Retinal Fascin: Functional Nature, Subcellular Distribution, and Chromosomal Localization

Abstract

purpose. To investigate the functional properties, subcellular localization, and chromosomal location of retinal fascin.

methods. Recombinant retinal fascin protein was prepared by using a baculovirus–insect expression system. Actin-binding and -bundling assays were performed with chick actin purified from skeletal muscle. Western blot analysis and immunohistochemistry were performed with a polyclonal antibody raised against bovine retinal fascin. A human retinal cDNA library was screened with an expressed sequence tag cDNA fragment. Chromosomal location was determined with fluorescent in situ hybridization.

results. The actin-binding and actin-bundling activities of retinal fascin were demonstrated by high- and low-speed centrifugation assays. Formation of filamentous (F)-actin bundles by retinal fascin in vitro was also morphologically confirmed by fluorescence microscopy and electron microscopy. Immunohistochemical analysis revealed that retinal fascin protein was localized specifically in the outer and inner segments of the photoreceptor cells in the retina. Two splicing variants of human retinal fascin cDNA were also located. One clone encoded 492 amino acids, and the other encoded 516 amino acids. The gene encoding retinal fascin was localized to human chromosome 17, region q24–25.

conclusions. These results suggest that retinal fascin may play a role in formation of unique morphologic structures of the photoreceptor cells and is a candidate gene for retinal degenerative disorders.

Fascins are a family of actin-bundling proteins, which are expressed in various types of tissues and organs. Fascins cross-link filamentous (F)-actin into bundles. Kane ¹ first isolated echinoid fascin from the extracts of the cytoplasmic proteins of sea urchin eggs. These extracts produced a gel that displayed a complex cross-banding pattern consisting of needlelike structures. ^{1 2} The cDNA encoding echinoid fascin was isolated and had no significant sequence similarities to known actin-binding proteins or three known actin-bundling proteins: fimbrin, villin, and a 30-kDa protein from Dictyostelium. ³ Homology searches have revealed significant similarity between echinoid fascin and the singed gene product in Drosophila that is required for mechanosensory bristle development and oogenesis. ^{3 4} Peptide sequence analysis shows that the 55-kDa actin-bundling protein isolated from HeLa cells was the human homologue of echinoid fascin and the Drosophila singed gene. ^{5 6} Fascins are widely distributed in various tissues and may play important roles in integration into cell adhesion complexes. ⁷  

We have previously isolated retinal fascin cDNA from a bovine retina cDNA library. The cDNA of retinal fascin encodes a protein of 492 amino acids with a calculated relative molecular mass of 55 kDa. Northern blot analysis showed that retinal fascin mRNA is expressed exclusively in the eye, not in other tissues. In situ hybridization revealed that retinal fascin is expressed specifically in photoreceptor cells. ⁸ Although retinal fascin shows significant homology to the fascin gene family, domains highly conserved in other members of the fascin family are not completely conserved in retinal fascin. To elucidate the function of retinal fascin, we analyzed the properties of retinal fascin protein using recombinant protein. We report the functional analysis of retinal fascin, subcellular localization of retinal fascin protein, and the cloning of a human homologue of retinal fascin.  

Materials and Methods

Animals

All animal procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Proteins

Actin was purified from chicken skeletal muscle as described previously. ⁹ The molecular masses used for calculating the molarity of protein concentrations were actin, 42,000, and retinal fascin, 55,000. 

Preparation of Recombinant Retinal Fascin

Full length bovine retinal fascin cDNA was subcloned into the glutathione-S-transferase (GST) fusion pVL 1392 vector (PharMingen, San Diego, CA) between EcoRI and BamHI sites. This plasmid and baculovirus DNA (Linearized BaculoGold; PharMingen) were cotransfected into Spodoptera frugiperda 9 (Sf9) insect cells (Invitrogen, Carlsbad, CA) according to a commercial system (BaculoGold Transfection Kit; PharMingen). Recombinant fusion proteins extracted with Triton X-100 were purified by chromatography on glutathione-agarose (Amersham Pharmacia Biotech, Piscataway, NJ). 

