Distribution of CESP-1 Protein in the Corneal Endothelium and Other Tissues

Abstract

purpose. The gene expression profile of human corneal endothelium (CE) was established with the gene signature system. A novel gene, GS3582, was abundantly transcribed in the CE compared with other tissues according to a human gene expression database. This protein was designated corneal endothelium–specific protein (CESP)-1. The tissue distribution and subcellular localization of CESP-1 was assessed in humans and mice, to investigate its physiological function.

methods. Rabbit and mouse CESP-1 cDNAs were cloned, and a polyclonal anti-human CESP-1 antibody (Ab) and anti-mouse N- or C-terminal ovary-specific acidic protein (OSAP)-1 Ab were produced. CESP-1 expression was investigated in human and mouse corneas by Western blot and/or immunohistochemical analysis. The distribution of CESP-1 in human tissues was also examined by Western blot analysis. To identify the subcellular localization of CESP-1, cultured human CE was colabeled with anti-human CESP-1 Ab and anti-cytochrome c monoclonal Ab or anti-GRP78 monoclonal Ab for confocal microscopy.

results. The rabbit and mouse CESP-1 cDNA sequences contained an open reading frame coding 242 and 283 amino acids, respectively. Mouse CESP-1 was entirely consistent with mouse OSAP. Western blot analysis showed that CESP-1 was expressed in the human corneal epithelium, CE, cultured CE, brain, testis, and ovary. Mouse CESP-1 was also expressed in mouse corneal epithelium and CE with anti-mouse C- but not N-terminal OSAP Ab according to immunohistochemical analysis. Subcellular localization of CESP-1 to the mitochondria was demonstrated in cultured human CE. The N-terminal of CESP-1, possessing a mitochondrial targeting sequence, may be processed after the protein is imported into the mitochondria.

conclusions. CESP-1 was distributed in the corneal epithelium, the CE and cultured human CE, as well as the brain, testis, and ovary. CESP-1 was localized in the mitochondria of cultured human CE. These findings may provide some clues about the physiological function of CESP-1.

The corneal endothelium (CE) is essential for maintenance of corneal transparency. Stromal hydration is regulated by the barrier and pump functions of CE. ^{1 2} CE injury due to trauma, disease, or intraocular surgery can result in irreversible opacity, because endothelial mitotic activity is extremely low in vivo and decreases with age. ^{3 4} The lack of knowledge about physiological mechanisms regulating CE functions such as mitosis makes it difficult to develop effective treatment for endothelial diseases. 

To clarify the physiological mechanisms regulating the CE, we investigated the gene expression profile of human CE by using the gene signature (GS) system, ^{5 6} since the pattern of gene expression should reflect the unique characteristics of this tissue. As a result, we identified a novel gene, GS3582, ⁷ which was one of the most abundant transcripts and showed a higher level of expression than in other tissues in a human gene expression database (BodyMap; http://bodymap.ims.u-tokyo.ac.jp/ provided in the public domain by the University of Tokyo). We designated the protein from this gene as corneal endothelium–specific protein (CESP)-1. The primary structure of human CESP-1 is characterized by a high proportion of acidic amino acids (46 glutamates and aspartates among 240 amino acids) and the theoretical isoelectric point is 4.24. The sequence of CESP-1 shows marked correspondence with that of human ovary-specific acidic protein (OSAP), ^{8 9 10} and CESP-1 transcripts have been detected in the pituitary, prostate, testis, and ovary among various human tissues tested. ⁷ Although the characteristic motif associated with physiological function has not been determined, the PSORT (http://psort.nibb.ac.jp/ provided in the public domain by the University of Tokyo) cellular localization prediction algorithm has suggested that CESP-1 may be localized in the mitochondria. ¹¹  

To obtain clues about the physiological function of this protein, we produced a polyclonal antibody (Ab) directed against human CESP-1. Then we investigated the expression of CESP-1 in the corneas of mice, rabbits, and humans, as well as the subcellular localization of CESP-1 in cultured human CE using anti-mouse and/or anti-human CESP-1 Abs. 

