Trimethylamine N-oxide alleviates the severe aggregation and ER stress caused by G98R αA-crystallin

Expression constructs and mutagenesis

A 548-bp EcoRI/XhoI fragment encompassing the full-length 519 bp open reading frame of CRYAA was ligated into the EcoRI/XhoI sites of a mammalian expression plasmid pcDNA6-His/myc (Invitrogen, Carlsbad, CA) to obtain the construct pHis/myc-CRYAA^WT. This cloned fragment was used as a template to generate specific mutations of CRYAA, using QuikChange II Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA) and oligonucleotides (Table 1). Wild-type (WT) and mutant constructs were verified by direct sequencing.

Cell culture and transfection

Human lens epithelial B3 cells (American Tissue Culture Collection, Manassas, VA) were cultured in Eagle’s Minimal Essential medium (Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen) and antibiotics (100 units/ml penicillin G and 100 μg/ml streptomycin sulfate, Invitrogen). Cells (10⁵ cells/cm²) were transfected with CRYAA constructs by using FuGene HD transfection reagent (Roche, Basel, Switzerland) at an optimized ratio of 3 μl FuGene per μg DNA in Opti-MEM^® I (Invitrogen) supplemented with GlutaMAX^TM-I (Invitrogen). Chemical chaperone treatments were started at 24 h after transfection.

Treatment by trimethylamine N-oxide and other chemical chaperone molecules

Trimethylamine N-oxide (TMAO) (25–300 mM; Sigma, St Louis, MO), 4-phenylbutyric acid (4-PBA; 0.2–2 mM; tributyrate; Triple Crown America Inc, Perkasie, PA), glycerol (1–5%; Sigma), or dimethylsulfoxide (DMSO; 0.5–1%; Sigma) was added to the transfected cells. Fresh culture medium containing drugs was replenished every 2 days.

Protein analysis

Cells were lysed at a concentration of 2.5×10⁶ cells/ml radioimmunoprecipitation (RIPA) buffer containing 50 mM Tris-HCl (Sigma), 150 mM sodium chloride, 1% Nonidet P-40 (Sigma), 0.25% sodium deoxycholate (Sigma), protease inhibitor cocktail (Roche), and 1 mM phenylmethyl sulfonylfluoride (PMSF; Sigma) for 30 min on ice. After centrifugation, the supernatant was collected and denatured in sample buffer with a final concentration of 2% sodium dodecyl sulfate (SDS; BioRad, Hercules, CA) and 50 mM DTT (Sigma). The cell pellet was washed with ice-cold PBS and denatured in SDS sample buffer containing 9 M urea (BioRad).

For Triton X-100 (Tx) solubility analysis, cells were washed twice with ice-cold PBS and added to lysis buffer, which contained 100 mM Tris-HCl (pH 7.4), 3 mM ethylene glycol tetraacetic acid (EGTA; Sigma), 5 mM MgCl₂, 0.5% Triton X-100 (Tx; Sigma), protease inhibitor cocktail, and 1 mM PMSF at 5×10⁶ cells/ml, for 2 min on ice. After centrifugation, the supernatant containing Tx-soluble protein was denatured in SDS buffer. The pellet containing Tx-insoluble protein was washed twice with ice-cold PBS, sonicated, and denatured in urea-SDS buffer [29]. The samples were analyzed with 10% SDS-polyacrylamide gel electrophoresis and western blotted using horseradish peroxidase (HRP)-conjugated antibodies against myc (BD Biosciences, San Jose, CA) recognizing CRYAA, glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Sigma), or β-actin (Sigma), or monoclonal antibodies against binding immunoglobulin protein (BiP; BD BioSciences), C/EBP homologous protein/growth arrest and DNA damage-inducible gene 153 (CHOP/GADD153; Santa Cruz Biotech, Santa Cruz, CA), caspase-3 (Santa Cruz Biotech), or phosphorylated ER kinase (PERK; Santa Cruz Biotech), followed by appropriate HRP-conjugated immunoglobulin (Ig) secondary antibodies. Signals were detected by enhanced chemiluminescence (Amersham, Bucks, UK).

