Detection and Quantification of Sulfated Disaccharides ... - CiteSeerX

Detection and Quantification of Sulfated Disaccharides from Keratan Sulfate and Chondroitin/Dermatan Sulfate during Chick Corneal Development by ESI-MS/MS Yuntao Zhang,1 Abigail H. Conrad,1 Elena S. Tasheva,1 Ke An,1 Lolita M. Corpuz,1 Yutaka Kariya,2 Kiyoshi Suzuki,2 and Gary W. Conrad1 PURPOSE. To identify and quantify changes in keratan sulfate (KS) and chondroitin/dermatan sulfate (CS/DS) sulfated disaccharides in the developing chick cornea using electrospray ionization tandem mass spectrometry (ESI-MS/MS). METHODS. Cryostat sections of fresh nonfixed corneas were obtained from White Leghorn embryonic day (E)8 to E20 chicks, and from 4- and 70-week-old chickens. Tissue sections on glass slides were incubated with selected glycosidase enzymes. Digest solutions were analyzed directly by ESI-MS/MS. RESULTS. The concentration of KS monosulfated disaccharide (MSD) Gal-␤-1,4-GlcNAc(6S) in E8 cornea equaled that at E20, declined to its lowest level by E10, increased to a second peak by E14, decreased to a second low by E18, peaked again by E20, and remained high in adult corneas. A similar concentration profile was observed for KS disulfated disaccharide (DSD) Gal(6S)-␤-1,4-GlcNAc(6S), and thus also for total sulfated KS disaccharides. The molar percent of DSD was higher than that of MSD from E8 to E18, equivalent at E20, and less than that of MSD in adult corneas. In contrast, total concentration of CS/DS ⌬di-4S plus ⌬di-6S decreases as development progresses and is lowest in adult corneas. Concentration and molar percent of ⌬di-6S is highest at E8, then decreases through development as the concentration and molar percent of ⌬di-4S increases from E8 and exceeds that of ⌬di-6S after E14. CONCLUSIONS. New, rapid, direct chemical analysis of extracellular matrix components obtained from sections from embryonic and adult chick corneas reveals heretofore undetected changes in sulfation characteristics of KS and CS/DS disaccharides during corneal development. (Invest Ophthalmol Vis Sci. 2005;46:1604 –1614) DOI:10.1167/iovs.04-1453

G

lycosaminoglycans (GAGs) are negatively charged, highmolecular-weight polysaccharides classified on the basis of their structures into several groups: hyaluronan (HA), CS/ DS, KS, heparan sulfate (HS), and heparin (Hep). With the exception of KS, GAGs are composed of alternating residues of uronic acid and N-acetylhexosamine. Sulfate groups are at-

From the 1Division of Biology, Kansas State University, Manhattan, Kansas; and the 2Central Research Laboratories, Seikagaku Corporation, Tokyo, Japan. Supported by National Institutes of Health Grants EY000952 (GWC) and EY13395 (GWC, EST). Submitted for publication December 10, 2004; revised January 19, 2005; accepted January 24, 2005. Disclosure: Y. Zhang, Seikagaku Corp. (F); A.H. Conrad, None; E.S. Tasheva, None; K. An, None; L.M. Corpuz, None; Y. Kariya, Seikagaku Corp. (E); K. Suzuki, Seikagaku Corp. (E); G.W. Conrad, 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. Corresponding author: Yuntao Zhang, Division of Biology, Kansas State University, Ackert Hall, Manhattan, KS 66506-4901; [email protected].

1604

tached to a limited number of hydroxyl or amino groups on GAGs and contribute greatly to their polyanionic properties. In cornea, KS, and CS/DS are the two major GAGs.1 KS was first isolated from bovine cornea.2 KS chains have been extracted from many tissues, and several KS-containing proteoglycans have been identified. KS chains have been classified according to their linkage to protein: KS-I for N-asparagine-linked chains from cornea and KS-II for O-serine-linked chains from skeletal tissues such as cartilage. A third type of KS, O-linked from mannose to serine or threonine, has been isolated from brain tissue.3,4 In contrast to other GAGs, KS does not contain uronic acids, and its repeating disaccharide unit is composed of alternating residues of D-galactose (Gal) and N-acetyl-D-glucosamine (GlcNAc) linked ␤-1,4 and ␤-1,3, respectively. During biosynthesis of GAG chains, a variety of factors influence sulfation, leading to appearance of both nonsulfated and variously sulfated domains along individual chains. In some domains of KS, the hydroxyl groups at the C-6 positions of both Gal and GlcNAc residues may be sulfated.5 Chains of CS/DS consist of N-acetylgalactosamine (GalNAc) residues alternating in glycosidic linkages with glucuronic acid or iduronic acid residues. The GalNAc residues are predominantly sulfated in the C-4 or C-6-hydroxyl position, interspersed with a few nonsulfated residues. CS is heterogeneous, containing variable proportions of nonsulfated chondroitin, chondroitin 4-sulfate, and/or chondroitin 6-sulfate chains in different species and tissues. CS type A (CSA) is sulfated primarily (90%) at the C-4-position of GalNAc, whereas CS type C (CSC) is sulfated primarily (90%) at the C-6 position. Within a given disaccharide unit, usually a single GalNAc residue carries only one sulfate, either on C-4 or C-6, but residues disulfated at both C-4 and C-6 of GalNAc are found occasionally. In addition, glucuronic acid may be sulfated at the C-2 and C-3 positions.6 Normally, corneal CS/DS represents approximately 30% of total GAGs in the stroma.1 Proteoglycans (PGs) and GAGs are involved in morphogenesis and differentiation of several tissues. The developing cornea represents one of the most thoroughly studied tissues in this regard.7–13 Proteoglycans containing KS and CS/DS chains play important roles in determining collagen spacing and transparency of the corneal stroma. In the developing embryonic chick cornea, neural crest-derived cells first form the endothelial layer on the back of the primary stroma by E4.14 Invasion of the stroma by more neural crest cells begins on E5.14 Over the next several days, these invading stromal fibroblastic cells assume the stellate morphology of keratocytes and secrete the connective tissue that characterizes the secondary or adult stroma. Beginning at about E14, under the influence of thyroxine, the stroma starts to dehydrate and become thinner and more transparent, with full transparency being achieved just before hatching on E20 to E21.15,16 It is thought that an abundance of keratan sulfate proteoglycan (KSPG) in the stroma is essential for corneal transparency, which has led to the use of KS as a marker for stromal differentiation. Although expressions of mRNAs for KSPG core proteins have been shown to reach peaks between E9 and E15,17 those core Investigative Ophthalmology & Visual Science, May 2005, Vol. 46, No. 5 Copyright © Association for Research in Vision and Ophthalmology

