Connexin43 interacts with βarrestin: a pre-requisite for osteoblast survival induced by parathyroid hormone

Abstract

Parathyroid hormone (PTH) promotes osteoblast survival through a mechanism that depends on cAMP-mediated signaling downstream of the G protein-coupled receptor PTHR1. We present evidence herein that PTH-induced survival signaling is impaired in cells lacking connexin43 (Cx43). Thus, expression of functional Cx43 dominant negative proteins or Cx43 knock-down abolished the expression of cAMP-target genes and anti-apoptosis induced by PTH in osteoblastic cells. In contrast, cells lacking Cx43 were still responsive to the stable cAMP analog dibutyril-cAMP. PTH survival signaling was rescued by transfecting wild type Cx43 or a truncated dominant negative mutant of βarrestin, a PTHR1-interacting molecule that limits cAMP signaling. On the other hand, Cx43 mutants lacking the cytoplasmic domain (Cx43(Δ245)) or unable to be phosphorylated at serine 368 (Cx43(S368A)), a residue crucial for Cx43 trafficking and function, failed to restore the anti-apoptotic effect of PTH in Cx43-deficient cells. In addition, overexpression of wild type βarrestin abrogated PTH survival signaling in Cx43-expressing cells. Moreover, βarrestin physically associated in vivo to wild type Cx43 and to a lesser extent to Cx43(S368A) ; and this association and the phosphorylation of Cx43 in serine 368 were reduced by PTH. Furthermore, induction of Cx43(S368) phosphorylation or overexpression of wild type Cx43, but not Cx43(Δ245) or Cx43(S368A) , reduced the interaction between βarrestin and the PTHR1. These studies demonstrate that βarrestin is a novel Cx43-interacting protein and suggest that, by sequestering βarrestin, Cx43 facilitates cAMP signaling, thereby exerting a permissive role on osteoblast survival induced by PTH.

Figure 1. Cx43 expression is required for PTH-mediated survival signals and transcription of cAMP-target genes in osteoblasts

(A) OB-6 cells were transfected with vector or the indicated constructs and treated with 50 nM PTH or the corresponding vehicle for 1 h, followed by 50 μM etoposide for 6 h. Apoptosis was assessed by evaluating nuclear morphology of transfected (fluorescent) cells as detailed under Experimental Procedures. Representative images of vehicle- and etoposide-treated cultures show alive (white arrows) and apoptotic (yellow arrow) nuclei. (B) Cx43 protein and mRNA expression were assessed by Western blotting and qPCR, respectively, in wild type OB-6 cells or cells infected with scramble shRNA or Cx43-specific shRNA. Values were normalized against the housekeeping protein ERK1/2 or to CHOB mRNA levels. N=3, * p < 0.05 versus scr cells. (C) OB-6 that were not infected (−) or stably infected with scramble shRNA (scr cells) or Cx43-specific shRNA [(Cx43(−) cells] were treated with 50 nM PTH, 100 μM DBA or the corresponding vehicle for 1 h, followed by 50 μM etoposide for 6 h. Apoptosis was assessed by Trypan blue uptake. Bars represent mean ± S.D. of triplicate determinations. * p < 0.05 versus vehicle-treated cultures. (D) Scr and Cx43 (−) cells were treated with vehicle (−) or the indicated concentrations of DBA for 1 h, followed by 50 μM etoposide for 6 h. Apoptosis was assessed by Trypan blue uptake. Bars represent mean ± S.D. of triplicate determinations. * p < 0.05 versus vehicle-treated cultures. (E) Messenger RNA expression levels of the indicated genes in scr and Cx43-deficient cells treated with 100 nM PTH, 150 μM DBA or the corresponding vehicle for 4 h, followed by RNA extraction and qPCR as described in Experimental Procedures. Bars represent mean ± S.D. of triplicate determinations. * p < 0.05 versus vehicle-treated cultures.

