Molecular Vision 2004; 10:933-942 <>
Received 28 June 2004 | Accepted 30 November 2004 | Published 13 December 2004

Nuclear and plasma membrane localization of SH3BP4 in retinal pigment epithelial cells

Kornnika Khanobdee,1,2 Jon B. Kolberg,1 Jane R. Dunlevy1

1Department of Anatomy and Cell Biology, School of Medicine and Health Sciences, University of North Dakota, Grand Forks, ND; 2Neuro-Behavioral Biology Center, Institute of Science and Technology for Research and Development, Mahidol University Salaya, Nakornpathom, Thailand

Correspondence to: Dr. Jane R. Dunlevy, Department of Anatomy and Cell Biology, University of North Dakota School of Medicine and Health Sciences, 501 North Columbia Road, Grand Forks, ND, 58203; Phone: (701) 777-2575; FAX: (701) 777-2477; email:


Purpose: The SH3BP4 protein contains domains belonging to the Eps15-Homology (EH) network family of endocytosis proteins and a C-terminal death domain. The purpose of this study was to determine the expression of SH3BP4 in ARPE-19, Y79 and COS-7 cell lines and to determine SH3BP4 subcellular localization within ARPE-19 cells.

Methods: A chicken anti-human SH3BP4 antibody was generated that specifically immunostains SH3BP4 fusion proteins and a corresponding endogenous protein band at 120 kDa. Protein expression of SH3BP4 was determined using western analysis of multiple cell lines and dissected retinal tissue. Intracellular localization of both endogenous SH3BP4 and SH3BP4 fusion proteins were determined using subcellular fractionation and microscopy studies using ARPE-19 and COS-7 cells.

Results: The retinal pigment epithelial (RPE) cell line ARPE-19 was found to express SH3BP4 protein at more than 7 fold the levels in Y79 retinoblastoma cells and more than 2.5 fold the levels in COS-7 cells. Both the RPE and neural retinal layers of the eye were also found to express the SH3BP4 protein. SH3BP4 endogenous and fusion proteins were found to localize to both membrane and nuclear fractions but not the cytosol in subcellular fractionation experiments. Subsequent microscopy analyses show that SH3BP4 fusion proteins localize to the plasma membrane and the nuclear periphery.

Conclusions: These studies show that SH3BP4 is expressed in the RPE and neural retina in vivo, and in ARPE-19, Y79, and COS-7 cell lines. Compared to other EH network and death domain proteins, SH3BP4 fusion proteins have an unusual intracellular localization to the plasma membrane and the nuclear periphery. The present demonstration of the suborganelle localization in conjunction with the unique domain combinations belonging to both endocytosis and cell death pathways suggests that SH3BP4 has physiological significance for RPE cells.


SH3BP4 was discovered by isolating a cDNA clone amplified in differential display studies on cultured human corneal fibroblasts [1]. SH3BP4 is encoded on human chromosome 2q37.1-q37.2 and has the official gene name of SH3 domain binding protein 4 (SH3BP4). Northern blot analysis showed that the 5.6 kb SH3BP4 message is expressed in several tissues including pancreas, heart, kidney, and placenta, and in cultured corneal fibroblasts. However, protein expression of SH3BP4 has not yet been reported.

The SH3BP4 message codes for a 963 amino acid polypeptide, which contains several interesting domains particularly unique in their combination (Figure 1A). The first of these are multiple domains typical of proteins in the Eps15-Homology (EH) network family of endocytosis, intracellular sorting, and cell fate determination proteins and includes three N-P-F amino acid repeats, a SH3-domain, a bipartite nuclear targeting signal, and a Tyr phosphorylation site [1,2]. Secondly, SH3BP4 contains a death domain close to the C-terminus of the molecule. This domain is present in proteins that are involved in programmed cell death and constitutes about 85 amino acids homologous to the C-terminus of TNF-R (tumor necrosis factor receptor) and Fas [3,4]. Currently, no other members of the EH network family of proteins are known to have a significant homology to a death domain. Therefore, the presence alone of both endocytosis and death domain domains within the same protein, SH3BP4, indicates fundamental significance to basic cell biological function.

The retinal pigment epithelium (RPE) has a unique role as an epithelium where it is responsible for phagocytosis of shed rod outer segments. In addition, the RPE provides physical support to many rods and cones so that apoptosis of a single RPE cell affects at least 10 photoreceptor cells [5,6]. The cell biological processes of endocytosis/phagocytosis and apoptosis were shown to be linked in the Drosophilia visual system where retinal cell endocytosis of persistent rhodopsin-arrestin complexes trigger apoptosis [7,8]. Therefore, the expression of a single protein that functions in both of these capacities would have an added significance in RPE cells.

