PH domain-mediated autoinhibition and oncogenic activation of Akt
Curation statements for this article:-
Curated by eLife
Evaluation Summary:
Bae et al. examine the regulatory role of the PH domain of the AKT Ser/ Thr kinase, finding a key set of interactions that mediate autoinhibitory interactions between PH and kinase domains. The work provides additional mechanistic understanding of the E17K mutation, a common variant in human cancers. This manuscript will be of great interest to scientists focused on protein kinase regulation and molecular mechanisms that control signal transduction.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
- Biochemistry and Chemical Biology (eLife)
Abstract
Akt is a Ser/Thr protein kinase that plays a central role in metabolism and cancer. Regulation of Akt’s activity involves an autoinhibitory intramolecular interaction between its pleckstrin homology (PH) domain and its kinase domain that can be relieved by C-tail phosphorylation. PH domain mutant E17K Akt is a well-established oncogene. Previously, we reported that the conformation of autoinhibited Akt may be shifted by small molecule allosteric inhibitors limiting the mechanistic insights from existing X-ray structures that have relied on such compounds (Chu et al., 2020). Here, we discover unexpectedly that a single mutation R86A Akt exhibits intensified autoinhibitory features with enhanced PH domain-kinase domain affinity. Structural and biochemical analysis uncovers the importance of a key interaction network involving Arg86, Glu17, and Tyr18 that controls Akt conformation and activity. Our studies also shed light on the molecular basis for E17K Akt activation as an oncogenic driver.
Article activity feed
-
-
Evaluation Summary:
Bae et al. examine the regulatory role of the PH domain of the AKT Ser/ Thr kinase, finding a key set of interactions that mediate autoinhibitory interactions between PH and kinase domains. The work provides additional mechanistic understanding of the E17K mutation, a common variant in human cancers. This manuscript will be of great interest to scientists focused on protein kinase regulation and molecular mechanisms that control signal transduction.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)
-
Reviewer #1 (Public Review):
In an effort to better understand the regulatory switch controlled by distinct phosphorylation sites in the C-tail of AKT (pS473 versus pS477/pT497) the authors set out to mutate basic sites on the AKT N-terminal PH domain. In so doing, they found that mutation of one basic residue (R86) reduced AKT enzymatic activity. This result prompted the hypothesis that loss of the R86 side chain might stabilize the autoinhibitory form of AKT (specifically the PH/Kinase interaction). Consistent with this hypothesis the authors report that the R86A mutation increases the binding of the PH domain to the AKT kinase domain. Subsequent solution NMR data suggest stabilization of the PH domain occurs upon mutation of R86 to Ala. The authors also make excellent use of segmental labeling combined with NMR spectroscopy to gain …
Reviewer #1 (Public Review):
In an effort to better understand the regulatory switch controlled by distinct phosphorylation sites in the C-tail of AKT (pS473 versus pS477/pT497) the authors set out to mutate basic sites on the AKT N-terminal PH domain. In so doing, they found that mutation of one basic residue (R86) reduced AKT enzymatic activity. This result prompted the hypothesis that loss of the R86 side chain might stabilize the autoinhibitory form of AKT (specifically the PH/Kinase interaction). Consistent with this hypothesis the authors report that the R86A mutation increases the binding of the PH domain to the AKT kinase domain. Subsequent solution NMR data suggest stabilization of the PH domain occurs upon mutation of R86 to Ala. The authors also make excellent use of segmental labeling combined with NMR spectroscopy to gain additional insight into how mutation of R86 to Ala affects the PH/kinase domain interaction. The findings underscore the importance of solution measurements in understanding the effects of dynamics, conformational heterogeneity, and structural pre-organization in regulating function (crystal structures of the wildtype and R86A PH domains are nearly identical).
