Antigen Retrieval to Improve the Immunocytochemistry Detection of Sigma-1 Receptors and ER Chaperones

Abstract

Molecular chaperones localized at the endoplasmic reticulum (ER) lumen constitutively or cellular stress-dependently associate with a variety of proteins to promote their proper folding or to inhibit protein misfolding. ER chaperones preferentially form large complexes with co-chaperones and/or misfolded proteins in a highly crowded cellular environment that often hampers their detection by immunocytochemistry (ICC). This study establishes an antigen retrieval (AR) protocol to improve the ICC detection of ER chaperones in cultured cells using widely available antibodies against synthetic peptides. Among ten different antigen retrieval/fixation conditions, only the AR with Tris-HCl (pH 9.5) containing 6 M urea (80 °C for 10 min) significantly improved the ICC detection of the novel ER chaperone sigma-1 receptor (Sig-1R) in Chinese hamster ovary cells. Extended fixation with 4% paraformaldehyde for 1 hr effectively preserved the morphology of the ER under the AR condition. This method greatly enhanced the signal-to-noise ratio in Sig-1R ICC, thus allowing for semi-quantitative detection of protein upregulation under ER stress. The AR similarly improved the ICC detection of a series of other major ER chaperones, including BiP/GRP78, GRP94, calnexin, calreticulin, ERp57, protein disulfide isomerase, and cyclophilin B. The improved ICC methodology using the urea AR at 80°C may improve ICC of ER molecules as well as visualization of ER structure and substructures.

Introduction

The endoplasmic reticulum (ER) is the main site for calcium (Ca²⁺) storage and for the synthesis, modification, and delivery of nearly all soluble and membrane proteins. Thus, the highest concentrations of folded and unfolded proteins reside in the ER lumen (Schroder and Kaufman 2005). The proper folding and post-translational modification of nascent polypeptides entering the ER is dependent on ER-resident enzymes and chaperone proteins (Schroder and Kaufman 2005; Yamamoto et al. 2008). The highly selective quality-control machinery in the ER allows the transport of only correctly folded proteins to the Golgi, while misfolded proteins are retained in the ER where they either complete the folding process or are marked for degradation (Yamamoto et al. 2008; Meusser et al. 2005). Disruption of any of these processes by alterations in redox state, Ca²⁺ levels, and/or failure to posttranslationally modify secretory proteins compromises the overall ability of the ER to produce properly folded proteins, and is collectively referred to as ER stress (Schroder and Kaufman 2005; Kopito and Ron 2000). ER stress causes the accumulation of unfolded and/or misfolded proteins and consequently activates the unfolded protein response (UPR) in an effort to restore normal ER functioning by reducing unfolded and/or misfolded protein accumulation (Yoshida et al. 2001; Schroder and Kaufman 2005; Kopito and Ron 2000).

Certain ER luminal chaperones contain a C-terminal KDEL sequence that serves as an ER-retention signal and constitutively associates with nascent polypeptides to facilitate proper folding by masking regions that might otherwise interact with each other and lead to misfolding or aggregation (Capitani and Sallese 2009). Under UPR, chaperone proteins are transcriptionally upregulated, thus allowing for more chaperone proteins to associate with accumulated unfolded proteins (Schroder and Kaufman 2005; Yoshida et al. 2001). Interestingly, UPR also causes the translocation of chaperone proteins from the ER lumen to extra-ER localizations (Sun et al. 2006; Johnson et al. 2001). Indeed, stress induces a significant amount of BiP to redistribute to the cytosol and the ER membrane (Sun et al. 2006). Particularly, when the ER becomes stressed due to Ca²⁺ depletion, BiP has been found to localize to the inter-membrane space, inner membrane, and matrix of the mitochondria and may allow for a change in function to modulate Ca²⁺ homeostasis (Sun et al. 2006). Likewise, upregulation of calreticulin has been implicated in Ca²⁺ signaling and other cellular activities such as cell adhesion through translocation to the cytosol, cell surface, and nucleus (Johnson et al. 2001). It remains unclear; however, how BiP and calreticulin escape the KDEL-mediated retention system in order to translocate to these extra-ER localizations (Johnson et al. 2001; Sun et al. 2006).

