Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
Received 15 June 2005 | Accepted 19 July 2005 | Published 30 August 2005
©2005 Molecular Vision
Stargardt-like macular dystrophy protein ELOVL4 exerts a
dominant negative effect by recruiting wild-type protein into
aggresomes
Vidyullatha Vasireddy,1 Camasamudram Vijayasarathy,2 Jibiao Huang,1 Xiaofei F. Wang,3 Monica M. Jablonski,3
Howard R. Petty,1 Paul A. Sieving,2,4 Radha Ayyagari1
1
Department of Ophthalmology and Visual Sciences, University of Michigan, Ann Arbor, MI; 2National Eye Institute and 4National
Institute on Deafness and Other Communication Disorders, National Institutes of Health, Bethesda, MD; 3University of Tennessee
Health Science Center, Hamilton Eye Institute, Memphis, TN
Purpose: Mutations in the gene Elongation of very long-chain fatty acids-4 (ELOVL4) have been shown to be associated
with autosomal dominant Stargardt-like macular dystrophy (STGD3). ELOVL4 is expressed in photoreceptors and encodes a putative transmembrane protein of 314 amino acids with an endoplasmic reticulum (ER) retention signal. A 5 bp
deletion in exon 6 of ELOVL4 observed in some STGD3 patients results in the truncation of the protein and loss of the ER
retention signal. To understand the disease mechanism underlying STGD3 we studied the intracellular trafficking of the
wild-type and a 5 bp deletion mutant of ELOVL4.
Methods: Wild-type and mutant ELOVL4 proteins with the N-terminal GFP/V5 tags were expressed in COS-7 cells.
Expression and the intracellular localization of the wild-type and mutant proteins were characterized by immunocytochemistry and western blot analysis using tag- and organelle-specific antibodies. Interaction between the wild-type and
mutant proteins was studied by two-dimensional gel electrophoresis and fluorescence resonance energy transfer (FRET)
analysis.
Results: The mutant ELOVL4 protein exerted a dominant negative effect when the wild-type and 5 bp deletion mutant
ELOVL4 proteins were co-expressed in COS-7 cells. Immunocytochemical analysis, two-dimensional gel electrophoresis and FRET revealed that the mutant ELOVL4 interacts with the wild-type protein, forming higher molecular mass
complexes that accumulate in aggresomes.
Conclusions: In the presence of mutant ELOVL4 protein, the wild-type protein was recruited into perinuclear cytoplasmic inclusions that resemble aggresomes. The interaction between the wild-type and mutant forms of ELOVL4 and the
resultant alteration in the trafficking of the wild-type ELOVL4 protein suggest a mechanism for the pathogenicity observed in patients with autosomal dominant STGD3.
shift leading to the loss of 51 amino acids; (2) a two-basedeletion, which also results in the loss of 51 terminal amino
acids; and (3) a nonsense mutation at codon 270. All three
mutations result in premature termination of protein translation and loss of the ER retention signal [3,4,6]. We previously
demonstrated that the dilysine ER-retention signal at the Cterminal end of the ELOVL4 protein is functional, and its loss
results in misrouting of the protein [7]. However, the mechanism underlying the degeneration process due to the mutations in the ELOVL4 gene is not known. Mutations present in
the heterozygous state could cause disease either due to
haploinsufficiency or a dominant negative effect exerted by
the mutant protein. Expression of the ELOVL4 mutant protein in COS-7 cells indicated that the mutant protein is localized to cytosolic juxtanuclear compartments [7]. Several
neurodegenerative diseases, including retinitis pigmentosa due
to mutations in the rhodopsin gene, are found to result from
the dominant negative effect of the misfolded mutant proteins
[8-10]. In order to evaluate whether the mutant ELOVL4 exerts similar dominant negative effect, we studied the cellular
fate of the wild-type and 5 bp deletion mutant ELOVL4 proteins following their co-expression in COS-7 cells. Our stud-
Stargardt-like autosomal dominant macular dystrophy is
characterized by progressive macular degeneration leading to
central macular and peripapillary retinal pigment epithelium
(RPE) atrophy and accumulation of flecks. The onset of the
disease is during the teenage years [1,2]. For these macular
degenerations, two loci, one on chromosome 6q (STGD3) and
the other on chromosome 13 (STGD2), have been mapped.
Mutations in the elongation of very long chain fatty acids-4
gene (ELOVL4) have been identified in families linked to the
locus on the 6q [3,4]. The ELOVL4 protein has sequence and
structural similarities with the ELO family of proteins described in yeast [5]. ELOVL4 encodes a protein with 314
aminoacids with a dilysine motif (KXKXX) at the C-terminus that is thought to signal ER retention. Three different
mutations have been identified in the ELOVL4 gene associated with Stargardt-like macular degeneration: (1) a 5 bp deletion mutation in exon 6 of ELOVL4, which causes a frameCorrespondence to: Radha Ayyagari, PhD, Department of Ophthalmology and Visual Sciences, W. K. Kellogg Eye Center, The University of Michigan, 1000 Wall Street, Room number 325, Ann Arbor,
MI, 48105; Phone: (734) 647-6345; FAX: (734) 936-7321; email:
ayyagari@umich.edu
665
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
©2005 Molecular Vision
ies suggest that the mutant protein interacted with the wildtype protein to form higher molecular weight aggregates. These
aggregates were localized to perinuclear inclusions that have
the characteristic features of aggresomes. Interaction of the
wild-type and mutant proteins, and the alteration in their cellular trafficking may explain the molecular pathology underlying the degeneration process due to mutations in the ELOVL4
gene.
