Enhanced detection method for corneal protein identification using shotgun proteomics
© Meade et al; licensee BioMed Central Ltd. 2009
Received: 20 April 2009
Accepted: 29 June 2009
Published: 29 June 2009
The cornea is a specialized transparent connective tissue responsible for the majority of light refraction and image focus for the retina. There are three main layers of the cornea: the epithelium that is exposed and acts as a protective barrier for the eye, the center stroma consisting of parallel collagen fibrils that refract light, and the endothelium that is responsible for hydration of the cornea from the aqueous humor. Normal cornea is an immunologically privileged tissue devoid of blood vessels, but injury can produce a loss of these conditions causing invasion of other processes that degrade the homeostatic properties resulting in a decrease in the amount of light refracted onto the retina. Determining a measure and drift of phenotypic cornea state from normal to an injured or diseased state requires knowledge of the existing protein signature within the tissue. In the study of corneal proteins, proteomics procedures have typically involved the pulverization of the entire cornea prior to analysis. Separation of the epithelium and endothelium from the core stroma and performing separate shotgun proteomics using liquid chromatography/mass spectrometry results in identification of many more proteins than previously employed methods using complete pulverized cornea.
Rabbit corneas were purchased, the epithelium and endothelium regions were removed, proteins processed and separately analyzed using liquid chromatography/mass spectrometry. Proteins identified from separate layers were compared against results from complete corneal samples. Protein digests were separated using a six hour liquid chromatographic gradient and ion-trap mass spectrometry used for detection of eluted peptide fractions. The SEQUEST database search results were filtered to allow only proteins with match probabilities of equal or better than 10-3 and peptides with a probability of 10-2 or less with at least two unique peptides isolated within the run along with default Xcorr values. These parameters resulted in the identification of over 350 proteins, including over 225 new proteins not previously detected in the cornea by mass spectrometry. In addition, corneal layer separation resulted in identification of nearly every protein that was identified in the complete cornea assay. The epithelium and endothelium each revealed many unique proteomes specific to each layer. In the endothelium, the protein olfactomedin-like 3 was identified for the first time in the cornea by this analysis. Olfactomedin-3 is a neuronal expressed protein also known as optimedin that stimulates formation of cell adherent and cell-cell tight junctions and its expression modulates cytoskeleton organization and cell migration. However, the function of this protein in rabbit corneal endothelium is currently unknown.
This manuscript presents a description of a more comprehensive proteomic profile for mammalian cornea compared to past methods. The use of simple dissection procedures of the tissue and the application of long chromatographic gradients, many more proteins can be identified.
The cornea is a transparent connective tissue that provides a majority of the refraction for the eye. In addition, the cornea also acts as a hydrated protective external barrier for the rest of the eye and provides clear optical components for image focusing on the retina. There are three main layers of the cornea: the epithelium, the center stroma, and the endothelium, with Descemet's membrane residing between the stroma and endothelium. The external epithelium the layer is 5–6 layers of stratified cells cumulatively ~50 μm thick. The epithelium provides the primary protective layer of the cornea and quickly regenerates new layers for maintenance of this external barrier function. The connective tissue derived stroma constitutes a majority of the mass and optical thickness of the cornea (~450 μm) and consists of parallel fibrils of collagen that provide optical clearness and the light refractivity of the cornea for image focus. The endothelium averages only one cell layer thick and is known to be physiologically responsible for assuring appropriate corneal hydration of the stroma by pumping fluid and nutrients in and out to the aqueous humor.
Previous proteomic investigations of the cornea have been performed on complete isolated cornea powder with such techniques as 2-D PAGE, 1-D PAGE and SCX fractionation of the extracted protein mixture prior to LC-MS/MS identification. [1, 2] A study of human corneal proteins by Karring et. al. showed the identification of 141 unique proteins using 2-D electrophoresis followed by spot LC-MS/MS analysis. While all techniques require one to reduce the complexity of the sample prior to LC-MS/MS analysis, low abundance proteins may not appear on the gel or may elute throughout the SCX fractionation, thus not allowing the component to be present for detection by the mass spectrometer.  Front end sample preparation methods with procedure simplicity and low amounts of sample handling steps are the key to providing the most successful approach for detecting the greatest protein diversity within a biological sample.