Several properties of ER chaperone proteins influence their successful immunocytochemistry (ICC) staining including the native 3-D structure of the chaperone, association of the chaperone with co-chaperones or client proteins, and conformational changes or translocation of the chaperone under stress. For example, more than 50% of cellular BiP associates with other chaperone proteins including calreticulin and sigma-1 receptor, and this complex may function as a store from which BiP is released to interact with misfolded proteins (Crofts et al. 1998; Hayashi and Su 2007). Indeed, epitopes at the C-terminus of BiP are masked when in complex with calreticulin, which are not masked when BiP is bound to an unfolded protein (Kleizen and Braakman 2004). To bind unfolded proteins, the ATPase domain of BiP coordinates ATP hydrolysis to induce a conformational change in its 3-D structure (Hendershot et al. 1995). Furthermore, BiP acts upon nascent polypeptides as a complex that may consist of GRP94, PDI, cyclophilin B, and/or ERp57 (Kleizen and Braakman 2004; Crofts et al. 1998). In many instances, ICC is unable to quantitatively assess the translocation and upregulation of chaperone proteins under stress (Sun et al. 2006; Hayashi and Su 2007); therefore, Western blotting is mainly utilized for these purposes [e.g., (Sun et al. 2006)].

Antibodies (Abs) for ICC can be produced using either synthetic (denatured) peptides or purified proteins preserving the native folding conformation (Van Regenmortel 1989). Synthetic peptide production is advantageous for its ease in obtaining antigen peptides. In many cases; however, the Ab raised against synthetic short peptides is not able to recognize the 3-D conformation of the protein or the surface of the native protein, and often protein associations or folding hinder antigen association with the Ab (Van Regenmortel 1989). In contrast, antibodies raised against folded functional proteins often work better in ICC recognizing the native conformation of proteins, although a tedious process is necessary to purify folded protein (Zolla-Pazner et al. 1999). Also, polyclonal Abs raised from folded proteins often fail to recognize denatured or conformationally changed proteins, which makes them problematic for Western blotting and ICC of misfolded proteins (Zolla-Pazner et al. 1999; Van Regenmortel 1989).

In order to overcome the drawbacks associated with Abs based on synthetic peptides, methods of antigen retrieval (AR) were introduced into immunohistochemistry to unmask the target protein epitopes of interest by these antibodies (D’Amico et al. 2009; MacIntyre 2001; Shi et al. 2001). One of the earliest methods of AR was proteolytic digestion, which was followed by heat-induced antigen retrieval protocols (D’Amico et al. 2009; MacIntyre 2001; Shi et al. 2001; Mori et al. 2002). Currently, the most popular heat- induced methods of antigen retrieval include using a microwave, steamer, or pressure cooker in combination with various retrieval buffers including citrate, EDTA, and Tris (Warembourg and Leroy 2000; Taylor et al. 1996; Taylor et al. 1994; Morgan et al. 1994). Indeed, ER chaperone proteins often require AR for immunostaining (Hendershot et al. 1995; Zhao et al. 2005). However, it is hard to predict which AR method is suitable for each particular protein, and few studies aimed at improving ICC of ER luminal chaperones have been systematically performed. Thus, the purpose of our study is to develop an AR protocol that improves immunocytochemical staining of ER chaperones and that semi-quantitatively measures upregulation of ER chaperones. We first focus on ICC of the novel ER chaperone sigma-1 receptor (Sig-1R), and then apply the established AR method to ICC of other ER chaperone proteins.