(GFP-Ewt) and EGFP-ELOVL4 5 bp deletion mutant (GFPEmut) constructs was reported earlier [7]. The V5-tagged
ELOVL4 wild-type (V5-Ewt) and V5 tagged ELOVL4 mutant (V5-Emut) constructs were prepared by standard two-step
method using Gateway cloning technology (Invitrogen Corp.,
Carlsbad, CA). Gateway is a universal cloning technology that
takes advantage of the site specific recombination properties
of bacteriophage lambda to provide a rapid and highly efficient method to move the gene of interest into multiple vector
systems.
The complete cDNA sequence for ELOVL4 was amplified from human brain cDNA library, using gene specific primers carrying BamHI cohesive ends using the forward (5'-CTC
GGA TCC GCG ATG GGG CTC CTG GAC-3') and reverse
(5'-GCG GAT CCC AGT TCA ATT TAA TCT CCT TTT GCT
TTT CC-3') primers and Pfx DNA polymerase. wild-type and
mutant ELOVL4 cDNAs were first cloned into the pENTERdirectional-TOPO entry vectors. These TOPO entry clones
were used for recombination reactions and thereby directional
cloning into the Gateway pcDNA 3.1/nV5 DEST mammalian
expression vector carrying a V5 tag. Sequencing confirmed
that the resulting clone, pcDNA 3.1/nV5 DEST-ELOVL4, contained the complete ELOVL4 wild-type or the mutant sequence
in frame and the N-terminal V5 extension.
Cell culture: COS-7 cells were cultured in DMEM with
glutamine and 10% fetal bovine serum with an atmosphere of
10% CO2 at 37 °C. The cells were seeded and grown to 8090% of confluence. After 24 h, the cells were transfected with
expression constructs or control vectors using Lipofectamine
Plus (Invitrogen) according to the manufacturer’s instructions.
METHODS
Reagents and antibodies: The following antibodies and chemicals were used in this study. Antibodies: mouse monoclonal
anti-vimentin (Sigma-Aldrich, Inc., St. Louis, MO, dilution
1:100), mouse monoclonal anti-γ-tubulin (Sigma-Aldrich, dilution 1:500), rabbit polyclonal anti-V5 (Sigma-Aldrich, dilution 2 µg/ml), mouse monoclonal anti-V5 (Invitrogen Corporation, Carlsbad, CA, dilution 1:500), mouse monoclonal
anti-GFP (Covance, Berkeley, CA, dilution 1:10,000), mouse
monoclonal anti-golgin-97 (Molecular Probes, Eugene, OR,
concentration 1 µg/ml), mouse monoclonal anti-trans Golgi
network-38 (TGN-38, Abcam, MA, dilution 1:1000).
Secondary antibodies: Alexa fluor-555 conjugated goat
anti-mouse IgG (Molecular probes, Eugene, OR, concentration 8 µg/ml), mouse monoclonal anti-rabbit IgG (Pierce Biotechnology, Rockford, IL, dilution 1:7500), goat monolclonal
anti-mouse IgG (Pierce Biotechnology, IL, dilution 1:7500).
Chemicals: proteasome inhibitor MG-132 and nocodozole
were purchased from Calbiochem, San Diego, CA. Fetal bovine serum was obtained from ATCC (Manassas, VA) and the
other chemicals were obtained form Sigma-Aldrich, Inc.
Cloning of ELOVL4 into GFP tagged and V5 tagged expression vectors: Preparation of EGFP-ELOVL4 wild-type
Figure 1. Expression of GFP tagged wild-type and mutant forms of
ELOVL4 in COS-7 cells. To determine the expression of Ewt and
Emut fusion proteins, lysates of cells trasfected with GFP-Ewt (lane
2), GFP-Emut (lane 3), GFP-Ewt and GFP-Emut (lane 4), V5-Ewt
(lane 5), V5-Emut (lane 6), GFP-Ewt, and V5-Emut (lane 7), V5Ewt, and GFP-Emut (lane 8), EGFP vector alone (lane 9), and mock
transfected COS-7 cells (lane 10) were labeled with antibody specific to the GFP tag. Lane 1 contained molecular weight markers.
Expected size bands were detected in cells transfected with GFP
tagged ELOVL4.
Figure 2. Immunoblot analysis of V5 tagged ELOVL4 fusion proteins. Western blot analysis of COS-7 cells expressing V5 tagged
ELOVL4 was carried out using V5 antibody. Lysates of cells expressing GFP-Ewt (Lane 2), GFP-Emut (Lane 3), co-expressed GFPEwt and GFP-Emut (Lane 4), V5-Ewt (Lane 5), V5-Emut (Lane 6),
co-expressed GFP-Ewt and V5-Emut (Lane 7), V5-ELOVL4-wt and
GFP-Emut (Lane 8), and mock transfected cells (Lane 9) were
immunobloted. Analysis of epitope tagged ELOVL4 protein expression demonstrated the presence of bands of the expected molecular
weights. Lane 1 contained molecular weight markers.
666
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
©2005 Molecular Vision
SDS-PAGE and blotted onto nitrocellulose membrane for
western analysis. The blot was developed by the chemiluminescent method using ECL reagents according to the
manufacturer’s instructions (Amersham Biosciences,
Piscataway, NJ).
The effects of proteasomal inhibition and Golgi disruption were
studied by treating the cells 48 h after transfection with 10
µM MG-132 for 12 h and 10 µg/ml nocodazole for 1 h, respectively, at 37 °C and then washed with PBS and
immunolabeled.
Immunofluorescence analysis of transfected COS-7 cells
by confocal microscopy: Cells were washed with cold PBS
48 h after transfection and fixed in absolute methanol at -10
°C for 15 min. Fixed cells were blocked for 1 h with goat
serum in PBS, incubated for 1 h with primary antibody followed by washing with PBS, and further incubation with secondary antibody. Subsequently, washed cells were mounted
with Vectashield anti-fade mounting medium containing DAPI.
Immunofluorescence was visualized with a Zeiss LSM 510
laser scanning confocal microscope.
