Precise and comparative pegylation analysis by microfluidics and mass spectrometry

Standard SDS-PAGE analysis of a pegylated protein was able to confirm an increase in its molecular size after reaction with an activated polyethylene glycol (PEG) but could do little to identify the extent of pegylation or to support characterization of the consistency of the modified protein. In this article, we demonstrate the utility of the capillary electrophoresis technology (using a microfluidic system) in analyzing the pegylation pattern of a recombinant protein over a range of 1-12 PEGs per polypeptide. Confirmatory data from mass spectrometry analysis of pegylated adducts are also presented. These allowed independent confirmation of the extent of pegylation. This electrophoretic analysis gives a robust, reproducible, and direct characterization of PEG adducts. We found that traditional estimation of PEG adducts by an indirect colorimetric (trinitrobenzene sulfonic acid) reaction, which detects loss of free amino groups, was quite erroneous for the recombinant protein in our study as well as several commercially available pegylated proteins. These results support the use of this capillary electrophoresis device for precise characterization of pegylated proteins.     

Regulation of transduction efficiency by pegylation of baculovirus vector in vitro and in vivo

In this study, poly(ethylene glycol) (PEG) was coupled to baculovirus to regulate transduction efficiency of baculovirus in vitro and in vivo. The degree of pegylation in virions was measured by the loss of free amines via a fluorescamine-based assay. The efficiency of green fluorescent protein (GFP) expression was used to monitor transduction efficiency. As the results, the transduction efficiency in pegylated baculovirus was decreased with an increase of pegylation in baculovirus in vitro and in vivo. Interestingly, the transduction efficiency of the pegylated baculovirus for the lung and brain was increased compared with baculovirus itself possibly owing to increased stability of baculovirus by pegylation.     

Ultrafiltration characteristics of pegylated proteins

There is growing clinical interest in the use of pegylated recombinant proteins with enhanced stability, half-life, and bioavailability. The objective of this study was to develop a quantitative understanding of the ultrafiltration characteristics of a series of pegylated proteins with different degrees of pegylation. Sieving data were compared with available theoretical models and with corresponding results for the partition coefficient in size exclusion chromatography (SEC). The sieving coefficients of the pegylated proteins depended not only on the protein size and the total molecular weight of the polyethylene glycol (PEG) but also on the number of PEG chains. This is in sharp contrast to the partition coefficient in SEC, which was uniquely determined by the total molecular weight of the PEG and protein. This difference is due to the deformation and/or elongation of the PEG chains caused by the convective flow into the membrane pores, an effect that is not present in SEC. These results provide important insights into the transport and separation characteristics of pegylated proteins.     

Lipid-induced conformational transitions of beta-lactoglobulin

Bovine beta-lactoglobulin (betaLG) provides an excellent model protein system for beta-to-alpha conformational change, but its behavior varies when the change is induced by alcohols, surfactants, or lipid vesicles. Here the interaction and orientation of betaLG in association with various artificial lipid vesicles at neutral and acidic pH have been studied by use of several complementary spectroscopic techniques. Circular dichroism (CD) and Fourier transform infrared (FTIR) spectra demonstrated that betaLG acquires a non-native alpha-helical structure upon binding with anionic lipids, while zwitterionic lipids do not have a significant effect on its conformation. The degree of induced alpha-helix depends on the lipid concentration and is strongly affected by the charge of the protein and lipids as well as the ionic strength of the solution. Near-UV CD and Trp emission spectra revealed that the tertiary structure of lipid-bound betaLG is highly expanded but not completely disrupted. Fluorescence quenching together with a Trp emission blue shift showed that the Trp residues remain largely shielded from the solvent when interacting with DMPG, which would be consistent with at least some portions of betaLG having been inserted into the lipid membrane. The orientations of the alpha-helix and beta-sheet axes in membrane-bound betaLG were found to be parallel and perpendicular, respectively, to the membrane film normal, as determined by use of polarized attenuated total reflection (ATR) FTIR spectra. Our findings reveal that the lipid-induced beta-to-alpha transition in betaLG, accompanied by a substantial disruption in tertiary structure, is mainly driven by strong electrostatic interactions. Once the tightly packed betaLG is disrupted, hydrophobic residues become exposed and available for insertion into the lipid bilayer, where hydrophobic interaction with the lipids may play a role in stabilizing the helical components.     

Photoimmobilization of organophosphorus hydrolase within a PEG-based hydrogel.