Materials and Methods

Animals

Inbred C57BL/6 female mice (7–8 weeks old) were purchased from Clea Japan (Tokyo, Japan). Adult female Japanese White rabbits were purchased from Saitama Experimental Animals, Inc. (Saitama, Japan). Eyes harvested from Japanese White rabbits were purchased from Funakoshi Co. (Tokyo, Japan). The animals were handled in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Cloning of Rabbit CESP-1 cDNA

After clear corneas were removed from five rabbit eyes with scissors, a sheet of CE with Descemet’s membrane was peeled from periphery to center with fine forceps. The CE was immediately homogenized (RNA STAT-60; Tel-Test. Inc., Friendswood, TX), and total RNA was extracted from the solution. ⁷ First-strand cDNA was synthesized using 0.4 μg of total RNA. For 5′-RACE PCR, a kit was used (SMART RACE cDNA Amplification Kit; BD-Clontech, Palo Alto, CA) that provides a “universal primer” at the 5′ end and gene-specific primer (GSP: 5′-CTC AGC AGG AGA CTA GCC TTG CGG GG-3′) designed from the expected sequence. cDNA was amplified (Advantage 2 Polymerase Mix; BD-Clontech, Palo Alto, CA) for first-round PCR. Touchdown PCR was performed as the first round PCR, with 5 cycles of 94°C for 5 seconds and 72°C for 3 minutes followed by 5 cycles of 94°C for 5 seconds, 70°C for 10 minutes and 72°C for 3 minutes, and then 25 cycles of 94°C for 5 seconds, 68°C for 10 minutes, and 72°C for 3 minutes in a PCR thermal cycler (MP; Takara, Kyoto, Japan). PCR products were separated on a 1% agarose gel containing 100 pg/mL ethidium bromide. A 900-bp band was excised and purified (GeneElute agarose spin column; Sigma-Aldrich, St. Louis, MO). The purified cDNA was amplified by PCR again, cloned (pBluscript; Stratagene, La Jolla, CA), and sequenced with an autosequencer (PRISM 310 Genetic Analyzer; Applied Biosystems, Inc., [ABI] Foster City CA). 

Cloning of Mouse CESP-1 cDNA

The eyes were harvested from a C57BL/6 mouse and first-strand cDNA was synthesized as described earlier. Mouse CESP-1 cDNA was amplified by PCR with a pair of primers (EcoRI-linked 5′-GGA ATT CAT GTA TCT CCG CAG GGC TGT-3′ and BamHI-linked 5′-CGG GAT CCA AGG GCT AAG GTC ACT AAA AAT ACA AA-3′) designed on the basis of the mouse OSAP sequence. ^{8 9} The amplified fragment was digested with EcoRI and BamHI, cloned (pBluscript; Stratagene) and sequenced with an autosequencer. 

Preparation of Anti-human CESP-1 Ab and Anti-mouse OSAP Ab

The portion of human CESP-1 cDNA corresponding to Met¹-Gly²⁴⁰ was amplified by PCR using a pair of primers that flanked the insert (5′-AAG GAT CCA TGT ATC TCC GCA GGG CG-3′ and 5′-CGG AAT TCA AAT GTC TAC CGG CTG GAG ATT AG-3′). The PCR product was subcloned into the a prokaryotic expression vector (pGEX-6P-1; GE Healthcare, Piscataway, NJ), which is a glutathione S-transferase (GST) fusion vector. GST-CESP-1 fusion protein expression was induced in Escherichia coli BL21 (DE3) pLysS by 1 mM isopropyl-β-d-thiogalactoside. Then the protein was purified by affinity chromatography (GFTrap FF; GE Healthcare, Piscataway, NJ), followed by anion-exchange chromatography (Resourse Q and PD-10 Desalting columns; GE Healthcare, Piscataway, NJ). The purified GST-CESP-1 fusion protein (1 mg/mL) was mixed with an equal volume of Freund’s complete adjuvant (Difco Laboratories, Sparks, MI), and 0.5 mg of protein was injected subcutaneously into an adult female rabbit. Three booster injections were administered at 3-week intervals and serum was harvested 2 weeks after the last dose. 