Figure 2. PTH-induced osteoblast survival requires phosphorylation of serine 368 within the cytoplasmic domain of Cx43 and it is abolished by overexpression of βarrestin

(A) Cx43-deficient cells were transiently transfected with the indicated constructs. Forty hours after transfection, apoptosis was assayed as detailed under Experimental Procedures. ^* p < 0.05 versus vehicle-treated cultures. (B) mRNA expression levels of βarrestin1 and 2 in scr and Cx43-deficient cells. (C) Scr and Cx43(−) cells were transiently transfected with wild type βarrestin1 and wild type Cx43 or βarrestin1^319–418, respectively, along with nGFP. Apoptosis was determined by evaluating nuclear morphology of transfected (fluorescent) cells as described in Experimental Procedures. Bars represent mean ± S.D. of triplicate determinations. * p < 0.05 versus vehicle-treated cultures; # p<0.05 versus vector-transfected, vehicle-treated scramble cells.

Figure 3. βarrestin association with Cx43 or the PTHR1 is modulated by Cx43^S368 phosphorylation

(A) Cx43-deficient cells transfected with wild type Cx43 or Cx43^S368A were lysed and total protein extracts were immunoprecipitated with anti-Cx43 antibodies. Western blot was performed with anti-Cx43 and anti-βarrestin1 antibodies. Bands were analyzed by densitometry and the ratio between immunoprecipitated βarrestin1 and Cx43 was calculated. (B) Representative images showing the sub-cellular distribution of Cx43-GFP and βarrestin1-RFP analyzed by confocal microscopy in wild type OB-6 cells transiently transfected with the indicated constructs. Yellow and white arrows point at areas of co-localization of the two proteins in the cytoplasm and the plasma membrane, respectively. (C) CHO cells were transfected with the indicated constructs and βarrestin binding to the PTHR1 was assayed as detailed in the Experimental Procedures section. Bars represent mean ± S.D. of six wells independently transfected. * p < 0.05 versus vector-transfected cells. Similar results were reproduced in two additional independent experiments. (D) 100 μM AGA, GA or DMSO as vehicle were added for 30 min to OB-6 cells and the levels of phosphorylated Cx43^S368 (pSer368) were evaluated by Western blotting. Bands were analyzed by densitometry and the ratio between pSer368 and total Cx43 was calculated. Values represent mean ± S.D. of three replicas. (E) CHO cells were treated with 100 μM AGA, GA or DMSO as vehicle 30 min prior to the addition of 50 nM PTH for 90 min. βarrestin binding to the PTH receptor was assayed as detailed in the Experimental Procedures section. Bars represent mean ± S.D. of six wells independently treated. * p < 0.05 versus vector-transfected cells. Similar results were reproduced in two additional independent experiments (F) 50 nM PTH was added for the indicated time points to wild type OB-6 cells and phosphorylation of Cx43^S368 was evaluated by Western blotting on total protein extracts. Bands were analyzed by densitometry and the ratio between pSer368 and total Cx43 was calculated. Values correspond to mean ± S.D. of three replicas. (G) Cx43-deficient cells transfected with wild type Cx43 were treated with 50 nM PTH or vehicle for 5 min. Cells were lysed and total protein extracts were immunoprecipitated with anti-Cx43 antibodies. Western blot was performed with anti-Cx43 and anti-βarrestin1 antibodies. Bands were analyzed by densitometry and the ratio between immunoprecipitated βarrestin1 and Cx43 was calculated. The mean ± S.D. of three independent experiments are shown. * p < 0.05 versus vehicle-treated cells.

Figure 4. Working model

In Cx43-expressing osteoblasts [Cx43(+) cells], a pool of βarrestin is sequestered by its interaction with phosphorylated serine 368 within the cytoplasmic domain of Cx43 (1), allowing PTH-dependent pro-survival signaling downstream of cAMP (2 and 3). PTH also induces the dephosphorylation of Cx43, likely by activating a protein phosphatase (PP) (4). This leads to the release of βarrestin from Cx43, binding of βarrestin to the PTHR1, inhibition of cAMP production, and internalization of PTHR1 (5). In Cx43-deficient osteoblasts [Cx43(−) cells], a larger pool of βarrestin is available to bind to PTHR1, thus blunting cAMP accumulation, transcription of cAMP target genes, and survival signaling induced by PTH.