Given the oncogenic mutant E17K drives AKT function, it is interesting that the only differences (outside of the site of mutation) between the crystal structures of WT and R86A mutant PH domains are differences in side chain orientations for E17 and Y18. This difference led the authors to further investigate the role of positions 17 and 18 in autoinhibition and to explore the connection to the R86 side chain. The findings support the conclusions that Y18 in particular is involved in autoinhibitory interactions with the AKT kinase domain and that loss of the R86 sidechain leads to preorganization of the PH domain (and the Y18 sidechain) for enhanced autoinhibition. These results, therefore, advance our understanding of AKT autoinhibition and extend previous reports of AKT regulation and the role of E17. Specifically, the authors suggest (based on the findings reported here and the previously determined crystal structure of the AKT E17K PH domain) that the E17K mutation drives AKT activity by engaging with Y18 in a manner that restricts the Y18 sidechain from mediating autoinhibitory contacts with the kinase domain. Moreover, their findings provide more detailed insight into the interplay between C-tail phosphorylation status and the effects of E17K on AKT catalytic activity. The new insights into the mechanism by which E17K drives AKT activity add to the previously described roles of E17K in augmenting affinity for phospholipid (Carpten et al. 2007, Landgraf et al. 2008, Truebestein et al. 2021).
Crystal structures of autoinhibited AKT have thus far only been obtained in the presence of exogenous ligands, either small molecule inhibitors or a covalently linked nanobody. While these structures are valuable, the field lacks high-resolution structural information for autoinhibited AKT without ligands and it is thought that bound ligands shift the conformational preferences and obscure the physiologically relevant autoinhibitory interfaces. The work presented in this manuscript makes process toward a better understanding of the physiologically relevant contacts between PH and kinase domains that control activity. Another gap in the information provided by existing crystal structures is the lack of electron density for the important activation loop in the kinase domain. Interestingly, the authors note that the AlphaFold-predicted AKT structure reveals a potential interaction between Y18 and F309 on the activation loop. This prompted experimental approaches (kinase assays and 19F NMR) the results of which largely support a regulatory role for F309. Thus, the solution-based experiments presented here provide significant insight beyond avail crystal structures.
Overall, the manuscript is very interesting - in particular the allosteric connections between R86 and the E17-Y18 dipeptide segment. The resulting insights into the role of the PH domain residue Y18 in autoinhibition and a mechanism by which the oncogenic E17K mutation might alter the side chain conformation of Y18 to interfere with autoinhibition add significantly to our understanding of AKT regulation.
-
Reviewer #2 (Public Review):
This paper sets out to understand the intra-domain regulation of AKT activity by investigating how differential phosphorylations of the C-tail influence interactions between the kinase and PH domain of AKT. During the course of the study, Bae et al identify an unanticipated hydrogen bonding network in the PH domain of AKT (R86/T18/E17) that stabilizes interactions between the PH domain and F309 in the activation loop of AKT that inhibits kinase activity. This previously unidentified interaction provides insight into the mode of activation by a well-established cancer-causing mutation in AKT (E17K) and builds upon a series of recent advances deciphering the mechanism of AKT regulation. While the role of the C-tail is still not fully understood, instead they shed insight into the mechanism of this …
Reviewer #2 (Public Review):
This paper sets out to understand the intra-domain regulation of AKT activity by investigating how differential phosphorylations of the C-tail influence interactions between the kinase and PH domain of AKT. During the course of the study, Bae et al identify an unanticipated hydrogen bonding network in the PH domain of AKT (R86/T18/E17) that stabilizes interactions between the PH domain and F309 in the activation loop of AKT that inhibits kinase activity. This previously unidentified interaction provides insight into the mode of activation by a well-established cancer-causing mutation in AKT (E17K) and builds upon a series of recent advances deciphering the mechanism of AKT regulation. While the role of the C-tail is still not fully understood, instead they shed insight into the mechanism of this cancer-causing mutation. This work will be of interest to those that study AKT and highlights the complexity of kinase regulation.
Strengths. The multi-pronged approach systematically investigates the interaction between PH:Kinase, stability/structure of PH domain, activity of AKT, and half-life of activation loop phosphorylation. Methodologically this work is a tour de force with the use of 3-way ligation to generate specific, stoichiometric post-translationally modified variants to study. The insight gained from studying these precise reagents highlights the limitations of studies using non-physiological variants.
Weaknesses: The main conclusion of the paper comes at the end of a series of detailed structural and functional experiments and was unanticipated. While such findings are exciting, a reframing of the work would help highlight the findings.
Detailed Comments:
• This paper is written for an expert and would benefit from more introduction.
• The paper is framed as a study in the role of different phosphorylations of the C-tail on AKT but ends with a result that is separate from the tail. At the end of the paper, I'm still pondering how the tail could/would influence this new autoinhibitory interaction.