Preparation of cell extracts and western blotting: Cells
transfected with the EGFP or V5 tagged ELOVL4 constructs
were lysed with buffer containing 25 mM Tris-HCl (pH 7.5),
1% (w/v) SDS, 1 mM DTT, glycerol, and protease inhibitor
cocktail (Sigma-Aldrich). Cell lysates were separated by 10%
TABLE 1. SUMMARY OF ELOVL4 PROTEIN EXPRESSION IN COS-7
CELLS
Constructs transfected
---------------------GFP-Ewt
GFP-Emut
GFP-Ewt and GFP-Emut
V5-Ewt
V5-Emut
GFP-Ewt and V5-Emut
GFP-Emut and V5-Ewt
Localization
in the cells
-----------ER
Perinuclear
Perinuclear
ER
Perinuclear
Perinuclear
Perinuclear
Size of the
fragment observed
on Immunoblots
----------------about 67 kDa
about 63 kDa
about 63 kDa
about 41 kDa
about 37 kDa
about 37 kDa
about 37 kDa
Cells transfected with wild-type and/or mutant ELOVL4 constructs
showed the presence of ELOVL4 fusion proteins of expected molecular mass. In presence of the mutant protein, the wild-type protein
is misrouted to perinuclear aggresomes.
Figure 3. Altered localization
of wild-type ELOVL4 in presence of the mutant ELOVL4.
To test the localization of wildtype ELOVL4 in the presence
of mutant ELOVL4, COS-7
cells transfected with GFPEwt and V5-Emut (A-C) and
GFP-Emut and V5-Ewt (D-F)
were immunostained with antiV5 antibody (red). In presence
of the mutant protein, localization of the wild-type protein
was altered. The scale bar represents 5 µm.
667
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
©2005 Molecular Vision
Figure 4. ELOVL4 mutant oligomerizes with wild-type ELOVL4. Whole cell lysates prepared from cells transfected with EGFP- or V5tagged ELOVL4 constructs were first subjected to CN-PAGE in the first dimension to separate the multiprotein complexes in accordance with
their molecular mass and state of aggregation. Following this first dimension separation, each sample lane (0.5 cm in width) of the first
dimension native gel was cut out and equilibrated with Laemmli Buffer (62.5 mM Tris-HCl pH 6.8, 2% (w/v) SDS, 2% (w/v) DTT, 5% (v/v)
glycerol) for 30 min at ambient room temperature. This SDS treatment is to break apart aggregates into single polypeptides that should migrate
in the second dimension gel according to the molecular mass of the polypeptide. The gel strip was placed horizontally (at right angles to the
first dimension) at the usual position for the stacking gel in the second dimension 10% SDS-polyacrylamide gel and overlaid with 0.5%
agarose in Laemmli buffer. After solidification of the agarose, the second dimension electrophoresis was conducted, and immunoblot analyses
were performed with GFP or V5 epitope specific antibodies according to standard protocols. A: GFP-vector. B: GFP-Ewt. C: GFP-Emut. D:
GFP-Ewt and GFP-Emut. E: GFP-Ewt and V5-E mut. F: V5-E wt and GFP-Emut. G: V5-E wt. H: V5-E wt and GFP-Emut.
668
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
©2005 Molecular Vision
Figure 5. FRET between wild-type and mutant ELOVL4 proteins. The physical interaction between the wild-type and mutant ELOVL4
proteins was explored by FRET. FRET measurements were made by exciting the donor at 485 nm and monitored the emission spectrum
between 400-800 nm. In these experiments, green fluorescence of EGFP served as a donor (A,D,G,J), and Alexa fluor 555 conjugated V5
antibody epitope served as an acceptor (B,E,H,K). In co-transfected cells (C,F) and cells transfected with GFP-Emut and V5-E mut (L), both
donors and acceptors were detected and FRET was observed. However, in cells carrying GFP-Ewt and V5-Ewt (I) no FRET was detected.
Intensity was expressed as counts per second (CPS). The excitation and emission wavelengths used for EGFP and Alexa fluor labeled anti-V5antibody are 490 nm, 509 nm, and 555 nm, 565 nm, respectively. The energy transfer observed in co-transfecetd cells indicates the close
proximity between wild-type and mutant ELOVL4 (less than a 100 Å distance) and their physical interaction. These data support that the
mutant and wild-type ELOVL4 proteins forms aggregates.
669
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
©2005 Molecular Vision
10% SDS-polyacrylamide gel, and the protein components
were identified by immunoblot analysis using epitope-specific
(anti-GFP or anti-V5) antibodies according to standard protocols [9].
FRET analysis to study the interaction between Ewt and
Emut proteins: The interaction of wild-type and mutant
ELOVL4 proteins was explored by fluorescence resonance
energy transfer (FRET), in which molecular proximity is determined by the energy transfer from a fluorescent donor to an
acceptor. COS-7 cells were transfected with GFP-Ewt and
GFP-Emut, V5-Ewt and GFP-Emut, GFP-Ewt and V5-Ewt,
or GFP-Emut, and V5-Emut constructs. After 24 h of transfection, cells were fixed with 2% paraformaldehyde,
permeabilized with 1% Brij-58, and fixed again with 2%
paraformaldehyde for 20 min at room temperature [11]. These
cells were labeled with Alexa fluor 555-conjugated anti-V5
antibodies, and were washed for 1 h with Hank’s Balanced
Salt solution with constant shaking at room temperature.