IOVS, May 2005, Vol. 46, No. 5 proteins, if translated, may or may not be posttranslationally modified with KS chains. However, those GAG chains appear to regulate many biological properties of the extracellular matrix. Studies of GAG biosynthesis by developing embryonic chick cornea or by its constituent cell populations using labeled precursors have been reported.10 Monoclonal antibodies to KS12 and a polyclonal antiserum to a KSPG core protein have also been used to characterize patterns of KS and KSPG accumulation during differentiation of neural crest cells in the stroma of embryonic chick corneas.12 Combining use of labeled precursors and HPLC techniques, KS and CS/DS sulfated disaccharides from GAGs labeled during explant culture have been examined during corneal development.13 However, the profiles of KS sulfated disaccharides from postembryonic corneas have never been reported. With the development of ionization methods, electrospray ionization (ESI)-MS/MS techniques have become more attractive and prominent for the analysis of oligosaccharides,18 because they provide high mass accuracy, structural information, and ability to quantify the fragments,19 –23 and because they do not require labeling of the GAGs in explant culture before analysis. It has also been demonstrated that MS/MS techniques can be used for compositional analysis of purified chondroitin sulfate (CS) and heparan sulfate (HS) polymers that have been hydrolyzed by enzymes.24 –27 More recently, the KS disaccharides Gal-␤-1,4-GlcNAc(6S) (monosulfated disaccharide; MSD) and Gal(6S)-␤-1,4-GlcNAc(6S) (disulfated disaccharide; DSD), released from bovine cornea, bovine nasal cartilage, mouse brain, and rat brain by enzymatic digestion and isolated by liquid chromatography, have been analyzed by turbo ion spray tandem mass spectrometry (LC-MS/MS).28 The CS/DS disaccharides ⌬di-4S and ⌬di-6S, from tumor tissue sections, have also been analyzed by LC-MS/MS.29 However, heretofore, investigation of sulfated disaccharides from KS and CS/DS during chick corneal development by ESI-MS/MS has not been reported. In this study, we use ESI-MS/MS techniques to directly identify and quantify KS and CS/DS sulfated disaccharides liberated from single frozen sections of developing chick cornea by enzymatic digestion, without purification of digestion products or chromatographic separation preceding mass spectral analysis. By such an approach, the proportions of KS and CS/DS sulfated disaccharides during chick corneal development were determined.

MATERIALS

AND

METHODS

Reagents KS Gal-␤-1,4-GlcNAc(6S) and Gal(6S)-␤-1,4-GlcNAc(6S) were kindly provided by Seikagaku Corporation (Tokyo, Japan). Disaccharides of CS/DS—⌬di-2S, which was also used as the internal standard for quantification of KS sulfated disaccharides, ⌬di-4S, ⌬di-6S, and chondroitinase ABC (from Proteus vulgaris, protease free, EC 4.2.2.4)—were purchased from Sigma-Aldrich (St. Louis, MO). ⌬UA-␤-1,4-GlcNS, which was used as the internal standard for quantification of CS/DS sulfated disaccharides, was purchased from EMD Biosciences, Inc. (San Diego, CA). Keratanase II (from Bacillus sp.) was purchased from Seikagaku America (Falmouth, MA). Ammonium acetate and ammonium sulfate were purchased from Sigma-Aldrich. Solvents used were of HPLC grade and were purchased from Fisher Scientific (Pittsburgh, PA). Centrifugal filter units (Ultra free-MC, 5000 NMWL; Millipore, Bedford, MA) were purchased from Fisher Scientific.

Preparation of Standard Solutions All disaccharide standards were diluted in a series of 10 nmol/␮L, 1 nmol/␮L, 100 pmol/␮L, and 10 pmol/␮L in water, and aliquots were stored at ⫺20°C. For qualitative analysis, 10 ␮L of the 1 nmol/␮L solution of each standard was diluted by adding 70 ␮L of MeOH, 5 ␮L

Corneal KS and CS/DS Sulfated Disaccharides

1605

of 5 mM ammonium acetate buffer (pH 7.5), 5 ␮L of 2 mM (NH4)2SO4, and 10 ␮L of water to make the solution 7:3 MeOH/H2O, containing 100 pmol/␮L of disaccharide standard sample. The addition of (NH4)2SO4 was helpful in suppressing sodiated adducts.24

Preparation of Tissue Sections All animals were used in accordance with the ARVO Statement for the Use of Animals for Ophthalmic and Vision Research. Corneas were dissected from White Leghorn embryonic chicks of ages ranging from embryonic day (E)8 to E20, as well as from 4- and 70-week-old chickens. Fresh corneas were washed three times in saline solution and then embedded in optimal cutting temperature polymer (Sakura Finetek USA. Inc., Torrance, CA) in liquid nitrogen. Sections of 10-␮m thickness were cut on a cryostat (Hacker-Bright, Huntington, UK) at ⫺24°C, collected on precleaned glass microscope slides (Fisherbrand Superfrost/Plus; Fisher Scientific), fixed with 100% methanol for 20 minutes at room temperature, and air dried at room temperature.30 Photographs of tissue sections were taken with a digital camera mounted on an inverted microscope (Nikon, Tokyo, Japan). The areas of tissue sections were calculated with NIH Image software (Scion Image 1.60; Scion, Frederick, MD), and volumes were calculated by multiplying tissue section areas by section thickness.

Enzymatic Digestion of KS and CS/DS Tissue sections (n ⫽ 5), each from a cornea of a separate chick embryo incubated to the same age, were placed on glass slides, encircled with a hydrophobic ring (PAP pen), and then covered with a droplet of digestion solution, as follows: for analysis of KS sulfated disaccharides, the digestion solution consisted of 25 ␮L of 0.1 M ammonium acetate buffer (pH 6.0), 0.1 mU/␮L keratanase II, and 50 pmol/␮L ⌬di-2S (as an internal standard) applied to each section. For analysis of CS/DS sulfated disaccharides, the digestion solution consisted of 25 ␮L containing 50 mM ammonium acetate buffer (pH 8.0), chondroitinase ABC (1 mU/␮L), and ⌬UA-␤-1,4-GlcNS (as an internal standard, 50 pmol/␮L) applied on each section. The slides were then incubated in a moist chamber at 37°C for 24 hours. The solution was carefully collected from each tissue section, and its enzyme activity was terminated by heating for 10 minutes at 100°C. Digest solution (10 ␮L) was diluted by adding 70 ␮L MeOH, 5 ␮L of 5 mM ammonium acetate buffer (pH 7.5), 5 ␮L of 2 mM (NH4)2SO4, and 10 ␮L of water,23 and then applied to a centrifugal filter unit (5000 NMWL, Ultra free-MC; Millipore) and centrifuged at 3800 rcf for 30 minutes at 4°C. The filtrate was used directly for MS analysis.

Mass Spectrometry Mass spectra were obtained using an electrospray ionization source on a quadrupole ion trap instrument (Esquire 3000; Bruker Daltonics, Billerica, MA). Mass spectra were obtained in negative ion mode. The spray voltage was 3.5 kV. Dry gas (nitrogen) flowed at 5.0 L/min, and the drying temperature was 180°C. The mass range scanned was m/z 50 to 600. These same instrument conditions were used for all standards, mock mixtures, and digestion samples. Data acquisition software was used to record the results (Data Analysis 3.0; Bruker Daltonics).

