Detection of N-glycans on small amounts of glycoproteins in tissue samples and sodium dodecyl sulfate-polyacrylamide gels

N-linked glycans harbored on glycoproteins profoundly affect the character of proteins by altering their structure or capacity to bind to other molecules. Specific knowledge of the role of N-glycans in these changes is limited due to difficulties in identifying precise carbohydrate structures on a given glycoprotein, which arises from the large amounts of glycoprotein required for N-glycan structural determination. Here, we refined a simple method to purify and detect trace amounts of N-glycans. During the N-glycan purification step, most contaminants were removed by two kinds of columns: a graphite carbon column and a cellulose column. N-Glycans were identified with a three-dimensional high-performance liquid chromatography (HPLC) system. Using our method, a global analysis of N-glycans from human muscle biopsy samples and mouse brain sections was possible. By combining sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with our method, we refined analytical procedures for N-glycans from SDS-PAGE gels using hydrazinolysis to achieve a high N-glycan recovery rate. N-Glycans on as little as 1 μg of the target protein transferrin or immunoglobulin G (IgG) were easily detected. These methods allowed us to efficiently determine glycoprotein N-glycans at picomole (pmol) levels.     

Structural analysis of N-linked sugar chains of human blood clotting factor IX

The structures of N-glycans of human blood clotting factor IX were studied. N-Glycans liberated by hydrazinolysis were N-acetylated and the reducing-end sugar residues were tagged with 2-aminopyridine. The pyridylamino (PA-) sugar chains thus obtained were purified by HPLC. Each PA-sugar chain was analyzed by two-dimensional sugar mapping combined with glycosidase digestion. The major structures of the N-linked sugar chains of human factor IX were found to be sialotetraantennary and sialotriantennary chains with or without fucose residues. These highly sialylated sugar chains are located on the activation peptide of the protein.     

Evaluation of laser microdissection as a tool in cancer glycomic studies

Laser microdissection (LMD) is a recent development that enables the isolation of specific cell populations from tissue sections. This study focuses on the potential of LMD as a tool in cancer glycomics using colon cancer as a model. LMD was performed on hematoxylin and eosin stained frozen tissue sections. Tumor cells and normal epithelial cells were selectively microdissected. N-Glycans from the LMD- and the bulk tissue-derived samples were liberated by hydrazinolysis and then labeled with 2-aminopyridine. After sialidase digestion, the resulting asialo-N-glycans were analyzed by normal and reversed phase HPLC combined with mass spectrometry. Comparison of the various N-glycan profiles with the aid of LMD identified seven characteristic N-glycans with significantly different expression profiles between normal and cancerous cells that could not be detected by conventional analysis. Thus, LMD is a potent and useful tool for analyzing variations in the expression of N-glycans by overcoming the problem of tissue sample heterogeneity.     

Comparison of protein solubilization methods suitable for proteomic analysis of soybean seed proteins

Extraction of soybean seed proteins for two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) and mass spectrometry analysis is challenging and inconsistent. In this study, we compared four different protein extraction/solubilization methods-urea, thiourea/urea, phenol, and a modified trichloroacetic acid (TCA)/acetone-to determine their efficacy in separating soybean seed proteins by 2D-PAGE. In all four methods, seed storage proteins were well separated by 2D-PAGE with minor variations in the intensity of the spots. The thiourea/urea and TCA methods showed higher protein resolution and spot intensity of all proteins compared with the other two methods. In addition, several less abundant and high molecular weight proteins were clearly resolved and strongly detected using the thiourea/urea and TCA methods. Protein spots obtained from the TCA method were subjected to mass spectrometry analysis to test their quality and compatibility. Fifteen protein spots were selected, digested with trypsin, and analyzed using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) and liquid chromatography mass spectrometry (LC-MS). The proteins identified were beta-conglycinin, glycinin, Kunitz trypsin inhibitor, alcohol dehydrogenase, Gly m Bd 28K allergen, and sucrose binding proteins. These results suggest that the thiourea/urea and TCA methods are efficient and reliable methods for 2D separation of soybean seed proteins and subsequent identification by mass spectrometry.     