Emission spectrophotometry: Energy transfer was exam-
Colorless native polyacrylamide gel electrophoresis (CNPAGE): CN-PAGE was employed to determine whether
ELOVL4 exists as a higher molecular mass oligomeric complex and was carried out essentially as described earlier with
slight modifications [8-10]. CN-PAGE is identical to BlueNative PAGE (BN-PAGE) but Coomassie dye is omitted from
sample and cathode buffer and sodium taurodeoxycholate is
used to induce the charge shift, which facilitates the migration
of protein complexes according to their molecular mass on a
nondenaturing gel. The whole cell lysates derived from cells
transfected with GFP and/or V5 tagged ELOVL4 constructs,
containing about 25 µg of protein were loaded onto the 613% native gel. The samples were electrophoresed at 80 V for
1 h and 200 V (current not exceeding 15 mA) for 2 h at 4 °C.
The relative migration rates of standards on native gel were
visualized by staining the first dimension native gel with Simply Blue safe stain (Invitrogen). The oligomeric complexes
separated initially by CN-PAGE were further resolved into
individual polypeptide components in second dimension on a
Figure 6. Formation of
aggresomes at the MTOC.
COS-7 cells were transfected
with GFP-Ewt (A-C), GFPEmut (D-F), GFP-Ewt and
GFP-Emut (G-I), and mock
transfected with transfection
reagents alone (K) and were
immunostained with antibodies to the centrosome marker,
γ-tubulin. Staining of γ-tubulin
is indicated with arrows (red).
Overlays of images from first
two columns on DAPI-stained
nuclei of the respective cells
(blue) were shown in panels C,
F, I, and L. Physically interacting wild-type and mutant
ELOVL4 formed aggresomes
near MTOC (microtubule organizing center). The scale bar
represents 5 µm.
670
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
©2005 Molecular Vision
described earlier [7]. COS-7 cells were transfected or co-transfected with the following constructs: (1) GFP-Ewt, (2) GFPEmut, (3) GFP-Ewt and GFP-Emut, (4) V5-E wt, (5) V5-E
mut, (6) GFP-Ewt and V5-E mut, (7) V5-E wt and GFP-Emut,
(8) GFP vector, and (9) mock transfection with transfection
reagents only. Cell lysates were analyzed on SDS-PAGE followed by western blot analysis with appropriate antibodies to
determine the expression of ELOVL4 fusion proteins. The GFP
antibody detected a prominent band migrating at about 64 kDa
in the lysates (Figure 1) of cells expressing the GFP-Ewt protein (lane 2), and GFP-Emut protein (lane 3). Cells co-expressing GFP-Ewt and GFP-Emut protein (lane 4), GFP-Ewt and
V5-Emut protein (lane 7), and V5-Ewt and GFP-Emut protein (lane 8) also showed a signal at about 64 kDa. The band
size was consistent with the expected molecular mass of 67
kDa (ELOVL4 contributing 37 kDa and GFP contributing 30
kDa) for wild-type protein and 63 kDa for mutant protein
(ELOVL4 contributing a mass of 33 kDa and GFP contributing a mass of 30 kDa). A 30 kDa band was detected in control
GFP vector transfected cell lysate (lane 9). No specific signal
was detected in lysates of cells transfected with V5-Ewt (lane
ined by means of a microscope spectrophotometer apparatus
[12,13]. Fluorescence emission spectra were collected from
single cells by a Peltier-cooled IMAX camera with a liquid
nitrogen cooled intensifier (Princeton Instruments, Princeton,
NJ) attached to a modified Zeiss Axiovert fluorescence microscope. Microspectrophotometry used a 485/22-nm narrow
band pass discriminating filter for excitation, a 510 nm
longpass dichroic mirror, and a 520 nm long-pass emission
filter. The emission spectrum was monitored between 400800 nm. Winspec software (Princeton Instruments) was used
to analyze spectrophotometric data. The excitation and emission wavelength pairs used for EGFP and Alexa fluor labeled
anti-V5-antibody are 490 nm, 509 nm, and 555 nm, 565 nm,
respectively.
RESULTS
Analysis of wild-type and mutant ELOVL4 expressed in COS7 cells: To express wild-type and mutant ELOVL4 proteins
in COS-7 cells, mammalian expression constructs carrying
EGFP-ELOVL4 wild-type and mutant, and V5-tagged
ELOVL4 wild-type and mutant constructs were generated as
Figure 7. Reorganization of
vimentin in the cytoskeleton.
COS-7 cells transfected with
GFP-Ewt (A-C), GFP-Emut
(D-F), GFP-Ewt and GFPEmut (G-I), and mock transfected control (J-L) were labeled with anti-vimentin antibody.
Formation
of
aggresomes is associated with
the reorganization of intermediary filament protein,
vimetin.Vimentin is redistributed from the normal reticular
pattern into a condensed pattern near the aggregated
ELOVL4 protein. Vimentin
distribution is shown in B, E,
H, and K (red). Nuclei were
stained with DAPI. Panels C,
F, I, and L are the overlays of
first two columns. The scale
bar represents 5 µm.
671
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
©2005 Molecular Vision
type and mutant ELOVL4 proteins in cells expressing both
these proteins, COS-7 cells were transfected with constructs
encoding ELOVL4 in four combinations: (1)GFP-Ewt and
GFP-Emut, (2) V5-E wt and V5-E mut, (3) V5-E wt and GFPEmut, and (4) GFP-Ewt and V5-E mut. In addition, COS-7
cells were transfected separately and individually with the
GFP-Ewt, GFP-Emut, V5-E wt, or V5-E mut construct. Mock
transfections were performed with transfection reagents only.
Transfected cells were labeled with Alexa flour 555 conjugated anti-V5 antibody, then observed using laser confocal
microscopy (Figure 3B,E). In cells expressing the wild-type
ELOVL4 fusion protein, the ER localization of ELOVL4 is
consistent with its In vivo localization in photoreceptors [7].
The localization of mutant protein to the juxtanuclear region
was consistent with our previous report (data not shown) [7].