Shotgun proteomics is a technique in which a complex mixture, such as a tissue in our study, is lysed, digested and separated by liquid chromatography on-line prior to MS/MS analysis. [4–7] Shotgun proteomics has the ability to identify many proteins from a complex mixture in a short period of time. The disadvantage of this approach is that the instrumentation may not be able to analyze every peptide as it elutes in a highly complex protein mixture. This problem can be overcome with the use of a longer chromatographic gradient afforded by ultrahigh pressure liquid chromatography (UPLC), which can reduce the rate of peptide elution entering the mass spectrometer and the complexity of subsequent analysis with minimal peak broadening, thus enabling the mass spectrometer and software to analyze cleanly more peptides. In this investigation, detection of corneal proteins was substantially enhanced by separating the epithelial and endothelial layers from the collageneous stroma layer. The extended chromatographic technique provides a larger dataset of proteins that are present in these layers and for the corneal tissue. In this paper we present a comparison of the protein sets identified in the epithelium, endothelium, and a complete pulverized cornea.
Corneal protein identification by shotgun proteomics
The cornea is a transparent tissue and the identification of more proteins by direct measurements can aid in future studies of corneal function. Our described LC/MS/MS shotgun proteomic method has the ability to probe the corneal tissue protein content beyond other techniques. The separation of the regions of the tissue and using a six hour liquid chromatography gradient were essential in probing the less abundant proteins. The use of a six hour separation gradient enabled the identification of twice the number of proteins compared to a three hour chromatographic separation (data not shown). The epithelium had the most proteins identified with 279, while the endothelium had 121 and the pulverized sample had 66 proteins identified. The epithelial and endothelial samples contained nearly every protein identified in the pulverized sample. The epithelium and endothelium also had many different proteins that were found in only one layer. As expected, many basement proteins, such as collagen IV, fibulins and fibronectin, among others were identified in the endothelium while the epithelium had many proteins identified as actively dividing tissue region, including eukaryotic translation and elongation factors, as well as, many ribosomal proteins.
Identified corneal proteins were found to be from all regions of the cell, including the extracellular matrix. Many proteins identified were responsible for the structural integrity of the cell including keratins, collagens and other proteins that interact with these proteins to form the cytoskeleton of the cell. Proteins involved in synthesis and folding comprised the largest number of identified proteins. This group includes proteins involved in the transcription, translation and maintenance of functional proteins in the cell. These proteins were found in all areas, but were significantly represented in the rapidly replenishing cellular epithelium layer. The epithelial fraction also contained many ribosomal protein subunits, while none were found in the other samples. Major metabolic proteins indentified included α-enolase and a number ofidentified dehydrogenases. Other proteins responsible for metabolism of carbohydrates and lipids such as pyruvate kinase to apolipoproteins, respectively, were also represented.
Another advantage to this method was that very large proteins such as membrane-associated proteins often lost in preparation for gel analysis were detected by the mass spectrometer. The membrane proteins detected were responsible for the passage of molecules through the membrane and extracellular adhesion molecules. The epithelium and endothelium had very different proteins that were identified with this function. Many of the epithelial membrane proteins were desmosomal proteins. Of the 32 total proteins identified, only 3 were found in both the epithelium and endothelium.
Immune-related proteins were found fairly homogeneously across all layers with the main proteins identified being immunoglobulins. Many other blood and plasma proteins identified were representative of major proteins found in blood. Serum albumin was by far the most abundant protein which is to be expected. Other major blood proteins such as transferrin and hemoglobin subunits were identified as well. These proteins may be associated with the tissue or could be associated from the excision of the tissues during isolation. Lastly, redox maintenance/balance proteins detected in the epithelium and endothelium were found with peroxiredoxins, gluthionone-S-transferases and superoxide dismustases identified in both layers.