In the present study we show that SH3BP4 is expressed in retinal tissues and two retinal cell lines, ARPE-19 (a spontaneously derived cell line of human RPE cells) and Y79 (a human retinoblastoma cell line), with higher protein expression present in the ARPE-19 cells. Using a combination of subcellular fractionation and fusion protein studies, we determined that SH3BP4 is localized to the plasma membrane and nuclear periphery in ARPE-19. The present study is the first report on the SH3BP4 protein.


SH3BP4 antibody

An SH3BP4 clone was ligated in frame with the maltose binding protein sequence in the pMAL-p2 prokaryotic expression vector (New England Biolabs, Beverly, MA) and the resulting fusion protein was used as the antigen to raise a chicken anti-human SH3BP4 antibody. Briefly, a SH3BP4 clone encoding for the first 763 amino acids of SH3BP4 was excised from pBluescript SK+ (Stratagene, La Jolla, CA) using BamHI. This insert, encoding for amino acids 5-763 of SH3BP4 was then treated with Klenow DNA polymerase I to generate blunt ends, followed by ligation of an EcoRI 10-mer phosphorylated linker (New England Biolabs, Beverly, MA). The insert was then digested with EcoRI and ligated into pBluescript (pBS). The 5'-end of the insert was DNA sequenced to confirm that the reading frame was maintained. The SH3BP4 insert was then excised from pBS using EcoRI and ligated into the pMAL-p2 vector that had been EcoRI digested and dephosphorylated [9]. This construct generates a fusion protein containing the maltose binding protein (MBP) sequence at the N-terminus followed by the EcoRI site and amino acids 5-763 of SH3BP4 (Figure 1B). XL10-Gold Escherichia coli cells transformed with the pMAL-SH3BP4 construct were grown to an A600 of about 0.5 followed by treatment with 0.3 mM IPTG for 1 or 2 h to induce fusion protein expression. The MBP-SH3BP4 fusion protein expressed by bacterial cells was found to be relatively insoluble and required a 2% SDS buffer to solubilize the fusion protein. Therefore, induced bacterial cells were lysed in 2% SDS in PBS with protease inhibitors, run on a 4-12% Bis-Tris gel and the 130-140 kDa region containing the MBP-SH3BP4 fusion protein was excised from the gel. The minced gel fragments were then used as the antigen in raising a chicken anti-human SH3BP4 antibody [10,11]. The yield of fusion protein was not quantitated due to insolubility; however, there was enough fusion protein that it was readily visible on a Coomassie stained gel (Figure 2A). Purified IgY samples, pre-immune and SH3BP4-immune, were provided by Aves Laboratories (Tigard, OR).

Cell culture and retinal tissue

The cell lines, ARPE-19, Y79, and COS-7 were purchased from American Type Tissue Culture Collection (Manassas, VA). Tissue culture reagents including fetal bovine serum (FBS) were purchased from Invitrogen/Gibco (Carlsbad, CA). Cells were cultured at 37 °C, 5% CO2 in the following media: ARPE-19 in DMEM-F12, Y79 in RPMI-1640, COS-7 in DMEM, and all media contained 10% FBS, 2% L-glutamine, and 0.5% antibiotic/antimyotic. For western analysis of SH3BP4 expression, cell layers were quickly rinsed in phenol red free DMEM-F12 catalog number 11039-021 Invitrogen (Carlsbad, CA) and catalog number SH30272.02 Hyclone (Logan, Utah), dissolved in 2% SDS in PBS plus proteases inhibitors. Cells were not counted prior to being dissolved as this would require trypsinization that can result in altering cellular levels of proteins. The monolayers were dissolved in 200 μl per well of a 6-well plate and 500 μl per T25 cm2 flask using the 2% SDS buffer, and briefly sonicated. This volume gives a protein concentration of 1-3 mg/ml of total protein depending on the cell line used and confluency of the culture. Total protein concentrations loaded onto gels ranged from 10-35 μg per lane. Rabbit retinas from young New Zealand White Rabbits were dissected into neural retina and RPE layers and quickly frozen on dry ice (Pel-Freeze Biologicals, Rogers, AR). Retina tissues were then dissolved in 2% SDS in PBS plus proteases inhibitors followed by tissue disruption using sonication.

