WDR5 is a highly conserved scaffolding protein and an essential core component of multiple chromatin-modifying complexes [1], including the mixed-lineage leukemia (MLL1-4) and SET1A/B histone methyltransferases [2], [3]. These complexes catalyze histone H3 lysine 4 di- and tri-methylation (H3K4me2/3), epigenetic marks critical for active transcription, cellular identity, and development [4], [5], [6]. WDR5′s functional versatility extends beyond histone methylation, encompassing roles in histone acetylation [7], spindle assembly [8], viral entry [9], [10], and stem cell maintenance [11], all governed by its ability to recruit diverse protein partners through two primary binding sites [12].

Structurally, WDR5 folds into a seven-bladed β-propeller, presenting two key interaction surfaces [12]. A shallow groove, the WDR5-binding motif (WBM) site, engages a limited set of partners like MYC [13], KANSL2 [7] and RbBP5 [14]. In contrast, the more promiscuous WIN site is a deep, acidic pocket that recognizes an arginine residue within a loosely conserved [A/C/S]-R-[T/A/S/C/K] sequence motif found in many nuclear proteins [3], [7], [15], [16], [17], [18], [19], [20]. The canonical WIN site interaction, exemplified by histone H3 and MLL1 [19], [21], is well-characterized. However, recent discoveries of non-canonical binders like MBD3C [16], [22], which engages the WIN site through atypical modes, hint at a broader and more complex recognition landscape than previously appreciated.

Given its central role in oncogenic processes, the WIN site has emerged as a promising therapeutic target in cancer [23], [24], [25]. Current drug discovery efforts, however, have been conceptually constrained by mimicking this single, canonical binding mode. While proof-of-concept was established with high-affinity peptidomimetics like the MM-102 [26], MM-401 [27], and MM-589 [28], which inhibit MLL-fusion leukemia cells, and recent work has yielded small molecules and PROTACs [29], [30], [31], [32], [33], [34], [35], no WIN site–targeting therapeutic has achieved clinical approval [29]. This underscores a critical need to explore alternative strategies that move beyond canonical mimicry.

A critical knowledge gap exists regarding the structural limits of WIN site plasticity. Its potential to accommodate non-canonical binding modes remains largely unexplored and could provide a novel foundation for inhibitor development. To systematically probe the conformational limits of this pocket, we conducted a structural and biochemical analysis using arginine-rich peptide probes. Here, we report high-resolution crystal structures of WDR5 bound to two such peptides, revealing unprecedented binding modalities. Our findings demonstrate that the WIN site can support an extended linear conformation that simultaneously engages the WIN and S7 pockets, and most strikingly, a completely reversed “trans-WIN” orientation, characterized by an inverted peptide backbone directionality relative to the canonical binding mode. These results fundamentally expand the mechanistic understanding of WDR5 interactions and provide a new structural framework for designing next-generation inhibitors that exploit these novel binding topologies.