Materials and methods

Antibody

Antibodies (Abs) used in the current study are listed in Table 1. The anti-Sig-1R Ab 5460 was produced in a rabbit against synthetic peptides corresponding to a.a. 143–165 of the rat Sig-1R(Hayashi and Su 2007). The rabbit anti-Sig-1R Ab-full was raised against purified full-length Sig-1Rs by Ruoho and colleagues(Mavlyutov and Ruoho 2007), and kindly provided. Other anti-Sig-1R Abs were purchased from AbCam (ab71446, ab53852, ab89655; Cambridge, MA) or Invitrogen (42–3300; Carlsbad, CA). Alexa fluorophore-conjugated goat anti-mouse or anti-rabbit IgG secondary Abs were purchased from Invitrogen.

Table 1.

Characteristics of antibodies used in this study

Antibody	Immunogen	Source
Sig-1R pAbs
• Sig-1R 5460	synthetic rat Sig-1R143–165	Hayashi and Su (2007)
• Sig-1R-full	purified full-length guinea-pig Sig-1R	Mavlyutov and Ruoho (2007)
• Sig-1R ab71446	synthetic N-terminus of human Sig-1R	AbCam
• Sig-1R ab53852	synthetic C-terminus of rat Sig-1R	AbCam
• Sig-1R ab89655	synthetic internal domain of human Sig-1R	AbCam
• Sig-1R 42–3300	synthetic C-terminus of rat Sig-1R	Invitrogen
BiP/GRP78 pAb SPA-826	synthetic C-terminus of rat BiP	StressGen
BiP/GRP78 mAb 610979	recombinant human BiP a.a. 625–628	Transduction Lab
GRP94 pAb ab13509	synthetic mouse GRP94 a.a 787–802	AbCam
Calnexin pAb sc11397	recombinant human calnexin a.a. 1–70	Santa Cruz
Calnexin mAb 610523	recombinant human calnexin a.a. 116–301	Transduction Lab
Calreticulin pAb sc6468	synthetic N-terminus of human calreticulin	Santa Cruz
PDI pAb sc20132	recombinant human PDI a.a. 211–370	Santa Cruz
PDI mAb SPA-891	synthetic C-terminus of rat PDI	StressGen
Cyclophilin B pAb ab16045	synthetic C-terminus of human cyclophilin B	AbCam
ERp57 mAb ab13506	recombinant full-length human ERp57	AbCam

Cell culture and transfection

Chinese hamster ovary (CHO) cells were maintained at 37 °C with 5% CO₂ in the minimum essential medium (MEM)-alpha containing 10% fetal bovine serum. Prior to transfection, CHO cells were seeded into 3-cm dishes or 12-well plates and allowed to grow for 24 hrs. A plasmid (4 μg)/Lifofectamine-200 (10 μl; Invitrogen) mixture in 200 μl of Opti-MEM (Invitrogen) was applied to cells with 2 ml of the culture medium for 8 hrs. Cells were then harvested with trypsin-EDTA, and cultured in new dishes with fresh growth medium. Expression vectors for control and Sig-1R siRNAs were constructed as previously described(Hayashi and Su 2004). To establish a CHO cell line stably expressing rat Sig-1Rs with a V5 epitope (Sig-1R-V5), CHO cells were transfected with pcDNA3.1-Sig-1R-V5/His expression vectors (Invitrogen) and maintained in MEM-alpha containing Geneticin for two months; positive colonies were screened four times.

Immunocytochemistry

CHO cells were grown on poly-D-lysine-coated 1-cm round coverslips at 50% confluency. After aspiration of medium, cells were immediately fixed by applying 1 ml of 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS: 10 mM phosphate buffer, pH 7.2, containing 0.15 M NaCl). After incubation at room temperature for 10–60 min, fixed cells were washed three times (10 min) with PBS. In some cases, cells were alternatively fixed using a 95% ethanol-5% acetic acid mixture (10 min at 4°C) or 100% methanol-NaOH (15 min at −20°C). Following washing, permeabilization was performed with 0.1% NP-40 in PBS at room temperature for 10 min and quenching was performed to remove free PFA by incubating samples with 10 mM glycine solution (pH 7.2) for 10 min at room temperature. After AR, cells were incubated with 10% non-fat milk in PBS at room temperature for 1 hr, and then incubated with primary Abs in Tris-buffered saline with Tween-20 (TBST: 0.025 M Tris-base, pH 7.4, 0.137 M NaCl, 0.003 M KCl, 0.05% Tween-20) containing 2% bovine serum albumin (BSA). Following three washes with TBST (10 min) fixed cells were incubated with secondary Abs (1:200) in PBS containing 2% BSA at room temperature for 1 hr, followed by two washes with TBST (10 min). Samples were mounted on a slide grass with ProLong Antifade solution (Invitrogen) before confocal microscopy. Confocal images were captured using a PerkinElmer Ultraview confocal system (PerkinElmer, Covina, CA) and analyzed with Volocity software (Improvision, Waltham, MA).