Organophosphorous hydrolase (OPH) was physically and covalently immobilized within photosensitive polyethylene glycol (PEG)-based hydrogels. The hydroxyl ends of branched polyethylene glycol (b-PEG, four arms, MW = 20,000) were modified with cinnamylidene acetate groups to give water-soluble, photosensitive PEG macromers (b-PEG-CA). The b-PEG-CA macromers underwent photocrosslinking reaction and formed gels upon UV irradiation (>300 nm) in the presence of erythrosin B. Native OPH was pegylated with cinnamylidene-terminated PEG chains (MW = 3400) to be covalently linked with the b-PEG-CA macromers during photogelation. The effect of pegylation on the stability of the enzyme was determined. Furthermore, the effect of enzyme concentration, wavelength of irradiation, and photosensitizer on the stability of the entrapped enzyme was also investigated. The pegylated OPH was more stable than the native enzyme, and the OPH-containing gels exhibited superior stability than the soluble enzyme preparations.     

Inactivation of horse liver alcohol dehydrogenase by modification of cysteine residue 174 with diazonium-1H-tetrazole.

Diazonium-1H-tetrazole was tested as a potential active-site-directed reagent for amino acid residues involved in catalysis by alcohol dehydrogenase. In a novel reaction with a protein, diazonium-1H-tetrazole inactivated the enzyme selectively, and almost stoichiometrically, but reacting with the sulfur of a cysteine residue, Cys-174. As a model compound, the tetrazole adduct of free cysteine was prepared. Elementary and spectral analyses of the adduct were consistent with the structure 5-tetrazoleazo-S-cysteine. The adduct absorbs light with a maximun at 316 nm, and is destroyed by irradiation at this wavelength. The inactivated enzyme still bound NADH as determined by difference spectroscopy, but did not enhance the fluorescence of the bound NADH as did native enzyme. X-ray crystallographic studies of free enzyme have shown that Cys-174 coordinates the zinc at the active site (Eklund, H., Nordstr?m, B., Zeppezauer, E., S?derlund, G., Ohlsson, I., Boiwe, T., and Br?ndén, C-I. (1974), FEBS Lett. 44, 200-204). The modified enzyme is probably inactive because the large, negatively charged tetrazole ring interferes sterically or electrostatically with the binding of substrates or with hydride transfer.     

Site of pegylation and polyethylene glycol molecule size attenuate interferon-alpha antiviral and antiproliferative activities through the JAK/STAT signaling pathway

Therapeutic pegylated interferon-alphas (IFN-alpha) are mixtures of positional isomers that have been monopegylated at specific sites on the core IFN-alpha molecule. The pegylation results in lower in vitro specific activity associated with the core IFN-alpha molecule that is related to the site of pegylation and size of polyethylene glycol (PEG) attached. We prepared purified, homogeneous, positional pegylation isomers of IFN-alpha2b that were monopegylated using 5-30-kDa linear PEG molecules attached at 7 primary reactive amino acid residues: Cys(1), His(34), Lys(31), Lys(83), Lys(121), Lys(131), and Lys(134). The isomers were evaluated for STAT translocation and antiviral and antiproliferative activity. The site of pegylation strongly influenced activity relative to an IFN-alpha2b control. The highest residual activity was observed with the His(34) positional isomers, and the lowest was observed with the Cys(1) positional isomers. The Lys positional isomers demonstrated intermediate activity, with a general order of Lys(134) > Lys(83) approximately Lys(131) approximately Lys(121) > Lys(31). The progressive relationship between decreased activity and increased PEG size suggests that pegylation may interfere with interaction and binding of IFN-alpha to the IFNAR1-IFNAR2 heterodimeric receptor. The higher specific activity associated with the His(34) positional isomer suggests that this site may be favorable for pegylating IFN-alpha2b molecules.     

Incorporation of beta-lactoglobulin in a lipid/porphyrin monolayer at the air--water interface

A catanionic lipid/porphyrin monolayer was formed at the air-water interface by the tetra-anionic porphyrin, tetra-sodium-meso-tetra(4-sulfonatophenyl)porphine (TSPP), mixed with the cationic lipid dioctadecyldimethylammonium bromide (DODAB) in a 1:4 molar ratio. This binary mixture (TSPP/4DODAB) was used as the incorporation matrix of beta-lactoglobulin (betaLG). Binary and ternary systems (TSPP/4DODAB/zbetaLG, where z stands for the number of protein residues per TSPP) were characterized by surface pressure versus area (pi-A) measurements and by Brewster angle microscopy (BAM) observation at the air-water interface. Pi-A measurements and BAM images show that protein is incorporated in the expanded regime of the monolayer and is gradually expelled upon compression at high surface pressures. The successive compression-expansion cycles indicate that the protein under adsorbed to the floating film is reincorporated after the expansion of the monolayer. At low subphase pH, TSPP tends to aggregate decreasing the interaction with DODAB molecules. Electrostatic and hydrophobic interactions are responsible for the presence of betaLG at the interfacial film.     