Quantitative Analysis of KS Sulfated Disaccharides A single-point normalization factor method, as described elsewhere,26,31,32 was used to determine the amounts of Gal-␤-1,4-GlcNAc (6S) and Gal (6S)-␤-1,4-GlcNAc (6S). Normalization factors for monosulfated disaccharide (RMSD) and disulfated disaccharide (RDSD) were determined with equation 1: R MSD or DSD ⫽

ITC ⫻ CIS , IIS ⫻ CTC

(1)

1606

Zhang et al.

IOVS, May 2005, Vol. 46, No. 5

TABLE 1. Contributions of Fragment Ions at m/z 281.8 and 299.9 in Total Ion Current TIC %

⌬di-4S ⌬di-6S ⌬di-4S:⌬di-6S ⫽ 1:1

m/z

1

2

3

4

5

6

Avg. % ⴞ SD (n ⴝ 6)

281.8 299.9 281.8 299.9 281.8 299.9

1.408 72.078 87.930 2.555 28.644 50.557

1.227 72.130 87.940 2.817 27.244 51.842

1.376 72.339 89.677 2.693 27.486 52.004

1.297 71.974 88.407 2.389 27.581 52.086

1.442 72.026 90.162 2.437 27.965 51.735

1.650 72.108 88.823 2.579 28.316 50.969

1.400 ⫾ 0.145 72.109 ⫾ 0.126 88.823 ⫾ 0.924 2.578 ⫾ 0.159 27.873 ⫾ 0.535 51.532 ⫾ 0.622

where ITC is the intensity of the target ion, IIS is the intensity of the internal standard, CIS is the molar concentration of the internal standard, and CTC is the molar concentration of the target compound. ⌬di-2S was chosen as an internal standard for the quantification of MSD and DSD, due to its similar molecular weight, chemical structure, and ionization efficiency. Also, it is not normally detected in cornea tissues. One sample containing equal molar concentrations (10 pmol/␮L) of both target compounds MSD and DSD, together with a constant molar concentration (5 pmol/␮L) of the internal standard, was run during each experiment, to determine the respective R values to be used for each unknown KS sulfated disaccharide. The RMSD and RDSD were 0.2 ⫾ 0.1 and 1.4 ⫾ 0.2, respectively. For each unknown sample analyzed, the molar concentration of each target compound (CTC) was then calculated with equation 2, using the mass spectrometric data collected and the appropriate R value:

normalization factors, the percentage contribution of each diagnostic ion was determined from the MS2 spectrum of a 1:1 mixture of ⌬di-4S and ⌬di-6S, as described in Table 1, and these observed percentages for C299.9 and C281.8 were substituted in equations 3 and 4, respectively. The equations were then solved for A (70.450) and B (30.270). Normalization factor R4S for m/z 299.9 is thus calculated to be 0.710 in the 1:1 mixture (50/70.450), and R6S for m/z 281.8 is calculated to be 1.652 in the 1:1 mixture (50/30.270). After normalizing, the observed values for A and B are then converted to the actual relative molar percentages of ⌬di-4S and ⌬di-6S in an unknown mixture by dividing each normalized observed value by the sum of the two normalized observed values.24 Experimental verification of the use of these equations for determining the molar percentages of ⌬di-4S and ⌬di-6S in a mixture are presented in the Results (see Fig. 5, Table 2).

RESULTS ITC ⫻ CIS C TC ⫽ . IIS ⫻ RMSD or DSD

(2)

Analysis of KS Standard Disaccharides

72.109A ⫹ 2.578B ⫽ C 299.9

(3)

The mass spectra of KS MSD and DSD are shown in Figure 1. MSD yielded one singly charged ion [M ⫺ H]⫺ at m/z 462.0 (Fig. 1A). DSD yielded both a singly charged ion [M ⫺ H]⫺ at m/z 541.9 and a doubly charged ion [M ⫺ 2H]2⫺ at m/z 270.4 (Fig. 1B). Established Domon and Costello nomenclature is used to describe fragment ions.34 The MS2 spectra of these KS sulfated disaccharides are shown in Figure 2. For MSD (Fig. 1A), using m/z 462.0 as a precursor ion, MS2 fragment ions observed were [M ⫺ H ⫺ H2O]⫺ at m/z 444.0, 0,2A2 at m/z 360.9, [0,2A2 ⫺ H2O]⫺ at m/z 342.9, [0,2A2 ⫺ HSO4)⫺ at m/z 264.0, [Z1 ⫺ H2O ⫺ SO3]⫺ at m/z 183.8, and [OCHCH2OSO3]⫺ at m/z 138.8 (Fig. 2A). For DSD, if the doubly charged state ion m/z 270.4 (Fig. 1B) was used as a precursor ion, MS2 fragment ions included [M ⫺ 2H ⫺ H2O]2⫺ at m/z 261.4, [Y1 ⫺ SO3]⫺ at 219.8, [OCHCH2OSO3]⫺ at m/z 138.8, and [HSO4]⫺ at m/z 96.9 (Fig. 2B). With m/z 541.9 as the precursor ion, MS2 fragment ions were observed at m/z 461.9, corresponding to [M ⫺ H ⫺ SO3]⫺, and at m/z 299.9, corresponding to Y1, representing the occurrence of glycosidic cleavage, with no ring cleavage found in this case (Fig. 2C). Therefore, combining MS1 and MS2 spectra provides definitive fingerprints to identify KS MSD and DSD. Indeed, simply from MS1 spectra, the molecular ions [M ⫺ H]⫺ at m/z 462.0 and [M ⫺ 2H]2⫺ at m/z 270.4 can be used to qualitatively identity and to quantify MSD and DSD, respectively.

1.400A ⫹ 88.823B ⫽ C 281.8,

(4)

Analysis of CS/DS Sulfated Disaccharides

Quantitative Analysis of CS/DS Sulfated Disaccharides The total molar concentration of ⌬di-4S plus ⌬di-6S was determined by quantitative analysis of the intensity of the characteristic molecular ion at m/z 458.1, using an internal standard, ⌬UA-␤-1,4-GlcNS and a singlepoint normalization factor, similar to the quantitative analysis of KS sulfated disaccharides just described. The R value determined for the total concentration of ⌬di-4S and ⌬di-6S was 1.2 ⫾ 0.3. The relative molar percentages of ⌬di-4S and ⌬di-6S isomers in a mixture were calculated from the observed relative intensities of the diagnostic ions for each isomer in the MS2 spectrum of that mixture, using a system of two equations that correct for the presence of additional ions in the MS2 spectrum, as described elsewhere.24,25,33 The relative intensities of each of the diagnostic ions for each isomer, m/z 299.9 for ⌬di-4S and m/z 281.8 for ⌬di-6S, in the MS2 spectra of each isomer disaccharide when analyzed alone (see Figs. 4B, 4C) were measured six times. The percentage ion contribution of each diagnostic ion to the total ion current (TIC) was then determined by exporting the mass list of the spectrum of interest into a computer program (SpecManager; ACD Labs, Toronto, Ontario, Canada; Table 1), and an average was computed. These average percent contributions were then used in the following two equations

where A and B are the apparent percentages of ⌬di-4S and ⌬di-6S in any unknown mixture, and C299.9 and C281.8 are the percentage contributions of the two diagnostic ions in the mixture. However, because many factors influence peak intensities in a mixture, including differences in the respective ionization efficiencies of the isomers,22 further normalization factors, R4S and R6S, must be used. To determine these

