Anal. Chem. 2005, 77, 2802-2809
Use of Multidimensional Lectin Affinity Chromatography in Differential Glycoproteomics Ruiqing Qiu and Fred E. Regnier*
Department of Chemistry, Purdue University, West Lafayette, Indiana 47907
This paper reports studies comparing the relative degree of sialylation among human serum glycoproteins carrying complex biantennary N-linked, hybrid, and high-mannose oligosaccharides. Comparisons were made by coupling lectin affinity selection with stable isotope coding of peptides from tryptic digests of serum. After proteolysis, samples were split and differentially acetylated with stable isotope coding agents according to either origin or the separation method by which they would be fractionated. A lectin column prepared from Sambucus nigra agglutinin (SNA) was used to select and compare the concentration of sialic acid containing glycopeptides. The relative standard deviation in quantification using this method was 4%. Using this method the concentration of sialic acid containing glycoproteins from a normal individual were compared to those in a pooled serum sample from a large number of normal individuals. It was found that sialylation varied less than 2-fold in all but four or five glycoproteins. Further studies were done on the degree of sialylation within glycoproteins. Samples labeled with the light isoform of the coding agent were applied to a set of serial lectin columns consisting of a concanavalin A (Con A) column coupled to an SNA column for selecting sialic acid appended to glycopeptides with complex biantennary N-linked, hybrid, and high-mannose glycans. In contrast, samples labeled with the heavy isoform of the coding agent were applied to a Con A lectin column alone to select glycopeptides containing complex biantennary N-linked, hybrid, and high-mannose glycans, without regard to sialylation. Glycopeptides thus selected were mixed, deglycosylated by PNGase F, and fractionated by reversed-phase chromatography (RPC). The RPC fractions were then analyzed by ESI-MS. The relative standard deviation of the method was 4%. All glycopeptides identified contained sialic acid except one. Peptides in which the relative abundance of isotopic isoforms was equal were considered to indicate that the protein parent was fully sialylated at that specific glycosylation site. Posttranslational modification (PTM) is an important feature of a proteome, frequently conveying a specific biological activity or role to a protein. Modifications can occur at a single or at multiple sites, often in varying forms. Among the more than 100 * To whom correspondence should be addressed. E-mail: purdue.edu.
2802 Analytical Chemistry, Vol. 77, No. 9, May 1, 2005
fregnier@
different types of PTM,1 each plays a different role. Clearly the diversity surrounding PTM considerably increases the complexity of the proteome and adds an additional level of intricacy and complication to proteomics. Among the many types of PTM, glycosylation appears to be the most complex. Proteins can be modified at one or more sites with numerous glycan variants at each site. Variation in the concentration of a single glycosyltransferase can alter the glycosylation profile of multiple proteins.2 Glycosylation patterns can vary with development, regulatory state, type of disease, and even with disease progression. Recent studies have shown that one way to deal with the complexity of the proteome in proteomics is to use affinity chromatography methods to select proteins or peptides with a common type of modification,3 as with carbonylation in the case of oxidative stress,4 phosphorylation in regulatory proteins,5 or glycosylation.6 The great advantage of affinity selection methods is that they can target a specific structural feature of a PTM, rapidly select the fraction of the proteome with the targeted PTM, and provide substantial simplification of the mixture in the process. Although lectins7 have been used in glycoprotein isolation and glycan structure analysis for decades, it is only recently that they are being applied to glycoproteomics.8 The high selectivity of the BS-II lectin from Bandeiraea simplicifolia and the agglutinin LTA from Lotus tetragonolobus have been exploited to select glycopeptides containing N-acetylglucosamine (GlcNAc)9 or R-fuc(1f6)β-glcNAc-Asn-,6,10 respectively. Coupling LTA affinity selection with recent advances in stable isotope coding has allowed comparative proteomics methods to be developed for recognizing aberrant glycosylation in the glycoproteins of cancer patients.10 Broader selectivity lectins in contrast are being used to define the glycoproteome by what is sometimes referred to as the “glycocatch method”.11 For example, soybean lectin has been used in (1) Gudepu, R. G.; Wold, F. In Proteins: Analysis and Design; Angeletti, R. H., Ed.; Academic: San Diego, 1998; pp 121-207. (2) Sburlati, A. R.; Umana, P.; Prati, E. G. P.; Bailey, J. E. Biotechnol. Progress 1998, 14, 189-192. (3) Mirzaei, H.; Regnier, F. E. J. Chromatogr. B 2005, 817, 23-24. (4) Yoo, B, Regnier, F. E. Electrophoresis. 