Limited proteolysis of branching enzyme from Escherichia coli

Branching enzyme is involved in determining the structure of starch and glycogen. It catalyzes the formation of branch points by cleavage and transfer of alpha-1,4-glucan chains to alpha-1,6 branch points. Branching enzyme belongs to the amylolytic family of enzymes containing four conserved regions in a central (alpha/beta)8-barrel. Limited proteolysis of the branching enzyme from Escherichia coli (84 kDa) by proteinase K produced a truncated protein of 70-kDa, which still retained 40-60% of branching activity, depending on the type of assay used. Amino acid sequencing showed that the 70-kDa protein lacked 111 or 113 residues at the amino terminal, whereas the carboxy terminal was still intact. We purified this truncated enzyme to homogeneity and analyzed its properties. The enzyme had a three- to fourfold lower catalytic efficiency than the native enzyme, whereas the substrate specificity was unaltered. Furthermore, a branching enzyme with 112 residues deleted at the amino terminal was constructed by recombinant technology and found to have properties identical to those of the proteolyzed enzyme.     

Relevance of charged groups for the integrity of the S-layer from Bacillus coagulans E38-66 and for molecular interactions.

In this paper, the importance of charged amino and carboxyl groups for the integrity of the cell surface layer (S-layer) lattice from Bacillus coagulans E38-66 and for the self-assembly of the isolated subunits was investigated. Amidination of the free amino groups which preserved their positive net charge had no influence on both. On the other hand, acetylation and succinylation, which converted the amino groups into either neutral or negatively charged groups, and amidation of carboxyl groups were accompanied by the disintegration or at least by the loss of the regular structure of the S-layer lattice. Treatment of S-layer monolayers with the zero-length cross-linker carbodiimide led to the introduction of peptide bonds between activated carboxyl groups and amino groups from adjacent subunits. This clearly indicated that in the native S-layer lattice the charged groups are located closely enough for direct electrostatic interactions. Under disrupting conditions in which the S-layer polypeptide chains were unfolded, 58% of the Asx and Glx residues could be amidated, indicating that they occur in the free carboxylic acid form. As derived from chemical modification of monolayer self-assembly products, about 90% of the lysine and 70% of the aspartic and glutamic acid residues are aligned on the surface of the S-layer protein domains. This corresponded to 45 amino groups and to 63 carboxyl groups per S-layer subunit. Labelling experiments with macromolecules with different sizes and charges and adsorption studies with ion-exchange particles revealed a surplus of free carboxyl groups on the inner and on the outer faces of the S-layer lattice. Since the carboxyl groups on the outer S-layer face were accessible only for protein molecules significantly smaller then the S-layer protomers or for positively charged, thin polymer chains extending from the surface of ion-exchange beads, the negatively charged sites must be located within indentations of the corrugated S-layer protein network. This was in contrast to the carboxyl groups on the inner S-layer face, which were found to be exposed on elevations of the S-layer protein domains (D. Pum, M. Sára, and U.B. Sleytr, J. Bacteriol. 171:5296-5303, 1989).     

Changes in peptic digestibility of bovine beta-lactoglobulin as a result of food processing studied by capillary electrophoresis and immunochemical methods

Digestion studies constitute a functional tool for allergen characterisation. This strategy for investigating allergenic proteins relates to the observation of increased proteolytic resistance of some proteins recognised to exhibit allergenic potential. beta-Lactoglobulin (betaLG) is one of the major whey proteins, a potent milk allergen and shows a high stability against peptic hydrolysis in its native form. In order to study the impact of milk fermentation process on its digestibility, two complementary analytical methods were applied: capillary zone electrophoresis (CZE) to quantitatively study proteolytic degradation of betaLG isolated from different fermented bovine milk products, and enzyme linked immunosorbent assay (ELISA) to assess differences in immunoreactivity. betaLG, isolated from either raw or pasteurised cow's milk (CM), as expected, showed only minimal digestibility (less than 10% in 2 h). However, when raw milk or pasteurised milk was fermented, the rate of peptic digestion of the protein significantly increased (up to 45% in 2 h). In accordance with changes in digestibility, the immunochemical response for all fermented samples was lower than that of non-fermented references. Raw and pasteurised milk "naturally" fermented in our laboratory only resulted in a slight reduction (betaLG detected is still in the range of milligrams per gram sample), whereas the industrially manufactured sour milk as well as the "Acidophilus milk" reflected a remarkably lower level of immunoreactivity (55-56 microg/g sample).     