Actin-Binding Assay

Recombinant retinal fascin at various concentrations (0–3.3 μM) was mixed with 7.1 μM F-actin in 100 mM NaCl and 20 mM Tris-HCl (pH 7.6). The reaction mixtures were incubated at room temperature for 30 minutes and centrifuged in a rotor (LP42TI; Beckman, Palo Alto, CA) at 140,000g for 40 minutes. Supernatants and resuspended pellets were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and densitometry. 

Actin-Bundling Assay

Actin-bundling activity was analyzed by low-speed centrifugation assay, fluorescence microscopy, and electron microscopy. In the low-speed centrifugation assay, retinal fascin at various concentrations (0–3.3 μM) was mixed with 7.1 μM F-actin in 100 mM NaCl and 20 mM Tris-HCl (pH 7.6). The reaction mixtures were incubated at room temperature for 30 minutes and centrifuged at 8000g for 20 minutes. Both supernatants and pellets were dissolved in an equivalent volume of SDS sample buffer and analyzed by SDS-PAGE and densitometry. For fluorescence microscopy, F-actin was labeled with rhodamine-phalloidin (Molecular Probes, Eugene, OR) as described elsewhere. ¹⁰ F-actin (7.1 μM, of which 10% was labeled) was then mixed with 1.4 μM retinal fascin. After incubation for 30 minutes at room temperature, the samples were observed with a fluorescence microscope (Axioplan; Carl Zeiss, Oberkochen, Germany) equipped with a ×100 oil lens (Plan-Neofluar; Zeiss). For electron microscopy, F-actin (7.1 μM) was mixed with retinal fascin (1.4 μM) and incubated for 30 minutes at room temperature. After incubation, the samples were stained with 1% uranyl acetate and observed with an electron microscope (JEM-2010; JEOL, Tokyo, Japan). 

Preparation of Anti-Retinal Fascin Antibody and Western Blot Analysis

Bovine retinal fascin-GST fusion protein was used as an antigen to generate polyclonal antibodies to bovine retinal fascin. Rabbits were immunized with a mixture of retinal fascin fusion protein and Freund’s adjuvant. A GST column was used to remove antibody against the GST fusion part. Bovine retinas and nonretinal tissues (spleen, heart, liver, brain, and small intestine) were harvested and then lysed in buffer containing 10 mM Tris-HCl (pH 7.4), 1% Nonidet P-40, 0.1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, 1 mM EDTA, and 10 μg/ml aprotinin (Sigma, St. Louis, MO). Lysates were centrifuged at 11,000g for 10 minutes to remove large tissue debris. Samples were boiled and loaded onto a 5% to 20% gradient SDS-polyacrylamide gel. After electrophoresis, gels were blotted onto membranes (Immobilon P; Millipore, Bedford, MA). After blocking with phosphate-buffered saline (PBS) containing 5% nonfat milk, the membranes were incubated with anti-retinal fascin antibody for 4 hours at room temperature. Detection of reactive bands was performed with 0.1% alkaline phosphatase-conjugated goat anti-rabbit IgG (Boehringer Mannheim, Mannheim, Germany) in PBS containing 0.1% Tween 20 (PBS-T) with 5% nonfat milk by the alkaline phosphatase method. 

Immunohistochemistry

Immunohistochemical analysis of the bovine retina was performed on 6-μm frozen sections of 4% paraformaldehyde-fixed tissues. The sections were blocked in both 5% normal goat serum (NGS) and 0.3% Triton X-100 in 0.1 M PBS for 30 minutes at room temperature and subsequently immunoreacted with a diluted solution of the anti-retinal fascin antibody (1:4000 in 0.1 M PBS containing 3% NGS and 0.3% Triton X-100) for 20 hours at 4°C. After the slides were washed with 0.1 M PBS, they were incubated with indocarbocyanine-conjugated anti-rabbit IgG (Jackson Laboratory, Bar Harbor, ME; 1:500 in 0.1 M PBS) for 2 hours at room temperature. After the slides were washed with 0.1 M PBS, specific staining was detected using a fluorescence microscope (Olympus, Tokyo, Japan). 