However, in the cells co-transfected with both wild-type and
mutant ELOVL4, both proteins were localized to cytoplasmic
aggregates in the perinuclear region (Figure 3). In these cells,
localization of the wild-type ELOVL4 protein was indistinguishable from the distribution of mutant protein by confocal
5), V5-Emut (lane 6) or mock-transfected cells (lane 10).
When these cell lysates were analyzed with the V5 antibody, a specific signal at a molecular mass of about 41 kDa
(ELOVL4 contributes a mass of 37 kDa and V5 a mass of 4
kDa, summing to an expected mass about 41 kDa, Figure 2)
was detected in cells expressing V5-Ewt protein (lane 5). A
polypeptide of about 37 kDa band was detected in lysates of
cells expressing V5-Emut (lane 6, ELOVL4 contributing 33
and V5 contributing 4 kDa) and in lysates of cells co-expressing GFP-Ewt and V5-Emut (lane 7), V5-Ewt and GFP-Emut
(lane 8). There was no specific band in mock transfected cell
lysates (lane 9), or cells transfected with GFP-Ewt protein (lane
2), GFP-Emut protein (lane 3), or GFP-Ewt and GFP-Emut
protein (lane 4). In summary, these results suggest that the
COS-7 cells transfected with wild-type and/or mutant ELOVL4
constructs carrying a GFP or V5 tag expressed ELOVL4 fusion proteins of the expected sizes (Figure 2, Figure 3, Table
1).
Intracellular misrouting of Ewt in presence of Emut protein: To determine the subcellular localization of the wild-
Figure 8. Localization of
ELOVL4 aggresomes in comparison to Golgi. Intracellular
localization of ELOVL4
aggresomes were compared
with the localization of the
Golgi marker, golgin-97. COS7 cells that express GFP-Emut
(A-C), GFP-Ewt and mutant
(D-F), mock transfected (G-I)
were immunolabeled with antibody to golgin-97 (red), and
cells expressing GFP-Ewt and
GFP-Emut (J-L) were labeled
with anti-TGN38 antibody.
Golgi staining is shown in red
(B,E,H,K). The scale bar represents 5 µm. Overlays of first
two columns show the partial
colocalization of ELOVL4
aggresomes with the Golgi
markers.
672
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
©2005 Molecular Vision
accordance with their expected monomeric molecular weight
indicating that these proteins exist as monomers. Lysates of
cells co-transfected with wild-type and mutant forms of
ELOVL4 constructs showed the presence of ELOVL4 protein on the left side of the gel (Figure 4D-F,H), suggesting
that these proteins are components of higher molecular mass
oligomeric complexes. The mutant ELOVL4 in lysates from
cells transfected with GFP-Emut construct also formed similar high molecular mass complexes on the left side of the gel
(Figure 4C). When co-expressed, both the wild-type and mutant forms of ELOVL4 showed a much broader distribution in
molecular weight of ELOVL4 protein (about 60-200 kDa) in
both horizontal and vertical directions (Figure 4D-F,H). This
is indicative of the protein being present both as a monomeric
species and as a higher molecular mass oligomeric complex
(Figure 4D-F). Thus, these results suggested that when expressed together, the wild-type ELOVL4 and its mutant form
tend to aggregate and form higher molecular mass oligomeric
complexes.
microscopy (Figure 3C,F). Furthermore, the wild-type protein did not co-localize with ER-specific marker protein disulfide isomerase (PDI; data not shown). This suggests that in
the presence of mutant protein, the wild-type protein also was
misrouted to cytoplasmic juxtanuclear aggregates along with
the mutant protein.
Mutant ELOVL4 oligomerizes with wild-type ELOVL4:
A combination of CN-PAGE and SDS-PAGE in a twodimensional approach was used to determine the interaction
between wild-type and mutant ELOVL4 proteins when they
are co-expressed. Immunoblotting of the second dimension
gel either with anti-GFP or anti-V5 identified the migration
pattern of ELOVL4 fusion proteins (Figure 4). Under the conditions employed, the monomeric forms of proteins migrate
towards the extreme right end of the second dimension gel
while high molecular mass proteins or complexes remain on
the left side of the gel. The GFP (Figure 4A), GFP-Ewt (Figure 4B), and V5-Ewt (Figure 4G) proteins appeared toward
the right side of the second dimension blot and migrated in
Figure 9. ELOVL4 aggresomes
are distinct from Golgi. Cells that
express the GFP-Emut (A-C),
GFP-Ewt and GFP-Emut (D-I),
control (J-L) were treated with
the microtubule disrupting drug
nocodazole. Treatment with
nocodazole disrupted both Golgi
and ELOVL4 aggresomes. Fluorescent signals from ELOVL4,
golgin-97, and TGN-38 were
found to be distributed as spots
in the cytoplasm in cells expressing GFP-Emut (A-C) and GFPEwt and GFP-Emut (D-I). The
ELOVL4 signal did not
colocalize with either golgin-97
or TGN-38 indicating that the
ELOVL4 aggresomes are distinct
from Golgi. Golgin-97 and
TGN38 staining is shown in red.
Panels C, F, I, and L are the overlays of GFP and Golgi images
with DAPI stained nuclei. The
scale bar represents 5 µm.