Two-dimensional gel electrophoresis of wool intermediate filament proteins

Two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) methods have been used to provide high-resolution separation of wool intermediate filament proteins (IFPs). An improved method of extraction was developed based on a previously published method. The improved method for extraction eliminates the use of dialysis and freeze-drying between the extraction and rehydration steps, allowing the extraction and rehydration for the first dimension gel to be achieved in one day. Improvements to the method for maintaining reducing conditions and chaotrope constitution, combined with low %T polyacrylamide gels, allowed the high-resolution separation of the two keratin IFP families and their individual family members. The IFPs were separated to produce a clearly defined spot pattern of higher intensity, with numerous minor spots not previously observed, and a marked improvement in the vertical resolution. Further work to analyse the composition of each of the protein spots has been made much easier by being able to separate the IFPs into discrete spots.     

Determination of the site-specific and isoform-specific glycosylation in human plasma-derived antithrombin by IEF and capillary HPLC-ESI-MS/MS

The glycan structures of the major and more than ten minor populated isoforms of antithrombin (AT) were determined after separation of the isoforms by IEF using IPG strips. The bands excised from the gel were reduced, derivatized by iodoacetamide and submitted to tryptic digestion. The digest was analyzed by RP-HPLC-ESI-MS equipped with a quadrupole ion-trap mass analyzer. MS/MS experiments allowed establishing the monosaccharide compositions in the glycopeptides. For the major isoform of alpha-AT four identical biantennary glycans with two terminal sialic acids (SA) each, a total of eight SA, were found in full agreement with the literature. In the IEF-band containing this major isoform (pI 5.18) a further, much less abundant, isoform was detected showing a fucosylation on the glycan attached to Asn155 but being of otherwise identical structure as described above. The isoforms with pI 5.10 were found to include one triantennary glycan, all antennas carrying terminal SA. The occurrence of triantennary structure is site specific, involving the peptides with Asn(135) and Asn(155), alternately. At pI 5.24 we found those four isoforms that carry the glycans like the main-isoform of alpha-AT but missing one terminal SA. There was no site specificity found for the mono-sialo structure. The isoform at pI 5.31 is the major isoform of beta-AT containing three identical biantennary structures being fully sialylated. No isoforms (above 0.5% abundance) with two glycans only or three glycans other than beta-AT were detected. Fucosylation was found in the main isoform with an abundance of about 5%, and as expected with all the other isoforms with a comparable abundance.     

An efficient extraction method to enhance analysis of low abundant proteins from soybean seed

Large amounts of the major storage proteins, β-conglycinin and glycinin, in soybean (Glycine max) seeds hinder the isolation and characterization of less abundant seed proteins. We investigated whether isopropanol extraction could facilitate resolution of the low abundant proteins, different from the main storage protein fractions, in one-dimensional polyacrylamide gel electrophoresis (1D-PAGE) and two-dimensional polyacrylamide gel electrophoresis (2D-PAGE). 1D-PAGE of proteins extracted by different concentrations (10%, 20%, 30%, 40%, 50%, 60%, 70% and 80%) of isopropanol showed that greater than 30% isopropanol was suitable for preferential enrichment of low abundant proteins. Analysis of 2D-PAGE showed that proteins which were less abundant or absent by the conventional extraction procedure were clearly seen in the 40% isopropanol extracts. Increasing isopropanol concentration above 40% resulted in a decrease in the number of less abundant protein spots. We have identified a total of 107 protein spots using matrix-assisted laser desorption/ionization time of flight mass spectrophotometry (MALDI-TOF-MS) and liquid chromatography-mass spectrometry (LC-MS/MS). Our results suggest that extraction of soybean seed powder with 40% isopropanol enriches lower abundance proteins and is a suitable method for 2D-PAGE separation and identification. This methodology could potentially allow the extraction and characterization of low abundant proteins of other legume seeds containing highly abundant storage proteins.     

Structural features of N-glycans linked to glycoproteins from oil palm pollen, an allergenic pollen*