Western blotting

Total CHO cell lysates were prepared in sodium dodecyl sulfate (SDS)-sample buffer as described previously(Hayashi and Su 2007, 2004). Protein concentrations in lysates were measured by a BCA kit (Thermo, Rockford, IL), and thirty μg of total proteins were resolved by SDS-PAGE. Proteins were transblotted onto a polyvinylidene fluoride membrane with Mini Trans-Blot Cell (BioRad; Hercules, CA). Following blocking with 10% non-fat milk in TBST for 1 hr, the membrane was incubated with specific primary Abs (1:1000 in TBST). After washing twice with TBST (10 min), the membrane was incubated with horseradish peroxidase-conjugated secondary Abs (Thermo) for 1 hr, and specific bands were visualized with a Kodak IS4000 MM Pro system (Carestream Health, Rochester, NY) by digitizing West Femto chemiluminescence (Thermo).

Results

While our previously developed polyclonal anti-Sig-1R Ab detects Sig-1Rs in ICC, there is a significant demand to optiomize Sig-1R ICC because of the following reasons: 1) A reliable signal-to-noise ratio is achieved only in limited cell types constitutively expressing high levels of Sig-1Rs; 2) Over-fixation with PFA (> 10 min) significantly reduces immunoreactivity; 3) Although all of our Abs developed against synthetic peptides worked excellently in immunoblotting, only two out of eight worked in ICC (unpublished data), which substantially limits the available sources of Abs; 4) The currently available Sig-1R Abs can rarely detect stress-induced acute protein upregulation in ICC; and 5) The Sig-1R pAb recently developed by another laboratory using purified full-length Sig-1Rs (Sig-1R-full Ab) worked excellently in ICC (Mavlyutov and Ruoho 2007), but barely detects Sig-1Rs in immunoblotting. Further, despite abundance of ER chaperones (e.g., BiP/GRP78 and PDI) in a large variety of cells, the ICC detection of these proteins often encounters similar problems (data not shown). In order to establish an AR protocol suitable for ICC of ER chaperones, we first focused on improving Sig-1R ICC using Sig-1R 5460 pAb by testing various fixation and AR procedures.

Although heat-induced citrate AR is a well-established protocol in immunohistochemistry, it failed to improve Sig-1R ICC in an affordable range of temperatures (55–95 °C). Rather, it caused a significant increase in background signals (Table 2). Fixation of cells with the 95% ethanol-5% acetic acid mixture, but not citrate AR, improved the ICC detection only when Sig-1Rs were overexpressed by vector transfection (Table 2). The heat-induced urea AR with Tris caused a slight destruction of ER membranes or cell morphology when fixation with 4% PFA was carried out for 7–15 min and the AR temperature exceeded 55 °C. Among the ten different fixation and AR protocols tested, only heat-induced AR with Tris (pH 9.5) containing 6 M urea significantly improved ICC of Sig-1R 5460 pAb over current method (Table 2). Indeed, the application of the urea AR equally improved ICC of both endogenous and transfected Sig-1Rs. Cell morphology was successfully preserved under heat-induced AR by extending the fixation time (Table 2). Cells fixed with 4% PFA for 60 min did not show any notable effect of the AR on ER and cell morphology even at 95 °C when assessed by confocal microscopy (data not shown).

Table 2.