cDNA Library Screening

A human retina cDNA library constructed in the λgt10 vector (Clontech, Palo Alto, CA) was screened as described elsewhere. ¹¹ The probe was made from an expressed sequence tag (EST) cDNA clone (accession number AA018572) which showed a high degree of homology with bovine retinal fascin at the amino acid level. Eight positive clones were isolated. λDNA was isolated according to the standard method. ¹² Inserts were excised by EcoRI digestion, subcloned into pBluescript II SK(−) (Stratagene, La Jolla, CA), and sequenced by the dideoxy chain termination method. ¹³  

Fluorescence In Situ Hybridization

Lymphocytes isolated from human blood were cultured in α-minimal essential medium (α-MEM) supplemented with 10% fetal calf serum and phytohemagglutinin at 37°C for 68 to 72 hours. The lymphocyte cultures were treated with 5-bromo-2′-deoxyuridine (0.18 mg/ml: Sigma) to synchronize the cell population. The synchronized cells were washed three times with serum-free medium to release the block and were recultured at 37°C for 6 hours in α-MEM with thymidine (2.5μg/ml: Sigma). Cells were harvested, and slides were made by using standard procedures (including hypotonic treatment), fixed, and air dried. The 1.5-kb cDNA probe was biotinylated with dATP using a labeling kit (15°C, 1 hour; BioNick; Gibco, Paisley, UK). ¹⁴ Fluorescence in situ hybridization (FISH) analysis was performed according to Heng et al. ¹⁴ and Heng and Tsui. ¹⁵ FISH signals and the 4′,6-diamidino-2-phenylindole (DAPI) banding pattern were recorded separately by taking photographs, and the assignment of the FISH mapping data with chromosomal bands was achieved by superimposing FISH signals with DAPI-banded chromosomes. ¹⁶  

Other Procedures

A protein concentration assay was performed based on the method of Bradford ¹⁷ using bovine serum albumin as the standard. SDS-PAGE with 12.5% polyacrylamide gels was performed as described elsewhere. ¹⁸ The buffer system used was that described by Laemmli. ¹⁹  

Results

Actin-Binding Activity of Retinal Fascin

We examined whether purified retinal fascin shows actin-binding activity in vitro by high-speed centrifugation assay as described. In the absence of F-actin, free retinal fascin remained in the supernatant (Fig. 1A ). In the presence of F-actin, however, retinal fascin formed pellets (Fig. 1A) . We measured actin-binding activity of retinal fascin at various concentrations to determine the binding profile in detail (Fig. 1B) . The binding was saturated at a molar ratio of 3:1 actin molecules to retinal fascin molecules. The dissociation constant, which is the free retinal fascin concentration under which half-maximum binding was obtained, was 3.0 × 10⁻⁷ M. 

Actin-Bundling Activity of Retinal Fascin

It is important to investigate whether retinal fascin has actin-bundling activity similar to that of other fascins. Fascin seems to have two actin-binding sites per molecule and forms tight, compact bundles of actin. ²⁰ First, we performed a low-speed centrifugation assay to examine the actin-bundling activity of retinal fascin. Bundles of F-actin formed a sediment in the low-speed centrifugation assay, whereas free F-actin remained in the supernatant. In the presence of retinal fascin, F-actin formed bundles and pellets (Fig. 2A ). In this assay, the amount of pelleted actin gradually increased according to the amount of retinal fascin present. Approximately 70% of the actin was pelleted when retinal fascin was added at 3.3 μM (Fig. 2B) . These observations confirmed the actin-bundling activity of retinal fascin. We also demonstrated formation of F-actin bundles by retinal fascin using fluorescence microscopy. In the absence of retinal fascin as a control, we detected no actin bundles (Fig. 3A ). In contrast, we found bright, thick F-actin bundles in the presence of retinal fascin. These F-actin bundles were straight and uniform (Fig. 3B) . Electron microscopy using the negative staining technique also revealed tight and compact bundles of F-actin formed by retinal fascin (Fig. 3C) . The results of fluorescence microscopy and electron microscopy confirmed those of the low-speed centrifugation assay. These observations revealed actin-bundling activity of retinal fascin biochemically and morphologically. 