The pollen of oil palm (Elaeis guineensis Jacq.) is a strong allergen and causes severe pollinosis in Malaysia and Singapore. In the previous study (Biosci. Biotechnol. Biochem., 64, 820-827 (2002)), from the oil palm pollens, we purified an antigenic glycoprotein (Ela g Bd 31 K), which is recognized by IgE from palm pollinosis patients. In this report, we describe the structural analysis of sugar chains linked to palm pollen glycoproteins to confirm the ubiquitous occurrence of antigenic N-glycans in the allergenic pollen. N-Glycans liberated from the pollen glycoprotein mixture by hydrazinolysis were labeled with 2-aminopyridine followed by purification with a combination of size-fractionation HPLC and reversed-phase HPLC. The structures of the PA-sugar chains were analyzed by a combination of two-dimensional sugar chain mapping, electrospray ionization mass spectrometry (ESI-MS), and tandem MS analysis, as well as exoglycosidase digestions. The antigenic N-glycan bearing alpha1-3 fucose and/or beta1-2 xylose residues accounts for 36.9% of total N-glycans: GlcNAc2Man3Xyl1Fuc1GlcNAc2 (24.6%), GlcNAc2Man3Xyl1GlcNAc2 (4.4%), Man3Xyl1Fuc1-GlcNAc2 (1.1%), GlcNAc1Man3Xyl1Fuc1GlcNAc2 (5.6%), and GlcNAc1Man3Xyl1GlcNAc2 (1.2%). The remaining 63.1% of the total N-glycans belong to the high-mannose type structure: Man9GlcNAc2 (5.8%), Man8GlcNAc2 (32.1%), Man7GlcNAc2 (19.9%), Man6GlcNAc2 (5.3%).     

Comparative biochemical study of N-linked glycans from skin of a squid, Todarodes pacificus

For comparative biochemical interest, we analyzed the structures of N-glycans in a squid belonging to the Lophotrochozoa, one of the protostome clades. N-Glycans were prepared from squid skin by hydrazinolysis and re-N-acetylation followed by fluorescent tagging with 2-aminopyridine. The labeled N-glycans were purified, and their structures were determined by the two-dimensional HPLC mapping method combined with glycosidase digestions and mass spectrometry. We found that high mannose-type glycans, paucimannose-type glycans and complex-type glycans with a type-1 structure (Galbeta1-3GlcNAc) were dominant in squid skin. The complex-type glycans detected in the squid were similar to those in vertebrates, but have not yet been found in the Ecdysozoa, which is another protostome clade. However, paucimannose-type glycans are commonly found in the Ecdysozoa. Thus, the N-glycan structures of the squid belonging to the Lophotrochozoa have features common to those in vertebrates and the Ecdysozoa including insects and nematodes.     

N-glycan structures from the major glycoproteins of pigeon egg white: predominance of terminal Galalpha(1)Gal

N-Glycans from major glycoproteins of pigeon egg white (ovotransferrin, ovomucoid, and ovalbumins) were enzymatically released and were reductively aminated with 2-aminopyridine, separated, and structurally characterized by mass spectrometry and a three-dimensional mapping technique using three different columns of high performance liquid chromatography (HPLC) (Takahashi, N., Nakagawa, H., Fujikawa, K., Kawamura, Y., and Tomiya, N. (1995) Anal. Biochem. 226, 139-146). Twenty-five major N-glycan structures, all of them hitherto unknown, were identified as pyridylamino derivatives. Of these, 13 were neutral, 10 were monosialyl, and 2 were disialyl oligosaccharides. All N-glycans contain from one to four Galalpha(1,4)Galbeta(1,4) sequences at the nonreducing terminal positions and are devoid of fucose residues. N-Acetylneuraminic acids were alpha(2,6)-linked only to beta-galactose. The HPLC profiles of the N-glycans from four different glycoproteins were qualitatively very similar to each other, but not identical in the peak distributions. Monosialyl glycans were most abundant in all four glycoproteins, followed by neutral glycans. Disialyl glycans were lowest in ovotransferrin, and highest in ovomucoid. Triantennary structures with bisecting GlcNAc were predominant in ovotransferrin, and tetra-antennary (with and without bisecting GlcNAc-containing) structures were predominant in other glycoproteins. Penta-antennary structures (with a sialic acid and without bisecting GlcNAc residue) were also found in small quantities in all four glycoproteins. In contrast to the chicken egg white counterparts, which contain mostly high mannose and hybrid types, all N-glycan structures in the major pigeon egg white glycoproteins are complex type.     

Actin-isoform pattern as a marker of normal or pathological smooth-muscle and fibroblastic tissues

The relative proportions of actin isoforms present in smooth-muscle (SM) and fibroblastic human and non-human tissue extracts were examined by densitometric evaluation of Coomassie-Blue-stained spots in two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) as well as by quantification of radiolabeled actin NH2-terminal peptide spots separated by two-dimensional paper electrophoresis. SM tissues contained alpha- and gamma-SM as well as beta- and gamma-cytoplasmic (CY) actins in different proportions in different organs. Species differences with respect to the ratios of the isoactins were also observed. Moreover, during pregnancy, both human and rat myometrium exhibited a changed actin-isoform pattern, there being an increased proportion of gamma-actin. Analysis of the NH2-terminal peptides showed that, in human myometrium, this was essentially due to an increase in the amount of the gamma-SM isoform. Fibroblastic tissues were found to contain only the beta- and gamma-CY isoforms, the ratio being approximately 2.6:1. Thus, the presence or absence of alpha-actin provides a reliable biochemical criterion for distinguishing between fibroblastic and SM cell populations and/or tissues. This distinction and the evaluation of changes in isoactin ratios may be useful in the study of differentiation as well as physiological and pathological phenomena, and for determining the origin of certain soft-tissue tumours.     