Immunofluorescence

Cells were fixed with 2% neutral buffered paraformaldehyde (Sigma) in 0.1 M PBS, permeabilized and detected with mouse monoclonal anti-myc (recognizing CRYAA; Sigma) or anti-protein disulfide isomerase (PDI) antibody (BD BioSciences) followed by appropriate fluorescence-conjugated IgG secondary antibody (Invitrogen) and 4'-6-diamidino-2-phenylindole (DAPI; Sigma) staining. For semiquantitative evaluation of cytoplasmic CRYAA aggregates, >200 cells per sample were analyzed. We graded CRYAA aggregation from 0 to 3+: 0, no aggregate; 1+, few dot-like aggregates; 2+, more dot-like aggregates, and 3+, extensive large aggregates. The staining index, indicative of phenotype severity, was calculated using the formula: 4(3+)% + 2(2+)% + 1(1+)% [30].

Transcription analysis

Total RNA was obtained by an RNA purification kit (RNeasy kit, Qiagen, Valencia, CA) and an on-column DNase digestion kit (RNase-free DNase kit, Qiagen). cDNA from 1 μg RNA, 10 ng/ml random hexanucleotide primer (Invitrogen), and reverse transcriptase (SuperScript III, Invitrogen) was amplified for CRYAA, hsp70, and hsp90 as well as other stress markers (Table 2).

Terminal apoptosis assay

Fixed cells were stained for myc and red X-conjugated secondary antibody (Jackson ImmunoRes Lab, West Gloves, PA), and nuclei were counterstained with DAPI. Samples were examined by fluorescence microscopy (DMRB microscope; Leica, Wetzlar, Germany) equipped with a color imaging system (Spot RT, Diagnostic Instruments, Sterling Heights, MI). Terminal apoptosis rate was represented as percentage of cells with fragmented nuclei. For each experiment (n=3), ten random images (40× objective) were analyzed.

Expression of cataract-causing mutant CRYAA variants in B3 cells

A repertoire of congenital cataract-causing CRYAA mutants (R12C, R21L, R49C, G98R, and R116C) and WT CRYAA were cloned into the pcDNA6-His/myc expression vector and verified by sequencing. Each construct was then transfected to human lens epithelial B3 cells. At 48 h, the cells were collected in 0.5% Tx lysis buffer and fractionated to Tx-soluble and Tx-insoluble proteins. By western blotting for myc, representing CRYAA, most His/myc-tagged CRYAA mutants were Tx-soluble, whereas His/myc-G98R CRYAA was predominantly Tx-insoluble (Figure 1A). By band densitometry analysis, the average Tx solubilities were 98% for WT CRYAA, 95% for R12C, 96% for R21C, 68% for R49C, 96% for R116C, and 1.6% for G98R CRYAA in B3 cells. Hence, the solubility of G98R CRYAA was substantially reduced when compared to WT and other known mutants. This insolubility was consistently observed upon transfection with different ratios of pHis/myc-CRYAA^G98R DNA to FuGene reagent. The transfection efficiency was similar among WT and all studied mutants, as verified by immunofluorescence of myc, representing CRYAA. The efficiency was maintained at about 50% at day 2 for sample collection.

Localization of G98R CRYAA in B3 cells

The immunofluorescence study showed that His/myc-G98R CRYAA formed large-sized cytoplasmic aggregates (Figure 1D), whereas WT CRYAA was located diffusely in the cytoplasm (Figure 1C). To localize the His/myc-G98R CYRAA, we performed confocal double immunofluorescence. The aggregated G98R CYRAA was intensively co-localized with PDI, an ER resident protein (Figure 1E-H). In contrast, WT CRYAA exhibited mild co-distribution with PDI. This demonstrated for the first time that mutant CRYAA formed insoluble aggregates in the ER. Hence, we suggested that the misfolded mutant G98R CRYAA-induced presenile cataract could be an ER storage disease.