In order to confirm and localize a number of proteins immunohistochemistry was employed and detected with a confocal fluorescent microscope. Three proteins were chosen and tissues stained for osteoglycin, perlecan and junction plakoglobin. Osteoglycin was identified in both the epithelium and endothelium by mass spectrometry and primary antibody staining show the presence of this protein in both regions. Perlecan is a major endothelial protein and was shown to be located mainly in the endothelium. However, there was some staining in the epithelium which corresponds to the mass spectral data that also identified the protein, although with only two unique peptides compared to over 50 unique peptides found in the endothelium. Junctional plakoglobin was found mainly in the epithelium, consistent with the mass spectrometric data. There was also junctional plakoglobin detected by immunohistochemistry in the endothelium although this protein was not identified with the mass spectrometer.
Although corneal proteomics has been performed in the past, this paper presents detection of many new proteins by a direct method. This manuscript proves that the removal and separate identification of proteins of the epithelium and endothelium combined with longer (six hour), chromatographic gradients identified many new proteins. This LC/MS/MS shotgun proteomic method has the ability to probe the complex corneal protein mixture with minimal protein sample preparation compared to other techniques. A total of 341 unique proteins were identified in mature rabbit cornea using this method for the epithelial and endothelial layers. Of these, 227 proteins (67%) have not previously been identified in the cornea by mass spectrometry. The ability of a chromatographic system to separate peptides over a long gradient time was crucial for the identification of a greater number of proteins in any complex mixture. The data shows that the separation of the corneal layers results in many more proteins identified and retains identification of over 90% of the proteins (60 of 66) in pulverized whole cornea. As a larger database of rabbit proteins is assembled, we feel confident that more proteins will be identified. This paper shows an expanded rabbit corneal protein library that may be useful in further corneal proteomic or deferential phenotype physiology studies.
Corneal protein extraction
Mature rabbit corneas were purchased from Pelfreez Biological (Rogers, AK). Corneas were stored at -20°C until analysis. The connective tissue around the cornea was removed prior to protein extraction. The cornea was cut in half along the sagittal plane with surgical scissors and washed briefly in 50 mM Tris-HCl (pH8.0). The cornea was then blotted to remove residual liquid. The epithelium and endothelium were removed from the stroma by gentle scraping with a surgical razor blade. The epithelium and endothelium are not bound tightly to the collagenous stroma and are easily removed. The toughness of the stroma, due to collagen, aids in facilitating this removal with minimal cross contamination from the stromal layer. To prepare the pulverized corneal powder, the complete cornea was frozen in liquid nitrogen and pulverized by a Bio-pulverizer by Biospec (Bartlesville, OK). The epithelial, endothelial and pulverized tissues were placed in a lysis buffer containing: 50 mM Tris-HCl (pH = 8.0), 8 M urea, 100 mM NaCl, and 10 mM DTT. Each sample was sonicated for one minute in an ice bath. After sonication, samples were additionally reduced with 15 mM DTT and alkylated with 50 mM iodoacetimide. Samples were diluted 3:1 with 50 mM ammonium bicarbonate prior to overnight digestion with trypsin at 37°C. The digested samples were spun at 14.5 k RPM for 60 min on a benchtop centrifuge to remove particulate matter. The supernatant was collected and desalted with a Michrom (Auburn, CA) C18 peptide trap prior to LC-MS/MS analysis according to the manufactureer's instructions. After desalting the samples were concentrated in vacuo to 10 μl.
