The allele-specific lipid sensing of KRAS mutants

Image

A hallmark of RAS pathophysiology is the compartmentalization of their oncogenic signaling to the proteolipid nanoclusters on the plasma membrane (PM). This suggests that RAS oncogenesis requires a unique local PM lipid environment. Significant efforts over the past several decades have revealed an intricate selectivity of RAS proteins for distinct lipids.

The enrichment of distinct lipids within RAS nanoclusters is pathologically important because RAS effectors possess their own lipid-binding/recognition motifs. Membrane binding, activation and signal transmission of these effectors require synergistic binding to both mutant RAS and specific lipids in the PM. This remarkable lipid specificity points to the possibility of targeting lipid metabolism as a viable strategy to inhibit RAS oncogenesis.

The selective lipid sensing function of the RAS C-terminal membrane-anchoring domains has been the primary focus of research efforts. For instance, with respect to HRAS, NRAS, and KRAS4A, combinations of their palmitoyl and farnesyl chains contribute to their distinct preferences for cholesterol / saturated lipids (main components of lipid rafts) or unsaturated lipids (enriched in non-raft domains) ¹. 

Image

KRAS4B is the only RAS isoform that does not have a palmitoyl chain, but its farnesylated polybasic domain provides intricate codes for the mixed-chain phosphatidylserine (PS) ^2-5. Historically the RAS globular domains (G-domains) are largely overlooked when discussing membrane association because they are thought to be mostly suspended off of membranes.

However, we consistently observe involvement of the G-domains in lipid sensing among small GTPases. For instance, Prior et al⁶ first reported intriguing GDP-/GTP-dependent spatial segregation of all RAS isoforms. Subsequent studies revealed that RAS isoforms, RAC1, RAP1A, and RAP1B display guanine nucleotide-dependent sensing of lipid headgroups ^2,3,7,8. 

Abankwa et al ^9,10 proposed that GDP-/GTP-binding alters the conformation of the HRAS G-domain, resulting in a shift of HRAS between the cholesterol-enriched lipid rafts and the more fluid non-raft domains. Jansen et al ¹¹ reported that the binding affinities of small molecules to the KRAS4B G-domain are markedly higher in the presence of PS lipids. Atomistic simulations by Goswami et al ¹² predicted that mutating KRAS4B G-domain to be NRAS-like or HRAS-like markedly alters its diffusion in cells. 

Morstein et al¹³ recently illustrated that treatment with the KRAS^G12C inhibitor Adagrasib, which covalently modifies the cysteine residue at the mutated G12 position distal from the PM, disrupts the PM association of KRAS^G12C without affecting that of KRAS^G12D. 

These findings strongly suggest that the RAS G-domains play important roles in facilitating membrane association, which has been puzzling because interactions between the membrane and G-domains are notably weak.

We recently characterized the G-domain-facilitated lipid tail sensing of HRAS, KRAS4A, and KRAS4B on intact cell PM¹⁴. Using electron microscopy (EM)-spatial analysis, we compared PS acyl chain preferences for the wild-type, oncogenic G12V mutant, and the minimal membrane anchor (no G-domain) of these isoforms/splice variants. 

For each isoform, despite sharing identical membrane anchoring domains, these constructs display markedly different preferences for PS species. In the case of KRAS4B, although all constructs prefer mixed-chain PS, the G12V mutant displays the highest selectivity and exclusively associates with the mixed-chain PS. The wild-type and the minimal anchor tK lose some selectivity and associate with the saturated PS (tK) and/or the dual-monounsaturated PS (KRAS4B wild-type), in addition to the mixed-chain PS. Thus, despite weakly associating with the membranes, the G-domains of RAS isoforms possess abilities to sense local environments within the membrane core.