Relative intensity of Sig-1R ICC staining with anti-Sig-1R 5460 pAb following various fixation or antigen retrieval protocols

Fixation	Antigen retrieval		Sig-1R staining
1. 4% PFA (7 min)		−	+
2. 4% PFA (60 min)		−	−
3. 4% PFA (7 min)	10 mM citrate + 1 mM EDTA (pH 6.0)
		55°C, 10 min	−*
		80°C, 10 min	−*
4. 4% PFA (60 min)	50 mM glycine-HCl (pH 3.5)
		55°C, 10 min	−*
		80°C, 10 min	−*
5. 4% PFA (60 min)	1% SDS	22°C, 10 min	−
6. 4% PFA (60 min)	0.05% trypsin/EDTA	22°C, 5min	−***
7. 4% PFA (60 min)	6M guanidine/50 mM Tris (pH 7.4)
		22°C, 10 min	−
8. 95% EtOH/5% AcA (10 min, 4°C)		−	++**
9. 100% MeOH/NaOH (−20°C, 15 min)		−	−
10. 4% PFA (7 min)	6M urea in 0.1M Tris (pH 9.5)
		55°C, 10 min	++**
		70°C, 10 min	−***
11. 4% PFA (60 min)	6M urea in 0.1M Tris (pH 9.5)
		55°C, 10 min	++**
		70°C, 10 min	++
		70°C, 15 min	+
		70°C, 20 min	+
		80°C, 10 min	++++
		80°C, 15 min	+++
		80°C, 20 min	++
		90°C, 10 min	++
		90°C, 15 min	+
		90°C, 20 min	+*

^*Strong background staining observed

^**Effective only in ICC of overexpressed Sig-1Rs, but not that of endogenous Sig-1Rs

^***Cells dislodged or destructed

PFA, paraformaldehyde; EtOH, ethanol; AcA, acetic acid; MeOH, methanol.

The effect of the urea AR on Sig-1R ICC was largely dependent on incubation temperature and time. Eighty degrees for 10 min gave the best result, whereas, incubation at either 70 or 90 °C (for 10 min) was less effective (Table 2). Indeed, longer incubations at 90 °C reduced the specific signal and showed an increase in background signals. In contrast, the optimized AR condition (Tris-urea at 80 °C for 10 min) increased the specific signal, but also decreased background signals as is evident from negligible signals observed in cells transfected with Sig-1R siRNA (Fig. 1). Of note, the AR protocol permitted the use of much lower concentrations of Abs (mostly 5~10 times lower concentration).

Figure 1. Improved signal-to-noise ratio in Sig-1R immunocytochemistry following urea antigen retrieval (AR).

Sig-1R siRNA was transfected with GFP cDNA to locate cells knocking down Sig-1Rs (in green; siSig-1R+). Sig-1R 5460 Ab were incubated at a 1:500 dilution followed by labeling with Alexa 590-anti rabbit goat secondary Ab (in red). Green and red fluorescence intensities along selected lines are displayed in the middle panels. The same confocal setting was used to capture images A and B for a better comparison of immunoreactivity. The graphs represent mean ± s.e.m. of red fluorescence (Sig-1R immunoreactivity) from 12–20 fields (>100 cells). Scale bar = 5 μm.

The modified AR protocol also improved Sig-1R ICC using some Abs from different sources, such as ab53852, ab89655, and 42–3300 (see Fig. 2). Interestingly, the urea AR (Fig. 2) or ethanol-acetic acid fixation (data not shown) largely reduced Sig-1R staining with Sig-1R-full pAb, indicating that the immunoreactivity of the antibody is highly dependent on the structural conformation of Sig-1Rs, and that the urea AR or ethanol-acetic acid fixation may cause an alteration of the conformation.

Figure 2. Immunocytochemical detection of Sig-1Rs with various Sig-1R pAbs in CHO cells: the effect of urea antigen retrieval (AR; 80°C, 10 min).