4 μl of the desalted peptide mixture diluted to 100 μl was injected onto the column. The peptide mixture was separated on a Symmetry BEH 130 C18 (1.7 μm, 75 μm × 250 mm)column using a Waters UPLC Nano-Acquity system (Milford, MA). The peptides were eluted at a flow of 0.250 μl/min using a linear gradient of acetonitrile with 0.1% formic acid from 2% to 50% over 320 minutes. Eluted peptides were ionized at 1.7 kV and the ions analyzed by a ThermoFisher LTQ-XL mass spectrometer (San Jose, CA). In the ion trap, the top 10 ions were selected for MS/MS by using a data dependent triple play scheme. Dynamic exclusion was set after two scans to facilitate identification of low abundance proteins.
The acquired data was searched in the Bioworks Browser with the SEQUEST search algorithm. Due to the limited size of the rabbit database, proteins from human, rat and rabbit from the NCBI non-redundant database were selected as the database. In order for a protein to be considered, the protein must have a probability of at least 0.001 in SEQUEST with at least two unique peptides with a peptide probability of 0.01. In addition, the peptide must have Xcorr scores above default values set by Bioworks (+1 = 1.5, +2 = 2.0, +3 = 2.5, +4 = 3.0). Proteins that met all of these requirements were determined to be selected as sufficient to be included as identified using MS. After identification, a literature search was performed to determine the function and location of the protein in the cell.
Previously fixed and mounted corneal slices were placed onto heating blocks at 55°C for ten minutes to melt the paraffin. Slides were deparaffinized in 2 washes of xylenes for 10 minutes. Slides were rinsed twice for 2 minutes in 100% alcohols (18:1:1, 100% ethanol:100% methanol:100% isopropanol) followed by two 2 minute rinses in a 95% solution of the 100% alcohols. Slides were placed in an 80% solution of the 100% alcohols for 2 minutes followed by 5–6 washes with deionized water for 2 minutes each. Slides were incubated in blocking solution that consisted of 10% normal bovine (for osteoglycin staining) or donkey (for gamma catenin and perlecan stainings) serum (Jackson Immunoresearch Laboratories, West Grove, PA, USA) in Tris Buffered Saline Buffer (TBS; 100 mM Tris pH 7.4, 138 mM NaCl, 27 mM KCl) overnight at +4°C. Antibody incubations were performed in TBSN supplemented with 10% of respective normal serum. Dilutions (1:100)were used for all primary and secondary antibodies. Incubation times were 2 hours with primary antibody and 1 hour with secondary antibody at room temperature on an orbital shaker at slow speed. Samples were washed 5× for 5 min. each in TBS containing 0.1% NonIdet P40 (TBSN) after antibody incubations. Primary antibodies sc-7900, sc-25848 or sc-47277 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) were used to stain against gamma catenin, perlecan or osteoglycin, respectively. Donkey anti-rabbit, FITC-labeled antibody or bovine anti-goat, TRITC-labeled antibody (Jackson Immunoresearch Laboratories, West Grove, PA, USA) were used to stain against gamma catenin, perlecan or osteoglycin, respectively. After the final wash, slides were quickly rinsed twice in TBS and dried at room temperature. A drop of ProLong Gold antifade reagent with DAPI (Invitrogen, Carlsbad, CA, USA) was applied onto the slides. Slides were covered with cover slips. Images were taken using BD Pathway 435 confocal bioimager (BD BioSciences, Rockville, MD, USA).
