Elucidating the Structure of Melanin and Its Structure-Property Correlation

Melanin, specifically eumelanin, is a brownish-black pigment abundant in nature. It is a heterogeneous biopolymer known most commonly for its spectacular photoprotection property in humans, animals, and plants. Numerous other properties of eumelanin have recently been identified, including radical scavenging, thermal regulation, charge transport, etc. While the presence and significance of melanin have long been accepted, its exact chemical makeup remains unclear. Melanin has an inherent diversity of structure and complexity, making understanding many of its properties challenging. Therefore, there are large gaps in our understanding of the structure-property relationship in melanin. On the other hand, due to its diverse properties and biocompatibility, there is significant interest in engineering such biomimetic devices and materials with targeted properties utilizing melanin-like scaffolds. Furthermore, recent efforts toward understanding the molecular origin of its properties are also based on synthetic melanin analogs. For all these reasons, understanding the structure of melanin and its structure-property relationship remains pivotal to much of the progress in this direction.Our research has revolved around providing new insights into the structure of melanin and its structure-function correlations for properties such as absorption spectra, electron transport, and photoprotection. We have used tools based on a combination of computational chemistry and machine learning to answer these questions. Many of these methods and protocols developed in the group for solving this problem can be utilized for other problems of similar complexity.The difficulties in elucidating these aspects of melanin chemistry lie in its inherent structural and chemical diversity. It is a biopolymer with heterogeneity in monomeric units, oxidation states, polymerization site, and significant conformational degrees of freedom. Many of the attractive properties of melanin are a direct consequence of this diversity. One such property is its broadband absorption spectra, which allow it to absorb light across the UV-visible range and, therefore, function as a photoprotective agent against solar radiation. The melanin absorption spectra are unusual in their monotonic and featureless form. It is now well-accepted that both structural and chemical heterogeneity are responsible for these unusual spectra. Therefore, the problem of isolating specific chromophores responsible for different parts of the spectra becomes a daunting task. While previous computational and experimental studies have shown that diversity is central to this property, i.e., the spectrum is an ensemble or average spectrum, they have not been able to identify specific structures that are responsible. Therefore, we have used machine learning and artificial intelligence concepts to identify patterns and reconstruct particular structures. We have shown, with the help of machine learning and computational chemistry, that data-driven approaches can identify specific structures or classes of structures that are important chromophores. This provides a significant direction in answering the questions about the structure-function correlation of melanin for its absorption spectra.For charge transport properties we have shown the importance of hydrogen bond networks and, therefore, the most successful structural motifs for efficient charge transport. Furthermore, with a bottom-up approach, we have also identified nonradiative pathways for the deactivation of photoactivated melanin.