Localization of Retinal Fascin Protein by Immunohistochemistry

Western blot analysis using a polyclonal antibody raised against bovine retinal fascin confirmed retinal fascin protein with a molecular weight of approximately 55 kDa only in extracts of bovine retinas, whereas retinal fascin protein was not detected in other tissues, including spleen, heart, liver, brain, and small intestine (Fig. 4A ). Immunohistochemical analysis using this antibody revealed that retinal fascin immunoreactivity was localized only in the outer and inner segments of the photoreceptor cells in the retina (Fig. 4C) . We detected no positive immunoreactivity in other parts of the retina. As a negative control, preabsorption of retinal fascin antiserum with 10⁻⁶ M recombinant bovine retinal fascin resulted in no staining in the retina (Fig. 4D) . 

Cloning of Human Retinal Fascin Splicing Variants

We isolated cDNA encoding the human homologue of retinal fascin to analyze the relationship between retinal fascin and hereditary photoreceptor diseases. We screened a human retina cDNA library using EST cDNA as a probe and isolated six independent positive clones. Restriction enzyme mapping and partial sequencing indicated the presence of at least two different clones. The long clone (human retinal fascin 2) contained a cDNA insert of 1782 bp with an open reading frame encoding a protein of 516 amino acids. The encoded protein was predicted to have a calculated relative molecular mass of 57.4 kDa. The short clone (human retinal fascin 1) contained a cDNA insert of 1710 bp with an open reading frame encoding a protein of 492 amino acids; 24 amino acids were deleted from human retinal fascin 2. The encoded protein of an alternatively spliced product was predicted to have a calculated relative molecular mass of 55.1 kDa. The sequence of human retinal fascin 1 was identical with that of human retinal fascin reported previously in GenBank (Tubb and Bryan accession number AF030,165). The nucleotide and deduced amino acid sequences of human retinal fascin 2 are shown in Figure 5A . There was no signal sequence in the N-terminal. The hydrophobicity profile of the primary amino acid sequence of the predicted protein revealed the absence of transmembrane domains. Sequence analysis of human retinal fascin 2 indicated that the deduced protein shared 55% amino acid identity with human fascin, 53% with murine fascin, 50% with Xenopus fascin, 40% with Drosophila singed gene, and 36% with echinoid fascin, whereas the protein shared 93% identity with bovine retinal fascin. Comparison of the deduced amino acid sequences of human retinal fascin and bovine retinal fascin is shown in Figure 5B . 

Chromosomal Localization of Retinal Fascin

Under the conditions used, the hybridization efficiency was approximately 72% for this probe (among 100 checked mitotic figures, 72 showed signals on one of the chromosome pairs). DAPI banding was used to identify the specific chromosome, and the signal was assigned to the long arm of chromosome 17. The detailed position was further determined based on the summary from 10 photographs (Fig. 6A ). No additional loci were found by FISH analysis under the conditions used. Therefore, the retinal fascin gene was located on human chromosome 17, region q24-q25 (Figs. 6B 6C) . 

Discussion

This study describes the characterization of retinal fascin. Members of the fascin family that share three highly conserved regions ⁷ have been isolated from sea urchin, ³ Drosophila, ²¹ Xenopus, ²² mouse, ²³ and human. ²⁴ These are widely expressed proteins found in a broad spectrum of tissues and organs. Retinal fascin was isolated from a bovine retina cDNA library and is specifically expressed in the retina. ⁸ Retinal fascin shares one highly conserved region with other fascins and shows moderate homology in two other regions that are also highly conserved in the fascin gene family. ⁸ These findings indicated that retinal fascin has a unique localization and structure in comparison with other fascins. 