2004, 25, 1334-1341. (5) Seeley, E.; Riggs, L.; Regnier, F. E. J. Chromatogr. B 2005, 817, 81-88. (6) Xiong, L.; Andrews, D.; Regnier, F. E. J. Proteome Res. 2003, 2, 618-625. (7) Lis, H; Sharon N; Katchalski E. Biochim. Biophys. Acta 1969, 192, 364366. (8) Ji, J.; Chakraborty, A.; Geng, M.; Zhang, X.; Amini, A.; Bina, M.; Regnier, F. E. J. Chromatogr., B 2000, 745, 197-210. (9) Geng, M.; Zhang, X.; Bina, M.; Regnier, F. E. J. Chromatogr., B 2001, 752, 293-306. (10) Xiong, L.; Regnier, F. E. J. Chromatogr., B 2002, 782, 405-418. (11) Hirabayashi, J.; Kasai, K.-I. J. Chromatogr., B 2002, 771, 67-87. 10.1021/ac048751x CCC: $30.25
© 2005 American Chemical Society Published on Web 03/31/2005
conjunction with two-dimensional gel electrophoresis to identify at least 22 periplasmic glycoproteins from the Gram-negative bacterium, Campylobacter jejuni.12 Concanavalin A (Con A) and galectin LEC-6 (GaL6) have been used in similar fashion for the identification of glycoproteins in Caenorhabditis elegans, where proteolytic digests were fractioned with two lectin columns. A Con A affinity column was used to capture peptides with high-mannose type N-glycans from extracts while the flow-through was sent to a GaL6 column to select peptides with complex-type N-glycans.13 As a result, 44 glycopeptides were captured by the Con A column and 23 by the GaL6 columns, allowing the identification of 32 and 16 proteins, respectively. A similar strategy has been used to examine extracellular domains from membrane proteins on mammalian cells.14 Following metalloprotease treatment, cleavage fragments were affinity selected, then deglycosylated, and finally fractionated by polyacrylamide gel electrophoresis. Relative quantification of the resulting peptides was achieved by isotope dilution. Similar to the targeting of an epitope by an antibody, lectins interact with a specific structural motif in a glycan. Moreover, the portion of the structure being probed is likely to differ between lectins. This has led to the use of lectins to examine glycan structure15 and the concept of serial lectin affinity chromatography (SLAC).16 Sequential binding to a serial set of lectin columns of different structural selectivity suggests the presence of a combination of structural elements in a glycan or glycoproteins without actually determining the whole structure.17 SLAC is a simple, but rapid way to explore glycan structure. It has been noted above that comparative glycoproteomics is a valuable tool in the search for biomarkers. But, comparative analyses require quantification. One approach to examining differential glycoproteomics has been to use differential staining in 2-D gel electrophoresis.18 A second has been with isotope dilution.15 Yet another is by stable isotope coding agents. Of the two stable isotope coding methods reported, the method based on global internal standard technology (GIST) seems to be of broadest utility. After all peptides in a proteolytic digest are isotopically coded in the GIST protocol,11,19 a wide variety of affinity chromatography columns can be used to select different types of glycopeptides for quantification. An alternative approach is to select all vicinal diol containing proteins from a proteome through periodate oxidation and Schiff base formation of the oxidized proteins with a hydrazide containing support. Captured glycoproteins are then subjected to proteolysis and isotope coding before the peptides are released and analyzed.20 (12) Young, N. Martin; Brisson, Jean-Robert; Kelly, John; Watson, David C.; Tessier, Luc; Lanthier, Patricia H.; Jarrell, Harold C.; Cadotte, Nicolas; St. Michael, Frank; Aberg, Erika; Szymanski, Christine M. J. Biol. Chem. 2002, 277, 42530-42539. (13) Hirabayashi, Jun; Hayama, Ko; Kaji, Hiroyuki; Isobe, Toshiaki; Kasai, KenIchi. J. Biochem. (Tokyo, Japan) 2002, 132, 103-114. (14) Guo, Lin; Eisenman, June R.; Mahimkar, Rajeev M.; Peschon, Jacques J.; Paxton, Raymond J.; Black, Roy A.; Johnson, Richard S. Mol. Cell. Proteomics 2002, 1, 30-36. (15) Wu, A. M.; Wu, J. H.; Watkins, W. M.; Chen, C.-P.; Tsai, M.-C. Biochim. Biophys. Acta 1996, 1316, 139-144. (16) Cummings, R. D; Kornfeld, S. J. Biol. Chem. 1984, 259, 6253-6260. (17) Sumi, S.; Arai, K.; Kitahara, S.; Yoshida, K.-I. J. Chromatogr., B 1999, 727, 9-14. (18) Lopez, M. F.; Berggren, K.; Chernokalskaya, E.; Lazarev, A.; Robinson, M.; Patton, W. F. Electrophoresis 2000, 21 (17), 3673-3683. (19) Regnier, F. E.; Riggs, L.; Zhang, R.; Xiong, L.; Liu, P.; Chakraborty, A.; Seeley, E.; Sioma, C.; Thompson, R. A. J. Mass Spectrom. 2002, 37 (2), 133-145.
