Human Metalloenzymes Are Disproportionately Linked to Disease: The HEAM Structural Annotation Resource

Metal-binding sites in human proteins are often targeted by drugs for treating diseases, but studying the relationship between metal function and disease has been hampered by a lack of comprehensive resources. The HEAM (Human Enzymatic Annotation of Metalloproteins) dataset addresses this gap by providing a structured collection of 1,933 human metalloenzymes and 1,171 non-catalytic metalloproteins.

This dataset incorporates metal annotations into protein structures at their respective binding sites and categorizes each site as either catalytic or non-catalytic using the MAHOMES II machine learning classifier.

HEAM expands the coverage of human site-level metalloenzymes by more than four times compared to the previously largest metalloenzyme database. Analyzing HEAM, researchers discovered that human metalloproteins are disproportionately represented in KEGG disease pathways compared to the overall human proteome. Moreover, metalloenzymes are notably enriched in ClinVar pathogenic variants, a distinction that prior datasets failed to highlight due to unresolved catalytic classification.

These pathogenic variants are primarily clustered in metabolic disorders, suggesting that metabolic dysfunction plays a significant role in diseases linked to metalloenzyme activity.

By integrating 3D structural annotation with catalytic classification, HEAM serves as a robust foundation for structure-guided drug design targeting metal-binding sites. Additionally, HEAM offers insights into the differential contributions of catalytic and non-catalytic metal functions to human disease.