SH3BP4 fusion proteins for eukaryotic expression

Three fusion proteins were generated for eukaryotic expression studies. A full length SH3BP4 clone encoding for amino acids 1-963 was ligated in frame with enhanced green fluorescence protein (GFP) at the C-terminus of the fusion protein (Figure 1C). Briefly, a full length SH3BP4 clone encoding for amino acids 1-963 was excised from pBluescript SK+ (Stratagene, La Jolla, CA) using a sequential digestion of KpnI then EcoRI. This insert was then ligated into the 5'-EcoRI to KpnI-3' site of the pEGFP-N2 vector (Clontech, Palo Alto, CA). PCR was then used to amplify SH3BP4 from amino acids 813-963 using a forward primer containing an endogenous XmnI site (5'-GAG AAC AAA GAA CGG AAG TCC TTC-3') and a reverse primer that replaced the stop codon with a KpnI site (5'-CCC GGT ACC CAA TCA CGA AGT CGT CCT-3'). This PCR fragment was then blunt end ligated into the pCR-script SK(+) vector (Stratagene, La Jolla, CA) and sequenced using T3 and T7 primers to confirm the integrity of the entire insert. This insert was then excised from pCR-script using XmnI and KpnI and the fragment encoding for amino acids 819-963 was purified. Likewise, the SH3BP4 insert in pEGFP-N2 was digested with XmnI and KpnI and the fragment contain the vector plus SH3BP4 amino acids 1-819 was purified. The PCR amplified fragment encoding for amino acids 819-963 was then ligated into the vector plus SH3BP4 (amino acids 1-819) and restriction enzyme digestions with XmnI and KpnI were used to confirm that the insert was correct. The resulting fusion protein, SH3BP4-GFP encodes for amino acid 1-963 of SH3BP4 followed by the KpnI site and the sequence for GFP (Figure 1C).

SH3BP4 was ligated in frame with two different vectors, one containing GFP (Figure 1D) and another containing the c-myc epitope tag (Figure 1E) both at the N-terminus of the fusion protein. The cloning strategy was the same for both constructs where BamHI was used to excise SH3BP4 from pBS and was cloned directly into the BamHI site of either the pEGFP-C2 (Clontech, Palo Alto, CA) or pCMV-Tag3C (Stratagene, La Jolla, CA) vector. The resulting fusion proteins are: GFP-SH3BP4 which encodes for GFP to the BamHI site followed by amino acids 6-963 of SH3BP4 (Figure 1D) and myc-SH3BP4 which encodes for c-myc epitope tag to the BamHI site followed by amino acids 6-963 of SH3BP4 (Figure 1E).

Transient transfections and cell staining

In transient transfection assays, cells were grown to 50-70% confluency for 36 h in 35 mm dishes or 6-well plates containing two serum coated 12 mm coverslips. Cells were then transfected for about 24 h with 2 μg of plasmid DNA in 12 μl of TransIT LT-1 transfection reagent (Mirus Corporation, Madison, WI) diluted in 120 μl of Opti-MEM (Invitrogen, Carlsbad, CA). Constructs used in transfection experiments were SH3BP4-GFP (Figure 1C), GFP-SH3BP4 (Figure 1D), and myc-SH3BP4 (Figure 1E). Cells grown on coverslips were fixed in 3.7% paraformaldehyde in PBS for 15 min while the remaining cell layer was lysed in 2% SDS in PBS with protease inhibitors and analyzed by immunoblot. Residual fixative was quenched with 0.1 M NH4Cl in the fixed cells on coverslips. Cells transfected with myc-tag constructs were stained with 1 μg/ml mouse anti-c-myc antibody (Calbiochem, La Jolla, CA) followed by a 1:500 dilution of goat anti-mouse IgG-FITC (Jackson ImmunoResearch Laboratories, West Grove, PA). For plasma membrane staining, GFP-SH3BP4 transfected cells were fixed with 3.7% paraformaldehyde in HBSS, quenched with 0.1 M NH4Cl, incubated for 5 min at 37 °C with 0.5 μM CellTracker CM-DiI (Molecular Probes, Eugene, OR) in HBSS followed by an additional 15 min incubation at 4 °C. All coverslips were mounted onto glass slides using VectaShield mounting medium (Vector Laboratories, Burlingame, CA). Digital photomicrographs were generated using either an Olympus Fluoview FV300 confocal laser scanning unit mounted on an Olympus IX-70 inverted fluorescence microscope or a Nikon TE300 inverted epifluorescence microscope with phase contrast optics and a Hamamatsu Orca-100 cooled digital camera.

