Abstract
Agmatine is a biogenic amine that functions as a neurotransmitter and exhibits anticonvulsant, antineurotoxic, and antidepressant properties. It can be metabolized into putrescine and urea by canonical agmatinases or by the agmatinase-like protein (ALP), which corresponds to the C-terminal region of the LIMCH1 protein. The amino acid sequence of ALP/LIMCH1 diverges significantly from that of canonical agmatinases and lacks the conserved residues typically required for coordination with Mn(2+), an essential cofactor for ureohydrolase activity. The three-dimensional structure of ALP/LIMCH1 remains unresolved, and predictive artificial intelligence algorithms such as AlphaFold have failed to model it reliably. As a result, the configuration of its active site and the identity of potential metal-coordinating ligands remain elusive. In this study, we purified recombinant full-length rat LIMCH1 (119.5 kDa) and a truncated ALP variant, ΔLIM-ALP (51 kDa), and analyzed their secondary structures using circular dichroism spectroscopy. Our results indicate that both proteins differ markedly from known ureohydrolases, exhibiting a high proportion of disordered regions (~60%) and β-structures (~30%). In contrast, Escherichia coli agmatinase displays a well-defined α/β/α sandwich fold. Despite these structural differences, ALP/LIMCH1 remain the only known mammalian proteins exhibiting agmatinase activity. To gain insight into the putative active site of ALP, we proposed candidate Mn(2+)-binding residues and generated single-point mutants (N213A, Q215A, D217A, E288A, K290A). Although these mutations did not significantly alter Mn(2+) binding or its overall content in the protein samples, four mutants exhibited a decreased K(m) for agmatine and a reduced V(max) when normalized to protein concentration.