A. Western blot analyses of CHO cell lysates with various Sig-1R pAbs. Total cell lysate were prepared from CHO cells transfected with control siRNA or Sig-1R siRNA. Note: for better comparison of immunoreactivity, all membranes were developed by exposing membranes for 1 min; except for the membrane with Sig-1R 5460 that was exposed for 10 sec. The extended exposure revealed clear Sig-1R bands with all Sig-1R Abs except for Sig-1R-full Ab (data not shown). Nucleoporin p62 serves as loading controls. B. Immunocytochemistry of endogenous Sig-1Rs in CHO cells. Sig-1R siRNA was transfected with GFP cDNA to identify cells knocking down Sig-1Rs (in green). All Sig-1R Abs were incubated at a 1:500 dilution followed by labeling with Alexa 590-anti rabbit goat secondary Abs (in red). Asterisks in red images indicate cells transfected with Sig-1R siRNA. Note that the same confocal setting was used during capturing all images for a better comparison of immunoreactivity, thus settings of confocal imaging may not be necessarily optimal in all displayed images. Bar = 5 μm.

Since the established urea AR protocol dramatically increased immunoreactivity of specific Abs as well as decreasing the background staining, we speculated that the protocol enables detection of protein upregulation with ICC. To investigate this aspect we first examined whether the AR provides ICC intensities proportional to protein levels of Sig-1Rs. To this end, we employed CHO cells stably overexpressing Sig-1R-V5 (CHO-Sig-1R-V5) where each cell expresses various levels of Sig-1R-V5 proteins in culture. Sig-1R-V5 expressed in individual cells was semi-quantified with highly selective anti-V5 monoclonal Abs (Invitrogen) in ICC. After double staining with V5 mAb and Sig-1R 5460 pAb, the relationship of ICC signals derived from each antibody was examined for each pixel of confocal images. Pixel intensities of V5 and Sig-1R 5460 Abs were plotted at x- and y-axes, respectively, to examine the correlation between the two labelings. As shown in Fig. 3, a linear correlation was observed between fluorescence intensities derived from two immunolabelings. Notably, the urea AR provided superior linearity and a steeper linear slope in the correlation plotting when compared to samples without AR [slope=0.2605 (AR−), =2.298 (AR+); R²=0.8340 (AR−), =0.9932 (AR+)].

Figure 3. Immunocytochemical detection of Sig-1Rs in proportion to the protein level.

CHO cells stably overexpressing Sig-1R-V5 were immunolabeled with anti-V5 mAb (in green) and Sig-1R 5460 pAb (in red). V5 immunocytochemistry demonstrates a range of Sig-1R-V5 protein levels expressed in each cell. Fluorescence pixel intensities from V5 (x-axis) and Sig-1R 5460 (y-axis) Abs are plotted to present the relationship between green and red ICC signals in confocal images. Note that the urea AR (+) provides better linearity with a steeper linear slope in the correlation plotting when compared to AR (−). In graphs, pixels where green fluorescence is greater than red are shown in green, whereas those with red intensity greater than green are shown in red. For analyses of red fluorescence, the minimum threshold was set to exclude signals from endogenous Sig-1Rs. Bar=5 μm.

Next, we examined whether the urea AR protocol is applicable for assessing ER stress-induced upregulation of endogenous Sig-1Rs. Immunoblotting of Sig-1Rs confirmed that thapsigargin (1 μM, 1 hr), an ER Ca²⁺/ATPase pump inhibitor that depletes ER Ca²⁺ stores thus promoting ER stress, significantly increased the Sig-1R protein level [Fig. 4 and (Hayashi and Su 2007)]. ICC without urea AR failed to detect the thapsigargin-induced increase of Sig-1R proteins [fluorescence intensity before thapsigargin (1 μM for 60 min) = 19.11 ± 1.45; after thapsigargin = 19.66 ± 1.83. N = 11~17 fields (total > 200 cells in each group)]. In contrast, the urea AR successfully detected the upregulation of endogenous proteins. Although, the conformation-sensitive Sig-1R-full pAb detected Sig-1Rs excellently in non-treated cells, the pAb hardly detects upregulation of Sig-1Rs, suggesting that transcriptionally upregulated nascent Sig-1Rs may include a 3-D conformation insensitive to the Sig-1R-full Ab or that Sig-1Rs under ER stress may change its conformation or associated partner proteins to impede recognition of the epitope by Sig-1R-full Ab.