Optimization of 2D-PAGE protocols for proteomic analysis of two nonaxenic toxin-producing freshwater cyanobacteria: Cylindrospermopsis raciborskii and Raphidiopsis sp.

Aims:  To optimize a protocol for the extraction and an in-depth analysis of the soluble protein fraction of two nonaxenic toxin-producing cyanobacteria Cylindrospermopsis raciborskii (hepatotoxin-producing), and Raphidiopsis sp. (neurotoxin-producing), using two-dimensional polyacrylamide gel electrophoresis (2D-PAGE).
Methods and Results:  The soluble protein fractions from strains of C. raciborskii and Raphidiosis sp. with different toxicity phenotypes were analysed by 2D-PAGE. Specific protocols were optimized specifically for each strain. Between 500 and 700 sharp protein spots were distinguished in a single 4–7 pH range 2D-PAGE for each cyanobacterium. Comparison of the protein maps of C. raciborskii CS-505 (a cylindrospermopsin-producing strain) and Raphidiopsis sp. D9 (saxitoxin-producing strain) against the nontoxic C. raciborskii strain CS-509 revealed many unique proteins in each protein map. We confirmed that the resolved proteins were cyanobacterial by identifying three randomly chosen protein spots from a Raphidiopsis sp. strain D9 2D-PAGE, using high-performance liquid chromatography (HPLC) tandem mass spectrometry (MS).
Conclusions:  The 2D-PAGE conditions presented here provide a robust protocol for proteomic studies in two CYN- and STX-producing model organisms, C. raciborskii and Raphidiopsis sp.
Significance and Impact of the Study:  We present the first protocols for proteomic analyses of Cylindrospermopsis raciborskii and Raphidiopsis sp.     

Separation and identification of soybean leaf proteins by two-dimensional gel electrophoresis and mass spectrometry

To establish a proteomic reference map for soybean leaves, we separated and identified leaf proteins using two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) and mass spectrometry (MS). Tryptic digests of 260 spots were subjected to peptide mass fingerprinting (PMF) by matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) MS. Fifty-three of these protein spots were identified by searching NCBInr and SwissProt databases using the Mascot search engine. Sixty-seven spots that were not identified by MALDI-TOF-MS analysis were analyzed with liquid chromatography tandem mass spectrometry (LC-MS/MS), and 66 of these spots were identified by searching against the NCBInr, SwissProt and expressed sequence tag (EST) databases. We have identified a total of 71 unique proteins. The majority of the identified leaf proteins are involved in energy metabolism. The results indicate that 2D-PAGE, combined with MALDI-TOF-MS and LC-MS/MS, is a sensitive and powerful technique for separation and identification of soybean leaf proteins. A summary of the identified proteins and their putative functions is discussed.     

Investigations into charge heterogeneity of wool intermediate filament proteins

Genomic studies have shown that there are four abundant type I and type II intermediate filament proteins (IFPs) in wool. When separated using 2D-PAGE, the type I IFPs separated into four clearly defined major rows. The type II IFPs separated into two distinct staggered rows. The large number of spots seen by 2D-PAGE has previously been attributed to charge heterogeneity caused by post-translational modification of the protein. However, analysis of wool IFPs by 2D-PAGE techniques and mass spectrometry suggested an absence of phosphorylation or glycosylation modifications. Investigations with both the type I and type II IFPs showed that when single protein spots from a 2D-PAGE separation are eluted, re-focused and re-electrophoresed, several spots are formed on both the acidic and basic side of the original spot. Amino acid analysis, mass spectrometry and Ellman's assay support the hypothesis that the proteins have the same sequence but vary in isoelectric charge, due to differences in exposure of charged residues on the molecular surface. The cause of IFP charge heterogeneity is thus proposed to be a conformational equilibrium between several different forms of the same protein in the rehydration solution used for the first dimension.     