- Karring H, Thogersen IB, Klintworth GK, Moller-Pedersen T, Enghild JJ: The human cornea proteome: bioinformatic analyses indicate import of plasma proteins into the cornea. Mol Vis 2006, 12: 451–460.PubMedGoogle Scholar
- Karring H, Thogersen IB, Klintworth GK, Moller-Pedersen T, Enghild JJ: A dataset of human cornea proteins identified by Peptide mass fingerprinting and tandem mass spectrometry. Mol Cell Proteomics 2005, 4: 1406–1408. 10.1074/mcp.D500003-MCP200PubMedView ArticleGoogle Scholar
- Nagele E, Vollmer M, Horth P: Improved 2D nano-LC/MS for proteomics applications: a comparative analysis using yeast proteome. J Biomol Tech 2004, 15: 134–143.PubMed CentralPubMedGoogle Scholar
- McDonald WH, Yates JR III: Shotgun proteomics and biomarker discovery. Dis Markers 2002, 18: 99–105.PubMed CentralPubMedView ArticleGoogle Scholar
- Nagele E, Vollmer M, Horth P: Improved 2D nano-LC/MS for proteomics applications: a comparative analysis using yeast proteome. J Biomol Tech 2004, 15: 134–143.PubMed CentralPubMedGoogle Scholar
- Wu CC, MacCoss MJ: Shotgun proteomics: tools for the analysis of complex biological systems. Curr Opin Mol Ther 2002, 4: 242–250.PubMedGoogle Scholar
- McDonald WH, Yates JR III: Shotgun proteomics: integrating technologies to answer biological questions. Curr Opin Mol Ther 2003, 5: 302–309.PubMedGoogle Scholar
- Shen Y, Tolic N, Masselon C, Pasa-Tolic L, Camp DG, Hixson KK, Zhao R, Anderson GA, Smith RD: Ultrasensitive proteomics using high-efficiency on-line micro-SPE-nanoLC-nanoESI MS and MS/MS. Anal Chem 2004, 76: 144–154. 10.1021/ac030096qPubMedView ArticleGoogle Scholar
- Kato T, Nakayasu K, Kanai A, Nishiyama T, Imamura Y, Hayashi T: Distribution and isoform characterization of type XII collagen in bovine cornea. Ophthalmic Res 2000, 32: 215–221. 10.1159/000055616PubMedView ArticleGoogle Scholar
- Jimenez-Asensio J, Gonzalez P, Zigler JS Jr, Garland DL: Glyceraldehyde 3-phosphate dehydrogenase is an enzyme-crystallin in diurnal geckos of the genus Phelsuma. Biochem Biophys Res Commun 1995, 209: 796–802. 10.1006/bbrc.1995.1570PubMedView ArticleGoogle Scholar
- Bateman RL, Rauh D, Tavshanjian B, Shokat KM: Human carbonyl reductase 1 is an S-nitrosoglutathione reductase. J Biol Chem 2008, 283: 35756–35762. 10.1074/jbc.M807125200PubMed CentralPubMedView ArticleGoogle Scholar
- Urade R, Kito M: Inhibition by acidic phospholipids of protein degradation by ER-60 protease, a novel cysteine protease, of endoplasmic reticulum. FEBS Lett 1992, 312: 83–86. 10.1016/0014-5793(92)81415-IPubMedView ArticleGoogle Scholar
- Giudice GJ, Cohen SM, Patel NH, Steinberg MS: Immunological comparison of desmosomal components from several bovine tissues. J Cell Biochem 1984, 26: 35–45. 10.1002/jcb.240260104PubMedView ArticleGoogle Scholar
- Blonder J, Terunuma A, Conrads TP, Chan KC, Yee C, Lucas DA, Schaefer CF, Yu LR, Issaq HJ, Veenstra TD, et al.: A proteomic characterization of the plasma membrane of human epidermis by high-throughput mass spectrometry. J Invest Dermatol 2004, 123: 691–699. 10.1111/j.0022-202X.2004.23421.xPubMedView ArticleGoogle Scholar
- Yamaguchi N, Sato N, Ko JS, Ninomiya Y: Cloning of alpha 1(IV) and alpha 2(IV) collagen cDNAs from rabbit corneal endothelial cell RNA. Invest Ophthalmol Vis Sci 1991, 32: 2924–2930.PubMedGoogle Scholar