G98R CRYAA aggregated in cytoplasm but reduced by trimethylamine N-oxide

We hypothesized that the misfolding of G98R CRYAA with the exposed hydrophobic peptide patches could be rectified by treatment with TMAO, which is an osmolyte with a hydrophobic nature and acts as a chemical chaperone to stabilize proteins with a tightly packed conformation. To authenticate this action, we also tested other small molecule chemicals, 4-PBA, glycerol, and DMSO, by virtue of the chaperoning activity reported previously [31-37]. After treatment for 2 days separately with these chemicals, we examined the distribution of transfected CRYAA. We found that both TMAO and 4-PBA treatments resulted in fewer cells exhibiting large-sized myc-positive aggregates (Figure 2B,C compared to untreated control in Figure 2A). The majority of these mutant cells had cytoplasmic punctate staining. A scoring analysis showed 10% of the TMAO-treated mutant cells and 12% of the 4-PBA-treated mutant cells with a 3+ level of aggregation (Figure 2D). This was highly contrasted to approximately 75% of mutant cells before treatment (graded as 3+). The staining index, representing the phenotype severity, was also dramatically reduced after TMAO or 4-PBA treatment, approaching that of WT-expressing cells (Figure 2E). Treatment with 0.5 or 1% DMSO was ineffective in reducing His/myc-G98R CRYAA aggregates. Also, toxicity was observed in cells treated with 1% or 5% of glycerol.

Trimethylamine N-oxide reduced Triton X-100-insoluble G98R CRYAA

Treatment with 100 mM TMAO or 1 mM 4-PBA for 2 days reduced Tx-insoluble His/myc-G98R CRYAA, whereas the expression of His/myc-WT CRYAA was unaffected (Figure 3A). The percentage of Tx-insoluble mutant protein after TMAO or 4-PBA treatment was similarly decreased to almost one-fourth of that in untreated cells (22% for TMAO treatment and 28% for 4-PBA; p<0.05, independent Student t test; Figure 3B). When examined at 5 days, 4-PBA-treated cells had notable morphological changes, whereas TMAO showed a slender fibroblast-like morphology at more than 5 days, which could be maintained for about 10 days. Due to the undesirable morphological changes, 4-PBA was not selected in further experiments. Treatment with 0.5% or 1% DMSO was again ineffective in reducing the insoluble His/myc-G98R CRYAA (Figure 3B), and 1% or 5% glycerol killed cells. Hence, DMSO and glycerol were not used in subsequent experiments.

A 2-day treatment with TMAO at concentrations from 25 to 300 mM led to a dose-dependent decrease of Tx-insoluble His/myc-G98R CRYAA (Figure 3C). TMAO at 25 mM reduced Tx-insoluble G98R CRYAA to 83% (compared to 98% for untreated mutant protein). Tx-insoluble G98R CRYAA was further reduced to 48% at 50 mM and to 24% at 100 mM treatment. The reduction was statistically significant (p<0.05, independent Student t test). Cell death became prevalent upon treatment with TMAO at concentrations higher than 200 mM. We further observed a time-dependent decrease of Tx-insoluble His/myc-G98R CRYAA when cells were treated with 100 mM TMAO (Figure 3D). At 12 h, the level decreased to 67% (compared to 98% for untreated mutant protein). It was further reduced to 28% at 24 h, 21% at 48 h, and 3% at 72 h, and the reduction was statistically significant (p<0.05, independent Student t test). It is notable that while the insoluble G98R CRYAA decreased by TMAO treatment, the level of soluble G98R did not increase. Hence, we postulated that TMAO could reduce the misfolded polypeptides of CRYAA via degradation rather than by refolding them.

Trimethylamine N-oxide induced G98R CRYAA degradation by ubiquitin-proteasome pathway (UPP)

