Matriptase and TMPRSS6 (also known as matriptase 2) are type II transmembrane serine proteases (TTSPs) that share significant structural similarity and substrate specificity, yet they exhibit distinct biological roles. Matriptase is ubiquitously expressed in epithelial tissues, where it plays a crucial role in maintaining epithelial integrity [1]. In healthy tissues, matriptase activity is tightly regulated through two independent mechanisms: autoactivation [2] and rapid inhibition by Kunitz-type hepatocyte growth factor activator inhibitors, HAI‐1 and HAI‐2 [3]. Increased expression of matriptase is observed in variety of epithelial cancers, such as prostate, breast, ovarian, cervical, and stomach [[3], [4], [5], [6]]. Noteworthy, its overexpression in breast and prostate cancers is associated with tumor grade and stage, making it a potential predictive factor for poor disease prognosis [[7], [8], [9], [10]]. TMPRSS6 displays a tissue-restricted expression pattern, being almost exclusively found in hepatocytes in normal tissues [11]. In liver, it is involved in iron homeostasis through the downregulation of the BMP/SMAD signalling pathway, leading to reduced hepcidin production and, ultimately, elevated of plasma iron levels [12,13]. Hence, TMPRSS6 has been identified as a therapeutic target in iron-overload diseases [14], but the role of its catalytic activity in iron regulation remains elusive [[15], [16], [17]]. Additionally, unlike matriptase, TMPRSS6 expression is positively correlated with favorable prognosis in breast cancer patients [[18], [19], [20]]. These contrasting functions of matriptase and TMPRSS6 underscore the need for potent and highly specific inhibitors to enable targeted therapeutic strategies.

Until now, only a few strong and moderately selective matriptase inhibitors have been reported including non-covalently acting derivatives of L-3-amidinophenylalanine [21,22] exhibiting up to 50-fold selectivity for matriptase over TMPRSS6, as well as the covalent inhibitor Arg-Gln-Ala-Arg-ketobenzothiazole [23] that binds to matriptase with a subnanomolar inhibitory constant (Ki = 0.011 nM) and 300-fold selectivity over TMPRSS6. Studies have also highlighted the successful use of nature-derived peptides as convenient scaffolds for developing novel inhibitors of TTSPs, including plant-derived trypsin inhibitors of the squash Momordica cochinchinensis [24] and sunflower seeds [[25], [26], [27], [28]]. For instance, our group reported a series of non-covalent peptide-based inhibitors of matriptase and TMPRSS6, derived from the primary structure of the 14-amino acids, cyclic sunflower seeds-derived trypsin inhibitor (SFTI-1) [26,28]. One of these, a bicyclic peptide containing a disulfide bridge and a covalent bond between the ε-amino group of N-terminal Lys and α-carboxyl moiety of C-terminal α-carboxyl Asp, exhibited high inhibitory activity against matriptase (Ki = 2.6 nM) and nearly 100-fold selectivity over TMPRSS6 [28]. Conversely, two other peptides of this series demonstrated moderate potency against TMPRSS6 (Ki = 0.43 μM and 0.28 μM) but remarkable 176- and 226-fold selectivity when compared to matriptase [26].

Amphibian skin is another significant source of nature-derived peptides with potential therapeutic applications [29]. Of particular interest is a peptide isolated from the skin secretion of the Chinese bamboo leaf odorous frog, Huia versabilis, which contains a Bowman-Birk-like protease inhibitor motif and exhibits strong anti-trypsin activity [30]. This monocyclic peptide, known as Huia Versabilis Bowman-Birk Inhibitor (HV-BBI), consists of 18-amino acids and features a disulfide-bridged, rigid protease-binding loop (an inhibitory loop) that inserts into a concave protease active site. The P1 position, the most crucial for interaction with enzymes, is occupied by Lys8, which determines both inhibitory activity and specificity against trypsin-like proteases [31]. Interestingly, although the HV-BBI loop is longer than plant-derived BBIs (11 vs 9 residues), its spatial structure is highly similar. Consequently, the frog-skin HV-BBI, as well as many other amphibian-originating inhibitors, is described as a BBI-like trypsin inhibitor [32]. Our group has recently reported that a truncated 13-amino acid HV-BBI variant (inhibitor 6) exhibits strong in vitro inhibitory activity against bovine β-trypsin (Ki = 151 nM), human plasmin (Ki = 120 nM), and human matriptase (Ki = 8.5 nM) [33]. Notably, the inhibitory activity against matriptase was associated with the presence of a C-terminal amide group. Hence, analogues with the same primary structure but featuring either a C-terminal carboxyl group (6∗) or a continuous cyclic backbone (6∗∗) were weaker matriptase inhibitors [33].

Building on these findings, our current study expands the investigation of the HV-BBI derived peptides as selective matriptase inhibitors. Their inhibitory activities against matriptase and other proteases, with a particular focus on the closely related TMPRSS6, were characterized both in vitro and in cellulo. To uncover the structural determinants of selectivity and inhibitory potency, alanine-scanning mutagenesis and C-terminal residue substitutions were performed. Finally, advanced molecular dynamics (MD) simulations and modeling was applied to analyze their binding behaviors in greater detail. This comprehensive analysis not only deepens our understanding of these peptides as matriptase selective protease inhibitors but also underscores their potential as scaffolds for developing highly specific therapeutic molecules.