673
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
©2005 Molecular Vision
Interaction of wild-type and mutant ELOVL4 proteins:
Physical interaction of wild-type and mutant ELOVL4 proteins was determined by FRET. This energy transfer between
the donor and acceptor takes place through a nonradiative,
dipole-dipole interaction. The transfer of energy depends upon
the distance and the orientation between the two fluorophores,
occurring over a donor to acceptor separation range of about
10-100 Å. Energy transfer in the nanometer scale makes the
FRET an ideal technique to study the wild-type and mutant
ELOVL4 protein interaction in intact cells. In our studies we
used EGFP as the donor and Alexa 555 as the acceptor. When
the two fluorophores are within a range of about 10-100 Å,
the energy of the donor (EGFP) will transfer to the acceptor
that leads to two peaks in the emission spectrum with one
peak originating from EGFP at 485 nm and a second peak
from Alexa 555 at 565 nm. We assessed the proximity of GFPEwt and V5-Emut (Figure 5A-C) in cells co-expressing wildtype and mutant proteins, and the emission spectra of the two
molecules are shown in Figure 5A,B, respectively. FRET between these molecules was detected (Figure 5C). Similar results were observed in cells co-expressing V5-Ewt and GFPEmut (Figure 5D-F), and FRET was also detected between
V5-Ewt and GFP-Emut (Figure 5F). As a negative control,
GFP-Ewt and V5-Ewt co-expressing cells were labeled with
Alexa 555-conjugated anti-V5 antibodies, and emission spectra are shown in Figure 5G-I. There was no FRET between
the two wild-type proteins (Figure 5I). However, as a positive
control, FRET (Figure 5L) was observed on GFP-Emut and
V5-Emut co-expressing cells (Figure 5J-L). Thus, these results provide evidence that the wild-type and mutant ELOVL4
are in close proximity in cells co-expressing both of them.
The close proximity, as indicated by FRET signals, suggested
that a mutant and wild-type polypeptide must be immediately
adjacent to each other, implying that the two types of polypeptides are together in a complex or aggregate.
ELOVL4 aggresomes are associated with the microtubule
organizing center (MTOC): COS-7 cells expressing either
the mutant ELOVL4 alone or co-expressing the wild-type and
mutant ELOVL4 proteins appeared to form intracellular, perinuclear inclusions that have characteristics of aggresomes.
Aggresomes are pericentriolar membrane-free, cytoplasmic
ubiquitinated inclusion bodies formed at the microtubule organizing center (MTOC), and ensheathed in a cage of intermediate filaments [14,15]. It is well established that misfolded
and aggregated proteins are sequestered into aggresomes
[14,15]. To test whether the wild-type and mutant ELOVL4
complex is associated with aggresomes [14,15], localization
of the mutant and wild-type ELOVL4 was compared with that
of γ-tubulin, a centrosomal protein (Figure 6). The wild-type
and mutant ELOVL4 protein complex was found to be colocalized with γ-tubulin (Figure 6G-I). Similar co-localization
of ELOVL4 with γ-tubulin was also observed in cells expressing the mutant ELOVL4 protein alone (Figure 6D-F), whereas
the wild-type ELOVL4 protein (Figure 6A-C) did not co-localize with γ-tubulin when it was expressed independently
(Figure 6A-C). The γ-tubulin distribution was found to be similar in cells transfected with ELOVL4 fusion constructs or mock
transfected cells (Figure 6J-L).
Reorganization of vimentin distribution in cells with
ELOVL4 aggresomes: We have tested the localization and
nature of intermediate filament network protein, vimentin, in
relation to the distribution of the perinuclear inclusions that
contained GFP-Ewt and GFP-Emut (Figure 7). Staining the
cells with antibody specific to vimentin showed a typical dispersed distribution of vimentin as a long cytoplasmic filamentous network, in wild-type ELOVL4 expressed cells and mock
transfected cells (Figure 7B,K). In contrast to its normal reticular distribution throughout the cell, vimentin showed
juxtanuclear localization in cells co-expressing wild-type and
mutant ELOVL4 proteins (Figure 7H). This altered distribution pattern of vimentin was reported to be associated with
the formation of aggresomes [16]. Similar altered distribution
of vimentin was observed in cells expressing the mutant
ELOVL4 protein alone (Figure 7D-F), but the distribution of
vimentin did not change in cells expressing the wild-type
ELOVL4 protein alone (Figure 7G-I).
To further evaluate the association of co-expressed mutant and wild-type ELOVL4 proteins with aggresomes, cells
transfected with the GFP-Ewt and GFP-Emut constructs were
treated with MG-132, a low molecular weight proteasome inhibitor [17]. The localization of ELOVL4 proteins was examined in comparison to the localization of vimentin in these
cells (data not shown). In cells co-expressing the wild-type
and mutant ELOVL4 proteins, vimentin was found to be completely redistributed from its normal reticular pattern to a
juxtanuclear region (data not shown). Though redistribution
of vimentin was observed in cells expressing the wild-type
ELOVL4 protein, the extent of vimentin redistribution was
significantly different from the pattern of vimentin distribution observed in cells expressing mutant and mutant and wildtype ELOVL4 proteins (data not shown). This suggests a possible impairment of proteasomal degradative pathway in cells
co-expressing wild-type and mutant ELOVL4 proteins. However additional analysis is needed to establish these observations.
ELOVL4 aggregates are distinct from Golgi: The
juxtanuclear localization of aggregates formed in cells expressing mutant ELOVL4 protein alone or co-expressing GFP-Ewt
and GFP-Emut proteins resembled the localization of Golgi.
To determine their specific localization, COS-7 cells expressing mutant ELOVL4 protein alone and cells co-expressing both
wild-type and mutant proteins were labeled for the Golgi markers, golgin-97 and TGN-38. Confocal microscopy revealed a
clear distinction between the localization of co-expressed Ewt
and Emut proteins and the distribution of golgin-97 (Figure
8D-F) and TGN-38 (Figure 8J-L). A similar difference in distribution was observed in the cells expressing the mutant protein alone (Figure 8A-C). golgin-97 signal in these cells was
not completely co-localized with the GFP fluorescence instead
there was a partial co-localization of golgin-97 signal with
mutant ELOVL4 (Figure 8A-C).
