Upregulation and maintenance of gap junctional communication in lens cells

Abstract

The cells of the lens are joined by an extensive network of gap junction intercellular channels consisting of connexins 43, 46, and 50. We have proposed, and experimentally supported, the hypothesis that fibroblast growth factor (FGF) signaling is required for upregulation of gap junction-mediated intercellular coupling (GJIC) at the lens equator. The ability of FGF to increase GJIC in cultured lens cells requires sustained activation of extracellular signal-regulated kinase (ERK). In other cell types, activation of ERK has been shown to block GJIC mediated by connexin43 (Cx43). Why ERK signaling does not block lens cell coupling is not known. Another unresolved issue in lens gap junction regulation is how connexins, synthesized before the loss of biosynthetic organelles in mature lens fiber cells, avoid degradation during formation of the organelle-free zone. We have addressed these questions using serum-free cultures (termed DCDMLs) of primary embryonic chick lens epithelial cells. We show that FGF stimulates ERK in DCDMLs via the canonical Ras/Raf1 pathway, and that the reason that neither basal nor growth factor-stimulated GJIC is blocked by activation of ERK is because it is not mediated by Cx43. In fibroblastic cells, the normally rapid rate of degradation of Cx43 after its transport to the plasma membrane is reduced by treatments that either directly (ALLN; epoxomicin) or indirectly (generation of oxidatively un/mis-folded proteins by arsenic compounds) prevent the ubiquitin/proteasome system (UPS) from acting on its normal substrates. We show here that Cx45.6 and Cx56, the chick orthologs of mammalian Cx50 and Cx46, behave similarly in DCDMLs. When organelles lyse during the maturation of fiber cells, they release into the cytosol a large amount of new proteins that have the potential to saturate the capacity, and/or compromise the function, of the UPS. This would serve to spare gap junctions from degradation during formation of the organelle-free zone, thereby preserving GJIC between mature fiber cells despite the lack of de novo connexin synthesis.

Figure 1. DCDMLs can be efficiently transiently transfected and cotransfected

One-day-old DCDML cultures were transfected with plasmids encoding EGFP, a constitutively active form of MEK (CA MEK), wild type MEK (WT MEK), or B-galactosidase (B gal) as indicated using Lipofectamine. (A) At 20 h, 2 d, 3 d, and 5 d after transfection, expression of EGFP was monitored by fluorescence microscopy. (B) 48 h after transfection, cultures were analyzed for activated ERK using anti-phosphoThr202/Tyr204-specific ERK antibodies detected either immunohistochemically (top row) or by immunoblotting (bottom row). One set of CA MEK-expressing cultures were incubated with the MEK inhibitor UO126 (UO) for 45 min prior to analysis. Inset, level of anti-phosphoERK immunoreactivity in untransfected cultures after a 4 h treatment with 15 ng/ml FGF-2. (C) 48 h after transfection with 0.1 μg/well CA MEK plasmid plus 0.03 ug/well B-galactosidase plasmid (a′ and b′), or 0.1 μg/well B-galactosidase only (c′), cultures were double-stained with anti-B-galactosidase and anti-pERK antibodies. Arrows denote doubly-transfected cells. More than 50% of the cells continued to express CA MEK or B-galactosidase on day 5 after transfection with 0.1 μg/well (not shown).

Figure 2. Role of Ras in ERK activation in DCDMLs

DCDMLs were transfected with the indicated plasmids, and 48 h later assessed for activation of either: ERK-FLAG by anti-FLAG immunoprecipation followed by anti-pERK immunoblotting (lanes 1–5), or endogenous ERK by anti-pERK immunoblotting of whole cell lysates (lanes 6–8). In some cases, the cultures were treated with 15 ng/ml FGF-2 for 15 min immediately prior to lysis. Asterisk, heavy chain of the immunoprecipitating anti-FLAG immunoglobulin as detected by the secondary antibody used to visualize anti-pERK.

Figure 3. Role of Raf-1 in ERK activation in DCDMLs

DCDMLs were cotransfected with plasmids encoding ERK-FLAG and either RapV12 or the catalytic subunit of protein kinase A (cPKA), both known activators of Rap-1 and inhibitors of Raf-1. 48 h later, cells were treated with or without 15 ng/ml FGF for 15 min prior to lysis and immunoprecipitation of ERK-FLAG. Activation of ERK-FLAG was assessed by anti-pERK immunoblotting. Asterisk, heavy chain of the immunoprecipitating anti-FLAG immunoglobulin as detected by the secondary antibody used to visualize anti-pERK.

Figure 4. Cx43 mediates intracellular dye transfer in CHO cells, but not in DCDMLs

Scrape-load GJIC assays were conducted with a mix of 0.25% Alexa594 (594; permeable to Cx43 gap junctional channels) and either 1% Lucifer yellow (LY; permeable to Cx43, Cx45.6, and Cx56 channels), or 1% FITC-dextran (FITC-dex; impermeable to all gap junctional channels) in: (A) CHO cells or (B) DCDMLs as described in Material and Methods. High magnification views of Alexa594, Lucifer yellow, or FITC-dextran transfer, and their overlay. In (B), DCDMLs were incubated for 48 h with no additions (control), 15 ng/ml FGF-2, 10 ng/ml BMP4, or vitreous body conditioned medium (VBCM). Each panel depicts a portion of the right half of the scrape/load wound.

Figure 5. The proteasome inhibitor ALLN slows the normally rapid turnover of gap junctions in DCDMLs

DCDMLs were incubated for 3.5 h either with no additions (control), with 6 μg/ml BFA, or with BFA plus 200 μM ALLN. The cultures were then immunostained for Cx45.6, Cx56, Cx43, or N-cadherin (A) or assayed for intercellular transfer of Lucifer yellow (B). Transfer of Lucifer yellow in BFA only-treated cells was 0.44 +/− 0.55 fold that in untreated control cultures, and transfer of Lucifer yellow in BFA plus ALLN-treated cells was 1.02 +/− 0.07 fold that in untreated control cultures (calculated from 12 measurements made in four independent experiments).

Figure 6. A mechanistically distinct, highly specific small molecular inhibitor of the proteasome mimics the effects of ALLN on DCMDL gap junctions

Following a 3.5 h incubation with BFA in the presence or absence of the irreversible proteasome inhibitor epoxomicin, DCDMLs were either assayed for intercellular transfer of Lucifer yellow (LY) and Alexa594 (594) after scrape-loading of the cells with a mixture of the two dyes, or immunostained for Cx45.6, Cx56, or Cx43.

Figure 7. Indirect perturbation of the ubiquitin-proteasome system by oxidative stress stabilizes gap junctions and GJIC in DCDMLs

DCDMLs were incubated for 3.5 h with BFA either in the absence or presence of 60 μM arsenic trioxide (ATO), which disrupts the folding of oxidatively sensitive proteins within the cytosol. DCDMLs were either immunostained for Cx45.6, Cx56, Cx43, or N-cadherin, or assayed for intercellular transfer of Lucifer yellow and Alexa594 as in Fig 6. Transfer of Lucifer yellow in BFA only-treated cells was 0.35 +/− 0.084 fold that in untreated control cultures, and transfer of Lucifer yellow in BFA plus ATO-treated cells was 0.94 +/− 0.1 fold that in untreated control cultures (calculated from 11 measurements made in three independent experiments).