Synthetic peptides corresponding to residues Gly¹⁹-Ser³⁵ and Ser²⁵⁴-Gly²⁸³ of mouse OSAP were prepared and designated as anti-N-terminal OSAP Ab and anti-C-terminal OSAP Ab, respectively. The reason we selected those regions of mouse OSAP as the antigen is that they have low homology among other species and high antigenicity—that is, high hydrophilicity and a high amount of α-helix in secondary structure. BALB/c mice were immunized intraperitoneally with the peptide Gly¹⁹-Ser³⁵ and Ser²⁵⁴-Gly²⁸³, which were coupled to keyhole limpet hemocyanin. Antigen in Freund’s complete adjuvant was used for the first immunization, followed by 3 boosts in Freund’s incomplete adjuvant. Postimmune sera were screened for their affinity toward purified recombinant OSAP. Splenocytes of the best responder mouse were fused with a mouse myeloma cell line according to standard procedures, and growing hybridomas were screened by an ELISA in which recombinant proteins were coated to the microtiter plate. A clone for each peptide with a strong and specific reaction with recombinant OSAP was selected. The IgG was isolated from the culture medium by protein G affinity chromatography. 

Cell Culture

All donor corneas with no history of corneal disease, infection, or intraocular surgery, obtained from the Central Florida Lions Eye and Tissue Bank, were kept in storage medium (Optisol GS; Chiron Vision, Irvine, CA) at 4°C and were used within 7 days of the donor’s death. The age of the donor was from 41 to 68 years. Primary human CE cultures were made as described elsewhere. ¹² Briefly, small explants from the endothelial layer, including Descemet’s membrane, were removed with sterile surgical forceps. The ∼1-mm² size explants were made of a cornea and placed endothelial cell-side down onto four 35-mm tissue culture dishes coated with bovine extracellular matrix (ECM). This coating dish was prepared by primary bovine CE culture and CE removal with trypsin-EDTA (EDTA). When a sufficient proliferating cell density was reached, the human CEs were passaged at ratios ranging from 1:1 to 1:4. Subsequent passages were done by the same method, but at a ratio of 1:16 in growth medium consisting of Dulbecco’s modified Eagle’s medium supplemented with 15% fetal bovine serum, 30 mg/L of l-glutamine, 2.5 mg/L of amphotericin B (Fungizone; Invitrogen Co., Carlsbad, CA), 2.5 mg/L of doxycycline, and 2 ng/mL of basic fibroblast growth factor. Culturing was done in a humidified incubator at 37°C under 5% CO₂, and the medium was replaced every second day. 

Preparation of Protein Extracts

After excising the transparent cornea with scissors, the corneal epithelium was scraped off the stroma and the CE with Descemet’s membrane was peeled off as described earlier. Human CEs were plated on 100-mm dishes for culture and then were washed three times with phosphate-buffered saline (PBS), harvested, and centrifuged. 

The corneal epithelium, CE, and cultured human CE were suspended in 200 μL of 20 mM Tris-HCl (pH 7.5), 500 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol (DTT), mammalian tissue protease inhibitor cocktail for 14 μM E-64 (trans-epoxysuccinyl-l-leucylamido (4-guanidino) butane), 21 μM leupeptin, 15 μM pepstatin A, 36 μM bestatin, 0.8 μM aprotinin, and 1.4 mM AEBSF (Sigma-Aldrich, St. Louis, MO) and 1% Triton X-100, and then were sonicated for 4 minutes (Sonifier; Branson, Danbury, CT) at output level 4 and the 50% duty cycle. After centrifugation at 20,000g for 20 minutes, the protein concentration of the supernatant thus obtained was determined by a protein assay (DC; Bio-Rad Laboratories, Hercules, CA). The supernatant was diluted 1:5 in sodium dodecyl sulfate (SDS) buffer (87.5 mM Tris-HCl [pH 6.8], 600 mM DTT, 10% SDS, 0.01% bromphenol blue, and 30% glycerol) and boiled for 5 minutes. Protein extracts were also obtained from the CE of a rabbit and a C57BL/6 mouse. 