It is important to determine whether retinal fascin has both the actin-binding and actin-bundling activities shown by other members of the fascin gene family despite its unique structure. We demonstrated that retinal fascin also showed actin-binding activity. The binding of retinal fascin to F-actin was saturated at an approximate stoichiometry of 1:3 molecules of retinal fascin to F-actin. The stoichiometries of the binding were reported as follows. The estimated molar ratios of fascin-actin were 1:4 to 5 for echinoid fascin, ²⁰ 1:4.3 for the Drosophila singed gene, ⁴ and 1:4.1 for murine fascin. ²³ Fascin purified from rat brain bound to F-actin with a stoichiometry of 1:2 or 3 molecules of fascin to F-actin. ²⁵ Sasaki et al. ²⁵ argued that the differences between rat brain fascin and other fascins could be due to determination of protein concentration. The differences in the stoichiometry of binding activity of retinal fascin and those of other fascins may also be explained in the same way ²⁵ or may be due to the unique structure of retinal fascin, because the domains highly conserved among other members of fascin gene family are not fully conserved in retinal fascin. Actin bundles generated by retinal fascin were directly visualized by fluorescence microscopy and were found to show a morphology similar to those of actin bundles made with fascin purified from rat brain. Electron microscopy by the negative staining technique showed numerous bundles of F-actin generated by fascin. ^{5 25} F-actin bundled with retinal fascin also formed numerous bundles. There were morphologic similarities between the actin bundles made with retinal fascin and those made with other fascins, despite the structural differences, suggesting that the common regions between retinal fascin and other fascins may play critical roles in the actin-bundling activity. 

To determine whether retinal fascin is associated with hereditary photoreceptor diseases, it is necessary to isolate the cDNA encoding a human homologue of retinal fascin. In Drosophila fascin, single amino acid mutations that change glycine 409 to glutamic acid and serine 289 to asparagine were shown to disrupt actin-bundling activity in vivo. ²⁶ These two amino acids are conserved in both bovine and human retinal fascin. Fascin should have two actin-binding domains per molecule, ⁵ one located in the C terminus and the other in the N terminus. ²⁷ Yamakita et al. ²⁸ reported that the actin binding of human fascin was regulated by phosphorylation and that one of the sites phosphorylated in human fascin was Ser-39. This residue is conserved in human retinal fascin, suggesting that it is important in the functions of human retinal fascin. There are multiple isoelectric variants in many fascins, including those of human, ⁵ mouse, ²³ and sea urchin. ⁷ Edwards et al. ²³ reported that fascin isoforms are not the result of alternative splicing. We isolated two variants of human retinal fascin. One clone was 24 amino acids longer than the other clone. We considered the two variants of human retinal fascin to be alternative splicing products because a FISH study demonstrated the presence of a single retinal fascin gene. This is the first report of the result of alternative splicing in retinal fascin. 

The chromosomal location of retinal fascin was determined by FISH. Retinitis pigmentosa (RP) is the term for a group of hereditary degenerative diseases of the retina that are characterized as progressive dysfunction in photoreceptors and other cell layers. Clinical features of RP include night blindness, peripheral visual field loss, and eventual total loss of vision. RP can be inherited as an autosomal dominant, autosomal recessive, or X-linked disorder. Autosomal dominant RP (adRP) has been mapped to nine loci on chromosomes 1p, 3q, 6p, 7p, 7q, 8q, 17p, 17q, and 19q. ^{29 30 31} Dryja et al. ³² reported that a point mutation of the gene encoding for rhodopsin on chromosome 3q caused one form of adRP. Some cases of adRP have been shown to be due to mutations in the peripherin–retinal degeneration slow gene on chromosome 6p. ^{33 34 35} The gene for the RP1 locus has been characterized. ³⁶ The six other adRP loci have not yet been identified. Genes encoding components of the phototransduction pathway, vitamin A (retinol) metabolism, and the structure of the disc membrane are good candidates for genes involved in RP. Immunohistochemical analysis revealed that retinal fascin is localized only in the outer and inner segments of the photoreceptor cells in the retina. The connecting cilium that joins the inner and outer segments of both rods and cones is unusual. ³⁷ A vertebrate photoreceptor cell contains a cluster of F-actin in its connecting cilium. ^{38 39 40 41 42} The plasma membrane of the distal connecting cilium evaginates to form new outer segment disks. ⁴³ F-actin may be involved in disc morphogenesis in vertebrate photoreceptors. ^{38 40 42} It may also be involved in stopping the growth of the nascent disks or in initiating the morphogenesis of a new disc. ^{37 44} Retinal fascin may play a role in formation of the unique morphologic structures of the photoreceptor cells such as the disc and connecting cilium. A dominant form of retinitis pigmentosa has been mapped to a locus on chromosome 17q. ³¹ Interestingly, the retinal fascin gene is also located on human chromosome 17, region q24–25. The RP17 locus has recently been reassigned to 17q22. Multiple recombination events in two RP17-linked families using novel intragenic polymorphisms in the retinal fascin gene provided evidence for the exclusion of retinal fascin as the disease-causing gene in this form of adRP. ⁴⁵ The more recent publication by den Hollander et al. ⁴⁶ describes another RP17-linked family in which the critical region is 7.7cM on 17q22, and retinal fascin could be the disease-causing gene in that adRP family. In fact, retinal fascin could be the disease-causing gene for any retinal degeneration, not just RP. 