The three CS/DS sulfated disaccharide isomers, ⌬di-2S, ⌬di-4S, and ⌬di-6S, yielded very simple MS1 spectra. The most abundant ion in each case (Figs. 3A–C) corresponds to molecular ion [M ⫺ H]⫺ for each disaccharide with one sulfate, all of which generate the same molecular ion [M ⫺ H]⫺ at m/z 458.1. Figure 3D shows the MS1 spectra of ⌬UA-␤-1,4-GlcNS, the CS/DS internal standard, which generates the molecular



1607

FIGURE 1. MS1 spectra of KS sulfated disaccharides, in the negative ionization mode. (A) Gal-␤-1,4GlcNAc(6S), or MSD, yielded a singly charged ion [M ⫺ H[⫺ at m/z 462.0. (B) Gal(6S)-␤-1,4-GlcNAc(6S), or DSD, produced both a singly charged ion [M ⫺ H]⫺ at m/z 541.9 and a doubly charged ion [M ⫺ 2H]2⫺ at m/z 270.4.

ion [M ⫺ H]⫺ at m/z 415.8. MS2 under identical conditions was therefore used to differentiate these three CS/DS sulfated disaccharide isomers (Figs. 4A–C). Figure 4A shows the MS2 spectrum for ⌬di-2S, which was used as the KS internal standard (see the Materials and Methods section). Fragment ions included [M ⫺ H ⫺ H2O ⫺ SO3]⫺ at m/z 360.0, 2,5A2 at m/z 341.9, [0,4A2 ⫺ H ⫺ H2SO4]⫺ at m/z 299.7, B1 at m/z 254.6, [B1 ⫺ H2O]⫺ at m/z 236.6, [B1 ⫺ H ⫺ HCOO]⫺ at m/z 192.7, and [B1 ⫺ H ⫺ H2SO4]⫺ at m/z 156.8. Figure 4B shows the MS2 spectrum for ⌬di-4S. Fragment ions included 2,5A2 at m/z 341.9, Y1 at m/z 299.9, Z1 at m/z 281.8, and [Z1 ⫺ H ⫺ SO3]⫺ at m/z 198.8. Figure 4C shows the MS2 spectrum for ⌬di-6S. Fragment ions included 2,5A2 at m/z 341.9, Y1 at m/z 299.9, Z1 at m/z 281.8, and [Z1 ⫺ H2SO4]⫺ at m/z 183.9. From the observed glycosidic cleavages, the identification of each isomer is established based on the retention of the sulfate on either the uronic acid or the galactosamine moiety. The MS2 ions corresponding to the glycosidic cleavage, where charge was retained on the portion containing the sulfate group, were m/z 236.6 ([B1 ⫺ H2O]⫺ ions) for ⌬di-2S (Fig. 4A), m/z 299.9

FIGURE 2. MS2 spectra of KS sulfated disaccharides, in the negative ionization mode. (A) For Gal-␤-1,4GlcNAc(6S) (MSD), using the MS1 m/z 462.0 as a previous ion (Fig. 1A), product ions were observed at m/z 444.0 corresponding to [M ⫺ H ⫺ H2O]⫺, m/z 360.9 corresponding to 0,2 A2, m/z 342.9 corresponding to (0,2A2 ⫺ H2O]⫺, m/z 264.0 corresponding to [0,2A2 ⫺ HSO4]⫺, m/z 183.8 corresponding to [Z1 ⫺ H2O ⫺ SO3]⫺, and m/z 138.8 corresponding to (OCHCH2OSO3)⫺. (B) For DSD, using the double charge state ion m/z 270.4 (Fig. 1B) as a precursor ion, the product ions included [M ⫺ 2H ⫺ H2O]2⫺ at m/z 261.4, [Y1 ⫺ SO3]⫺ at m/z 219.8, [OCHCH2OSO3]⫺ at m/z 138.8, and [HSO4]⫺ at m/z 96.9. (C) Using m/z 541.9 (Fig. 1B) as a precursor ion, fragment ions were observed at m/z 461.9, corresponding to [M ⫺ H ⫺ SO3]⫺, and m/z 299.9, corresponding to Y1.

(Y1 ions) for ⌬di-4S (Fig. 4B), and m/z 281.8 (Z1 ions) for ⌬di-6S (Fig. 4C). Major MS2 fragment ions at m/z 236.6, 299.9, and 281.8, therefore, can be used as diagnostic ions to differentiate these three isomers. Figure 4D shows the MS2 spectrum for ⌬UA-␤-1,4-GlcNS, the CS/DS internal standard. Fragment ions included [Z1 ⫺ H]⫺ at m/z 239, [Z1 ⫺ H ⫺ H2O]⫺ at m/z 221.7, and 0,2X0 at m/z 137.8. To verify the utility of the equations presented in the Materials and Methods section for quantitative analysis of ⌬di-4S and ⌬di-6S, a range of mixtures with specific concentrations of each of the isomers was analyzed. Figures 5A–E show the MS2 spectra of ⌬di-4S and ⌬di-6S mixtures with various molar ratios. The relative abundance of ions at m/z 299.9 increased with increasing ⌬di-4S concentration, whereas the relative abundance of m/z 281.8 ions decreased with decreasing ⌬di-6S concentration. The calculated results are shown in Table 2. The errors in all the samples were very small (0.2%– 6.2%), thus establishing the utility of the equations for the quantitative analysis of the isomeric disaccharides in CS/DS.

1608

Zhang et al.


FIGURE 3. MS1 spectra of CS/DS disaccharides and ⌬UA-␤-1,4-GlcNS, in the negative ionization mode. (A) For ⌬di-2S, molecular ion [M ⫺ H]⫺ at m/z 458.1 was observed; (B) for ⌬di-4S, molecular ion [M ⫺ H]⫺ at m/z 458.1 was observed; (C) for ⌬di-6S, molecular ion [M ⫺ H]⫺ at m/z 458.1 was observed; and (D) for ⌬UA-␤-1,4-GlcNS, molecular ion [M ⫺ H]⫺ at m/z 415.8 was observed.