Grant et al ¹⁵ and Ostrem et al ¹⁶ first proposed that the dynamic G-domain of KRAS4B undergoes allosteric reorientation when anchored to membranes, rolling between two main orientation states (OS): OS₁ and OS₂. Abankwa et al ^9,10 also proposed similar conformational shifts of the HRAS G-domain. Each OS state presents a unique membrane interface on the G-domain. It is, thus, possible that the G-domain allostery facilitates lipid sensing. Indeed, we show that mutating lipid-contacting residues, such as Arg 73 / Arg 102 (contacting PS lipids in the OS₂), or Arg135 (contacting PS in the OS₁), differentially alters PS acyl chain preferences of KRAS4B wild-type vs. G12V. Taken together, allosteric reorientation of the KRAS4B G-domain in part mediates the sensing of lipid acyl chains.

The key hotspots of KRAS4B oncogenic mutations—residues G12, G13, and Q61—are mostly positioned distal to membranes. Recent structural studies propose that the G-domains of KRAS4B oncogenic mutants possess allele-specific OS₁/OS₂ equilibria ^17-20. It is possible that KRAS4B oncogenic mutants may sense lipids in distinct manners via distinct G-domain allosteries. We compared lipid preferences of KRAS4B^G12C, KRAS4B^G12D, KRAS4B^G13D, KRAS4B^Q61H, and KRAS4B^G12V. While G12D, G12V, and Q61H share a similar preference for the non-rafty mixed-chain PS, G12C and G13D prefer the rafty saturated PS, cholesterol and/or phosphoinositol 4,5-bisphosphate (PIP₂). 

While we had always assumed KRAS4B as a non-rafty protein, our new data suggest that, despite sharing identical membrane anchors, KRAS4B mutants segregate into spatially distinct membrane regions, with G12C and G13D possessing high preferences for the lipid rafts.

An important role of the specific lipid enrichment of KRAS4B nanoclusters is to facilitate effector recruitment. KRAS4B effector recruitment and MAPK signaling occur more efficiently in the cholesterol-poor non-raft regions. Our EM-bivariate co-clustering analysis showed that binding of a main effector CRAF is weaker for G12C and G13D when compared with G12D, G12V, and Q61H. Concordantly, Wei et al ²¹ compared genetically engineered mouse models with pancreatic ductal adenocarcinoma (PDAC) driven by KRAS4B^G12C or KRAS4B^G12D, and demonstrated that tumor cells expressing KRAS4B^G12C contain lower MAPK signaling and less proliferation than G12D tumor cells. G12C tumors also possess fewer lesions and smaller lesion areas than mice with KRAS4B^G12D tumors, and mice with G12C tumors have a longer lifespan.

It is well-known that KRAS signaling requires anchoring to the cell surface. However, this unique property has not been extensively utilized because of the traditional perception that membrane association lacks specificity. Our work uncovers remarkable selectivity, with which KRAS oncogenic mutants enrich lipids. Our recently reported allele-specific raft preferences suggest that targeting lipid metabolism may perturb oncogenic activities of KRAS4B mutants in distinct manners.

Citations

1          Hancock, J. F. Ras proteins: different signals from different locations. Nat Rev Mol Cell Biol 4, 373-384 (2003). https://doi.org:10.1038/nrm1105 nrm1105 [pii]

2          Zhou, Y. et al. Signal Integration by Lipid-Mediated Spatial Cross Talk between Ras Nanoclusters. Mol Cell Biol34, 862-876 (2014). https://doi.org:10.1128/MCB.01227-13 MCB.01227-13 [pii]

3          Zhou, Y. et al. SIGNAL TRANSDUCTION. Membrane potential modulates plasma membrane phospholipid dynamics and K-Ras signaling. Science 349, 873-876 (2015). https://doi.org:10.1126/science.aaa5619

4          Zhou, Y. et al. Lipid-Sorting Specificity Encoded in K-Ras Membrane Anchor Regulates Signal Output. Cell 168, 239-251 e216 (2017). https://doi.org:10.1016/j.cell.2016.11.059

5          Zhou, Y., Prakash, P. S., Liang, H., Gorfe, A. A. & Hancock, J. F. The KRAS and other prenylated polybasic domain membrane anchors recognize phosphatidylserine acyl chain structure. Proc Natl Acad Sci U S A 118 (2021). https://doi.org:10.1073/pnas.2014605118

6          Prior, I. A., Muncke, C., Parton, R. G. & Hancock, J. F. Direct visualization of Ras proteins in spatially distinct cell surface microdomains. J Cell Biol 160, 165-170 (2003).