Reduction and fluorescent labeling of cyst(e)ine-containing proteins for subsequent structural analyses

Procedures which allow rapid, quantitative, and selective fluorescent labeling of protein cyst(e)ine residues prior to electrophoresis by reaction with 4-(aminosulfonyl)-7-fluoro-2,1,3-benzoxadiazole (ABD-F) under mild conditions are described. After labeling, the protein(s) of interest is easily monitored throughout electrophoresis and subsequent electroblotting or electroelution procedures. The stoichiometry of labeling and therefore the number of cysteine and/or half-cystine residues can be measured spectrophotometrically or fluorometrically and the derived cyst(e)ine adduct can also be quantitated by amino acid analysis and identified in protein sequencing. N-terminal blockage is not observed under the conditions utilized, nor are any other amino acid side chains modified. The procedures described allow complete, rapid, and facile reduction and alkylation of proteins with simultaneous incorporation of a fluorophore, permitting sensitive detection in subsequent manipulation of the proteins. Quantitative fluorescence prelabeling also allows the generation, purification, and sequencing of peptide fragments containing cyst(e)ine residues for determination of internal sequences and residues involved in disulfide bonds.     

Kinetic folding and unfolding of staphylococcal nuclease and its six mutants studied by stopped-flow circular dichroism

Kinetics of refolding and unfolding of staphylococcal nuclease and its six mutants, each carrying single or double amino acid substitutions, are studied by stopped-flow circular dichroism measurements. A transient kinetic intermediate formed within 10 ms after refolding starts possesses a substantial part of the N-domain core β-structure, whereas helices are formed at the later stages. The structure of the kinetic intermediate is less organized than the structure that is known to be formed by a nuclease 1-136 fragment. Only the refolding kinetics are affected by the mutations in all the mutants except two in which the mutations have changed the native structure. From this result and also from the locations of the mutation sites, the major N-terminal domain of the nuclease in the transition state of folding has a structure nearly identical to the native one. On the other hand, the minor C-terminal domain has previously been shown to be still disorganized in the transition state. The effects of the amino acid substitutions on the stability of the native and the transition states are in good agreement with the changes in the hydration free energy, expected for the corresponding amino acid replacements in the unfolded polypeptide. Since side chains of all the mutated residues are not accessible to solvent in the native structure, the result suggests that it is the unfolded state that is mainly affected by the mutations. © 1995 Wiley-Liss, Inc.     

Metal binding specificity in carbonic anhydrase is influenced by conserved hydrophobic core residues.

The role of highly conserved aromatic residues surrounding the zinc binding site of human carbonic anhydrase II (CAII) in determining the metal ion binding specificity of this enzyme has been examined by mutagenesis. Residues F93, F95, and W97 are located along a beta-strand containing two residues that coordinate zinc, H94 and H96, and these aromatic amino acids contribute to the high zinc affinity and slow zinc dissociation rate constant of CAII [Hunt, J. A., and Fierke, C. A. (1997) J. Biol. Chem. 272, 20364-20372]. Substitutions of these aromatic amino acids with smaller side chains enhance the copper affinity (up to 100-fold) while decreasing the affinity of both cobalt and zinc, thereby altering the metal binding specificity up to 10(4)-fold. Furthermore, the free energy of the stability of native CAII, determined by solvent-induced denaturation, correlates positively with increased hydrophobicity of the amino acids at positions 93, 95, and 97 as well as with cobalt and zinc affinity. Conversely, increased copper affinity correlates with decreased protein stability. Zinc specificity is therefore enhanced by formation of the native enzyme structure. These data suggest that the hydrophobic cluster in CAII is important for orienting the histidine residues to stabilize metals bound with a distorted tetrahedral geometry and to destabilize the trigonal bipyramidal geometry of bound copper. Knowledge of the structural factors that lead to high metal ion specificity will aid in the design of metal ion biosensors and de novo catalytic sites.     