- Kobayashi N, Kostka G, Garbe JH, Keene DR, Bachinger HP, Hanisch FG, Markova D, Tsuda T, Timpl R, Chu ML, et al.: A comparative analysis of the fibulin protein family. Biochemical characterization, binding interactions, and tissue localization. J Biol Chem 2007, 282: 11805–11816. 10.1074/jbc.M611029200PubMedView ArticleGoogle Scholar
- Hopfer U, Fukai N, Hopfer H, Wolf G, Joyce N, Li E, Olsen BR: Targeted disruption of Col8a1 and Col8a2 genes in mice leads to anterior segment abnormalities in the eye. FASEB J 2005, 19: 1232–1244. 10.1096/fj.04-3019comPubMedView ArticleGoogle Scholar
- Tang CH, Lee JW, Galvez MG, Robillard L, Mole SE, Chapman HA: Murine cathepsin F deficiency causes neuronal lipofuscinosis and late-onset neurological disease. Mol Cell Biol 2006, 26: 2309–2316. 10.1128/MCB.26.6.2309-2316.2006PubMed CentralPubMedView ArticleGoogle Scholar
- Kruth HS: The fate of lipoprotein cholesterol entering the arterial wall. Curr Opin Lipidol 1997, 8: 246–252. 10.1097/00041433-199710000-00002PubMedView ArticleGoogle Scholar
- Saghizadeh M, Kramerov AA, Tajbakhsh J, Aoki AM, Wang C, Chai NN, Ljubimova JY, Sasaki T, Sosne G, Carlson MR, et al.: Proteinase and growth factor alterations revealed by gene microarray analysis of human diabetic corneas. Invest Ophthalmol Vis Sci 2005, 46: 3604–3615. 10.1167/iovs.04-1507PubMed CentralPubMedView ArticleGoogle Scholar
- Lee JE, Kim Y: A tissue-specific variant of the human lysyl oxidase-like protein 3 (LOXL3) functions as an amine oxidase with substrate specificity. J Biol Chem 2006, 281: 37282–37290. 10.1074/jbc.M600977200PubMedView ArticleGoogle Scholar
- Huang Y, Dai J, Tang R, Zhao W, Zhou Z, Wang W, Ying K, Xie Y, Mao Y: Cloning and characterization of a human lysyl oxidase-like 3 gene (hLOXL3). Matrix Biol 2001, 20: 153–157. 10.1016/S0945-053X(01)00124-XPubMedView ArticleGoogle Scholar
- Evans MD, Steele JG: Multiple attachment mechanisms of corneal epithelial cells to a polymer – cells can attach in the absence of exogenous adhesion proteins through a mechanism that requires microtubules. Exp Cell Res 1997, 233: 88–98. 10.1006/excr.1997.3523PubMedView ArticleGoogle Scholar
- Underwood PA, Bennett FA: The effect of extracellular matrix molecules on the in vitro behavior of bovine endothelial cells. Exp Cell Res 1993, 205: 311–319. 10.1006/excr.1993.1091PubMedView ArticleGoogle Scholar
- Torrado M, Trivedi R, Zinovieva R, Karavanova I, Tomarev SI: Optimedin: a novel olfactomedin-related protein that interacts with myocilin. Hum Mol Genet 2002, 11: 1291–1301. 10.1093/hmg/11.11.1291PubMedView ArticleGoogle Scholar
- Rabinowitz YS, Dong L, Wistow G: Gene Expression Profile Studies of Human Keratoconus Cornea for NEIBank: A Novel Cornea-Expressed Gene and the Absence of Transcripts for Aquaporin 5. Investigative Ophthamology 2005,46(4):1239–1246. 10.1167/iovs.04-1148Google Scholar
- Messent AJ, Blissett MJ, Smith GL, North AJ, Magee A, Foreman D, Garrod DR, Boulton M: Expression of a single pair of desmosomal glycoproteins renders the corneal epithelium unique amongst stratified epithelia. Invest Ophthalmol Vis Sci 2000, 41: 8–15.PubMedGoogle Scholar
- Sundarraj N, Fite D, Belak R, Sundarraj S, Rada J, Okamoto S, Hassell J: Proteoglycan distribution during healing of corneal stromal wounds in chick. Exp Eye Res 1998, 67: 433–442. 10.1006/exer.1998.0540PubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.