 Submitted for publication August 26, 1999; revised February 1, 2000; accepted February 7, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Binding of retinal fascin to F-actin. (A) Retinal fascin (3.3 μM) was incubated without (−) or with (+) F-actin (7.1μ M) for 30 minutes. The reaction mixtures were centrifuged at 140,000g for 40 minutes. Both supernatants (s) and resuspended pellets (p) were analyzed by SDS-PAGE and densitometry. In the presence of F-actin, retinal fascin formed pellets. (B) The amount of retinal fascin bound to F-actin was plotted against the concentration of free retinal fascin.

Figure 2.

 

F-actin-bundling activity of retinal fascin. (A) F-actin (7.1 μM) were incubated without (−) or with (+) retinal fascin (3.3μ M) for 30 minutes. The reaction mixtures were centrifuged at 8000g for 20 minutes. Both supernatants (s) and pellets (p) were analyzed by SDS-PAGE and densitometry. In the presence of retinal fascin, F-actin bundles and pellets were formed. (B) The amounts of pelleted actin were plotted against the concentration of retinal fascin. Data are means ± SD (bars; n = 3).

Figure 3.

 

Direct visualization of actin-bundling activity of retinal fascin. F-actin (7.1 μM) labeled with rhodamine-phalloidin was mixed with 1.4μ M retinal fascin. (A) Actin only. (B) In the presence of retinal fascin, F-actin bundles were observed by fluorescence microscopy. (C) F-actin (7.1 μM) was mixed with 1.4 μM retinal fascin and stained with 1% uranyl acetate. The reaction mixtures were observed by electron microscopy. Bars, (A, B) 10 μm; (C) 200 nm.

Figure 4.

 

(A) Western blot analysis of retinal fascin in the bovine tissue extracts. The tissue extracts from retina, spleen, heart, liver, brain, and small intestine were separated by SDS-PAGE and then transferred onto a membrane. The filter was incubated with the antibody and then stained by the alkaline phosphatase method. Sizes of standard proteins are shown in kilodaltons. The arrow shows the immunoreactive band of retinal fascin. Lane 1, spleen; lane 2, heart; lane 3, liver; lane 4, brain; lane 5, retina; and lane 6, small intestine. (B) Light microscopic view of bovine retina. (C) Localization of retinal fascin by immunohistochemistry in the bovine retina. Intense retinal fascin immunoreactivity was seen only in the outer segments and inner segments of the photoreceptor cells in the retina. (D) Preabsorption of retinal fascin antiserum with 10⁻⁶ M recombinant bovine retinal fascin resulted in no staining in the retina. PL, photoreceptor cell layer; OS, outer segment; IS, inner segment; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer. Bar, 50 μm.

Figure 5.

 

(A) Nucleotide and deduced amino acid sequences of human retinal fascin 2. The translational termination codon is marked by an asterisk. This clone contained a cDNA insert of 1782 bp with an open reading frame encoding a protein of 516 amino acids. (B) Alignment of amino acid sequences encoding human and bovine retinal fascin. Identical amino acid residues in each peptide are indicated by asterisks.

Figure 6.

 

(A) FISH mapping results for retinal fascin. Each dot represents the double FISH signals detected on human chromosome 17. (B, C) Example of FISH mapping of retinal fascin: (B) FISH signals on chromosome; (C) same mitotic figure stained with DAPI to identify chromosome 17.