morphoscanner: A New Tool for Protein-Peptide Structural Characterization

Affiliations
Free PMC article

Elucidating Self-Assembling Peptide Aggregation via Morphoscanner: A New Tool for Protein-Peptide Structural Characterization

Gloria A A Saracino et al. Adv Sci (Weinh). .
Free PMC article

Abstract

Self-assembling and molecular folding are ubiquitous in Nature: they drive the organization of systems ranging from living creatures to DNA molecules. Elucidating the complex dynamics underlying these phenomena is of crucial importance. However, a tool for the analysis of the various phenomena involved in protein/peptide aggregation is still missing. Here, an innovative software is developed and validated for the identification and visualization of b-structuring and b-sheet formation in both simulated systems and crystal structures of proteins and peptides. The novel software suite, dubbed Morphoscanner, is designed to identify and intuitively represent b-structuring and b-sheet formation during molecular dynamics trajectories, paying attention to temporary strand-strand alignment, suboligomer formation and evolution of local order. Self-assembling peptides (SAPs) constitute a promising class of biomaterials and an interesting model to study the spontaneous assembly of molecular systems in vitro. With the help of coarse-grained molecular dynamics the self-assembling of diverse SAPs is simulated into molten aggregates. When applied to these systems, Morphoscanner highlights different b-structuring schemes and kinetics related to SAP sequences. It is demonstrated that Morphoscanner is a novel versatile tool designed to probe the aggregation dynamics of self-assembling systems, adaptable to the analysis of differently coarsened simulations of a variety of biomolecules.

Keywords: coarse‐grained molecular dynamics; multilayer graph theory; pattern recognition; self‐assembling peptides; β‐structures.

Figures

Figure 1
Figure 1
Morphoscanner validation on different protein structures. A series of PDB structures were analyzed with STRIDE‐based R script and Morphoscanner. In the first column, the reference PDB structures are represented as cartoon using VMD. In the second column, CG structures are visualized highlighting β‐sheets identified through Morphoscanner. β‐strand percentages calculated via R script (%S*) and Morphoscanner (%MS) are shown in the third column. In addition, shift profiles were used to quantify strand displacement in each structure. In the last column we depicted just the predominant shift profile. P = parallel alignment; A+ = antiparallel alignment with positive shift; A‐ = antiparallel alignment with negative shift. The analyses of 2mxu (SL = 32, S = 12) were in agreement and showed that strands were parallel aligned. In 2fkg (SL = 9, S = 35) strands were preferentially antiparallel aligned. The same conclusions were reached for 1d2s (SL = 10, S = 34) and 3bep (SL = 6, S = 122) analysis.
Figure 2
Figure 2
β‐interactions and β‐structuring of SAPs in CG‐MD simulations with extended SS parameters. The onset of β‐interactions does not warrant the formation of β‐sheet structures. This is clearly evident from the comparison among peptides 2,4 and B26. The above‐mentioned SAPs reached the same number of β‐interactions, but B26 had the lowest degree of β‐structuring propensity, followed by 2 and 4. Such features are attributable to their sequences and, in particular, to N‐terminal functionalization.
Figure 3
Figure 3
Analysis of mutual alignment of peptides featuring diverse self‐assembling propensities. Peptides mutual alignment shift profiles of SAP 2, B24, and 30 which were simulated with extended secondary structure parameters (see Table 1). P refers to parallel alignment, A+ to antiparallel alignment with positive shift, A‐ to antiparallel alignment with negative shift. BMHP1‐derived SAPs preferentially shifted by one residue in P alignment, but (LDLK)3 and CAPs showed much stronger alignment in both P and A‐ alignments at one residue shift. This feature was likely due to the electrostatic interactions among their complementary charged side‐chains. On the other hand, the mutation of Pro and Ser with Ala increased the number of β‐interactions in B24 and 30 assemblies if compared to SAP 2 (see Table 1). Biotinylation also slightly improved β‐sheet structuration propensity in B24 in respect to 30. Notably, CAPs and (LDLK)3 showed less β‐interactions than BMHP1‐derived SAPs. This was due to the different shapes of supramolecular aggregates; (LDLK)3 and CAPS formed bilayered β‐sheet‐rich aggregates. BMHP1‐derived SAPs formed ovoid aggregates where peptide strands could simultaneously interact with multiple surrounding peptides.
Figure 4
Figure 4
Structural characterization of B24 molten particles at different timeframes. B24 showed good β‐sheet propensity (AI) characterized by parallel out‐of‐register β‐strands (AII). Parallel β‐sheets shift profiles became wider between 2500 and 4500 ns: this was matched by changings in β‐sheet topology (BI–III) and influenced the identification of oligomers (BIV–VI). P2 was calculated for the identified oligomers (CI–III). Same colors between BIV–VI and CI–III point at the same oligomers identified at the selected timeframes. P2 values of the identified oligomers were calculated for all timeframes. The identified oligomers ranged from 8‐mer to 25‐mer aggregates. Interestingly, oligomers (BIV–VI) featuring higher order (or P2 values in CI–III) showed a large presence of β‐sheets (BI–III).
Figure 5
Figure 5
Structural characterization of 30 molten particles at different timeframes. SAP 30 had a good β‐structuring propensity (AI) and peptides were mutually aligned in parallel out‐of‐register by one residue within β‐sheets (AII). Shift profiles of parallel β‐sheets became sharper after 2500 ns but did not vary as extensively as in B24. The topology of β‐sheets changed slightly (BI–III): this was reflected in a modest variation of the identified oligomers at different timeframes (BIV–VI). Same colours between BIV–VI and CI–III point at the same oligomers identified at the selected timeframes. P2 values of the identified oligomers were calculated for all timeframes. More ordered oligomers (CI–III) were characterized by stronger presence β‐sheet structures (BI–III). The oligomers identified at 4500 ns were more heterogeneous and with higher P2 values (CI–III): big oligomers identified in previous timeframes were here split in two or more subgroups.
Figure 6
Figure 6
Structural characterization of CAPs (LDLD)3 + (LKLK)3 molten particles at different timeframes. CAPs established less ß‐interactions (AI) than in BMHP1‐derived SAPs and β‐strands were preferentially aligned in parallel out‐of‐register by one residue throughout the simulations (AII). CAPs formed stable β‐sheet structures (BI–III) mainly matching oligomers distribution (BIV–VI). β‐sheets paired into bilayered aggregates but with different orientations. Same colours between BIV–VI and CI–III point at the same oligomers identified at the selected timeframes. P2 values of the identified oligomers were calculated for all timeframes. The identified oligomers displayed a superior order (values of P2) and a slow but ongoing trend of increments toward more ordered assemblies (CI–III).

Similar articles

See all similar articles

References

    1. Smith‐Carpenter J. E., Li S., Goodwin J. T., Mehta A. K., Lynn D. G., From Matter to Life (Eds: alker S. I. W, editor; , avies P. W. C. D, editor; , llis G. F. R. E, editor. ), Cambridge University Press, Cambridge: 2017, Ch. 5.
    1. Krishnamurthy R., Acc. Chem. Res. 2017, 50, 455. - PubMed
    1. Mathis C., Bhattacharya T., Walker S. I., Astrobiology 2017, 17, 266. - PubMed
    1. Higgs P. G., J. Mol. Evol. 2017, 84, 225. - PubMed
    1. Mathis C., Ramprasad S. N., Walker S. I., Lehman N., Life 2017, 7, 38.

LinkOut - more resources

Feedback