Proteomic study of plasma from moderate hypercholesterolemic patients

Proteomics is a technology to detect and identify several proteins and their isoforms in a single sample. We used proteomics to analyze modifications in the protein map of plasma after simvastatin treatment of moderate hypercholesterolemic patients. Plasma from hypercholesterolemic patients (n = 9) was compared before and after 12 weeks of simvastatin treatment (40 mg/day). Patients with similar cardiovascular risk factors were used as controls (CR group). By using two-dimensional electrophoresis and mass spectrometry, we identified the different protein isoforms. The plasma expression of three fibrinogen gamma chain isoforms (FGG) was enhanced, whereas the expression of two isoforms of the fibrinogen beta chain (FGB) was reduced in the hypercholesterolemic patients compared with the CR group. The expression of apolipoprotein A-IV and three haptoglobin isoforms was higher in hypercholesterolemic patients. Simvastatin treatment modified the plasma expression of FGG chain isoform 1, FGB chain isoforms 1 and 2, vitamin D binding protein isoform 3, apo A-IV, and haptoglobin isoform 2. The modification of FGG chain isoform 1 and FGB chain isoforms 1 and 2 was positively correlated with total plasma cholesterol level. Proteomic analysis of plasma may help to know more in depth the molecular mechanism modified by simvastatin treatment.     

Mass spectrometric assay for analysis of haptoglobin fucosylation in pancreatic cancer

A mass spectrometric method was developed to elucidate the N-glycan structures of serum glycoproteins and utilize fucosylated glycans as potential markers for pancreatic cancer. This assay was applied to haptoglobin in human serum where N-glycans derived from the serum of 16 pancreatic cancer patients were compared with those from 15 individuals with benign conditions (5 normals, 5 chronic pancreatitis, and 5 type II diabetes). This assay used only 10 μL of serum where haptoglobin was extracted using a monoclonal antibody and quantitative permethylation was performed on desialylated N-glycans followed by MALDI-QIT-TOF MS analysis. Eight desialylated N-glycan structures of haptoglobin were identified where a bifucosylated triantennary structure was reported for the first time in pancreatic cancer samples. Both core and antennary fucosylation were elevated in pancreatic cancer samples compared to samples from benign conditions. Fucosylation degree indices were calculated and show a significant difference between pancreatic cancer patients of all stages and the benign conditions analyzed. This study demonstrates that a serum assay based on haptoglobin fucosylation patterns using mass spectrometric analysis may serve as a novel method for the diagnosis of pancreatic cancer.     

Sequential analysis of N- and O-linked glycosylation of 2D-PAGE separated glycoproteins

A robust method has been developed that allows analysis of both N- and O-linked oligosaccharides released from glycoproteins separated using 2D-PAGE and then electroblotted to PVDF membrane. This analysis provides efficient oligosaccharide profiling applicable to glycoproteomic analysis. The method involves the enzymatic release of N-linked oligosaccharides using PNGase F followed by the chemical release of O-linked oligosaccharides using reductive beta-elimination and analysis using LC-ESI-MS. Oligosaccharides from the major plasma glycoproteins with a pI between 4 and 7 were characterized from the glycoforms of haptoglobin, alpha2-HS-glycoprotein, serotransferrin, alpha1-antitrypsin, and alpha1-antichymotrypsin. It was shown that the separation of protein glycoforms evident in 2D-PAGE is partially due to the combined sialylation of the O-linked and N-linked oligosaccharides. Bi-, tri- and tetra-antennary N-linked structures, which had differing levels of sialylation and fucosylation, were found to be present on the glycoproteins analyzed, together with O-linked oligosaccharides such as mono-, and disialylated T-antigen and a disialylated core type 2 hexasaccharide. In addition, N-linked site-specific information was obtained by MALDI-MS analysis using tryptic digestion after PNGase F release of the oligosaccharides.     

N-linked oligosaccharides of glycoproteins from Ginkgo biloba pollen, an allergenic pollen.