SDS-PAGE, Coomassie staining, and western blot analysis

For SDS-PAGE and western blot analyses, all cell lysate samples were adjusted to 2% SDS, loaded by equal volume and run under reducing conditions on 4-12% Bis-Tris NuPAGE gels with MOPS running buffer (Invitrogen, Carlsbad, CA) using the Novex Gel System. Similarly for retinal tissue western blot analyses, 30-33 μg of total protein from the tissue lysates were loaded per well and run under reducing conditions on 4-12% Bis-Tris gels. Broad range pre-stained SDS-PAGE standards from BioRad Labartories (Hercules, CA) were used as molecular weight markers and estimated sizes of proteins were based on the calibrated molecular weights provided. For gel staining, gels were incubated in 0.2% Coomassie blue R-250 (BioRad Laboratories, Hercules, CA) followed by destaining in 10% acetic acid, 20% methanol. For western blot analysis, gels were transferred to 0.45 μm nitrocellulose, blocked in 5% (w/v) dehydrated skim milk in PBS-Tween 20, incubated with primary then secondary antibody diluted in PBS with 1% (w/v) milk, and developed using the chemiluminescent ECL kit and Hyperfilm ECL (Amersham Biosciences, Piscataway, NJ) or Kodak X-OMAT film (Eastman Kodak, Rochester, NY). Primary antibodies were chicken anti-human SH3BP4 IgY diluted 1:5000, 1:10,000, or 1:50,000 and chicken pre-immune IgY diluted 1:10,000 (Aves Laboratories, Tigard, OR); mouse anti-PCNA (proliferating cell nuclear antigen) clone PC10 diluted 1:3000 and mouse anti-chicken α-tubulin, clone DM 1A diluted 1:2000 (Sigma, St. Louis, MO); sheep anti-human TGN46 diluted 1:2000 (Serotec, Raleigh, NC); mouse anti-human lamin B diluted 1:1000 (Calbiochem, La Jolla, CA); mouse anti-RPE65 diluted 1:5000 and rabbit anti-rhodopsin diluted 1:2000 (Novus Biologicals Inc., Littleton, CO). Secondary antibodies containing a horseradish peroxidase tag were used at a 1:10,000 dilution and included anti-chicken IgY (Aves Laboratories, Tigard, OR), anti-sheep/goat IgG (Serotec, Raleigh, NC), anti-mouse IgG and anti-rabbit IgG (Amersham Biosciences, Piscataway, NJ). Polyclonal primary antibody blots were rinsed in PBS-0.5% or -0.3% Tween 20 and monoclonal primary antibody blots were rinsed in PBS-0.05% Tween 20. Total protein concentrations were determined using the BCA Protein Assay Kit (Pierce, Rockford, IL). Immunoblots were quantitated using NIH image software to determine the integrated optical density (IOD) for each band and these data were graphed relative to μg of total protein using DeltaGraph software (SPSS Inc., Chicago, IL).

Subcellular fractionation

Subcellular fractionations were performed using two separate techniques, either using NE-PER reagent or by differential centrifugation.

The nuclear (NF) and post-nuclear fractions (PNF) were obtained using NE-PER [12] nuclear and cytoplasmic extraction kit (Pierce, Rockford, IL) according to the manufacturer suggestions. Briefly, 2x106 cells were trypsinized, pelleted, resuspended in 100 μl CER I plus protease inhibitors and incubated on ice 10 min. CER II, 5.5 μl, was added followed by vortexing and incubation on ice for 1 min. Samples were re-vortexed and centrifuged at 2,000x g for 10 min at 4 °C. The PNF supernatant was then re-centrifuged twice. The nuclei pellet was washed twice in 50 μl CER I with protease inhibitors plus 2.75 μl CER II followed by resuspension in 50 μl NER II with protease inhibitors and vortexing 15 s every 10 min over 40 min. The nuclei were then centrifuged at 16,000x g for 10 min. All samples were desalted using protein desalting spin columns (Pierce, Rockford, IL) and adjusted to 2% SDS before being analyzed by western blot and BCA assay.