The focus of the work described in this paper was to evaluate the utility of parallel and serial lectin affinity chromatography13 in conjunction with relative quantification by GIST8,17 for differential glycoproteomics. Following the development and validation of a lectin affinity based quantification method, two comparative proteomics problems were examined. The first was the deviation in individual sialylation of glycopeptides from normal. The second problem examined was the degree to which glycans at specific glycosylation sites in proteins were sialylated. MATERIALS AND METHODS Materials. Human serum, bovine fetuin, tris(hydroxymethyl)aminomethane (Tris base), tris(hydroxymethyl)aminomethane hydrochloride (Tris acid), L-1-tosylamide-2-phenylethyl chloromethyl ketone (TPCK)-treated trypsin, iodoacetic acid, cysteine, dithiothreitol (DTT), N-tosyl-L-lysine chloromethyl ketone (TLCK), manganese chloride, calcium chloride, N-hydroxysuccinimide, and lactose were purchased from Sigma (St. Louis, MO). Acetic anhydride-d0 and -d6 was purchased from Aldrich (Milwaukee, WI). Peptide-N-glycosidase F (PNGase F) and methyl R-D-mannopyranoside were obtained from Calbiochem (San Diego, CA). HPLC-grade acetonitrile (ACN) and hexane were purchased from Mallinckrodt Baker (Phillipsburg, NJ). Human serum from the individual subject was provided by Dr. Slentz. Con A agarose was purchased from Amersham Biosciences, (Piscataway, NJ), and Sambucus nigra (SNA) agarose was purchased from Vector laboratories (Burlingame, CA). Proteolysis. A 1-ml aliquot of human serum was reduced in 0.1 M phosphate buffer (pH 8.0) with 8 M urea and 10 mM DTT. After incubation at 37 °C for 2 h, iodoacetic acid was added to a concentration of 20 mM and incubated in darkness on ice for additional 2 h. The alkylation reaction was quenched for 30 min at room temperature by the addition of cysteine to a final concentration of 20 mM. After the sample was diluted with 0.1 M phosphate buffer to a final concentration of 2 M urea, TPCKtreated trypsin was added to the sample at a 50:1 trypsin to protein mass ratio. The sample was incubated at 37 °C for 24 h. Proteolysis was terminated by adding TLCK at a molar concentration exceeding that of trypsin by 2-fold. The sample was then frozen in liquid nitrogen for 10 min. Synthesis of N-Acetoxy-d3-succinimide. A solution of 4.0 g (34.8 mmol) of N-hydroxysuccinimide in 11.4 g (105 mmol) of acetic-d6-anhydride was stirred at room temperature. White crystals began to deposit in 10 min. After 15 h, the reaction mixture was filtered. The crystals were washed with hexane and then dried under vacuum. The product yield was 5.34 g (100%), mp 133-134 °C. Acetylation of the Peptides. A 5-fold molar excess of N-acetoxysuccinamide and N-acetoxy-d3-succinimide was added individually to two equal aliquots of tryptic digests. The reaction was allowed to proceed for 4-5 h at room temperature. NHydroxylamine was then added to the mixture in excess, and the pH was adjusted to 11-12. Incubation with hydroxylamine was allowed to proceed for 30 min. The function of the hydroxylamine reaction was to hydrolyze esters that might have been formed during the acetylation reaction. The labeled samples were adjusted to pH 7-8 before affinity selection. (20) Zhang, H.; Li, X.-j.; Martin, D. B.; Aebersold, R. Nat. Biotechnol. 2003, 21, 660-666.