The pollen of Ginkgo biloba is one of the allergens that cause pollen allergy symptoms. The plant complex type N-glycans bearing beta1-2 xylose and/or alpha1-3 fucose residue(s) linked to glycoallergens have been considered to be critical epitopes in various immune reactions. In this report, the structures of N-glycans of total glycoproteins prepared from Ginkgo biloba pollens were analyzed to confirm whether such plant complex type N-glycans occur in the pollen glycoproteins. The glycoproteins were extracted by SDS-Tris buffer. N-Glycans liberated from the pollen glycoprotein mixture by hydrazinolysis were labeled with 2-aminopyridine and the resulting pyridylaminated (PA-)N-glycans were purified by a combination of size-fractionation HPLC and reversed-phase HPLC. The structures of the PA-sugar chains were analyzed by a combination of two-dimensional sugar chain mapping, IS-MS, and MS/MS. The plant complex type structures (GlcNAc2Man3Xyl1Fuc1GlcNAc2 (31%), GlcNAc2Man3Xyl1GlcNAc2 (5%), Man3Xyl1Fuc1GlcNAc2 (13%), GlcNAc1Man3Xyl1Fuc1GlcNAc2 (8%), and GlcNAc1Man3Xyl1GlcNAc2 (17%)) have been found among the N-glycans of the glycoproteins of Ginkgo biloba pollen, which might be candidates for the epitopes involved in Ginkgo pollen allergy. The remaining 26% of the total pollen N-glycans have the typical high-mannose type structures: Man8GlcNAc2 (11%) and Man6GlcNAc2 (15%).     

High-performance liquid chromatographic mapping and structural characterization of neurophysin isoforms

Both ion-exchange and reverse-phase HPLC protocols for micromapping of neurophysins have been examined and the structural relationships among the major isoforms identified in the maps have been characterized. Reverse-phase HPLC was found to be especially useful for obtaining fingerprints of the isoforms within each of the two major families of neurophysins, I (oxytocin-related) and II (vasopressin-related), for both bovine and human neurophysins from posterior pituitary sources. From fractionation of the bovine proteins on octylsilyl columns, at least four neurophysins I were identified, one of which corresponds to the intact sequence of 93 residues and three of which vary from the parent by various degrees of carboxyl-terminal truncation. For bovine neurophysin II, two isoforms were identified in the reverse-phase HPLC maps, both of 95 residues, which vary from one another by the residue, either Ile or Val, at position 89. Isoforms were also detected for human neurophysins, including a carboxyl-terminal truncated form of human neurophysin II. All of the major neurophysin isoforms and several of the minor forms were shown to be functionally active as expressed by their binding to peptide ligand affinity matrices. Reverse-phase HPLC mapping on the octylsilyl matrix allowed neurophysin fingerprinting of crude tissue extracts by providing a narrow "window" within which the neurophysins elute but many other polypeptides expected to be present are excluded. The reverse phase HPLC method provides a useful way to obtain isolated neurophysin isoforms for physicochemical characterizations now usually carried out with mixtures of isoforms obtained by ion-exchange chromatography. The method also has characteristics amenable both for high-sensitivity fingerprinting of neurophysin isoforms, from different species and anatomical sources, and as a prelude to microstructural and -functional characterization of the isoforms so isolated.     

Distinct chromatographic forms of human hemi-myeloperoxidase obtained by reductive cleavage of the dimeric enzyme. Evidence for subunit heterogeneity

The enzyme myeloperoxidase (MPO) is a functionally important glycoprotein of neutrophilic granulocytes and occurs in three major isoforms (forms 1, 2, and 3) that are dimeric structures composed of two heavy subunit-light subunit protomers, each of which is associated with a chlorine-like prosthetic group. In the present study, highly purified MPO isoforms were obtained from the cells of a single normal donor, and each protein was subjected to reductive alkylation under nondenaturing conditions. The resulting enzymatically active protomers were separated from unreacted dimer using gel filtration chromatography. Use of a fast protein liquid chromatography cation exchange system with a Mono S matrix revealed heterogeneity of the protomers, and allowed essentially complete resolution of the protomers of MPO form 2. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the two resolved protomeric species under reducing conditions revealed small but reproducible differences in the Mr of their heavy subunits (59,000 and 57,000). Treatment with either endo-beta-N-acetylglucosaminidase or peptide N-glycohydrolase F reduced the Mr of each heavy subunit by approximately 3000 but did not change their relative electrophoretic mobilities. Heavy and light subunits were prepared from each of the MPO isoforms by reductive alkylation under conditions that allowed full retention of the prosthetic group with the heavy subunit. Reverse-phase chromatography and amino-terminal sequencing showed that each MPO isoform contained one major species of light subunit and several minor species. No differences in peroxidatic activity or inhibition by salicylhydroxamic acid were observed among any of the MPO isoforms or resolved protomers, but the latter were considerably more heat labile than dimeric forms of the enzyme and a monomeric form isolated from HL-60 cells. This is the first report of the isolation and partial characterization of distinct protomers from a single isoform of human MPO and suggests that the structure of MPO is more complex than considered previously.