The second approach for generating nuclear, membrane, and cytoplasmic subcellular fractions employed a modified differential centrifugation protocol by Coda et al. [13]. Cells were rinsed three times with PBS containing 0.5 μM sodium orthovanadate, scrapped into hypotonic lysis buffer (HBL; 10 mM Tris-HCl) pH 7.4, 10 mM NaCl, 3 mM MgCl2, 50 mM sucrose, 1 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 10 mM NaF, protease inhibitors (Roche Applied Science, Indianapolis, IN) and incubated on ice for 4.5 min. The final concentration of sucrose in each sample was adjusted to 250 mM followed by passing samples through a 22 gauge needle approximately 10 times. Samples were first centrifuged at 1,000x g to remove any intact cells. Samples were then centrifuged at 2,000x g for 10 min at 4 °C and the pellet containing nuclei were washed three times in HBL containing 250 mM sucrose and 0.1% Nonidet P-40. The supernatant was re-centrifuged followed by centrifugation at 100,000x g for 45 min at 4 °C which generated cytosol (supernatant) and membrane (pellet) fractions. The cytosolic fraction was re-centrifuged three times, while the membrane pellet was washed three times with HBL containing 250 mM sucrose. All fractions were then adjusted to 50 mM Tris-HCl pH 7.4, 20 mM NaCl, 3 mM MgCl2, 250 mM sucrose, 10 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 50 mM NaF, 2% SDS, and protease inhibitors.


Characterization of the SH3BP4 Antibody

XL10-Gold Escherichia coli cells, transformed with the pMAL-SH3BP4 construct, were treated with IPTG to induce fusion protein expression and lysates were run on a 4-12% Bis-Tris gel followed by Coomassie Blue staining (Figure 2A). The MBP-SH3BP4 fusion protein is seen as a 130-140 kDa doublet in cells induced for 1 or 2 h while absent in uninduced cells. The smaller band is most likely a proteolytically cleaved fusion protein since the XL10-Gold cells are not a protease deficient cell strain. The 130-140 kDa fusion protein was used as the antigen for raising a chicken anti-human SH3BP4 antibody. Subsequent immunoblot analysis of bacterial lysates showed that the SH3BP4 antibody strongly recognized the fusion protein antigen at 130-140 kDa (Figure 2B, lane 1). Proteolytically cleaved fragments of the fusion protein, a result of protease activity from the XL-10 Gold cells which are not a protease deficient strain, are also detected by the antibody. SH3BP4 antibody specificity is shown by a lack of recognition of any endogenously expressed bacterial proteins from the lysates of uninduced pMAL-SH3BP4 transformed cells (Figure 2B, lane 2). Additionally, the pre-immune IgY control did not specifically bind to either endogenously expressed bacterial proteins or the MBP-SH3BP4 fusion protein (Figure 2B, lanes 4 and 3, respectively).

SH3BP4 antibody recognition of eukaryotically expressed protein was examined in multiple cell lines including ARPE-19 human RPE, Y79 human retinoblastoma, and COS-7 African green monkey kidney epithelial cells (Figure 3). SH3BP4 immune IgY was found to react with a 120 kDa band that was not recognized by pre-immune IgY (Figure 3A). The deduced molecular weight of SH3BP4 is 107.5 kDa. The protein runs slightly higher on SDS-PAGE gels than the deduced molecular weight. It is not unusual for proteins to run higher or lower than their deduced molecular weight. The ability of the SH3BP4 antibody to recognize SH3BP4 fusion proteins was examined by transiently transfecting ARPE-19 with GFP-SH3BP4 or myc-SH3BP4 constructs followed by immunoblot analysis (Figure 3A). The SH3BP4 antibody was found to specifically recognize both the 150 kDa GFP-SH3BP4 fusion protein and the 120 kDa myc-SH3BP4 fusion protein. Full length SH3BP4 and truncated fusion proteins of SH3BP4 (data not shown) all immunoblot to the correct relative size according to the endogenously expressed 120 kDa band. Interestingly, a large differential expression of SH3BP4 was observed between the cell lines in the immunoblot shown in Figure 3A. Quantitation of immunoblots from three sets of cell lysates show that ARPE-19 cells expressed significantly greater levels of SH3BP4 protein 0.665±0.063 (±SD) than Y79 cells 0.0883±0.026 and COS-7 cells 0.247±0.044 (p<0.001, analysis of variance with Student-Newman-Keuls for each comparison). These results show that ARPE-19 cells have a 7.5 fold increase in SH3BP4 levels compared to Y79 cells and a 2.7 fold increase in SH3BP4 levels compared to COS-7 cells (Figure 3B).

The expression of SH3BP4 in retinal tissues was also examined by western analysis. Rabbit retinas were dissected into RPE and neural retinal layers and immunoblotted for SH3BP4 and for the control proteins RPE65 and rhodopsin (Figure 4). The 120 kDa SH3BP4 protein was found to be expressed in both the RPE and neural retinal layers while the RPE specific protein RPE65 was detected in the RPE only and the 35-40 kDa monomeric rhodopsin was found mainly in the neural retina.