Analytical Chemistry, Vol. 77, No. 9, May 1, 2005
2803
Chromatography. All chromatographic separations were performed using a Biocad 60 workstation from PerSeptive Biosystems (Framingham, MA). (a) SNA Affinity Selection. SNA agarose beads (Vector Laboratory) were packed into a 50 × 10 mm i.d. PEEK column. A 1-mL sample was applied to the SNA column after it had been equilibrated with 10 mM PBS. Unbound peptides were washed away with PBS, and the bound glycopeptides were eluted from the column with 0.2 M lactose in 10 mM PBS. (b) Con A Affinity Selection. Con A agarose beads (Amersham Biosciences) were packed into a 50 × 10 mm i.d. PEEK column. A 1-mL sample was applied to the Con A column after it had been equilibrated with 0.2 M Tris (pH 7.5) loading buffer containing 0.1 M NaCl, 1 mM CaCl2, and 1 mM MnCl2. Unbound peptides were washed away with loading buffer, and the bound glycopeptides were eluted from the column with 0.2 M methyl R-D-mannopyranoside in a solution of 0.1 M Tris (pH 7.5) containing 0.1 M NaCl. (c) Con A and SNA serial affinity selection. After a 1-mL sample was applied to the equilibrated Con A column, unbound peptides were washed away with loading buffer. The bound glycopeptides were transferred from Con A column to SNA column directly with 0.2 M methyl R-D--mannopyranoside. Con A column was then disconnected from SNA column. The 10 mM PBS was used to wash away the unbound peptides from SNA column. The bound glycopeptides were eluted from the SNA column with 0.2 M lactose in 10 mM PBS. (d) HPLC Fractionation. Fractionation of peptides on a 250 × 4.6 mm i.d. Vydac C18 column was achieved with a 120-min gradient from 100% buffer A (5% ACN, 95% H2O containing 0.1% TFA) to 60% buffer B (95% ACN, 5% H2O containing 0.1% TFA) at a flow rate of 1 mL/min. Eluted peptides were monitored at 215 nm, and fractions were manually collected for ESI-MS analysis. Deglycosylation by PNGase F. Glycopeptides were vacuumdried and redissolved in 100 µL pf 100 mM sodium phosphate buffer (pH 7.5). The reaction mixture was incubated with 2 µL (10 units) of PNGase F for 24 h at 37 °C. ESI-MS Analysis. Electrospray mass spectra were obtained on a QSTAR quadruple time-of-flight mass spectrometer (Applied Biosystems, Framingham, MA) equipped with an API ion source. The instrument was operated in positive ion TOF mode (m/z 3002000). Samples were dissolved in 100-µL of MeOH/H2O (50/50, v/v) containing 1% (v/v) formic acid and infused at 8 µL/min. The ion spray voltage used was 5500 V. Tandem Mass Spectrometry. Tandem MS analysis was performed by transmitting the appropriate parent ion from the quadruple to the collision cell. Collision gas used was nitrogen at a pressure of (4-6) × 10-6 Torr. Peptides were fragmented with appropriate collision energy (25-60 eV). Proteins were identified by searching the MS/MS spectra against the NCBI database (www.ncbi.nlm.nih.gov) using the Mascot search engine (Matrix Science, London, U.K.). RESULTS AND DISCUSSION Quantifications in the comparative studies of glycosylation reported here were based on stable isotope coding of samples according to sample origin or chromatographic methods used in glycopeptide fractionation. GIST was used to globally code peptides in tryptic digests (Figure 1). After sample coding, samples 2804 Analytical Chemistry, Vol. 77, No. 9, May 1, 2005
Figure 1. Differential acetylation of peptides from control and experimental samples with N-acetoxysuccinamide and deuterated N-acetoxysuccinamide.