Figure 4. Immunocytochemical detection of upregulated endogenous Sig-1Rs.

CHO cells were treated with vehicle (Con) or thapsigargin (TG; 1 μM, 60 min) followed by immunocytochemistry with Sig-1R 5460 or Sig-1R-full Abs (1:500). Sig-1R Abs were labeled with Alexa 590 anti-rabbit IgG goat secondary Ab (in red). The urea antigen retrieval protocol was applied, AR(+), for immunodetection with Sig-1R 5460 Ab. The same confocal setting was used to capture all four images for a better comparison of immunoreactivity. The graphs represent mean ± s.e.m. of red fluorescence (Sig-1R immunoreactivity) from 10–15 fields (>200 cells). Inset: a representative immunoblotting with Sig-1R 5460 pAbs showing TG-induced upregulation of Sig-1Rs in CHO cell total lysates. Nucleoporin p62 serves as a loading control.

Lastly, we tested if our urea AR protocol is applicable to ICC of other ER chaperones. The urea AR increased the ICC detection of different classes of ER chaperones, such as BiP/GRP78, GRP94, cyclophilin B, ERp57, PDI, calreticulin, and calnexin. Under the urea AR, all ICC exerted typical reticular staining of the ER together with some punctate staining. More strikingly, the AR protocol dramatically improved the background problems seen in some samples. For example, in ICC with ERp57 mAb (ab13506), calnexin pAb (sc11397) and PDI mAb (SPA-891), urea retrieval reduced relative background intensity in nuclei and also abolished background signals derived from large cytoplasmic structures.

Discussion

In contrast to the widely accepted AR techniques for immunohistochemistry of formalin-fixed, paraffin-embedded tissues, efforts to establish AR for free floating sections or even for fixed culture cells are much less developed. This deficiency is due in part to limitations in the use of heat-based AR in ICC, such as heat-induced background problems and deleterious effects on morphology (D’Amico et al. 2009; Robinson and Vandre 2001). Heat-based AR using urea was introduced in the early 1990’s for unmasking antigen on formalin-fixed, paraffin-embedded tissues (Shi et al. 1993; MacIntyre 2001; D’Ambra-Cabry et al. 1995; Shi et al. 1996; Taylor et al. 1994). Although urea AR is not as popular as other methods, such as citrate and EGTA, there are several reports demonstrating the unmasking effect of urea on formalin-fixed, paraffin-embedded tissues (Taylor et al. 1994; Shi et al. 1996). For example, the urea AR was successfully introduced to detect carbonic anhydrase II and mitochondrial cytochrome c oxidase subunit I in ICC of cultured cells (Chen et al. 2008; Janes et al. 2004). However, studies testing urea AR in ICC are limited in number, and few systematic studies have been performed to evaluate the overall benefit of urea AR in ICC.

The major potential problem in introducing heat-based ARs to ICC is increased morphological instability of fixed culture cells when compared to fixed tissues in immunohistochemistry (D’Amico et al. 2009). In many cases, fixed cultured cells cannot endure conventional heat-based ARs (D’Amico et al. 2009; Robinson and Vandre 2001). In fact, heat incubation over 55°C deteriorated cell morphology of CHO cells fixed with 4% PFA for 10 min (Table 2). Further, heat incubation with citrate considerably increased background staining in the same sample (Table 2). However, by optimizing multiple factors, these shortcomings of heat-based AR in cultured cells were overcome in the present study. Our protocol combines urea AR with extended PFA fixation (60 min) in order to preserve cell morphology during heat-based AR, and employs a water bath for mild heat incubation (80 °C) to avoid tools such as a microwave or high-pressure jar that may deteriorate cell morphology (Jiao et al. 1999). Although the urea AR will not be of benefit for all ICC, our protocol should be applicable to a wide range of applications in cell culture-based ICC.

