Disordered Regions Flanking Ordered Domains Modulate Signaling Transduction

Main Text

Signals generated by cell surface receptors are transduced and relayed to downstream intracellular signaling pathways typically through proteins in which folded domains are flanked by regions that are unstructured or only partially structured. Such signaling systems are often multidomain proteins in which disordered linkers provide an effective means to covalently tether and colocalize globular modules (1). However, in several regulatory proteins, linkers have been reported to serve not only as passive covalent tethers, but also as active modulators of the function elicited by adjacent domains (2, 3, 4). Although functionally relevant, such linker-domain interactions are often elusive and dissecting the mechanism underlying the cross talk between folded modules and adjacent partially unfolded regions is emerging as a major challenge in the structural biology of signaling.

In this issue of the Biophysical Journal, Chavan et al. (5) investigate how the flexible C-terminal hypervariable region (HVR) of K-Ras4B, a small GTPase that controls cell survival and proliferation, interacts with and modulates the adjacent catalytic domain. Using comparative NMR analyses (i.e., wild-type versus HVR deletion mutant and GTP- versus GDP-bound forms of K-Ras4B), bioassays, molecular dynamics simulations, peptide and nanodisk binding studies, Chavan et al. (5) propose an unanticipated model positing that the HVR of GDP-bound K-Ras4B is sequestered into its catalytic domain active site. While a nuclear Overhauser effect- or PRE-based structure with the HVR docked into the GDP-bound catalytic domain is not yet available as of this writing, the hypothesized nucleotide-dependent sequestration of the HVR into the catalytic site of K-Ras4B has several transformative corollaries.

First, it means that GDP, by preferentially recruiting the HVR to the active site of K-Ras4B more efficiently than GTP, weakens several interactions between K-Ras4B and effectors necessary to modulate downstream Ras-controlled signaling. These interactions include those between K-Ras4B and the plasma membrane, which is required for subcellular localization (6), and those between K-Ras4B and the Raf Kinase, which is activated by GTP-bound K-Ras4B (5). Hence, the hypothesis proposed by Chavan et al. (5) has the potential to impact at a fundamental level the current understanding of Ras autoinhibition and activation, which critically relies on the GDP-versus-GTP exchange. For example, the model of Chavan et al. predicts that the HVR-mediated subcellular localization of K-Ras to the plasma membrane may be nucleotide-dependent, prompting new targeted in vivo experiments to specifically test this corollary.

Second, the proposed HVR-catalytic domain interaction provides a viable mechanism to explain how dynamic regions modulate function in other protein systems as well (3, 4). For instance, a dynamic linker in the regulatory subunit of the ubiquitous protein kinase A has been shown to serve as an integral allosteric element, which controls kinase activation by tuning the autoinhibitory equilibrium through weak but functionally significant state-selective interactions (4). Such linker-domain interactions are likely to be a general phenomenon. Thus, the work presented here by Chavan et al. (5) has potential to impact not only the Ras community, but also the structural biology field at large.

Third, the model reported here may open new opportunities to treat tumor-related K-Ras4B mutations through peptide analogs of the HVR. This is particularly remarkable considering that the catalytic domain of K-Ras was previously considered to be an undruggable target. The authors challenge this paradigm by showing that synthetic peptides mimicking the HVR bind the catalytic domain of K-Ras4B with an affinity that is approximately two orders-of-magnitude higher in the GDP-bound state, compared to the GTP-bound state, thus promoting the inhibition of K-Ras4B (5). In this respect, the work of Chavan et al. (5) complements well a recent landmark contribution by Mazhab-Jafari et al. (7), who monitored nanodisk binding of K-Ras4B by solution NMR.

Mazhab-Jafari et al. (7) showed that membrane sequestration of the effector-binding site autoinhibits wild-type K-Ras4B, but tumor-related K-Ras4B mutations disrupt such autoinhibition. This major finding revealed the mechanism for enhanced signaling by K-Ras mutations and opened up a new therapeutic strategy for targeting K-Ras by stabilizing the autoinhibitory interaction with the membrane. These recent NMR studies of KRAS dynamics at the membrane (7) suggest that the proposed G-domain/HVR interaction is dynamic and transient. Yet, RAS uses the HVR as a way to modulate properties of the GTPase-domain, either by its direct interaction with the functional core or by dictating the orientation of the core with respect to the negatively charged membrane surface. In addition, the membrane potential may exert further control of K-Ras signaling (8).

Lastly, considering the poor conservation of HVRs across different Ras isoforms, the HVR sequestration mechanism proposed by Chavan et al. (5) may also rationalize previously unexplained differences between Ras isoforms. Overall, the model of Chavan et al. (5) will prompt future investigations that will further capitalize on the synergies between experimental and computational approaches to thoroughly map the free energy landscape of allosteric transitions. For example, chemical-shift-based analyses and simulations have been shown to efficiently and reliably sample conformational ensembles for poorly structured systems (9, 10), such as the HVR of Ras investigated by Chavan et al. (5). The hypothesis proposed by Chavan et al. (5) adds a new critical dimension to our understanding of the complex regulatory mechanisms in Ras-controlled signaling.

Author Contributions

J.A.B. and G.M. analyzed relevant literature and wrote the article.

Editor: Jeff Peng.