In vitro translation of human and mouse CESP-1 was performed with the TNT coupled reticulocyte lysate system (Promega Co., Madison, WI), according to the manufacturer’s protocol. Because in vitro translated protein is commonly unmodified such as by phosphorylation, they were used as a positive control for immunoblot analysis. The protein was dissolved in an equal volume of SDS buffer (31 mM Tris-HCl; [pH 6.8], 4% SDS, 0.01% bromphenol blue, 20% glycerol, and 280 mM-β-mercaptoethanol and boiled for 5 minutes. 

Immunoblot Analysis

Samples were subjected to SDS-PAGE on 7% or 10% polyacrylamide gel, and then transferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore, Bedford, MA). Then the membranes were blocked for 60 minutes at room temperature in blocking solution (20 mM Tris-HCl [pH 7.5], 500 mM NaCl, 0.05% Tween 20, and 5% skim milk), and incubated overnight at 4°C with diluted anti-CESP-1 Ab, anti-N-terminal OSAP Ab, or anti-C-terminal OSAP Ab (1:1000 to 1:5000 dilution of Abs, 500 mM NaCl and 0.05% Tween 20 in 20 mM Tris-HCl [pH 7.5]). After washing, incubation was done with horseradish peroxidase-conjugated donkey anti-rabbit immunoglobulin (1:500 dilution of 500 mM NaCl and 0.05% Tween 20 in 20 mM Tris-HCl [pH 7.5]; GE Healthcare). Bound Abs were detected by a chemiluminescence assay (ECL; GE Healthcare). 

To investigate the distribution of CESP-1 in human tissues, we used a ready-for-use Western blot system with human tissue lysates (INSTA-Blot; Imgenex, San Diego, CA). Approximately 10 μg per lane of each human tissue lysate was resolved on a membrane (Immobilon; Millipore). 

Immunohistochemistry

The primary Abs used for immunostaining were rabbit anti-mouse OSAP Ab. Control sections were incubated with rabbit immunoglobulin (Sigma-Aldrich) in place of the primary Ab. Frozen tissues were cut into 10-μm sections on a cryostat, air-dried, fixed in cold acetone for 10 minutes, and then washed with PBS. After the sections were blocked with 3% bovine serum albumin, the primary Ab was added, and the slides were allowed to stand for 30 minutes at room temperature. After three washes in PBS, the sections were incubated for 30 minutes at room temperature with phycoerythrin (PE)- or FITC-conjugated anti-rabbit Abs (DakoCytomation, Carpinteria, CA). The plates then were examined under a fluorescence microscope (BH2-RFL-T3 or BX50; Olympus, Tokyo, Japan). 

Immunocytochemistry

Human CE plated on 35-mm poly-lysine-coated glass dishes was fixed for 30 minutes at room temperature with 4% paraformaldehyde in 0.1 M phosphate buffer. The fixed cells were washed three times with ice-cold PBS and then were treated with blocking solution (0.2% Triton X-100 and 4% skim milk in PBS) for 60 minutes at room temperature. After they were washed twice with PBS, the cells were incubated with the primary Ab solution (0.5% BSA and 0.02% Triton X-100 in PBS) for 60 minutes at 37°C. Then the cells were washed three times with PBS and incubated with the secondary Ab solution (1:500 dilution of Abs, 0.5% BSA, and 0.02% Triton X-100 in PBS) for 60 minutes at 37°C. In situ staining of human CE was carried out using anti-CESP-1 antiserum and goat anti-rabbit immunoglobulin conjugated with Cy3 (GE Healthcare) as the primary and secondary Abs, respectively. For colabeling of human CE, we used an anti-cytochrome c monoclonal Ab (6H2.B4; BD Pharmingen, San Diego, CA) to label mitochondria or an anti-glucose-regulated protein (GRP) 78 monoclonal Ab (Stressgen Biotechnologies Co., BC, Canada) to label endoplasmic reticulum (ER) and Oregon Green-conjugated anti-mouse immunoglobulin (Invitrogen) as the primary and secondary Abs, respectively. The cells were observed under a scanning confocal microscope (Micro Radiance; Bio-Rad Laboratories). Excitation and emission wavelengths for Cy3 or Oregon Green were 543 and 570 nm or 488 and 530 nm, respectively. ¹³  