• Earlier work from this lab suggests that the linker between the PH and kinase domain influences inhibition. As this region was missing in experiments using the PH domain in trans, could the authors address how the lack of the linker would influence those results?
• Arguments that the PH domain of R86A retains the same fold as WT but is more stable/less dynamic are indirectly supported. While this claim is not critical to the later arguments, could the author provide data showing that the variants had varying stability over time? If the WT PH domain did degrade during the course of the NMR experiments and the authors could only assign ~60% of residues, how does the lack of coverage and degrading protein affect the assignments from this experiment? The analysis of the B-factor of the mutant crystal structure does not provide more insight into dynamics without a comparison to the b-factor distribution of the WT PH domain. Further a loop (#42-49) is missing in the WT structure but is resolved in the structure of the mutant PH domain bound to PIP3 - what significance does the additional disorder mean? This disordered loop along with the varying stability does suggest that the claim the structure of the two variants is identical may be an overstatement.
• Can the authors speculate at the precise step that the PH domain is inhibiting AKT? The Alpha Fold 2 structure looks to have the activation loop in the out (ie activated state). If the Y18:F309 interaction only happens when the activation loop is phosphorylated, is the autoinhibition by the PH domain only at the step of C-tail activation and does that interaction require activation loop phosphorylation? Does the PH domain interaction happen equally as strongly with a variant in which the activation loop is unphosphorylated? In cells Y18A AKT lacks C-tail phosphorylation. As upstream kinases are responsible for these phosphorylations, how does the PH domain:Kinase domain interaction block C-tail phosphorylation?
-
Reviewer #3 (Public Review):
The manuscript by Bae et al describes the role of a point mutation in the PH domain of Akt that changes the inhibition by the PH domain. The data underlying the manuscript appear to be done at a high technical level. The discovery that the R86A mutant has an enhanced inhibitory interface with the kinase domain is intriguing. Although this residue is not at the putative interface, it forms an electrostatic interaction with the Glu17 in the PH domain and causes a reorientation of the loop including the Y18. Analysis of Y18 and E17 mutants can reverse this effect, revealing a molecular mechanism of R86 increased inhibition.
My main concern with the manuscript is that the conclusions as currently written do not appear to be fully supported by the data. Mainly on the role of the pi-pi stacking of the 309-18 …
Reviewer #3 (Public Review):
The manuscript by Bae et al describes the role of a point mutation in the PH domain of Akt that changes the inhibition by the PH domain. The data underlying the manuscript appear to be done at a high technical level. The discovery that the R86A mutant has an enhanced inhibitory interface with the kinase domain is intriguing. Although this residue is not at the putative interface, it forms an electrostatic interaction with the Glu17 in the PH domain and causes a reorientation of the loop including the Y18. Analysis of Y18 and E17 mutants can reverse this effect, revealing a molecular mechanism of R86 increased inhibition.
My main concern with the manuscript is that the conclusions as currently written do not appear to be fully supported by the data. Mainly on the role of the pi-pi stacking of the 309-18 interface. This paper requires a major rewrite. There also could be additional validation data included to verify the stability and phosphorylation state of the different proteins purified.
Major concerns
1. There are concerns about the validation of the proteins used.
2. The authors note on page 9 that they analyzed the alphafold structure to look at the PhH-kinase interface.
From the analysis of the alphafold model, it does not seem appropriate for this analysis, as the alphafold predicted aligned error (taken from alphafold protein structure database, https://www.alphafold.ebi.ac.uk/entry/P31749) validation clearly shows that there is only limited predictive value of the inter-domain interfaces. I am not sure the mutant data on the predicted pi stacking interaction can be supported by alphafold here as strongly as the authors describe, as these mutants may be working through a separate mechanism. The alphafold model also appears to be templated on the 4ekk phosphorylated structure/mutant of 308 and 473, which seems to go against the authors' hypothesis that 473 phosphorylation disrupts the PH domain interface.
The best model for interpreting the Ph-kinase interface seems to be the nanobody-bound X-ray structure, and this region is disordered at F309 in this structure. While the authors' data clearly shows a role for the Y18 reorientation in changing Ph domain binding, and they also show that mutation of F309L also changes binding, they are basing their molecular model on an alphafold model with limited predictive ability for inter-domain contacts.
-
-
-
-
-