Figure 2. Protocol used to quantify differences between samples in the concentration of sialylated glycoproteins.
were mixed and the glycopeptides were selected by lectin affinity chromatography prior to isotope ratio measurements and computation of relative concentration. Variation in Sialylated Glycoproteins. Sialylation is a common feature of blood glycoproteins. Sialic acid is a 9-carbon sugar occurring in more than 30 forms that is commonly found in glycoproteins at the outermost end of N-glycans and O-glycans. One of the more abundant types is N-acetylated at the 5-carbon position. Less frequent modifications of hydroxyl groups at the 4-, 7-, 8-, and 9-carbons with O-acetyl, O-methyl, O-sulfate, and O-phosphate groups are also seen. Further diversity is generated by different R-linkages from the 2-carbon to underlying sugar chains, the most common being to the 3- or 6-position of galactose
Figure 3. Reversed-phase chromatograms of peptides from human serum: (a) total tryptic digest; (b) differentially acetylated peptides selected by SNA followed by PNGase F deglycosylation. Peptides were eluted from a C18 column in a 120-min gradient ranging from 100% buffer A (5% ACN, 95% H2O containing 0.1% TFA) to 60% buffer B (95% ACN, 5% H2O containing 0.1% TFA).
or 6-position of GalNAc. Sialylation plays an important role in cell adhesion and determining the half-life of glycoproteins in blood. The amount of R-2-6 linked sialic acid has also been noted to increase in the case of malignant progression.21 Increasing sialic acid content of glycoproteins at the surface of cancer cells apparently plays a role in metastasis. It is for this reason that the development of methods to examine differential sialylation is very important. Sialylation is generally studied by deglycosylating proteins and examining the structure of the released glycans. Because many proteins are glycosylated and there can be multiple glycosylation sites within a single protein, this approach does not allow correlation of glycan structural features with specific glycosylation (21) Sata, T.; Roth, J.; Zuber, C.; Stamm, B.; Heitz, P. U. Am. J. Pathol. 1991, 139, 1435-1448.
sites on proteins. The objective of the study reported here was to begin the development of methods that would allow an examination of changes in sialylation at specific sites in glycoproteins across a glycoproteome through the use of sialic acid specific lectins. The GIST/lectin selection method described in Figure 2 was chosen for this study. SNA binds to peptides carrying sialic acid coupled to the underlying sugar chains through an R-2-6 linkage. One of the issues with a multiple-step quantification method such as this is reproducibility. Reproducibility was tested using bovine fetuin. Two bovine fetuin samples (2.5 mg/mL) were digested and coded individually with heavy and light forms of the coding agent using the methods described in the Material and Methods section. Equal amounts of digest labeled with heavy and light forms of the coding agent were mixed, and 100 µL of the mixture was Analytical Chemistry, Vol. 77, No. 9, May 1, 2005
2805
Table 1. Identification of Glycopeptides Containing Sialic Acid along with Individual Variations in Their Concentration SWISS-PROT accession no.
protein name
peptide sequence
P02763 P01871 P19827 P02749 P10909
orosomucoid 1 Ig µ chain precursor, membrane-bound (clone 201) inter-R (globulin) inhibitor β-2-glycoprotein I precursor apolipoprotein J
P02790
hemopexin precursor
P00737
haptoglobin-1 precursor
P01008 P08603
antithrombin III complement H factor
P05546 P02748 P25311 P02765
leuserpin 2 (hLS2) precursor complement 9 Zn-R2-glycoprotein R2-HS glycoprotein
P01860 P00450
Ig γ-3 heavy chain disease proteins ferroxidase precursor
P04196 P01009 P01023 P01019 P01877
histidine-rich glycoprotein precursor R-1-antitrypsin R 2 macroglobulin precursor angiotensinogen Ig R-2 chain C region
P19652 P01011 P02750 P02787
orosomucoid 2 R-1-antichymotrypsin precursor leucine-rich R-2-glycoprotein transferrin
CANLVPVPITNATLDQITGK YKNNSDISSTR ANLSSQALR VYKPSAGNNSLYR LANLTQGEDQYYLR MLNTSSLLEQLNEQFNWVSR SWPAVGNCSSALR ALPQPQNVTSLLGCTH NGTGHGNSTHHGPEYMR NLFLNHSENATAK VVLHPNYSQVDIGLIK MVSHHNLTTGATLINEQWLLTTAK LGACNDTLQQLMEVFK MDGASNVTCINSR IPCSQPPQIEHGTINSSR NLSMPLLPADFHK AVNITSENLIDDVVSLIR DIVEYYNDSNGSHVLQGR AALAAFNAQNNGSNFQLEEISR KVCQDCPLLAPLNDTR EQQFNSTFR ENLTAPGSDSAVFFEQGTTR EHEGAIYPDNTTDFQR ELHHLQEQNVSNAFLDK IADAHLDRVENTTVY YLGNATAIFFLPDEGK VSNQTLSLFFTVLQDVPVR HLVIHNESTCEQLAK TPLTANITK LSLHRPALEDLLLGSEANLTCTLTGLR HYTNSSQDVTVPCR CANLVPVPITNATLDR FNLTETSEAEIHQSFQHLLR KLPPGLLANFTLLR CGLVPVLAENYNK
a
isotope ratio (d0/d3)%a 178 172 129 118 99 74 97 95 69 94 75 67 91 89 62 89 88 86 85 83 79 75 74 40 74 70 65 53 51 47 S•Hb 50 49 46 33
The RSD of this method is 4% based on reproducibility studies with fetuin. b S•H represents heavily labeled singlet.