Novel human light chain V kappa segment: serologic and structural analyses of the kappa III-like Bence Jones protein and IgG kappa light chain REE

Immunochemical and sequence analyses of kappa light chain REE (Bence Jones protein REE and the light chain isolated from IgG kappa myeloma protein REE) revealed antigenic and structural features not previously described for human kappa-chains. Although closely related to proteins of the V kappa III subgroup, light chain REE is readily distinguished from light chains classified serologically as members of the kappa IIIa or kappa IIIb sub-subgroups. Light chains REE (Bence Jones protein REE and light chain REE) are identical in sequence and differ from kappa III proteins by at least 10 uncommon amino acid substitutions in the first three framework regions. Further, kappa-chain REE is unique by virtue of a four-residue deletion in the third complementarity-determining region. The deletion encompasses the three carboxyl-terminal residues in the V kappa-encoded segment and the first residue at the site of V-J recombination. Urine specimens from patient REE also contained a light chain fragment that lacked the first (amino-terminal) 85 residues of the native light chain but otherwise was identical in sequence to the light chain REE. The extensive amino acid differences and unique length of the V kappa segment in light chain REE indicate that this kappa-chain is the product of an unusual V kappa III gene or, alternatively, represents a rarely expressed and novel human V kappa gene.     

Electrostatic and hydrophobic interactions governing the interaction and binding of beta-lactoglobulin to membranes

The role of electrostatic and hydrophobic interactions in the binding and penetration of beta-lactoglobulin (betaLG) to preformed lipid membranes was studied using various phospholipid micelles and vesicles. Zwitterionic lysophospholipid micelles are able to induce the beta-sheet to alpha-helix transition, as judged by circular dichroism (CD), but the degree of transition is dramatically below and the amount of lipid required above that for anionic phospholipids with equivalent hydrocarbon chains. Anionic phospholipids with short hydrocarbon chains induce only low alpha-helical content in betaLG as compared to phospholipids with the same head group but longer hydrocarbon chains. These results suggest that both electrostatic and hydrophobic interactions are indispensable in betaLG-lipid interaction. Furthermore, air-water interface monolayer surface pressure and fluorescence anisotropy studies reveal that the membrane insertion of betaLG strongly depends on the nature of phospholipids, given the identical headgroup, particularly lipid packing. These results are supported by urea denaturation and acrylamide fluorescence quenching tests and by the FTIR-ATR polarization results for betaLG in multilayers on a surface. Under the same experimental conditions, the membrane binding and insertion of betaLG as well as the stability of the betaLG-lipid complexes can be enhanced by lowering the pH. Collectively, electrostatic interactions play a crucial role in all the processes involved in the betaLG-lipid interaction, while the presence of hydrophobic interaction remains necessary. Finally, betaLG biological function in the transport of fatty acids was tested by demonstrating the release of 2-AS from a 2-AS-betaLG complex on binding to lipids.     

Identification of the major positional isomer of pegylated interferon alpha-2b

Interferons display a wide range of antiviral, antiproliferative, and immunomodulatory activities on a variety of cell types and have been used to treat many diseases including hairy-cell leukemia and hepatitis B and C and have also been applied to other therapeutic areas. To improve the pharmacological properties of interferon (IFN) alpha-2b, a long-acting pegylated form (PEG-IFN) has been developed [PEG, monomethoxy poly(ethylene glycol) with average molecular mass of 12 000 Da]. PEG-IFN is a mixture of pegylated proteins with differing sites of PEG attachment. To identify the major positional isomer in the pegylated material [PEG-IFN(His-34)], NMR studies were conducted on a subtilisin-digested N-acetylated peptide of the major positional isomer [PEG-IFN(His-34)dig], synthetic peptide analogues containing His-34, as well as unmodified IFN and PEG-IFN(His-34). Our studies reveal a novel interferon-polymer attachment site as a histidine-linked interferon conjugate. We show that the major component of PEG-IFN is pegylated in the imidazole side chain of histidine-34. Chemical shift data suggest that pegylation occurs mainly at the N(delta)(1) position in the imidazole side chain of this residue. This positional isomer, PEG-IFN(His-34), comprises approximately 47% of the total pegylated species when PEG-IFN is synthesized under the current experimental conditions at pH 6.5 with an electrophilic derivative of PEG, succinimidyl carbonate PEG. The reversibility of the histidine modification was examined. The PEG-imidazole adduct in the intact protein, PEG-IFN(His-34), is labile but much more stable than in the peptide, PEG-IFN(His-34)dig. Apparently, the tertiary structure of the intact protein protects the His(34)-imidazole ring from depegylation.     