7          Maxwell, K. N., Zhou, Y. & Hancock, J. F. Rac1 nanoscale organization on the plasma membrane is driven by lipid binding specificity encoded in the membrane anchor. Mol Cell Biol (2018). https://doi.org:10.1128/MCB.00186-18

8          Liu, J. et al. Glycolysis regulates KRAS plasma membrane localization and function through defined glycosphingolipids. Nat Commun 14, 465 (2023). https://doi.org:10.1038/s41467-023-36128-5

9          Abankwa, D., Gorfe, A. A., Inder, K. & Hancock, J. F. Ras membrane orientation and nanodomain localization generate isoform diversity. Proc Natl Acad Sci U S A 107, 1130-1135 (2010). https://doi.org:10.1073/pnas.0903907107

10        Abankwa, D. et al. A novel switch region regulates H-ras membrane orientation and signal output. EMBO J 27, 727-735 (2008). https://doi.org:10.1038/emboj.2008.10

11        Jansen, J. M. et al. Inhibition of prenylated KRAS in a lipid environment. PLoS One 12, e0174706 (2017). https://doi.org:10.1371/journal.pone.0174706

12        Goswami, D. et al. Membrane interactions of the globular domain and the hypervariable region of KRAS4b define its unique diffusion behavior. Elife 9 (2020). https://doi.org:10.7554/eLife.47654

13        Morstein, J. et al. Direct Modulators of K-Ras-Membrane Interactions. ACS Chem Biol 18, 2082-2093 (2023). https://doi.org:10.1021/acschembio.3c00413

14        Arora, N., Mu, H., Liang, H., Zhao, W. & Zhou, Y. RAS G-domains allosterically contribute to the recognition of lipid headgroups and acyl chains. J Cell Biol 223 (2024). https://doi.org:10.1083/jcb.202307121

15        Grant, B. J. et al. Novel allosteric sites on Ras for lead generation. PLoS One 6, e25711 (2011). https://doi.org:10.1371/journal.pone.0025711

16        Ostrem, J. M., Peters, U., Sos, M. L., Wells, J. A. & Shokat, K. M. K-Ras(G12C) inhibitors allosterically control GTP affinity and effector interactions. Nature503, 548-551 (2013). https://doi.org:10.1038/nature12796

17        Mazhab-Jafari, M. T. et al. Oncogenic and RASopathy-associated K-RAS mutations relieve membrane-dependent occlusion of the effector-binding site. Proc Natl Acad Sci U S A 112, 6625-6630 (2015). https://doi.org:10.1073/pnas.1419895112

18        Vatansever, S., Erman, B. & Gumus, Z. H. Comparative effects of oncogenic mutations G12C, G12V, G13D, and Q61H on local conformations and dynamics of K-Ras. Comput Struct Biotechnol J 18, 1000-1011 (2020). https://doi.org:10.1016/j.csbj.2020.04.003

19        Pantsar, T. KRAS(G12C)-AMG 510 interaction dynamics revealed by all-atom molecular dynamics simulations. Sci Rep 10, 11992 (2020). https://doi.org:10.1038/s41598-020-68950-y

20        Lu, H. & Marti, J. Predicting the conformational variability of oncogenic GTP-bound G12D mutated KRas-4B proteins at zwitterionic model cell membranes. Nanoscale14, 3148-3158 (2022). https://doi.org:10.1039/d1nr07622a

21        Wei, Y. et al. KRAS inhibition activates an actionable CD24 'don't eat me' signal in pancreas cancer. bioRxiv (2023). https://doi.org:10.1101/2023.09.21.558891