Subcellular localization of endogenous SH3BP4

Subcellular fractionation using two different protocols, NE-PER reagent and differential centrifugation, were used to determine the subcellular localization of SH3BP4. The NE-PER reagent was used to distinguish between nuclear and post-nuclear (membranes plus cytosol) subcellular localization of SH3BP4 in ARPE-19 and COS-7 cells. SH3BP4 was found in both the nuclear and post-nuclear fractions of ARPE-19 and COS-7 cells (Figure 5A). These data were qualitatively analyzed and are shown as integrated optical density (IOD) per μg of total protein (Figure 5B) and IOD alone (Figure 5C). Results show that SH3BP4 is enriched in the nuclear fraction from ARPE-19 cells with a 3.0 fold increase in the nuclear compared to the PNF fractions relative to total protein. Similarly, COS-7 had a 2.1 fold higher level of SH3BP4 in nuclear compared to the PNF fractions relative to total protein (Figure 5B). The distribution of total cellular levels of SH3BP4 were also compared and results showed that 45% of total cellular SH3BP4 was present in the nuclear fraction and 55% was in the PNF of ARPE-19 cells. In COS-7 cells, 23% of the total cellular level of SH3BP4 was found in the nuclear fraction and 77% was found in the PNF. Since SH3BP4 was present in the PNF, a differential centrifugation protocol was used to determine SH3BP4 presence in either membranes or the soluble cytoplasm. Results show that SH3BP4 is present in the nuclear and membrane fractions of ARPE-19 (Figure 5A) and COS-7 cells (data not shown). Qualitative analysis of the immunoblot show that relative to total protein, there is a 7.9 fold increase in SH3BP4 in the nuclear fraction compared to the membrane fraction (Figure 5D). There was an undetectable level of SH3BP4 in the cytosol.

The subcellular fractions generated using these two protocols were analyzed for fraction purity using standard markers (Figure 6). The DNA replication factor PCNA was used as a nuclear marker based on the protocol by Coda et al. [13]; however, in ARPE-19 and COS-7 cells this small 36 kDa protein diffused readily out of the nucleus accounting for low total protein concentrations in the nuclear fractions. Since SH3BP4 was consistently found in the nuclear fraction, the nuclear matrix intermediate filament protein lamin B was used as a positive control where about 100% of the 100 kDa lamin B protein was found associated with the nuclear fraction. Both PNF and membrane fractions contained 97-100% of the 110 kDa trans-Golgi network protein, TGN46, while 100% of the 55 kDa α-tubulin protein was found associated with the PNF and cytosol fractions.

Subcellular localization of SH3BP4 fusion proteins

The subcellular localization of GFP-SH3BP4 and myc-SH3BP4 fusion proteins was determined using subcellular fractionation and confocal microscopy. ARPE-19 cells were transiently transfected with GFP-SH3BP4 or myc-SH3BP4 constructs followed by subcellular fractionation using differential centrifugation. Both GFP-SH3BP4 and myc-SH3BP4 were found to localize to the nuclear and membrane fractions similar to endogenous SH3BP4 (Figure 7A). 100% of the control marker α-tubulin was found in the cytosol fraction with no detectable levels in membrane or nuclear fractions. Qualitative analysis of immunoblots show that compared to total protein, the majority of the SH3BP4 fusion proteins localize to the nucleus with 6.9 fold and 2.9 fold increases of GFP-SH3BP4 and myc-SH3BP4, respectively, in the nuclear fractions compared to the membrane fractions (Figure 7B). There was no detectable level of either SH3BP4 fusion protein in the cytosol.

In confocal microscopy studies, SH3BP4-GFP, GFP-SH3BP4, and myc-SH3BP4 fusion proteins all had a very similar intracellular localization patterns in the plasma membrane and nuclear periphery (Figure 8A). Some structures that appear to be membrane vesicles and membrane blebs, ruffles, and spikes that contained SH3BP4 fusion proteins were visualized (data not shown).

Purified nuclei from ARPE-19 cells transiently transfected with GFP-SH3BP4 were visualized by confocal, epifluorescence, and phase microscopy (Figure 8B). Results show that GFP-SH3BP4 localizes to the nuclear periphery and appears to encase the nucleus. Additionally, APRE-19 cells transiently transfected with GFP-SH3BP4 followed by plasma membrane staining with CM-DiI were analyzed by confocal microscopy (Figure 8C). GFP-SH3BP4 fusion protein at the cell periphery was found to co-localize completely with CM-DiI, shown by the yellow color in the merged image. Studies of ARPE-19 cells transiently transfected with GFP-SH3BP4 followed by staining with mitotracker red show no co-localization of SH3BP4 fusion protein with the mitochondrial marker (data not shown).