Although heat is considered the most important factor for antigen unmasking (MacIntyre 2001; D’Amico et al. 2009; Shi et al. 2001), the effects of unmasking solutions diverged widely under the same operating temperature. Only urea showed a promising unmasking effect in Sig-1R ICC (Table 2). Our data indicates that an additional mechanism is involved in improving antigen unmasking, requiring a combination of high temperature and exposure to the denaturing reagent. The exact mechanism of the urea AR is unknown. However, since ethanol-acetic acid fixation, guanidine, and SDS failed to unmask the Sig-1R epitope in apparent contrast to their successful unmasking of certain cytoplasmic proteins (Robinson and Vandre 2001; Peranen et al. 1993), there seems to be a denaturizing process that is unique to urea in the ER that is not accomplished with other denaturants.

In the current study, the urea AR increased specific signals, allowed much higher dilution magnitude of primary Abs, and often decreased background signals, resulting in a significant enhancement in the signal-to-noise ratio of ICC. Improved accessibility of Abs may be achieved by denaturization of antigenic epitopes that are inaccessible in the original native 3-D conformation. Since antigen recognition of the Ab raised against a synthetic peptide likely relies more on the amino acid sequence comprising the short antigen peptide and less on the native conformation of the protein, a larger number of Abs will be applicable to ICC with AR with urea denaturing as evident in the robust improvement of ICC using commercially available Abs in our study (Table 1, Fig. 5). However, it should be noted that the denaturing effect of urea AR also has certain shortcomings. For example, we often observed a significant reduction of GFP fluorescence after urea AR. In ICC studies performed in conjunction with expression of GFP-tagged proteins, immunolabeling of GFP might be necessary to enhance green fluorescence in some instances. Furthermore, future studies need to clarify if the urea AR method preserves microstructures of the cell to the degree applicable to high-resolution microscopic observations.

Figure 5. Immunocytochemistry of ER chaperones in CHO cells: the effect of urea antigen retrieval (AR).

All primary Abs were incubated at 1:300 except for calnexin 610523 (1:100) and cyclophilin B (1:500). Primary Abs were labeled with Alexa 480 anti-rabbit, anti-mouse or anti-goat Abs. For better comparisons of immunoreactivity, the setting of confocal imaging was kept constant throughout observations of all samples. Bar = 5 μm. N; nucleus with background staining. Note that AR significantly increases ER staining as well as lowering relative background labeling of nuclei or cytoplasmic structures in PDI SPA-891, ERp57 ab13506, and calnexin sc11397 immunostainings.

The demand for reliable ICC detection of ER chaperones is expanding as an increasing variety roles of ER chaperones in signal transduction, cell survival, and human diseases are unveiled (Schroder and Kaufman 2005; Yoshida et al. 2001; Kopito and Ron 2000). ER chaperones are also used as ER markers to visualize ER structures, which provides further justification for optimizing ICC detection of ER chaperones. It is also noteworthy that recent studies have demonstrated that certain classes of ER chaperones are highly clustered at specialized subdomains of ER membranes (Hayashi et al. 2009; Hayashi and Fujimoto 2010). For example, Sig-1Rs and BiP are shown to cluster at ER subdomains physically associated with mitochondria (i.e., the mitochondria-associated ER membrane; MAM) (Hayashi and Su 2007; Hayashi et al. 2009). Thus, ICC of MAM proteins often shows a punctate staining along with reticular structures of ER membranes (Hayashi and Su 2007; Hayashi and Fujimoto 2010). The MAM plays important roles in regulation of the IP₃ receptor-induced Ca²⁺ influx in mitochondria, bioenergetics, lipid metabolism, and cell survival (Rizzuto et al. 1993; Hajnoczky et al. 2003; Hayashi et al. 2009; Szabadkai and Duchen 2008). Thus the improved ICC detection of ER chaperones may help across a broad range of cell biology research involving visualization and characterization of the ER and its subdomains.