applied to the SNA column. Bound glycopeptides were then eluted, deglycosylated by PNGase F, and fractionated on a C18 reversedphase chromatography column. The peptide with the sequence KLCPDCPLLAPLNDSR was collected and applied to a QSTAR mass spectrometer for peptide identification and quantification of the isotope ratio. Theoretically the isotope ratio of peptides treated in this manner should be 1. Isotope ratios were calculated using peak areas derived from the total ion current of the mass spectrometer and dividing the area of the light isoform by the area of the heavy. Multiplication by 100 converted isotope ratios into percentage difference. The measurement error associated with this method was determined in a three-trial replicate of the protocol illustrated in the flowchart shown in Figure 2. The average of isotope ratio was 103% while the relative standard deviation (RSD) of the method was 4%. Differences in the concentration of sialylated proteins, or more precisely glycosylation sites, were observed by comparing concentrations from a population with those from an individual. An SNA affinity chromatography column was used to capture glycopeptides with this structural feature from trypsin digests of blood serum samples. This method however does not differentiate between N-glycosylated and O-glycosylated peptides and those that differ in number of sialic acid residues. A commercial serum sample pooled from many individuals was used to determine the 2806 Analytical Chemistry, Vol. 77, No. 9, May 1, 2005
average concentration of sialylated glycoproteins in a population of people. The experimental sample was taken from an active, normal 50-year-old male subject. The tryptic digest from the pooled sample was labeled with the heavy form of the isotopic coding agent while the sample from the individual subject was coded with the light form of the reagent. These specifically coded samples were then mixed, affinity selected with SNA, deglycosylated with PNGase F, separated by reversed-phase chromatography, and analyzed by tandem mass spectrometry. One of the great advantages of the lectin approach is that it greatly simplifies tryptic digests (Figure 3a) before mass spectral analysis. Fractionation of deglycosylated peptides on a C18 reversed-phase chromatography column is shown in Figure 3b. Although this method showed differences in sialic acid content in multiple glycoproteins, it did not exceed 2-fold in most cases (Table 1). It is the experience of this laboratory that up to a 2-fold variation from the average is normal in serum proteins. But exceptions to this were seen in transferrin, leucine-rich R2-glycoprotein, R1-antichymotrypsin precursor, ferroxidase precursor, and R2 immunoglobulin C-chain. In the case of R2 immunoglobulin C-chain, the amount of sialic acid on the glycopeptide LSLHRPALEDLLLGSEANLTCTLTGLR from the individual subject was 47% of that from the pooled sample while sialylation on the peptide HYTNSSQDVTVPCR from the individual subject was undetectable. In similar fashion, sialylation
Figure 4. Strategy used to recognize degree of sialylation in glycopeptides selected by Con A.