To further evaluate the localization of ELOVL4 cytoplasmic juxtanuclear inclusions in comparison to Golgi, cells expressing either the wild-type or mutant ELOVL4 or co-ex674
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
©2005 Molecular Vision
pressing wild-type and mutant ELOVL4 proteins were incubated with nocodazole and immunostained with antibody to
golgin-97 or TGN-38 (Figure 9). As nocodazole is a microtubule disrupting agent, and intact microtubules are required to
form and maintain aggresomes and the Golgi apparatus, treatment with nocodazole disrupts both the aggresome and the
Golgi [14]. Immunolabeling of nocodazole-treated cells with
Golgi specific antibodies will distinguish disrupted Golgi from
aggresomes. As reported previously [7], following nocodazole
treatment, the GFP fluorescence of GFP-Emut did not co-localize with either golgin-97 (Figure 9A-F) or TGN-38 (Figure 9G-I), and appeared as spots at multiple places in the cytoplasm, instead of a large, juxtanuclear aggregate. This suggests that the perinuclear inclusions containing mutant
ELOVL4 protein alone (Figure 8A-C) or co-expressed wildtype and mutant ELOVL4 proteins (Figure 8D-L) were associated with microtubules of the centrosome and were probably distinct from Golgi.
Formation and accumulation of aggresomes were shown
to result in adverse biochemical events that involve the
proteasomal system and apoptosis [19]. The degeneration of
retina observed in patients with ELOVL4 mutations is progressive with onset in the second decade of life [1]. Recruitment of the wild-type protein by the mutant protein may result in the loss of functional ELOVL4. However, it is likely
that a residual amount of wild-type ELOVL4 protein escapes
the interaction with mutant ELOVL4, to reach its target site in
the ER. This residual amount of functional ELOVL4 may be
sufficient for the normal development and maintenance of
retina in early stages of life. However, accumulation of the
wild-type protein with the mutant protein in aggresomes may
cause a cumulative effect leading to degeneration of ELOVL4
expressing photoreceptors over time, as observed in patients
with STGD3. Alternatively, the degeneration observed in these
patients could be due to the lack of sufficient functional
ELOVL4 later in life, as normal adult mouse retinal tissue
shows higher levels of ELOVL4 expression with maturation
indicating greater demand for this protein in older age [20].
Hetero-oligomerization of mutant protein with wild-type
protein, their misrouting and formation of aggresomes has been
found to be a common feature observed in many
neurodegenerative disorders including retinitis pigmentosa
[21-23]. The possible mechanism of protein aggregation in
these cells could be the inability of cells to process misfolded
and aggregated proteins. The mechanism by which the ER
retention signals present in the wild-type ELOVL4 of the aggregates are overridden, resulting in altered trafficking of the
complexes, remains to be elucidated. This may provide important clues towards understanding the degenerative process
and in developing therapeutic strategies to treat protein trafficking diseases. In this regard, further evaluation of aggresome
formation in cells co-expressing mutant and wild-type
ELOVL4 may serve as a model system for studying the cellular fate of mutant proteins and their disease mechanism.
DISCUSSION
We have reported that the mutant ELOVL4 protein, which
lacks an ER retention signal, is misrouted into the perinuclar
region [7]. Previously, we speculated that the 5 bp deletion in
ELOVL4 observed in families with autosomal dominant macular degeneration (STGD3) exerts a dominant negative effect
operating at the protein level [7]. Here we demonstrate that
the mutant ELOVL4 protein altered the trafficking of the wildtype protein by direct physical interactions using three independent approaches. Immunocytochemical analysis showed
the co-localization of the wild-type ELOVL4 protein with the
mutant ELOVL4 protein as cytoplasmic perinuclear aggregates. When co-expressed, wild-type and mutant ELOVL4
proteins were detected as high molecular weight aggregates
as shown by a combination of CN-PAGE and SDS-PAGE in a
two-dimensional electrophoretic approach. These findings and
the results of FRET analysis suggested a direct physical interaction between the wild-type and mutant forms of ELOVL4
proteins. Presently, it is not known whether this interaction
involves other proteins.
Although COS cells are distinct from highly specialized
photoreceptor cells, they are useful tools in understanding the
basic biochemical mechanisms that underlie the mutations
linked to diseases such as retinitis pigmentosa. It is known
that when overexpressed in heterologous systems, proteins tend
to aggregate; however, in this study overexpression of wildtype ELOVL4 alone in COS-7 cells did not cause aggregation, a finding that validated the specificity of the observations. The juxtanuclear location of aggregates containing wildtype and mutant ELOVL4 proteins, the co-localization of centrosome components with these aggregates, the redistribution
of intermediate filaments (IF), and the acceleration of ELOVL4
clustering in response to proteasome inhibition indicated that
these aggregates were sequestered in aggresomes [18]. The
altered trafficking of the wild-type ELOVL4 protein in the
presence of mutant ELOVL4 and their accumulation in the
perinuclear region suggested that this physical interaction underlies macular degeneration in patients with STGD3.
ACKNOWLEDGEMENTS
We thank Bruce Donohoe (University of Michigan) for helping in the operation of the confocal microscope, Austra Liepa
and Regina Belaney (University of Michigan) for assistance
in manuscript preparation, and Swaroop Mahaveer Bhojani
(University of Michigan) for helpful discussions. This work
is supported by grants to RA from NIH (R01EY13198) and
the Foundation Fighting Blindness, a research grant to RA
and an unrestricted grant to the Department of Ophthalmology, UTHSC from Research to Prevent Blindness, Inc., and a
NIH core grant to UTHSC (P30EY13080) and two core grants
to the University of Michigan, Department of Ophthalmology
(P30EY007003) and Vision Research (P30EY07060). This
work was presented, in part, at the Association for Research
in Vision and Ophthalmology (ARVO) meeting, May 2005,
Fort Lauderdale, FL.
REFERENCES
1. Griesinger IB, Sieving PA, Ayyagari R. Autosomal dominant
macular atrophy at 6q14 excludes CORD7 and MCDR1/PBCRA
675
Molecular Vision 2005; 11:665-76 <http://www.molvis.org/molvis/v11/a79/>
loci. Invest Ophthalmol Vis Sci 2000; 41:248-55.