Detection of Sulfated Disaccharides from KS of Embryonic Chick Corneas Figure 6A presents the MS1 spectra of the products released by digesting an E18 chick cornea section with keratanase II. Molecular ions at m/z 270.4 and m/z 462.0 were found. The ion at m/z 458.1 corresponds to the internal standard ⌬Di-2S, which was added at a final concentration of 50 pmol/␮L to the digest. We used mass isolation techniques to isolate the molecular ions at m/z 462.0 and m/z 270.4, and we used each as a precursor ion for MS2 experiments. MS2 spectra for the ion at m/z 462.0, shown in Figure 6B, included fragment ions at m/z 444.0, 360.9, 342.9, 264.0, 183.8, and 138.8. MS2 spectra for m/z 270.4, shown in Figure 6C, included fragment ions at m/z 261.3 and 219.8. Comparison of these fragment ions with MS2 spectra of MSD and DSD (Figs. 2A, 2C) confirms that the molecular ion m/z 462.0 in Figure 6A corresponds to MSD and the molecular ion at m/z 270.4 corresponds to DSD.

Detection of Sulfated Disaccharides from CS/DS of Embryonic Chick Corneas Figure 7A shows the MS1 spectrum of the products released by digesting an E18 chick cornea section with chondroitinase

ABC. The characteristic molecular ion at m/z 458.0 was observed and used as a precursor ion for MS2 experiments. The MS2 spectrum for m/z 458.1 is shown in Figure 7B. Fragment ions at m/z 341.9, 299.9, 281.8, and 198.7 were observed. Comparison with the MS2 spectra of the three chondroitin MSD isomers (Figs. 4A–C), indicate that the sample did not contain ⌬Di-2S, because of the absence of the ion at m/z 236.6. The significant ions at m/z 281.8 and 299.9 suggest that both ⌬di-4S and ⌬di-6S, respectively, are present in the CS/DS of chick cornea.

Content of the KS Sulfated Disaccharides MSD and DSD The tissue concentrations of the sulfated disaccharides, calculated on the basis of a cubic millimeter of tissue, in chick corneas from E8 to E20 embryos, 4-week-old chick, and 70week-old adult chickens are shown in Figure 8. These results show that the high concentration of MSD released from corneal stroma on E8 was very similar to that at E20 and 70 weeks. The MSD concentration declined sharply by E10; then increased with advancing embryo age to a peak on E14, when corneal transparency just begins to increase; decreased tran-



1609

FIGURE 4. MS2 spectra of CS/DS disaccharides, and ⌬UA-␤-1,4-GlcNS, in the negative ionization mode. (A) For ⌬di-2S, using m/z 458.1 (Fig. 3A) as a precursor ion, fragment ions included [M ⫺ H ⫺ H2O ⫺ SO3]⫺ at m/z 360.0, 2,5A2 at m/z 341.9, [0,4A2 ⫺ H ⫺ H2SO4]⫺ at m/z 299.7, B1 at m/z 254.6, [B1 ⫺ H2O]⫺ at m/z 236.6, [B1 ⫺ H ⫺ HCOO]⫺ at m/z 192.7, and [B1 ⫺ H ⫺ H2SO4]⫺ at m/z 156.8. (B) For ⌬di-4S, using m/z 458.1 (Fig. 3B) as a precursor ion, fragment ions included 2,5A2 at m/z 341.9, Y1 at m/z 299.9, Z1 at m/z 281.8, and [Z1 ⫺ H ⫺ SO3]⫺ at m/z 198.8. (C) For ⌬di-6S, using m/z 458.1 (Fig. 3C) as a precursor ion, fragment ions included 2,5A2 at m/z 341.9, Y1 at m/z 299.9, Z1 at m/z 281.8, and [Z1 ⫺ H2SO4]⫺ at m/z 183.9. (D) For ⌬UA-␤-1,4-GlcNS, using m/z 415.8 (Fig. 3D) as a precursor ion, fragment ions included [Z1 ⫺ H]⫺ at m/z 239.7, [Z1 ⫺ H ⫺ H2O]⫺ at m/z 221.7, and 0,2X0 at m/z 137.8.

siently during E16 and E18, when the cornea becomes more transparent; and then increased again by E20, when the cornea reaches maximum transparency, to the high levels of E8 and

FIGURE 5. MS2 spectra of the mock mixtures of ⌬di-4S and ⌬di-6S, in the negative ionization mode. (A) ⌬di-4S: ⌬di-6S ⫽ 1:9; (B) ⌬di-4S:⌬di-6S ⫽ 3:7; (C) ⌬di-4S:⌬di-6S⫽ 5:5; (D) ⌬di-4S:⌬di-6S ⫽ 7:3; and (E) ⌬di-4S: ⌬di-6S ⫽ 9:1. Characteristic marker fragments for identifying and quantifying ⌬di-4S and ⌬di-6S are shown. Note that the relative abundance of ion at m/z 299.9 increased with increasing ⌬di-4S concentration, whereas the relative abundance of m/z 281.8 decreased with decreasing ⌬di-6S concentration. Conversely, the relative abundance of m/z 341.9 and m/z 198.8 did not change markedly.

E14. After the chick hatched, the concentration of MSD increased further by 4 weeks and then stabilized near the E20 peak. For DSD, a similar profile of changes was observed,

1610

Zhang et al.


TABLE 2. Authentication of the Method for Quantifying Individual CS/DS Disaccharides in Mixtures of Widely Different Ratios Calculated %* ⌬di-4S:⌬di-6S

1

2

3

4

5

6

Avg. % ⴞ SD (n ⴝ 6)

Actual %

1:9

9.3 90.7 26.4 73.6 72.0 28.1 92.2 7.8

8.6 91.4 26.7 73.3 73.2 26.8 92.7 7.3

8.8 91.2 26.9 73.1 72.9 27.1 91.4 8.6

8.8 91.2 26.7 73.3 72.7 27.3 91.4 8.6

8.8 91.2 26.4 73.6 72.9 27.2 91.4 8.6

9.0 91.0 26.6 73.4 71.9 28.1 92.2 7.8

8.9 ⫾ 0.2 91.0 ⫾ 0.2 26.6 ⫾ 0.2 73.4 ⫾ 0.2 72.6 ⫾ 0.5 27.4 ⫾ 0.5 91.9 ⫾ 0.6 8.1 ⫾ 0.6

10.0 90.0 30.0 70.0 70.0 30.0 90.0 10.0

3:7 7:3 9:1

* Percentages calculated with equations 3 and 4.

except that after the chick hatched, the concentration of DSD declined by 4 weeks and remained at the reduced level (Fig. 8A). Consequently, the combined concentration of KS MSD⫹DSD displayed the same developmental profile, with the posthatch concentrations remaining high (Fig. 8B). The percentage that each type of KS sulfated disaccharide represented in relation to total KS sulfated disaccharides in the cornea during embryonic development and after hatching is shown in Figure 9. The molar percent of DSD was significantly higher than that of MSD during the embryonic period, reached equivalence at the time of hatching, and then declined sharply below that of MSD after hatching.