The results of this study showed that the protein SH3BP4 is expressed in both the RPE and neural retina from rabbit eyes and in three cells lines including ARPE-19 retinal pigment epithelial, Y79 retinoblastoma, and COS-7 kidney epithelial cells. Results also showed that ARPE-19 cells express SH3BP4 at a seven fold higher level, relative to total protein, than the retinoblastoma cell line Y79. ARPE-19 is a spontaneously derived cell line from human RPE cells grown from a 19 year old male donor [14]. These cells have been previously shown to express the RPE specific markers CRALBP and RPE65, exhibit morphological polarization with the formation of tight junctions and have phagocytosis capabilities including engulfing rod outer segments [14-16]. These previous studies show that ARPE-19 cells have multiple properties that are unique to the RPE, and we showed in this study that the EH network and death domain protein, SH3BP4, is expressed in these cells and in RPE tissue.

SH3BP4 contains domains belonging to both the EH network family of endocytosis, intracellular sorting, and cell fate determination molecules and to the death domain family of apoptosis proteins. Apoptosis domains in SH3BP4 were mentioned in this paper. We have begun to examine if the subcellular compartmentalization of SH3BP4 is altered during oxidative stress induced apoptosis. However, these results are preliminary and extend into the focus of another study that is currently in progress. This combination of domains has not been reported in other proteins to date and the presence alone of both endocytosis and apoptosis domains encoded within the same protein indicates a functional significance in RPE cells. Although ARPE-19 cells have not yet been shown to express other EH network proteins, Weigel et al. [17] identified the message for intersectin1 (SH3P17) by microarray analysis.

Standard cellular markers were used to evaluate two subcellular fractionation techniques, and results showed that the PNF, membrane, and cytosol fractions were >97% pure. In nuclear fractions, large nuclear matrix proteins were retained with almost 100% selectivity, while smaller proteins were found to diffuse from the nucleus, contributing to low total protein concentrations in these fractions. ARPE-19 and COS-7 cells were used to further characterize SH3BP4 distribution, and results showed that SH3BP4 endogenous and fusion proteins localized to both the nuclear and membrane fractions but not cytoplasmic fractions. Subsequent confocal and epifluorescence microscopy analyses showed that SH3BP4 is localized mainly to the plasma membrane and the periphery of the nucleus. Although multiple cytoplasmic and plasma membrane associated proteins, including the EH network protein Eps15R [13] and the death domain protein FADD [18,19], have been shown to localize to the nucleus, most localize in a diffuse pattern throughout the nucleus. SH3BP4 associated with the plasma membrane and also localized to the periphery of the nucleus. Interestingly, the epidermal growth factor receptor appears to have a similar localization pattern as SH3BP4 where it localizes to the plasma membrane and to the nuclear periphery [20]. The domains of SH3BP4 contain multiple nuclear localization signals including a bipartite nuclear targeting signal at amino acids 678-695, and two pat7 nuclear targeting signals at amino acids 237-243 and 927-933 [1,21]. Therefore, it is expected to find SH3BP4 associated with the nucleus.

The initial study on SH3BP4 reported the identification of this novel gene, its cDNA sequence, and its localization on human chromosome 2q37.1-q37.2 [1]. This corollary study is the first to report on the SH3BP4 protein and showed that this protein is expressed at significantly higher levels in the APRE-19 cell line in comparison to the Y79 retinoblastoma and the COS-7 kidney epithelial cell lines. SH3BP4 was also found to be expressed in rabbit retinal tissues in both the RPE and neural retina layers. Subsequently, localization studies using two subcellular fractionation techniques and microscopy analysis of SH3BP4 fusion proteins showed that SH3BP4 localizes to the plasma membrane and nuclear periphery. These studies provide the ground work for future analyses of SH3BP4 in the RPE including the effects of SH3BP4 over-expression and the role of SH3BP4 in oxidative stress. The unique combination of endocytosis and death domains in SH3BP4, its unusual intracellular distribution, and its high expression in RPE cells, strongly indicates functional importance in the cellular physiology of the retina.


This work was supported by grants from the North Dakota EPSCoR EPS-0132289 and the University of North Dakota Faculty Research Council to JRD. The authors would like to thank Dr. Don Sens for his valuable advice and comments regarding the manuscript, Bridgette L. Berryhill for her outstanding technical assistance in generating the MBP-SH3BP4 antigen, and Dr. Michael Atkinson and the Basic Sciences Imaging Center for microscopy assistance.