Production and characterization of a plant alpha-hydroxynitrile lyase in Escherichia coli

The coding sequence of the cyanogenic alpha-hydroxynitrile lyase gene of Manihot esculenta Crantz (cassava) was cloned in the plasmid vector pMal-c2 and expressed in Escherichia coli strain JM105. DNA sequencing showed that the recombinant plasmid contained the same sequence as the cDNA clone pHNL10. Peptide sequencing of the recombinant protein showed that the N-terminus was heterogeneous, with either four or six additional amino acid residues compared with the native protein. Circular dichroism spectra indicated similar secondary structure contents for both proteins. Enzyme assays showed that specific activity of native and recombinant proteins were 0.24 and 0.26 mmol CN(-)/mg/min, respectively; that both proteins had optimal activity at 40 degrees C and pH 5.5; and that both proteins were inhibited by the serine protease inhibitor phenyl-methane sulfonyl flouride (PMSF). Isoelectric focusing of native and recombinant protein revealed multiple isoforms for both proteins; the recombinant protein had a more basic mean isoelectric point (pl) (5.1) than the native protein (4.5). (c) 1997 John Wiley & Sons, Inc. Biotechnol Bioeng 53: 332-338, 1997.     

Antigen-antibody interaction. Synthetic peptides define linear antigenic determinants recognized by monoclonal antibodies directed to the cytoplasmic carboxyl terminus of rhodopsin

The specificities of four monoclonal antibodies rho 1D4, 1C5, 3A6, and 3D6 prepared by immunization of rod outer segments containing rhodopsin have been defined using synthetic peptides. All of these antibodies interact within the 18 residues at the COOH terminus of rhodopsin and recognize linear antigenic determinants of 4-11 residues. Twenty-seven synthetic peptide analogs of varying lengths of native sequence or containing single amino acid substitutions at each position of the COOH-terminal 18 residues have provided some insight into the mechanism of antigen-antibody binding. Our results clearly demonstrate that antibodies can be highly specific at key positions as shown by the loss of binding on single amino acid substitutions in the binding site. In contrast single amino acid substitutions at other positions in the binding site only affect affinity for some antibodies. Ionic interactions can dominate immunogenic determinants. Immunogenic determinants are not restricted to highly charged hydrophilic regions on the surface of a protein and may be dominated by hydrophobic interactions. Although certain side chains can dominate the interaction of the antigen with antibody, our results are in agreement with the interpretation that the free energies of all the contact points are additive and a certain free energy must be present to achieve binding. Antibodies with different specificities directed to the same region of the protein antigen can be produced in an immune response. Peptide antigens representing regions of a protein antigen bind best to the anti-protein antibody when the sequence is shortened to contain only those residues binding to the specificity site in the antibody. Cross-reactivity between protein antigens can be explained by conservation of the critical residues in the combining site.     

PDB-REPRDB: a database of representative protein chains from the Protein Data Bank (PDB) in 2003

PDB-REPRDB is a database of representative protein chains from the Protein Data Bank (PDB). Started at the Real World Computing Partnership (RWCP) in August 1997, it developed to the present system of PDB-REPRDB. In April 2001, the system was moved to the Computational Biology Research Center (CBRC), National Institute of Advanced Industrial Science and Technology (AIST) (http://www.cbrc.jp/); it is available at http://www.cbrc.jp/pdbreprdb/. The current database includes 33 368 protein chains from 16 682 PDB entries (1 September, 2002), from which are excluded (a) DNA and RNA data, (b) theoretically modeled data, (c) short chains (1<40 residues), or (d) data with non-standard amino acid residues at all residues. The number of entries including membrane protein structures in the PDB has increased rapidly with determination of numbers of membrane protein structures because of improved X-ray crystallography, NMR, and electron microscopic experimental techniques. Since many protein structure studies must address globular and membrane proteins separately, this new elimination factor, which excludes membrane protein chains, is introduced in the PDB-REPRDB system. Moreover, the PDB-REPRDB system for membrane protein chains begins at the same URL. The current membrane database includes 551 protein chains, including membrane domains in the SCOP database of release 1.59 (15 May, 2002).