Content of CS/DS Sulfated Disaccharides The tissue concentrations of the CS/DS sulfated disaccharides ⌬di-4S and ⌬di-6S, released by chondroitinase ABC from sections of developing chick cornea, are shown in Figure 10. The concentration of ⌬di-6S was at a high level on E8, declined sharply by E10, then decreased gradually with advancing embryonic age to a low level on E20, and remained stable at that

low level after hatching (Fig. 10A). Conversely, the concentration of ⌬di-4S was at a low level on E8, increased gradually until E20, and then declined after hatching to a stable lower level by 4 weeks (Fig. 10A). Thus, the total concentration of ⌬di-4S and ⌬di-6S gradually decreased as embryonic development progressed (Fig. 10B) and, in contrast to KS sulfated disaccharides, continued to remain low after hatching. The molar percent of ⌬di-4S increased with developmental age from 22% at E8 to 87% at E20, a high level that was maintained after hatching and into adulthood (Fig. 11). In contrast, the molar percent of ⌬di-6S decreased from 78% at E8 to 13% at E20, its future posthatch level. The proportions of ⌬di-4S and ⌬di-6S were equal at E14. These results indicate that ⌬di-6S is the major component of CS/DS at developmental times before E14, whereas ⌬di-4S becomes the major component after E14, as the cornea becomes transparent. Comparing Figure 10B to Figure 8B, it can be seen that the concentration of sulfated CS/DS disaccharides relative to sulfated KS disaccharides was about equivalent at E8, became approximately 3:2 at E10, was even again at E12, and was 2:3 by E14. After E14, the concen-

FIGURE 6. Mass spectra of products released by keratanase II digestion of an E18 chick corneal section, in the negative ionization mode. (A) MS1 spectra: molecular ions at m/z 462.0 and m/z 270.4 were found. The ion at m/z 458.1 corresponded to the KS internal standard ⌬Di-2S. (B) MS2 spectra of the precursor ion at m/z 462.0 (marker ion for MSD): fragment ions at m/z 440.0, 360.9, 342.9, 264.0, 183.8, and 138.8 were observed. (C) MS2 spectra of the precursor ion at m/z 270.4 (marker ion for DSD). Fragment ions at m/z 261.3 and 219.8 were observed.



1611

FIGURE 7. Mass spectra of products released by chondroitinase ABC digestion of an E18 chick cornea section, under negative ionization mode. (A) MS1 spectra: a molecular ion at m/z 458.1 was observed. The ion at m/z 415.8 corresponded to the CS/DS internal standard, ⌬UA-␤-1,4GlcNS. (B) MS2 spectra of the precursor ion at m/z 458.1. Fragment ions at m/z 341.9, 299.9, 281.8, and 198.7 were observed.

tration of sulfated CS/DS disaccharides was always less than the concentration of sulfated KS disaccharides, reaching approximately 1:3 after hatching.

DISCUSSION In this study, ESI-MS/MS techniques were used to directly detect and quantify KS and CS/DS sulfated disaccharides that were liberated from single frozen sections of developing chick corneas and analyzed without chromatographic purification. The data provide new insights into the changes occurring during biosynthesis and accumulation of these extracellular matrix molecules as this tissue becomes transparent before hatching. Our data demonstrate that the concentration of sulfated KS disaccharides in developing chick corneas changes during embryogenesis, as the cornea enlarges in diameter and becomes transparent. Sulfated KS disaccharide concentrations were quite high at E8, at the time that neural crest cells are still migrating into the corneal stroma; at E14, as the cornea begins to dehydrate and become transparent; and at E20 and into adulthood, when the cornea reaches maximum transparency. In contrast, the concentration of sulfated KS disaccharides is low at E10, just as corneal nerves, which have encircled the cornea from E5 to E10, begin to penetrate the stroma35 anterior to the zone of progressing extracellular sulfated KS accumulation.12,36 By E14, when the concentration of sulfated KS disaccharides reaches its second peak, and the wave of progressing extracellular sulfated KS accumulation is reaching the anterior edge of the corneal stroma,12 corneal nerves first penetrate from the corneal stroma into the corneal epithelium.36 Conceivably, pathways chosen by corneal nerve growth cones may be influenced by distributions and concentrations of sulfated KS or CS/DS, or by the exact structural positions of

their sulfate moieties. Finally, the molar percent of DSD was greater than that of MSD from E8 to E18, became equal to it at E20, and then was less than that of MSD after hatching. The molar ratios of MSD to DSD in 4- and 70-week-old adult chicken corneas are 2.2 and 1.5, respectively, which are similar to the ratios in adult bovine and human cornea.37 These changes may be significant for first attaining and then maintaining proper corneal hydration after hatching, as the corneal surface is first exposed to air. Compared to tissue concentrations of KS sulfated disaccharides, the tissue concentrations of CS/DS ⌬di-4S and ⌬di-6S in cornea gradually decreased during embryonic development and maintained a low level in the adult cornea. The total concentration of CS/DS ⌬di-4S and ⌬di-6S became markedly lower than that of KS sulfated disaccharides by E20 and after hatching (compare Fig. 8B with 10B), suggesting that the change in proportions of the two major GAGs with advancing embryonic age may be mechanistically involved in facilitating either the changes in corneal stroma hydration and thickness, and/or in the maintenance of transparency. Moreover, our results indicate that the proportion of ⌬di-4S increased with advancing embryonic age, whereas that of ⌬di-6S decreased (Fig. 11). Proportions of ⌬di-4S and ⌬di-6S were almost equal on E14, the day when the corneal stroma begins to undergo dehydration and become transparent.15,16 The proportion of ⌬di-6S then remained constantly low from E20 through hatching and adulthood, comprising ⬍15% of the total concentration of ⌬di-4S and ⌬di-6S (Fig. 11). The disaccharide ratios detected here may reflect changes in specific activities of chondroitin sulfotransferases in the developing chick cornea.6,38 Thus, our results suggest that in early embryonic chick corneas, the activity of chondroitin 6-O-sulfotransferase is higher than that of chondroitin 4-O-sulfotransferase, but that after the chicks hatch, these activities are reversed.

1612

Zhang et al.


FIGURE 8. (A) Tissue concentrations of MSD and DSD derived from KS during chick corneal development. Five individual sections were analyzed separately for each data point. Sections were digested with a solution containing keratanase II and internal standard ⌬di-2S, and concentrations were calculated. The digests were then analyzed directly by ESI-MS/MS under the negative ionization mode. (B) The total concentrations of MSD plus DSD, during chick corneal development.