on the peptide ENLTAPGSDSAVFFEQGTTR from ferroxidase precursor in the individual subject was 75% that of pooled sample while the amount of sialic acid on the peptide ELHHLQEQNV-
SNAFLDK was 40%. The fact that this degree of variation in sialylation occurs within a single protein is surprising. This strongly suggests that sialylation is independently regulated at the various sites in these proteins. Serial Lectin Affinity Chromatography. Lectins are of both broad and narrow selectivity. SNA is an example of a lectin with narrow selectivity. Con A in contrast is a relatively broad selectivity lectin that binds glycoproteins and glycopeptides with N-glycans of high-mannose, hybrid, and complex biantennary types. Nonglycosylated peptides along with peptides carrying O-glycans and tri- or tetraantennary complex glycans are not bound. It has been shown in previous studies that by using broad selectivity and narrow selectivity lectin columns in tandem it is possible to select broad classes of glycoproteins and examine specific structural features they might carry.17 A multidimensional chromatographic protocol combining serial and parallel lectin affinity chromatography was used with the GIST protocol to assess the relative degree of sialylation in Con A selectable glycopeptides from human blood serum (Figure 4). Reproducibility of this method was tested using equal aliquots of tryptic digests from the pooled serum sample described above after GIST coding. Glycopeptides from the sample coded with the heavy isotope version of the GIST reagent were selected by Con A affinity chromatography alone. In contrast, glycopeptides from the sample coded with the light form of the reagent were serial selected with the Con A column first and then the SNA column (Figure 4). Glycopeptides selected by these two protocols were then mixed, deglycosylated with PNGase F, separated by reversedphase chromatography (Figure 5), and analyzed by tandem mass spectrometry. The protocol shown in Figure 4 was repeated three times, and the isotope ratios of five peptides were used to determine reproducibility of the method (Table 2). The average RSD among the five peptides was 4%.
Figure 5. Reversed-phase chromatogram of deglycosylated peptides from pooled human serum analyzed according to the protocol illustrated in Figure 4. A serum digest labeled with nondeuterated acetate was selected sequentially with Con A and SNA columns. An equal amount of serum digest labeled with the heavy isotope form of the coding agent was selected with the Con A column alone. Glycopeptides selected by these two protocols were then mixed and deglycosylated with PNGase F. Peptides were eluted from a C18 column in a 120-min gradient ranging from 100% buffer A (5% ACN, 95% H2O containing 0.1% TFA) to 60% buffer B (95% ACN, 5% H2O containing 0.1% TFA).
Analytical Chemistry, Vol. 77, No. 9, May 1, 2005
2807
Table 2. Error Analysis of the Serial Lectin Method Described in Figure 4 SWISS PROT accession no.
protein name
P01009 P00450 P02765 P00737 P01861
R-1-antitrypsin ferroxidase precursor R2-HS glycoprotein haptoglobin-1 precursor Ig γ-4 chain C region
peptide sequence
average of isotope ratios (d0/d3)%
relative standard deviation (RSD) (%)
YLGNATAIFFLPDEGK ENLTAPGSDSAVFFEQGTTR AALAAFNAQNNGSNFQLEEISR NLFLNHSENATAK EEQFNSTYR
96 89 79 52 24
3 3 2 5 7
Table 3. Identification of Con A Selectable Glycopeptides along with the Degree of Sialylation. The Protocol Used To Generate This Data Is Outlined in Figure 4a SWISS-PROT accession no.
protein name
P01019 P02765
angiotensinogen R2-HS glycoprotein
P00737
haptoglobin-1 precursor
P04196 P01591 P03952 P08709 P19652 P05546 P02749
histidine-rich glycoprotein precursor Ig J chain plasma kallikrein B1 precursor coagulation factor VII precursor orosomucoid 2 leuserpin 2 (hLS2) precursor β-2-glycoprotein I precursor
P02750 P02763 P25311 P43652 P04217 P00450
leucine-rich R-2-glycoprotein orosomucoid 1 Zn-R2-glycoprotein afamin precursor R-1-B-glycoprotein ferroxidase precursor
P01871
Ig µ chain precursor, membrane-bound (clone 201)
P01009 Q96PD5 P02679 P01877
a-1-antitrypsin N-acetylmuramoyl-L-alanine amidase precursor fibrinogen γ-A chain precursor Ig R-2 chain C region
P02790