1. Dunlevy JR, Berryhill BL, Vergnes JP, SundarRaj N, Hassell JR. Cloning, chromosomal localization, and characterization of cDNA from a novel gene, SH3BP4, expressed by human corneal fibroblasts. Genomics 1999; 62:519-24.

2. Mayer BJ. Endocytosis: EH domains lend a hand. Curr Biol 1999; 9:R70-3.

3. Cleveland JL, Ihle JN. Contenders in FasL/TNF death signaling. Cell 1995; 81:479-82.

4. Feinstein E, Kimchi A, Wallach D, Boldin M, Varfolomeev E. The death domain: a module shared by proteins with diverse cellular functions. Trends Biochem Sci 1995; 20:342-4.

5. Young RW. The renewal of photoreceptor cell outer segments. J Cell Biol 1967; 33:61-72.

6. Jin GF, Hurst JS, Godley BF. Rod outer segments mediate mitochondrial DNA damage and apoptosis in human retinal pigment epithelium. Curr Eye Res 2001; 23:11-9.

7. Alloway PG, Howard L, Dolph PJ. The formation of stable rhodopsin-arrestin complexes induces apoptosis and photoreceptor cell degeneration. Neuron 2000; 28:129-38.

8. McDonald PH, Lefkowitz RJ. Beta-Arrestins: new roles in regulating heptahelical receptors' functions. Cell Signal 2001; 13:683-9.

9. Sambrook J, Fritsch EF, Maniatis T. Molecular cloning: a laboratory manual. 2nd ed. Cold Spring Harbor (NY): Cold Spring Harbor Press; 1989.

10. Harlow E and Lane D. Antibodies: a laboratory manual. Cold Spring Harbor (NY): Cold Spring Harbor Press; 1988.

11. Zhang YJ, O'Neal WK, Randell SH, Blackburn K, Moyer MB, Boucher RC, Ostrowski LE. Identification of dynein heavy chain 7 as an inner arm component of human cilia that is synthesized but not assembled in a case of primary ciliary dyskinesia. J Biol Chem 2002; 277:17906-15.

12. Smirnova IV, Bittel DC, Ravindra R, Jiang H, Andrews GK. Zinc and cadmium can promote rapid nuclear translocation of metal response element-binding transcription factor-1. J Biol Chem 2000; 275:9377-84.

13. Coda L, Salcini AE, Confalonieri S, Pelicci G, Sorkina T, Sorkin A, Pelicci PG, Di Fiore PP. Eps15R is a tyrosine kinase substrate with characteristics of a docking protein possibly involved in coated pits-mediated internalization. J Biol Chem 1998; 273:3003-12.

14. Dunn KC, Aotaki-Keen AE, Putkey FR, Hjelmeland LM. ARPE-19, a human retinal pigment epithelial cell line with differentiated properties. Exp Eye Res 1996; 62:155-69.

15. Dunn KC, Marmorstein AD, Bonilha VL, Rodriguez-Boulan E, Giordano F, Hjelmeland LM. Use of the ARPE-19 cell line as a model of RPE polarity: basolateral secretion of FGF5. Invest Ophthalmol Vis Sci 1998; 39:2744-9.

16. Finnemann SC, Bonilha VL, Marmorstein AD, Rodriguez-Boulan E. Phagocytosis of rod outer segments by retinal pigment epithelial cells requires alpha(v)beta5 integrin for binding but not for internalization. Proc Natl Acad Sci U S A 1997; 94:12932-7.

17. Weigel AL, Handa JT, Hjelmeland LM. Microarray analysis of H2O2-, HNE-, or tBH-treated ARPE-19 cells. Free Radic Biol Med 2002; 33:1419-32.

18. Screaton RA, Kiessling S, Sansom OJ, Millar CB, Maddison K, Bird A, Clarke AR, Frisch SM. Fas-associated death domain protein interacts with methyl-CpG binding domain protein 4: a potential link between genome surveillance and apoptosis. Proc Natl Acad Sci U S A 2003; 100:5211-6.

19. Sheikh MS, Huang Y. The FADD is going nuclear. Cell Cycle 2003; 2:346-7.

20. Lin SY, Makino K, Xia W, Matin A, Wen Y, Kwong KY, Bourguignon L, Hung MC. Nuclear localization of EGF receptor and its potential new role as a transcription factor. Nat Cell Biol 2001; 3:802-8.

21. Hicks GR, Raikhel NV. Protein import into the nucleus: an integrated view. Annu Rev Cell Dev Biol 1995; 11:155-88.

Khanobdee, Mol Vis 2004; 10:933-942 <>
©2004 Molecular Vision <>
ISSN 1090-0535