types of sulfated disaccharides either from KS or from CS/DS, from a single frozen section. Keratanase II (from Bacillus sp.) is an endo-␤-N-acetylglucosaminidase which cleaves the N-acetylglucosamine linkage of the KS chain, releasing Gal-␤-1,4-GlcNAc disaccharides with mono- or disulfates, both of which contain an N-acetyllactosamine structure,39 and a smaller proportion of tetrasaccharides.23,37,40 Keratanase II does not produce nonsulfated disaccharides. Our previous studies have shown that the major ion in the MS1 spectrum of a KS trisulfated tetrasaccharide occurs at m/z 324.4, and the most abundant ion in the MS1 spectrum of a KS tetrasulfated tetrasaccharide occurs at m/z 266.0.23 Because neither of these ions was significantly present in the MS1 spectrum of keratanase II-digested chick corneas (see Fig 6A), we conclude that the major keratanase II digestion products of chick cornea are sulfated disaccharides. The chondroitinase ABC we used is protease free, which means that it cleaves ␤-1,4-galactosaminidic bonds between N-acetylgalactosamine and either glucuronic acid or iduronic acid, to release ⌬di-4S or ⌬di-6S disaccharides, nonsulfated and disulfated disaccharides, and tetra- or higher chondroitin sulfate oligosaccharides.41,42 The procedure that renders chondroitinase ABC protease-free eliminates a second enzyme, present in many chondroitinase ABC preparations, which specifically splits off disaccharides from the nonreducing end of large oligosaccharides produced by protease-free chondroitinase ABC. We do not know what the major ions of the MS1 spectra of such hypothetical sulfated oligosaccharides are. Therefore, we cannot estimate the possible contributions of such constructs to the total sulfated CS/DS content of the chick cornea at this time. However, changes in the ratios of sulfated to nonsulfated KS and CS/DS disaccharides over time during development may be important in regulating cell movement within the cornea. We are currently developing methods to analyze nonsulfated KS and CS/DS disaccharides to obtain this information. Compared with previous studies of GAGs in the developing chick cornea, this is the first study to use single sections of embryonic corneas and the first to digest them with an appropriate glycosidase to liberate disaccharides selectively for direct MS analysis without preliminary chromatographic separations. Previously, radiolabeling methods used metabolically labeled KS produced in vitro by corneal explants or cells to monitor the KSPG synthesized in the cornea.10,11,13 However,

Our results are consistent with the possibility that chondroitin6-sulfate may be more important than chondroitin-4-sulfate for organizing the stroma before the initiation of the steps in dehydration, whereas chondroitin-4-sulfate may be more important for the steps that initially produce and then stabilize transparency after hatching, when the corneal surface is exposed to air and covered with a tear film. In the present study, the profiles of the concentrations of KS sulfated disaccharides and the concentrations of CS/DS sulfated disaccharides released from single 10-␮m sections of chick cornea were achieved using a single-point normalization factor method. In addition, a system of two equations was used successfully to determine the relative amounts of ⌬di-4S and ⌬di-6S isomers released from sections of the chick corneas digested with chondroitinase ABC and then analyzed directly by MS. This procedure therefore represents an advance over previous studies, which required substantial sample sizes, as well as chromatographic purification or separation of disaccharides, to quantify the proportion of ⌬di-4S and ⌬di-6S.24,28,29 Our mass spectrometric method is sufficiently sensitive to allow chemical analysis and structural determination of two

FIGURE 9. Molar percentages of KS sulfated disaccharides derived from KS during chick cornea development. Concentrations of MSD plus DSD (Fig. 8B) were divided by the concentrations of MSD or DSD (Fig. 8A) at each age.

IOVS, May 2005, Vol. 46, No. 5 KSPG made in vitro differs from that found in vivo in average chain length and in the degree of sulfation.43,44 Therefore, the method described herein allows a more accurate chemical characterization of the repeating units of GAGs present in the developing chick cornea. Fluorophore-assisted carbohydrate electrophoresis (FACE) provides a simple, highly sensitive, and quantitative tool for determination of the content and composition of KS and CS/DS,37,44,45 but the GAG preparations are still complex to analyze and require minimum amounts of sample tissue that may still be difficult to obtain from very early embryos. In contrast, the mass spectrometric method used in the present study is rapid, simple, accurate, and sufficiently sensitive, so that even the amount of extracellular matrix present in portions of single sections of single embryonic corneas provides sufficient material to allow structural and chemical analysis. Our data suggest that the described method can be used to analyze the respective sulfated disaccharides released from KS and CS/DS in various biological tissues other than cornea. An important potential application of this method may be in the diagnosis of diseases related to KS and CS/DS, and in response to surgical procedures, such as the laser-


1613

FIGURE 11. Molar percentages of sulfated disaccharides derived from CS/DS, ⌬di-4S, and ⌬di-6S, during chick cornea development. Concentrations of ⌬di-4S plus ⌬di-6S (Fig. 10B) were divided by the concentrations of ⌬di-4S or ⌬di-6S (Fig. 10A) at each age.

assisted intrastromal keratomileusis (LASIK) and laser-assisted subepithelial keratomileusis (LASEK) protocols in the cornea.

References

FIGURE 10. (A) Tissue concentrations of sulfated disaccharides derived from CS/DS, ⌬di-4S, and ⌬di-6S, during chick cornea development. Five individual sections were analyzed separately for each data point. Sections were digested with a solution containing chondroitinase ABC and internal standard ⌬UA-␤-1,4-GlcNS. The digests were then analyzed directly by ESI-MS/MS under negative ionization mode, and concentrations were calculated. (B) The total concentrations of ⌬di-4S plus ⌬di-6S, during chick cornea development.

1. Smolin G, Thoft RA. The Cornea. New York: Little, Brown; 1994: 47– 67. 2. Meyer K, Linker A, Davidson EA, Weissmann B. The mucopolysaccharides of bovine cornea. J Biol Chem. 1953;205:611– 616. 3. Krusius T, Finne J, Margolis RK, Margolis RU. Identification of an O-glycosidic mannose-linked sialylated tetrasaccharide and keratan sulfate oligosaccharides in the chondroitin sulfate proteoglycan of brain. J Biol Chem. 1986;261:8237– 8242. 4. Krusius T, Reinhold VN, Margolis RK, Margolis RU. Structural studies on sialylated and sulphated O-glycosidic mannose-linked oligosaccharides in the chondroitin sulphate proteoglycan of brain. Biochem J. 1987;245:229 –234. 5. Conrad GW. Biological role of keratan sulfate proteoglycan. In: Ernst B, Hart GW, Sinay¨ P, eds. Carbohydrates in Chemistry and Biology. Part II. Biology of Saccharides. Vol 4. Wiley-VCH: New York. 2000:717–727. 6. Sugumaran G, Vertel BM. Biosynthesis of chondroitin sulfate and dermatan sulfate proteoglycans. In: Ernst B, Hart GW, Sinay¨ P, eds: Carbohydrates in Chemistry and Biology. Part II. Biology of Saccharides. Vol 3. Wiley-VCH: New York. 2000:375–395. 7. Conrad GW. Collagen and mucopolysaccharide biosynthesis in the developing chick cornea. Dev Biol. 1970;21:292–317. 8. Conrad GW. Collagen and mucopolysaccharide biosynthesis in mass cultures and clones of chick corneal fibroblasts in vitro. Dev Biol. 1970;21:611– 635. 9. Conrad GW, Dorfman A. Synthesis of sulfated mucopolysaccharides by chick corneal fibroblasts in vitro. Exp Eye Res. 1974;18: 421– 433. 10. Hart GW. Biosynthesis of glycosaminoglycans during corneal development. J Biol Chem. 1976;251:6513– 6521. 11. Hart GW. Biosynthesis of glycosaminoglycans by the separated tissues of the embryonic chick cornea. Dev Biol. 1978;62:78 –98. 12. Funderburgh JL, Caterson B, Conrad GW. Keratan sulfate proteoglycan during embryonic development of the chicken cornea. Dev Biol. 1986;116:267–277. 13. Nakazawa K, Suzuki S, Wada K, et al. Proteoglycan synthesis by corneal explants from developing embryonic chicken. J Biochem. 1995;117:707–718. 14. Johnston MC, Noden DM, Hazelton RD, et al. Origins of avian ocular and periocular tissues. Exp Eye Res. 1979;29:27– 43. 15. Coulombre AJ, Coulombre JL. Corneal development. I. Corneal transparency. J Cell Comp Physiol. 1958;51:1–11.

