Caught in the act - protein adaptation and the expanding roles of the PACS proteins in tissue homeostasis and disease

Abstract

Vertebrate proteins that fulfill multiple and seemingly disparate functions are increasingly recognized as vital solutions to maintaining homeostasis in the face of the complex cell and tissue physiology of higher metazoans. However, the molecular adaptations that underpin this increased functionality remain elusive. In this Commentary, we review the PACS proteins - which first appeared in lower metazoans as protein traffic modulators and evolved in vertebrates to integrate cytoplasmic protein traffic and interorganellar communication with nuclear gene expression - as examples of protein adaptation 'caught in the act'. Vertebrate PACS-1 and PACS-2 increased their functional density and roles as metabolic switches by acquiring phosphorylation sites and nuclear trafficking signals within disordered regions of the proteins. These findings illustrate one mechanism by which vertebrates accommodate their complex cell physiology with a limited set of proteins. We will also highlight how pathogenic viruses exploit the PACS sorting pathways as well as recent studies on PACS genes with mutations or altered expression that result in diverse diseases. These discoveries suggest that investigation of the evolving PACS protein family provides a rich opportunity for insight into vertebrate cell and organ homeostasis.

Fig. 1.

Disorder prediction for PACS-1 and PACS-2. The PrDOS server (http://prdos.hgc.jp/cgi-bin/top.cgi; Ishida and Kinoshita, 2007) was used to predict natively disordered regions from the amino acid sequences of the human PACS-1 (UniProt Q6VY07) and PACS-2 (UniProt Q86VP3) proteins [false discovery rate (FDR)=2%]. The disorder probabilities for each residue were plotted as a function of length and the graphical profiles were juxtaposed with the predicted secondary structures, which were obtained using an improved self-optimized prediction method (SOPMA) on a set of aligned members of the PACS-1 or PACS-2 protein families (lower plots). The PACS-2 nuclear localization signal (NLS) and Akt site, which binds 14-3-3 proteins, together with the corresponding sequences in PACS-1 are shown and predicted to reside in disordered regions. ARR, atrophin-1-related region; FBR, furin (cargo)-binding region; MR, middle region; CTR, C-terminal region. Red dots, phosphorylation sites [as predicted by PhosphoSitePlus (http://www.phosphosite.org/; Hornbeck et al., 2015)].

Fig. 2.

Phylogenetic analysis of the PACS genes. Non-redundant protein sequences of the PACS family members were obtained from the UniProt and NCBI databases. The protein sequences were aligned using Muscle and were manually examined/modified for their accuracy within the non-conserved domains that flank conserved domains. The program Gblocks was used to curate and eliminate poorly aligned positions and divergent regions with the protein alignment prior to the phylogenetic analysis (Castresana, 2000). The program PhyML was used to estimate the maximum likelihood phylogenies from alignments of amino acid sequences (Guindon et al., 2005). The tree (A) and multiple alignments of the NLS and Akt sites (B) were visualized using Mega7. Black diamonds indicate invertebrate PACS proteins expressed from a single gene. Gray diamonds indicate cyclostome PACS proteins, which may be precursors to PACS-1 and PACS-2 and expressed from duplicated genes. Purple and magenta diamonds represent subfunctionalized PACS-1 and PACS-2 paralogs, respectively, that are expressed from duplicated genes in jawed vertebrates. A black background to the amino acid residue indicates identical residues, a red background to the amino acid residue indicates similar residues, and a yellow background divergent residues.

Fig. 3.

Protein traffic steps mediated by PACS-1 and PACS-2. (A) PACS-1 mediates the sorting of client proteins from late endosomes (LE) to the TGN, from early endosomes (EE) to the plasma membrane (PM), as well as delivery to the primary cilium. PACS-2 mediates the localization of cargo proteins to the ER, from early endosomes to the TGN or plasma membrane, and also promotes MAM integrity. (B) HIV-1 Nef usurps the sorting steps mediated by PACS-2 and PACS-1 to downregulate the levels of cell surface MHC-I in CD4⁺ T-cells. Nef binds to Akt-phosphorylated PACS-2 on early endosomes (Atkins et al., 2008; Dikeakos et al., 2012, and L.T. and G.T., unpublished results). This allows Nef to traffic to the TGN region where it binds and activates a Src family kinase (SFK; Hck, Src or Lyn). The Nef–SFK complex then recruits ZAP-70 (in T-cells) or Syk (in monocytes and other cell types) and a class I PI3K, which increases the level of PIP₃ (maroon circles) at the plasma membrane. This recruits an ARF6 GEF that accelerates MHC-I internalization through an ARF6-regulated endocytic pathway. Nef diverts the internalized MHC-I molecules from their local recycling compartment (dashed line) and combines with AP-1 and PACS-1 to transport MHC-I through early and late endosomes and sequester it in the TGN. The identity of the precise compartment containing Nef, MHC-I, AP-1 and PACS-1 is under investigation. This MHC-I downregulation pathway protects HIV-1 from CD8⁺ T-cell killing thereby allowing the virus to evade immune surveillance. Thus, small-molecule inhibition of the multi-kinase complex re-exposes MHC-I on the cells surface and sensitizes HIV-1-infected cells to CD8⁺ T-cell killing. The steps shown here depict the ‘signaling mode’ of HIV-1 Nef-induced immune evasion, which HIV-1 implements during the first 48 h post infection. A detailed discussion of Nef-induced immune evasion is presented elsewhere (Dikeakos et al., 2010; Pawlak and Dikeakos 2015).

