Novel mTORC2/HSPB4 Interaction: Role and Regulation of HSPB4 T148 Phosphorylation

Abstract

HSPB4 and HSPB5 (α-crystallins) have shown increasing promise as neuroprotective agents, demonstrating several anti-apoptotic and protective roles in disorders such as multiple sclerosis and diabetic retinopathy. HSPs are highly regulated by post-translational modification, including deamidation, glycosylation, and phosphorylation. Among them, T148 phosphorylation has been shown to regulate the structural and functional characteristics of HSPB4 and underlie, in part, its neuroprotective capacity. We recently demonstrated that this phosphorylation is reduced in retinal tissues from patients with diabetic retinopathy, raising the question of its regulation during diseases. The kinase(s) responsible for regulating this phosphorylation, however, have yet to be identified. To this end, we employed a multi-tier strategy utilizing in vitro kinome profiling, bioinformatics, and chemoproteomics to predict and discover the kinases capable of phosphorylating T148. Several kinases were identified as being capable of specifically phosphorylating T148 in vitro, and further analysis highlighted mTORC2 as a particularly strong candidate. Altogether, our data demonstrate that the HSPB4-mTORC2 interaction is multi-faceted. Our data support the role of mTORC2 as a specific kinase phosphorylating HSPB4 at T148, but also provide evidence that the HSPB4 chaperone function further strengthens the interaction. This study addresses a critical gap in our understanding of the regulatory underpinnings of T148 phosphorylation-mediated neuroprotection.

Keywords: HSPB4; chaperone; kinase; neuroprotection; phosphorylation; sHSP; αA-crystallin.

Figure 1

(A) Graphical representation of the PhAXA experimental process (adapted from [47]). (B) Representative PhAXA blot showing the band profiles in both the MIO-M1 and R28 cell line experiments. (X—marker; WT—wild-type).

Figure 2

Kinases reproducibly identified via PhAXA in R28 (A,C) or MIO-M1 (B,D) cells. (A) R28 PSM fold-change; (B) MIO-M1 PSM fold-change; (C) R28 abundance fold-change; (D) MIO-M1 abundance fold-change. (n = 3 for each cell line) Error bars represent SD.

Figure 3

In vitro kinase screen of a T148-containing HSPB4 peptide (residues 142–154) against (A) 38 kinases (100 nM kinase, n = 2) or (B) 13 kinases that showed relatively high activity at 100 nM (10 nM kinase, n = 3 [ATP only] or n = 4 [ATP+peptide]). Data for the 10 nM screen were analyzed using an unpaired T-test using Graphpad Prism. (Error bars represent SD; ns: not significant; *: p < 0.05; **: p < 0.01; ***: p < 0.001; ****: p < 0.0001).

Figure 4

In vitro kinase screen against full-length recombinant HSPB4. (A) Recombinant full-length wild-type (WT) and non-phosphorylatable (T148A) HSPB4 were screened against 10 kinases (100 nM kinase, n = 4. (B) The screen was then repeated using a lower kinase concentration (10 nM, n = 4). (Error bars represent SD; ***: p < 0.001; ****: p < 0.0001).

Figure 5

Identification and verification of mTOR–HSPB4 protein–protein interaction. (A) Number of peptides corresponding to mTORC1/2 component proteins identified in R28 and MIO-M1 PhAXA experiments. Error bars represent SD. (B) Representative FLAG IP blot validating the MS/MS results. (C) Representative Rictor and Raptor IP blots and corresponding input protein levels, confirming the bidirectionality of the interaction (EV—empty vector; WT—wild-type; NL—no lysate control; NA—no antibody control; TNF+HG—TNFα and high-glucose treatment).

Figure 6

Characterization of the novel HSPB4-mTORC2 interaction by Rictor IP in the presence of a cell-permeable DSP crosslinker using several HSPB4 mutants. (EV—empty vector; WT—wild-type; NA—no antibody control; X—marker; TNF+HG—TNFα and high-glucose treatment; DSP—crosslinker treatment). For quantification, the HSPB4 IP band intensity was normalized to the HSPB4 input band, and the Rictor IP band intensity was normalized to the Rictor input band. An HSPB4/Rictor ratio was obtained by dividing these normalized values and then normalizing to the WT ratio.