1614

Zhang et al.

16. Coulombre AJ, Coulombre JL. Corneal development. III. The role of the thyroid in dehydration and the development of transparency. Exp Eye Res. 1964;3:105–114. 17. Dunlevy JR, Beales MP, Berryhill BL, et al. Expression of the keratan sulfate proteoglycans lumican, keratocan and osteoglycin/ mimecan during chick corneal development. Exp Eye Res. 2000; 70:349 – 62. 18. Dell A, Morris HR. Glycoprotein structure determination by mass spectrometry. Sci. 2001;291:2351–2356. 19. Zaia J, Costello CE. Compositional analysis of glycosaminoglycans by electrospray mass spectrometry. Anal Chem. 2001;73:233–239. 20. Zaia J, McClellan JE, Costello CE. Tandem mass spectrometric determination of the 4S/6S sulfation sequence in chondroitin sulfate oligosaccharides. Anal Chem. 2001;73:6030 – 6039. 21. Schulz BL, Packer NH, Karlsson NG. Small-scale analysis of Olinked oligosaccharides from glycoproteins and mucins separated by gel electrophoresis. Anal Chem. 2002;74:6088 – 6097. 22. Williams D, Young MK. Analysis of neutral isomeric low molecular weight carbohydrates using ferrocenyl boronate derivatization and tandem electrospray mass spectrometry. Rapid Commun Mass Spectrom. 2000;14:2083–2091. 23. Zhang Y, Kariya Y, Conrad AH, et al. Analysis of keratan sulfate oligosaccharides by electrospray ionization tandem mass spectrometry. Anal Chem. 2005;77:902–910. 24. Desaire H, Leary JA. Detection and quantification of the sulfated disaccharides in chondroitin sulfate by electrospray tandem mass spectrometry. J Am Soc Mass Spectrom. 2000;11:916 –920. 25. Desaire H, Sirich TL, Leary JA. Evidence of block and randomly sequenced chondroitin polysaccharides: sequential enzymatic digestion and quantification using ion trap tandem mass spectrometry. Anal Chem. 2001;73:3513–3520. 26. Saad OM, Leary JA. Compositional analysis and quantification of heparin and heparan sulfate by electrospray ionization ion trap mass spectrometry. Anal Chem. 2003;75:2985–2995. 27. Pope RM, Raska CS, Thorp SC, et al. Analysis of heparin sulfate oligosaccharides by nano-electrospray ionization mass spectrometry. Glycobiology. 2001;11:505–513. 28. Oguma T, Toyoda H, Toida T, et al. Analytical method for keratan sulfates by high-performance liquid chromatography/turbo-ionspray tandem mass spectrometry. Anal Biochem. 2001;290:68 –73. 29. Oguma T, Toyoda H, Toida T, et al. Analytical method of chondroitin/dermatan sulfates using high performance liquid chromatography/turbo ionspray ionization mass spectrometry: application to analyses of the tumor tissue sections on glass slides. Biomed Chromatogr. 2001;15:356 –362. 30. Koshiishi I, Horikoshi E, Imanari, T. Quantification of hyaluronan and chondroitin/dermatan sulfates in the tissue sections on glass slides. Anal Biochem. 1999;267:222–226.

IOVS, May 2005, Vol. 46, No. 5 31. Ge X, Sirich TL, Beyer MK, et al. A strategy for the determination of enzyme kinetics using electrospray ionization with an ion trap mass spectrometer. Anal Chem. 2001;73:5078 –5082. 32. Pi N, Armstrong JI, Bertozzi CR, et al. Kinetic analysis of NodST sulfotransferase using an electrospray ionization mass. Biochem. 2002;41:13283–13288. 33. Desaire H, Leary JA. Utilization of MS3 spectra for the multicomponent quantification of diastereomeric N-acetylhexosamines. J Am Soc Mass Spectrom. 2000;11:1086 –1094. 34. Domon B, Costello CE. A systematic nomenclature for carbohydrate fragmentations in FAB-MS/MS spectra of glycoconjugates. Glycoconjugate J. 1988;5:397– 409. 35. Hay ED. Development of the vertebrate cornea. Int Rev Cytol. 1979;63:263–322. 36. Bee JA. The development and pattern of innervation of the avian cornea. Dev Biol. 1982;92:5–15. 37. Plaas AH, West LA, Midura RJ. Keratan sulfate disaccharide composition determined by FACE analysis of keratanase II and endo␤-galactosidase digestion products. Glycobiology. 2001;11:779 – 790. 38. Silbert JE, Sugumaran G. Biosynthesis of chondroitin/dermatan sulfate. IUBMB Life. 2002;54:177–186. 39. Yamagishi K, Suzuki K, Imai K, et al. Purification, characterization, and molecular cloning of a novel keratan sulfate hydrolase, endobeta-N-acetylglucosaminidase, from Bacillus circulans. J Biol Chem. 2003;278:25766 –772. 40. Tai GH, Nieduszynski IA, Fullwood NJ, et al. Human corneal keratan sulfates. J Biol Chem. 1997;272:28227–28231. 41. Hamai A, Hashimoto N, Mochizuki H, et al. Two distinct chondroitin sulfate ABC lyases: an endoeliminase yielding tetrasaccharides and an exoeliminase preferentially acting on oligosaccharides. J Biol Chem. 1997;272:9123–9130. 42. Lauder RM, Huckerby TN, Nieduszynski IA. A fingerprinting method for chondroitin/dermatan sulfate and hyaluronan oligosaccharides. Glycobiology. 2000;10:393– 401. 43. Funderburgh JL, Funderburgh ML, Mann MM, et al. Synthesis of corneal keratan sulfate proteoglycans by bovine keratocytes in vitro. J Biol Chem. 1996;271:31431–31436. 44. Funderburgh JL, Mann MM, Funderburgh ML. Keratocyte phenotype mediates proteoglycan structure: a role for fibroblasts in corneal fibrosis. J Biol Chem. 2003;278:45629 – 45637. 45. Plaas AH, West LA, Thonar EJ, et al. Altered fine structures of corneal and skeletal keratan sulfate and chondroitin/dermatan sulfate in macular corneal dystrophy. J Biol Chem. 2001;276:39788 – 39796.