hemopexin precursor
P04004
vitronectin
P01008 P04114 P02748 P10909
antithrombin III apolipoprotein B-100 complement 9 apolipoprotein J
P02787 P08603
transferrin complement H factor
P19827 P01023 P36955 P05090 P01861 P01860 P51884 P01024
inter-R (globulin) inhibitor a 2 macroglobulin precursor EPC-1 gene product apolipoprotein D Ig γ-4 chain C region Ig γ-3 heavy chain disease proteins lumican complement component 3 precursor
peptide sequence HLVIHNESTCEQLAK KVCQDCPLLAPLNDTR AALAAFNAQNNGSNFQLEEISR VVLHPNYSQVDIGLIK MVSHHNLTTGATLINEQWLLTTAK NLFLNHSENATAK IADAHLDRVENTTVY ENISDPTSLR IYSGILNLSDITK YPHKPEINSTTHPGADLQENFCR CANLVPVPITNATLDR NLSMPLLPADFHK VYKPSAGNNSLYR LGNWSAMPSCK KLPPGLLANFTLLR CANLVPVPITNATLDQITGK DIVEYYNDSNGSHVLQGR DIENFNSTQK AIFYETQPSLWAESESLLKPLANVTLTCQAR ELHHLQEQNVSNAFLDK EHEGAIYPDNTTDFQR ENLTAPGSDSAVFFEQGTTR YKNNSDISSTR THTNISESHPANTF YLGNATAIFFLPDEGK GFGVAIVGNYTAALPTEAALR DLQSLEDILHQVENK TPLTANITK HYTNSSQDVTVPCR SVTWSESGQNVTAR LSLHRPALEDLLLGSEANLTCTLTGLR ALPQPQNVTSLLGCTH SWPAVGNCSSALR NGTGHGNSTHHGPEYMR NGSLFALR NISDGFDGIPDNVDAALALPAHSYSGR LGACNDTLQQLMEVFK FNSSYLQGTNQITGR AVNITSENLIDDVVSLIR LANLTQGEDQYYLR MLNTSSLLEQLNEQFNWVSR CGLVPVLAENYNK IPCSQPPQIEHGTINSSR MDGASNVTCINSR ANLSSQALR GCVLLSYLNETVTVSASLESVR VTQNLTLIEESLTSEFIHDIDR ADGTVNQIEGEATPVNLTEPAK EEQFNSTYR EQQFNSTFR LHINHNNLTESVGPLPK TVLTPATNHMGNVTFTIPANR
isotope ratio (d0/d3)%a 103 101 79 100 95 52 100 100 100 99 99 99 98 97 98 98 98 96 96 96 93 89 96 76 96 96 94 94 91 88 74 93 89 70 93 79 92 92 90 89 80 87 86 85 83 77 77 33 24 22 16 S•Hb
a The relative standard deviation in these measurements is assumed to be 4% based on the test of this method seen in Table 2. b S•H represents heavily labeled singlet.
2808 Analytical Chemistry, Vol. 77, No. 9, May 1, 2005
The relative amount of sialylation among specific Con A selectable peptides from human serum (Table 3) is seen to be quite high. Of all the peptides identified, only five were less than 40% sialylated. Peptides with isotope ratios in the range 96-104% should be considered fully sialylated. Based on the fact that the peptide TVLTPATNHMGNVTFTIPANR from compliment component 3 precursor appeared as a heavily labeled singlet it was concluded that it carried no detectable sialylation. It is interesting that there can be substantial variation in sialylation between sites in the same protein. For example, sialylation in haptoglobin-1 precursor peptide VVLHPNYSQVDIGLIK was complete while that in the peptide NLFLNHSENATAK was 52%. Other examples can be seen in R2-HS glycoprotein, Ig µ chain precursor, Ig R-2 chain C region, and hemopexin precursor. It is highly likely that sialylation on various sites in the same protein is independently regulated.
A second conclusion is that the degree of variation in sialylation among glycoproteins is probably no greater than the variation in concentration between nonglycosylated proteins where 2-fold variations in protein concentration between subjects are common. Finally, it can be concluded that using broad selectivity and narrow selectivity lectins in the combined modes of serial lectin affinity chromatography along with parallel lectin affinity chromatography allows the degree of a specific type of glycosylation, in this case R-2-6 sialylation, to be assessed at each glycosylation site in a proteome. These methods should be a valuable asset in the study of glycopathologies relating to diseases such as cancer where the question is whether the aberration is due to changes in the total concentration of glycoproteins or the type of glycosylation at a particular site in a small number of proteins.
CONCLUSIONS Results from this study allow several conclusions to be made about lectin based comparative glycoproteomics. One is that lectins can be used in combination with reversed-phase chromatography, mass spectrometry, and stable isotope coating to quantitatively examine differences in glycosylation of glycoproteins involving a particular structural feature of glycans across a glycoproteome.
ACKNOWLEDGMENT The authors gratefully acknowledge financial support from NIH Grant GM-59996.
Received for review August 21, 2004. Accepted January 19, 2005. AC048751X
Analytical Chemistry, Vol. 77, No. 9, May 1, 2005
2809