Results

Significant Homology among CESP-1 Species, and Similarity of the N-Terminal Region between CESP-1 and Apoptosis-Inducing Factor

We cloned rabbit and mouse CESP-1 cDNAs by PCR based on homology information from the BLAST program. The rabbit and mouse CESP-1 cDNA sequences contained an open reading frame coding 242 and 283 amino acids, respectively. The amino acid sequences of human, rabbit, and mouse CESP-1 are shown in Figure 1 . Comparison among these orthologues revealed significant resemblance of the N-terminal sequence. The C-terminal region of mouse CESP-1 has a large insertion with three tandem repeats (EGADTSQ). Mouse CESP-1 is entirely consistent with mouse OSAP. 

BLAST analysis demonstrated significant similarity (E value of 4e⁻⁴) between the N-terminal region of CESP-1 and apoptosis-inducing factor (AIF), which possesses a mitochondrial localization signal (Fig. 2) . ^{14 15 16 17}  

Detection of CESP-1 Expression in Various Samples with Anti-human CESP-1 Ab

Figure 3shows Western blot analysis of human CESP-1 in human corneal epithelium (4 μg of protein per lane), human CE (4 μg of protein per lane), cultured human CE (4 μg of protein per lane), and in vitro translated human CESP-1, using the anti-CESP-1 Ab. Immunoreactive proteins were detected in all the samples. The molecular weight of human CESP-1 shown by 7% polyacrylamide gel electrophoresis was 40 kDa in human corneal epithelium, human CE, and cultured human CE, which was larger than the predicted molecular weight of 25.4 kDa, suggesting that the difference was due to CESP-1 being an acidic protein. An immunoreactive band was found at 43 kDa on examination of in vitro translated human CESP-1. This discrepancy of molecular weight may have resulted from processing, considering the boundary of the mitochondrial targeting sequence shown by the PSORT prediction algorithm. 

To determine the distribution of CESP-1 in human tissues, we used Western blot analysis with anti-human CESP-1 Ab. Immunoreactive bands of 38 kDa were detected in the brain, ovary, and testis (Fig. 4) . Because the brain tissue samples included the pituitary and thalamus, the result was consistent with our previous report ⁷ that human CESP-1 transcripts existed in the human pituitary, ovary, and testis. 

We also performed Western blot analysis of mouse CESP-1 in mouse CE and in vitro translated mouse CESP-1 using anti-N- or C-terminal OSAP Abs that we produced. An immunoreactive band was detected only in the extract of in vitro translated mouse CESP-1 by the anti-N-terminal OSAP Ab, but both in the extracts of mouse CE (46 kDa) and in vitro translated mouse CESP-1 (49 kDa) by the anti-C terminus OSAP Ab (Fig. 5) . These results indicated that the N-terminal sequence of CESP-1 undergoes processing. 

Anti-human CESP-1 Ab showed a cross-reaction with rabbit CESP-1 (0.4 μg of protein per lane) or mouse CESP-1 (1.1 μg of protein per lane), but the reaction with mouse CE was weak (Fig. 6) . The molecular weight of human CESP-1 was 36 kDa, according to 10% polyacrylamide gel electrophoresis, whereas the molecular weights of rabbit and mouse CESP-1 were 36 and 46 kDa, respectively. 

The localization of CESP-1 in the mouse cornea was also examined with anti-N- and C-terminal OSAP Abs. Strong staining of the CE and weak epithelial positivity were detected not with the anti-N-terminal OSAP Ab (Fig. 7A) , but with the anti-C-terminal OSAP Ab (Fig. 7B) . No positive staining was observed with control rabbit serum (Fig. 7C) . These findings were consistent with the Western blot analysis data for mouse CESP-1 (shown in Fig. 5 ). 

Anti-human CESP-1 Ab reacted specifically with CESP-1, and no bands of nonspecific immunoreactivity were detected by immunoblot analysis, except in the rabbit cornea. Therefore, the Ab was used to study the subcellular localization of CESP-1 in cultured human CE. Figure 8shows immunofluorescence images obtained by confocal microscopy. Staining with anti-CESP-1 Ab and anti-cytochrome c Ab was colocalized (Figs. 8A 8B 8C) , indicating that CESP-1 was expressed in the mitochondria. Human CESP-1 transiently expressed in HeLa cells was also localized to the mitochondria (not shown). In contrast, expression of CESP-1 was not seen in the ER (Figs. 8D 8E 8F) . These findings indicate that CESP-1 is mainly localized to the mitochondria in cultured human CE. 