We next investigated how TMAO reduces Tx-insoluble His/myc-G98R CRYAA and resolves protein aggregates in the ER. In the presence of 10 μM MG132, a reversible inhibitor of UPP, in the transfected cell culture for 8 h, a substantial increase of Tx-insoluble His/myc-G98R CRYAA protein was observed (Figure 4A), and the reduced level of Tx-insoluble G98R CRYAA after TMAO (100 mM, 2 days) was reversed by MG132 (10 μM, 8 h). This indicated that G98R CRYAA protein could be degraded via UPP. Blocking of UPP resulted in more insoluble protein and augmented the mutant CRYAA aggregation in cells, as demonstrated by immunofluorescence (Figure 4D, compared to non-MG132-treated G98R cells in Figure 4B). After treatment with 100 mM TMAO for 2 days, a low level of Tx-insoluble mutant was again observed (Figure 4A). The levels were similar as in previous experiments. Further incubation of these TMAO-treated cells with 10 μM MG132 for 8 h increased Tx-insoluble CRYAA protein. This was not significantly different from the untreated mutant cells. Immunofluorescence showed that His/myc-G98R CRYAA aggregates reappeared inside cells (Figure 4E, compared to TMAO-treated cells in Figure 4C). We semiquantified the cells with visible intracellular G98R CRYAA aggregates. Nearly all MG132-treated mutant cells showed His/myc-CRYAA aggregates (Figure 4F). The simultaneous treatment with TMAO and MG132 resulted in 73% of cells with aggregates, which was similar to the untreated mutant cells (81%). Only TMAO treatment significantly reduced the cells with CRYAA aggregates to 36% (p<0.05 when compared to untreated cells or cells with both TMAO and MG132 treatments). On the other hand, treatment with 10 μM 3-methyladenine (3-MA, an inhibitor of autophagy) did not show any change of His/myc-G98R CRYAA solubility and subcellular aggregation (Figure 4G), indicating that CRYAA was not a substrate of autophagy. In addition, the solubility and subcellular aggregation of WT CRYAA was not affected by treatment (data not shown).

Endoplasmic reticulum stress and apoptosis caused by aggregated G98R CRYAA were alleviated by trimethylamine N-oxide

We examined if TMAO affected the expression of ER stress response proteins, including PERK and BiP, in transfected B3 cells. These proteins were upregulated upon His/myc-G98R CRYAA expression when compared to WT CRYAA (Figure 5A). The expressions of PERK and Bip in mutant-expressing cells were increased about twofold compared to the WT-expressing cells (PERK from 13% to 22%; BiP from 51% to 92%). B3 cells incubated with a transfection reagent had no observable changes. There was, therefore, induction of ER stress due to mutant protein expression. With TMAO (100 mM) treatment for 2 days, the ER stress marker expression was substantially reduced to a level about one-half of the untreated cells (p<0.05; Figure 5A). The expression of PERK decreased from 22% to 7% and that of Bip decreased from 92% to 47%. WT-expressing cells were not affected by TMAO treatment, indicating that this chemical chaperone acted on proteins that were not natively folded.

We further explored the consequence on apoptotic cell death. Western blot analysis showed that CHOP/GADD153 and caspase-3 were upregulated in His/myc-G98R CRYAA-expressing B3 cells nearly twofold compared to WT-expressing cells (CHOP from 13% to 19%, caspase-3 from 20% to 29%; p<0.05; Figure 5A,B). This was accompanied by an increased percentage of cells undergoing apoptosis and showing fragmented nuclei revealed by double immunofluorescence of myc (representing CRYAA) and nuclear dye DAPI (G98R CRYAA cells 22% versus WT CRYAA cells 6%; Figure 5C). Treatment with TMAO (100 mM) for 2 days resulted in reduced expression of both CHOP/GADD153 and caspase-3 (p<0.05; Figure 5A,B). The expression of CHOP was decreased from 19% to 6% and that of caspase-3 from 2% to 18%. Consistent results were obtained in experiments done in triplicate. The suppression of apoptotic proteins was supported by a lower percentage of cells with fragmented nuclei in TMAO-treated G98R CRYAA cells (13% versus 22% in untreated G98R CRYAA cells; Figure 5C). TMAO treatment of WT cells did not induce changes in apoptotic protein expression or in amounts of apoptotic cells (Figure 5B,C).

Effect of trimethylamine N-oxide treatment on cell stress signaling

We sought to examine if TMAO treatment (100 mM, 48 h) affected pathways of stress signaling on G98R CRYAA cells. We studied gene expressions involved in heat-shock responses, including Hsp70 and 90α, and cell stress responses. By semiquantitative PCR we confirmed an upregulation of Hsp70 by using two different pairs of specific primers, Hsp70 (1) and Hsp70 (2) (Figure 6). No change was observed for Hsp90α expression. In nontransfected cells, Hsp70 and Hsp90 expression was not altered by TMAO treatment (Figure 6). On the other hand, gene expression of cell stress signaling (superoxide dismutase SOD and αB-crystallin CRYAB for oxidative stress, metallothionein MT for heavy metal-induced stress, and interleukin-6 IL6 for inflammatory-associated stress) was unaffected by TMAO.