Fig. 4.

The PACS-2 Akt site and NLS together modulate membrane traffic, TRAIL-induced apoptosis, MAM integrity and the response to DNA damage. (A) Protein trafficking. Akt-phosphorylated pSer⁴³⁷-PACS-2 (pPACS-2) interacts with ADAM17 on early endosomes (EE) and mediates delivery of the protease to the cell surface where it sheds EGF ligands to stimulate EGFR signaling. In the absence of PACS-2, ADAM17 is degraded in lysosomes (Lys.). (B) TRAIL-induced MOMP. TRAIL triggers dephosphorylation of PACS-2 Ser⁴³⁷, which mediates two trafficking steps required for MOMP. In one trafficking step, PACS-2 binds full-length Bid and translocates Bid to mitochondria. In the other trafficking step, PACS-2 forms a complex with Bim and Bax on lysosomes called the PIXosome, which is required for lysosome membrane permeabilization to release cathepsin B (cath. B). (C) MAMS. Top panel: insulin or growth factors trigger activation of mTORC2 on mitochondria-associated membranes (MAMs; green shading at the ER–mitochondria contact site), which activates Akt to phosphorylate PACS-2. In turn, pPACS-2 increases MAM contacts, which may modulate ER–mitochondria exchange and support increased lipogenesis. The ? denotes signaling pathways that may lead to Akt-dependent phosphorylation of PACS-2 independent of MAM-localized TORC2. Bottom panel: in starved cells or cells treated with TRAIL, Akt is inhibited and PACS-2 Ser⁴³⁷ is dephosphorylated by a protein phosphatase (PPase). Dephosphorylated PACS-2 in turn remodels MAMs (red shading at the ER–mitochondria contact site), which may reduce lipogenesis but increase ER–mitochondrial Ca²⁺ exchange as well as induction of autophagy. (D) DNA damage response. Top panel: to support induction of the NF-κB and Bcl-xL anti-apoptotic pathway, cytoplasmic PACS-2 interacts with a pool of ATM released from the nucleus and maintains the DNA damage kinase in the cytoplasm. The cytoplasmic ATM then triggers activation of the canonical IκBα–NF-κB pathway that leads to induction of anti-apoptotic Bcl-xL. Bottom panel: to support induction of the p53–p21 cell cycle arrest pathway, pPACS-2 traffics to the nucleus where it binds and inhibits SIRT1 to protect acetylation of p53 bound to the p21 promoter, promoting p21 induction and cell cycle arrest. Green arrows, pro-survival-anabolic pathways mediated by pPACS-2. Red arrows, apoptotic or catabolic pathways mediated by dephosphorylated PACS-2. Ac, acetylation; DDR, DNA-damage response.

Fig. 5.

Possible models for the regulation of Bid translocation to mitochondria by PACS proteins and vMIA. (A) Example of an experiment showing that vMIA prevents translocation of Bid–GFP to mitochondria (L.T., T.S., J.A. and G.T., unpublished results). MCF-7 cells expressing Bid–GFP were left untreated (control, top) or transfected with a vector expressing vMIA (blue) followed by treatment with anti-Fas antibody (1 µg/ml) plus cycloheximide (CHX, 20 µg/ml) for 3 h (bottom) to induce Bid translocation. Mitochondria were then labeled with Mitotracker Red. Image analysis showed that anti-Fas antibody concentrated Bid–GFP staining on mitochondria in untransfected cells (yellow asterisk) but not in cells expressing vMIA (white asterisks). Scale bar: 20 µm. (B) Conventional model of Bid regulation. Dephosphorylation of full-length Bid exposes a cleavage site for caspase-8. Caspase-8-mediated cleavage generates tBid, which is then myristoylated and traffics to mitochondria to promote MOMP. (C) Alternative model of Bid regulation. Dephosphorylation of full-length Bid exposes a binding site for PACS-2 or PACS-1. Binding to PACS-2 promotes Bid translocation to mitochondria (solid lines), whereas binding of Bid to PACS-1 interrupts its translocation to mitochondria (upper dashed lines). In HCMV-infected cells, vMIA sequesters PACS-2 to mitochondria (lower dashed lines), thereby preventing Bid recruitment and, ultimately, MOMP. TRAIL-R, TRAIL receptor.