Discussion

In the present study, we showed, by immunocytochemistry and/or Western blot analysis, that endogenous CESP-1 is strongly expressed in the CE and cultured human CE. These findings are consistent with our previous report that CESP-1 is a more abundant gene transcript in the human CE than in other tissues. ⁷ We also showed that CESP-1 was expressed in the corneal epithelium. Moreover, CESP-1 expression in cultured human CE was localized to the mitochondria. This finding was confirmed by the specific mitochondrial expression of human CESP-1 when it was transiently expressed in HeLa cells (data not shown). 

BLAST analysis indicated significant similarity between the N-terminal regions of CESP-1 and AIF, which possesses a mitochondrial localization signal. The AIF region is removed after import into the mitochondria, and the mature protein is released to the nucleus during apoptosis. ¹⁶ We found that tissue CESP-1 was smaller than in vitro–translated CESP-1 on Western blot analysis, and mouse CESP-1 was detected by the anti-C-terminal Ab, but not the anti-N terminus Ab, on both Western blot analysis and immunohistochemical analysis. Moreover, the PSORT prediction algorithm showed that the boundary of the mitochondrial targeting sequence in human CESP-1 was at the 39th amino acid. ¹⁸ Although we did not directly demonstrate the processing of CESP-1, these findings strongly suggest that the N-terminal region of CESP-1 possesses a mitochondrial targeting sequence and is processed after being imported into the mitochondria. 

The molecular weights of human, rabbit, and mouse CESP-1 were larger than the predicted molecular weight in all cases. Generally, the protein migration distance decreases as the acrylamide concentration of the gel increases. ¹⁹ The molecular weight of human CESP-1 was 40 kDa on 7% polyacrylamide gel electrophoresis and 36 kDa on analysis with 10% polyacrylamide gel. These results suggest that CESP-1 may restrict the binding of SDS to itself because it is an acidic protein. ²⁰  

The physiological functions of CESP-1 were not determined by our study, but its possible role can be estimated from the tissue distribution and subcellular localization. First, CESP-1 and AIF may move together or compete for binding to proteins in the mitochondria, because there was significant similarity between the N-terminal regions of these proteins. Second, the CE is metabolically active and requires various nutrients, so reactive oxygen species are produced. CESP-1 may reduce the damage due to oxidative stress in the mitochondria to compensate for the extremely low in vivo mitotic activity of the CE and thus may contribute to prolongation of its survival. Third, the ovary, testis, brain, and cornea, in which CESP-1 expression was detected, are all immune-privileged tissues. ²¹ In these tissues, the Fas-FasL system of cell death is involved in immunosuppression and expression of specific molecule(s) may be associated with their similarities in terms of immunity. Fourth, CESP-1 expression was not only found in human CE, but also in human corneal epithelium, rabbit CE, and mouse CE. This suggests that the physiological role of CESP-1 may not be related to suppression of mitotic activity. Fifth, pump function associated with Na⁺, K⁺-adenosine triphosphatase (ATPase) is important for maintaining normal corneal transparency. ^{22 23} Na⁺, K⁺-ATPase is a plasma membrane enzyme and exchange of Na⁺ and K⁺ across the membrane establishes a low internal Na⁺ and high internal K⁺ concentration. Because CESP-1 is localized in the mitochondria, it may indirectly contribute to pump function by producing ATP. 

In summary, we produced a polyclonal anti-human CESP-1 Ab and anti-mouse N- or C-terminal OSAP Abs, because CESP-1 is an abundant transcript in human CE compared with other tissues in the Bodymap human gene expression database. We then investigated the distribution and localization of CESP-1 in human tissues and animal corneas. Our findings revealed that CESP-1 is expressed in mouse, rabbit, and human CE, as well as in the human brain, testis, and ovary. CESP-1 is localized to the mitochondria in cultured human CE, and its N-terminal seems to be processed after mitochondrial import. These findings may provide some clues about the physiological role of this novel protein in the CE. 

 Supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

 Submitted for publication May 16, 2005; revised October 31, 2005; accepted February 16, 2006.

 Disclosure: R. Kinouchi, None; T. Kinouchi, None; T. Hamamoto, None; T. Saito, None; A. Tavares, None; T. Tsuru, None; S. Yamagami, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Alignment of the human, rabbit, and mouse CESP-1 sequences. Human, rabbit, and mouse CESP-1 consist of 240, 242, and 283 amino acids, respectively. Conserved amino acid residues among two or three of the CESP-1 species are indicated by shaded or reversed letters, respectively. Mouse CESP-1 has a large insertion in the C-terminal region that includes three tandem repeats (EGADTSQ).

Figure 2.

 

Amino acid sequences of CESP-1 and AIF. There is significant similarity (E value of 4e⁻⁴) between the N-terminal region of CESP-1 and AIF. Shaded letters: identical amino acids; gray lines: similar amino acids. Arrow: boundary of the mitochondrial targeting sequence according to the PSORT prediction algorithm.

Figure 3.

 

Detection of CESP-1 expression in human corneal epithelium, human CE, cultured human CE, and in vitro translation products by Western blot analysis with anti-CESP-1 antiserum (1:5000 dilution). The molecular weight of CESP-1 from human corneal tissue and in vitro translation is 40 kDa (filled arrowhead) and 43 kDa (open arrowhead), respectively, on 7% polyacrylamide gel electrophoresis.

Figure 4.

 

Tissue distribution of human CESP-1 analyzed by Western blot analysis. An immunoreactive band of 38 kDa is detected in the brain, ovary, and testis by anti-CESP-1 antiserum (1:5000 dilution; arrowhead). A positive band at 31 kDa detected in skeletal muscle sample is a nonspecific band by secondary antibody.

Figure 5.

 

Detection of CESP-1 expressions in mouse CE and in vitro translation products by Western blot analysis. (A) Western blot analysis with anti-N-terminal OSAP antibody (1:2000 dilution). The immunoreactive band is only detected in the in vitro translated mouse CESP-1 (open arrowhead). (B) Western blot analysis with anti-C-terminal OSAP antibody (1:2000 dilution). The immunoreactive bands are detected in the in vitro–translated mouse CESP-1 (open arrowhead) and mouse CE (filled arrowhead). The higher molecular weight bands in the in vitro translation lanes in both blots are nonspecific bands, which are found in the reticulocyte lysate loaded as a negative control (data not shown).

Figure 6.

 

Cross-reactivity of anti-human CESP-1 antiserum. Western blot analysis was performed on a 10% polyacrylamide gel with a 1:1000 dilution of anti-CESP-1 antiserum. Filled arrowheads: the molecular masses of human (36 kDa), rabbit (36 kDa), and mouse (46 kDa) CESP-1. Anti-human CESP-1 antiserum reacted strongly with human or rabbit CESP-1 and weakly with mouse CESP-1. Open arrowheads: nonspecific rabbit immunoglobulin bands.

Figure 7.

 

Immunohistochemical staining of mouse CESP-1 in the cornea. (A) No staining was observed with the anti-N-terminal OSAP antibody. (B) There was intense staining of the CE with the anti-C-terminal OSAP antibody, whereas the epithelium was stained weakly. (C) Control rabbit serum did not stain any layer of the cornea. Original magnification, ×200.

Figure 8.

 

Localization of CESP-1 in cultured human CE. (A, D) CESP-1 is immunostained by anti-CESP-1 antiserum in cultured human CE (red). (B) Mitochondria in the cultured cells were positive for anti-cytochrome c monoclonal antibody (green). (C) A combined image of (A) and (B) is shown. Double-positive staining (yellow) shows that expression of CESP-1 is localized to the mitochondria. (E) ER is immunostained by anti-GRP 78 antibody (green) in cultured human CE. (F) Combination of images in (D) and (E) reveals that positive staining did not overlap. Original magnification, ×600.