2. Stone EM, Nichols BE, Kimura AE, Weingeist TA, Drack A,
Sheffield VC. Clinical features of a Stargardt-like dominant progressive macular dystrophy with genetic linkage to chromosome
6q. Arch Ophthalmol 1994; 112:765-72.
3. Bernstein PS, Tammur J, Singh N, Hutchinson A, Dixon M, Pappas
CM, Zabriskie NA, Zhang K, Petrukhin K, Leppert M, Allikmets
R. Diverse macular dystrophy phenotype caused by a novel complex mutation in the ELOVL4 gene. Invest Ophthalmol Vis Sci
2001; 42:3331-6.
4. Zhang K, Kniazeva M, Han M, Li W, Yu Z, Yang Z, Li Y, Metzker
ML, Allikmets R, Zack DJ, Kakuk LE, Lagali PS, Wong PW,
MacDonald IM, Sieving PA, Figueroa DJ, Austin CP, Gould
RJ, Ayyagari R, Petrukhin K. A 5-bp deletion in ELOVL4 is
associated with two related forms of autosomal dominant macular dystrophy. Nat Genet 2001; 27:89-93.
5. Oh CS, Toke DA, Mandala S, Martin CE. ELO2 and ELO3, homologues of the Saccharomyces cerevisiae ELO1 gene, function in fatty acid elongation and are required for sphingolipid
formation. J Biol Chem 1997; 272:17376-84.
6. Maugeri A, Meire F, Hoyng CB, Vink C, Van Regemorter N, Karan
G, Yang Z, Cremers FP, Zhang K. A novel mutation in the
ELOVL4 gene causes autosomal dominant Stargardt-like macular dystrophy. Invest Ophthalmol Vis Sci 2004; 45:4263-7.
7. Ambasudhan R, Wang X, Jablonski MM, Thompson DA, Lagali
PS, Wong PW, Sieving PA, Ayyagari R. Atrophic macular degeneration mutations in ELOVL4 result in the intracellular
misrouting of the protein. Genomics 2004; 83:615-25.
8. Camacho-Carvajal MM, Wollscheid B, Aebersold R, Steimle V,
Schamel WW. Two-dimensional Blue native/SDS gel electrophoresis of multi-protein complexes from whole cellular lysates:
a proteomics approach. Mol Cell Proteomics 2004; 3:176-82.
9. Schagger H, Cramer WA, von Jagow G. Analysis of molecular
masses and oligomeric states of protein complexes by blue native electrophoresis and isolation of membrane protein complexes by two-dimensional native electrophoresis. Anal Biochem
1994; 217:220-30.
10. Vijayasarathy C, Damle S, Lenka N, Avadhani NG. Tissue variant effects of heme inhibitors on the mouse cytochrome c oxi-
©2005 Molecular Vision
dase gene expression and catalytic activity of the enzyme complex. Eur J Biochem 1999; 266:191-200.
11. Pedley KC, Jones GE, Magnani M, Rist RJ, Naftalin RJ. Direct
observation of hexokinase translocation in stimulated macrophages. Biochem J 1993; 291:515-22.
12. Petty HR, Worth RG, Kindzelskii AL. Imaging sustained dissipative patterns in the metabolism of individual living cells. Phys
Rev Lett 2000; 84:2754-7.
13. Kindzelskii AL, Yang Z, Nabel GJ, Todd RF 3rd, Petty HR. Ebola
virus secretory glycoprotein (sGP) diminishes Fc gamma RIIIBto-CR3 proximity on neutrophils. J Immunol 2000; 164:953-8.
14. Johnston JA, Ward CL, Kopito RR. Aggresomes: a cellular response to misfolded proteins. J Cell Biol 1998; 143:1883-98.
15. Riley NE, Bardag-Gorce F, Montgomery RO, Li J, Lungo W,
Lue YH, French SW. Microtubules are required for cytokeratin
aggresome (Mallory body) formation in hepatocytes: an in vitro
study. Exp Mol Pathol 2003; 74:173-9.
16. Kopito RR, Sitia R. Aggresomes and Russell bodies. Symptoms
of cellular indigestion? EMBO Rep 2000; 1:225-31.
17. Tawa NE Jr, Odessey R, Goldberg AL. Inhibitors of the
proteasome reduce the accelerated proteolysis in atrophying rat
skeletal muscles. J Clin Invest 1997; 100:197-203.
18. Wigley WC, Fabunmi RP, Lee MG, Marino CR, Muallem S,
DeMartino GN, Thomas PJ. Dynamic association of proteasomal
machinery with the centrosome. J Cell Biol 1999; 145:481-90.
19. Bence NF, Sampat RM, Kopito RR. Impairment of the ubiquitinproteasome system by protein aggregation. Science 2001;
292:1552-5.
20. Mandal MN, Ambasudhan R, Wong PW, Gage PJ, Sieving PA,
Ayyagari R. Characterization of mouse orthologue of ELOVL4:
genomic organization and spatial and temporal expression.
Genomics 2004; 83:626-35.
21. Ross CA, Poirier MA. Protein aggregation and neurodegenerative
disease. Nat Med 2004; 10 Suppl:S10-7.
22. Junn E, Lee SS, Suhr UT, Mouradian MM. Parkin accumulation
in aggresomes due to proteasome impairment. J Biol Chem 2002;
277:47870-7.
23. Tanaka M, Kim YM, Lee G, Junn E, Iwatsubo T, Mouradian
MM. Aggresomes formed by alpha-synuclein and synphilin-1
are cytoprotective. J Biol Chem 2004; 279:4625-31.
The print version of this article was created on 30 Aug 2005. This reflects all typographical corrections and errata to the article through that
date. Details of any changes may be found in the online version of the article. α
676